From korg at darkqueen.org  Tue Jul  1 01:09:48 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 11:09:48 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14 msgs
In-Reply-To: Your message of "Tue, 01 Jul 2003 03:53:01 +1000."
 <20030630175301.20310.73173.Mailman@numbat.object-craft.com.au>
Message-ID: <200307010109.h6119m713416@firewall.darkqueen.org>



From: Dave Cole <djc@object-craft.com.au>:
  | 
  | Matt> On Sun, 2003-06-29 at 10:42, Dave Cole wrote:
  | >> I am at the end of the largest outstanding documentation task - the
  | >> tag reference.  Now would be a good time to hit me with issues
  | >> before I launch into another task.
  | 
  | Matt> Is that a hint? ;-)
  | 
  | Maybe...

Now I feel silly for commenting on the docs. :-/



From: "Michael C. Neel" <neel@mediapulse.com>:
  | From: Dave Cole [mailto:djc@object-craft.com.au]=20
  |      > Subject: Re: [albatross-users] New tag reference
  | =20
  | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-ge
  | neric.html

        if user:
            ctx.locals._user = user
            ctx.add_session_vars('_user')
                                  ^
                             remove the '_' ??

  | =20
  | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-im
  | age.html

Oooooh, nifty!!


So, having said I feel stoopid for commenting on the docs...  I notice 
that the page renders as:

  <al-input type="image" nextpage="m" srcicons/right.gif" whitespace>
                                      ^
                                 missing " again.

  | =20
  | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-fi
  | le.html

   "a flag indicating whether or not the list (5.2.2.4) attribute was
   present."

is directly followed by:

   "The request merging allows the user to submit more than one file in a
   file input field. To simplify application code the Request always
   returns a list of FileField objects for file inputs"

So, 'request merging' allows submitting more than one file in a single
al-imput, but the list attribute allows for multiple al-inputs of the
same name? ?? 

It seems unclear.



From: "Eric S. Johansson" <esj@harvee.org>:
  | Subject: Re: [albatross-users] Renaming application and execution context
  |  classes
  | 
  | 
  | To me, simple context is far more descriptive of what the class does than 
  | template context.  This is more apparent when you contrast simple applicati
  | on 
  | versus client session page object application.

I didnt find it so as a newbie. I thought simple applied to the
complexity of the object, not how the data was stored.

That could be because Im dumb tho :) :) :)

  | 
  | On a personal note, I find StudleyCaps quite disturbing.  In 1994 after bei
  | ng a 
[...]
  | again, this is a long-winded way of saying that ordinary_words 
  | separated_by_the_underscore_character is preferable to me and that I subscr
  | ibe 

Im with you there, man. :) 

Long descriptive names and '_'s up the whazoo for me :)

I want going to comment because I didnt write albatross, so I didnt
feel that I could comment on coding style, but now you mention it...

Now, if we cant use generic classes with initiators like:
   Context(storage=Albatross.file)
then Id vote for

 2| SimpleContext                   Context_stored_nowhere(?)
                                    Context_no_store(?)
 2| AppContext
 2|
 2| SimpleAppContext                Context_stored_client
 2| SessionAppContext               Context_stored_server_process
 2| SessionFileAppContext           Context_stored_server_file

(note, typing this, was suddenly unsure of SessionAppContext again,
before I worked out which one it was by process of elimination. Id
argue that the name isnt immediatly obvious, so renaming would be
helpful)

And the same but longer for the Apps (although again, Id prefer
the app not to need to know how the context was stored, but maybe
thats just me).


--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
"I changed my headlights the other day. I put in strobe lights instead! Now
when I drive at night, it looks like everyone else is standing still ..."
		-- Steven Wright

From esj at harvee.org  Tue Jul  1 02:50:21 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 30 Jun 2003 22:50:21 -0400
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14
 msgs
In-Reply-To: <200307010109.h6119m713416@firewall.darkqueen.org>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
Message-ID: <3F00F6ED.5050009@harvee.org>

Cameron Blackwood wrote:

> Now I feel silly for commenting on the docs. :-/

shouldn't.  (Too much)
> 
> Im with you there, man. :) 
> 
> Long descriptive names and '_'s up the whazoo for me :)
> 
> I want going to comment because I didnt write albatross, so I didnt
> feel that I could comment on coding style, but now you mention it...
> 
> Now, if we cant use generic classes with initiators like:
>    Context(storage=Albatross.file)
> then Id vote for
> 
>  2| SimpleContext                   Context_stored_nowhere(?)
>                                     Context_no_store(?)
>  2| AppContext
>  2|
>  2| SimpleAppContext                Context_stored_client
>  2| SessionAppContext               Context_stored_server_process
>  2| SessionFileAppContext           Context_stored_server_file

I like this series of proposed names because it's something that makes sense.

> And the same but longer for the Apps (although again, Id prefer
> the app not to need to know how the context was stored, but maybe
> thats just me).

there is a couple things about albatross that has plagued me and excessive 
visibility of what should be internals is one of them.

The documentation looks like it is very good if you have already absorbed the 
gestalt of albatross.  Unfortunately, I've been reading the documentation and 
trying to understand the gestalt without much success.  Goodness knows I've 
written my fair share of train-by-telepathy documentation so there is no blame 
by me on that point, just observation.

one of the gestalt points is communicating information from the python side to 
the HTML side and back.  I've been mucking with it for about a week and I am no 
closer to having a useful knowledge of information transfer issues than I was 
before I started.  The black magic surrounding context.local really isn't clear 
nor is the usage of attributes.  I've been using Python for the better part of 
the year and a half now and I've never used attributes (knowingly).  Maybe it's 
some sort of watershed one must cross in order to become an anointed 
pythoninista, I don't know.

In any case, this gestalt has driven me to the point of admitting failure.  I've 
surveyed a fair number of template/CGI tools and I think albatross is one of the 
better overall and has the best potential for being small and really effective. 
  However, I have been unable to learn enough to do I need to do and I cannot 
ask any more from the community.  You have been more than generous with your 
time  already and for that I am grateful.

on the documentation front, take a look at CherryPY and pso.  while I'm not 
happy with either of them for different reasons, they both have good 
documentation teaching people how to use their system.

If you can accept a suggestion, I would suggest creating an very simple 
user-friendly layer.  It would reflect the simple nature of CGI environment.  It 
would consist of a couple of functions/methods.  The first would be used for 
generating output.  It would associate a dictionary with an albatross HTML 
template.  The keys of the dictionary would map to variable names in the 
template.  The matching value to a given key would be substituted into the 
variables.  Obviously, nested lists, dictionaries etc. would be permitted so 
that for loops etc. in the templates could be used.

the second function would extract information back from CGI forms.  I haven't 
thought long on the form it would take but I suspect it would also be some form 
of dictionary based structure based on an albatross HTML template.

obviously these ideas aren't fully formed but they are a reflection of my 
confidence that that the whole process can be simplified even further than it 
has been to date.  I also believe it should be possible to create a layer over 
the current albatross API to generate this kind of simple interface.

Again, I want to thank everyone who has been so very helpful and I appreciate 
the effort you have expended in helping me.

---eric





From korg at darkqueen.org  Tue Jul  1 05:42:05 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 15:42:05 +1000
Subject: [albatross-users] al-input type="image":  internal bug/obscured traceback??
Message-ID: <200307010542.h615g5N15836@firewall.darkqueen.org>


Halp...


Hi, Im playing with clickable images in 1.10pre1 and I try:
----------- vv -----------
<html>
<head><title>Oh crap</title></head>
<body>

<al-form method="post">

<al-input type="image" name="pos" src="http://10.0.0.254/cgi-bin/20/map_image" whitespace>

<al-input type="submit" name="submit" value="never!">

</al-form>
</body>
</html>
----------- ^^ -----------


which generates some thing like (neatened):
----------- vv -----------
<html>
<head><title>Oh crap</title></head>
<body>

<form method="post" action="/cgi-bin/20">

<input type="image" src="http://10.0.0.254/cgi-bin/20/map_image" name="pos">

<input type="submit" name="submit" value="never!">

<input type="hidden" name="__albform__" value="eJzLkl9YfuDO1a06y2cGbxZYUFlbyKgRylZcmpSbWVLI5M0QylyQX1zI7M1QqgcAnRIQMw==">

</form>
</body>
</html>
----------- ^^ -----------


but when I click on the image, I get:
----------- vv -----------
Template traceback (most recent call last):

Traceback (most recent call last):
  File "/usr/lib/python2.1/site-packages/albatross/app.py", line 221, in run
    self.merge_request(ctx)
  File "/usr/lib/python2.1/site-packages/albatross/app.py", line 301, in merge_request
    ctx.merge_request()
  File "/usr/lib/python2.1/site-packages/albatross/context.py", line 279, in merge_request
    text = self.request.field_value('__albform__')
  File "/usr/lib/python2.1/site-packages/albatross/cgiapp.py", line 26, in field_value
    if isinstance(field, list):
TypeError: isinstance() arg 2 must be a class or type
----------- ^^ -----------

Which doesnt look like it's in my code at all. :-/ (or if it is, how
to I get a real traceback or some more useful feedback?)


Infact clicking on the submit button also generates the above error.
Hmmm?

Its probably a problem in my actual code, but it's kinda hard to find
it with the above traceback (which makes it look like an albatross
internal problem). ??


cheese :),
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
TAILFINS!! ... click ...

From korg at darkqueen.org  Tue Jul  1 06:30:28 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 16:30:28 +1000
Subject: [albatross-users] Simple/Class based register?
Message-ID: <200307010630.h616USN16233@firewall.darkqueen.org>

Im sick of forgetting to register my classes. :-/

Can anyone see a problem with doing something like this? 
(Besides speed, which is obviously going to be slower than hardcoding).


----------- 8< also.py >8 -----------
import p

class   page_also1(p.AlbatrossClass): pass
class   page_also2(p.AlbatrossClass): pass
----------- 8< also.py >8 -----------


----------- 8< p.py >8 -----------

class   app:
  def   __init__(s):
    startpage=None
    for name,value in globals().items():
      is_albatross_start= is_albatross= None
      try:
        if value not in [AlbatrossClass, AlbatrossStartClass]:
          is_albatross_start=value.is_albatross_start_class
          is_albatross=      value.is_albatross_class
      except: pass
#     if startpage and is_albatross_start: squid=squid-donut() # error!
      if is_albatross_start:  startpage=name
      if is_albatross:        print 'I should register',name,value
    print 'Start page is:',startpage

class   AlbatrossClass:
  is_albatross_class=1
  is_albatross_start_class=0
class   AlbatrossStartClass(AlbatrossClass):
  is_albatross_start_class=1

class   page1(AlbatrossStartClass): pass
class   page2(AlbatrossClass): pass
class   page3(AlbatrossClass): pass


def     main():
  MyApp=app()

if __name__=="__main__": main()
----------- 8< p.py >8 -----------


Running   python p.py   gives me

----------- 8< output >8 -----------
register page_also1 also.page_also1
register page_also2 also.page_also2
register page3 __main__.page3
register page2 __main__.page2
register page1 __main__.page1
start page is: page1
----------- 8< output >8 -----------


cheers,
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
There is no such thing as fortune.  Try again.

From korg at darkqueen.org  Tue Jul  1 07:31:07 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 17:31:07 +1000
Subject: [albatross-users] Re: Simple/Class based register?
In-Reply-To: Your message of "Tue, 01 Jul 2003 16:30:28 +1000."
Message-ID: <200307010731.h617V7T16859@firewall.darkqueen.org>

"Cameron Blackwood" writes:
  | Im sick of forgetting to register my classes. :-/
  | 
  | Can anyone see a problem with doing something like this? 
  | (Besides speed, which is obviously going to be slower than hardcoding).

I can see a problem with it :)

I cant avoid needing to pass  globals()  to my wrapper library and I
think thats ugly, so ignore me.

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
But like the Good Book says... There's BIGGER DEALS to come!

From korg at darkqueen.org  Tue Jul  1 08:43:07 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 18:43:07 +1000
Subject: [albatross-users] @#*(&$@!! Waste of a day... :(
In-Reply-To: Your message of "Tue, 01 Jul 2003 17:31:07 +1000."
Message-ID: <200307010843.h618h7K17558@firewall.darkqueen.org>


http://www.object-craft.com.au/projects/albatross/albatross/inst-pre.html


Says: 
   "2.1 Prerequisites

      * Python 2.1 or later."

But...


cgiapp.py:26 reads:

        if isinstance(field, list):


Which breaks on list....


Python 2.1.1 (#1, Aug 30 2001, 17:36:05)
[GCC 2.96 20000731 (Mandrake Linux 8.1 2.96-0.61mdk)] on linux-i386
Type "copyright", "credits" or "license" for more information.
>>> type(list)
<type 'builtin_function_or_method'>

because a function isnt the same as:

Python 2.2.2 (#2, Feb  5 2003, 10:40:08)
[GCC 3.2.1 (Mandrake Linux 9.1 3.2.1-5mdk)] on linux-i386
Type "help", "copyright", "credits" or "license" for more information.
>>> type(list)
<type 'list'>


So I assume 1.10pre1 needs atleast 2.2?


Shouldnt setup.py not install albatross for non working versions of
python?  (Especially for it works until you acutally submit a form
type errors).


--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
I just got out of the hospital after a speed reading accident.
I hit a bookmark.
		-- Steven Wright

From matt at pollenation.net  Tue Jul  1 09:23:57 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 01 Jul 2003 10:23:57 +0100
Subject: [albatross-users] Simple/Class based register?
In-Reply-To: <200307010630.h616USN16233@firewall.darkqueen.org>
References: <200307010630.h616USN16233@firewall.darkqueen.org>
Message-ID: <3F01532D.6030401@pollenation.net>

Cameron Blackwood wrote:

>Im sick of forgetting to register my classes. :-/
>
<snip lots of code>

Is this just searching for page classes and registering them 
automatically? If so, have you thought about using the PageModuleMixin 
or, more likely, one of the ModularXxx prepackaged application classes 
instead? With those you do not need to register your classes - Albatross 
will load the module dynamically on first request.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From matt at pollenation.net  Tue Jul  1 09:37:26 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 01 Jul 2003 10:37:26 +0100
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14
 msgs
In-Reply-To: <3F00F6ED.5050009@harvee.org>
References: <200307010109.h6119m713416@firewall.darkqueen.org> <3F00F6ED.5050009@harvee.org>
Message-ID: <3F015656.2040606@pollenation.net>

Hi Eric,

I realise you have posted quite a bit to the list recently (which is 
definitely not a problem!) and got some answers but could you quickly 
and clearly summarize the problems you are still facing with Albatross. 
Perhaps we could then address each issue again and see if we can help you.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From djc at object-craft.com.au  Tue Jul  1 10:14:17 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 20:14:17 +1000
Subject: [albatross-users] Re: Renaming application and execution context classes
In-Reply-To: <200306300717.h5U7Hs105061@firewall.darkqueen.org>
References: <200306300717.h5U7Hs105061@firewall.darkqueen.org>
Message-ID: <m3vfumpn9i.fsf@ferret.object-craft.com.au>

> http://www.object-craft.com.au/projects/albatross/albatross/app-popview3.html
> "Adding Server-Side Session Support to Popview"
> 
> "The new application prologue looks like this:
> 
> #!/usr/bin/python
> from albatross import SimpleSessionApp, SessionAppContext
> "
> 
> and then:
> 
> "    def create_context(self):
>         return AppContext(self)"
> 
> but that should be:
> 
> "    def create_context(self):
>         return SessionAppContext(self)"

Once you look at the code in the samples/popview3 directory it is a
bit clearer what is going on.  It certainly looks like the AppContext
class should be included in the documentation though :-)

I have added the AppContext to the documentation.

Thanks for pointing out the problem.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul  1 10:18:07 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 20:18:07 +1000
Subject: [albatross-users] Re: Renaming application and execution context classes
In-Reply-To: <200306300721.h5U7Lv605110@firewall.darkqueen.org>
References: <200306300721.h5U7Lv605110@firewall.darkqueen.org>
Message-ID: <m3r85apn34.fsf@ferret.object-craft.com.au>

> http://www.object-craft.com.au/projects/albatross/albatross/app-proto-new.html
> 
> looks kinda weird:
>   ? <p>
>   ? <div class="verbatim"><pre>
>   ? "new " appid " " age CRLF
>   ? 
>   ? </pre></div>
>   ? 
>   ? <p>
>   ? Successful response will be a line of the form:
>   ? 
>   ? <p>
>   ? <div class="verbatim"><pre>
>   ? "OK " sesid CRLF
>   ? </pre></div>
>   ? 
> 
> I assume the "'s are ment to be more clear?

A Python representation of the above would look like this:

        'new %s %s\r\n' % (appid, age)

        'OK %s\r\n' % sesid

Is that how you read the text?

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul  1 10:30:26 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 20:30:26 +1000
Subject: [albatross-users] Renaming application and execution context classes
In-Reply-To: <3F007B12.3070107@pollenation.net>
References: <C0FC22C08B82074A88B50061764157775C8A@johnson.mediapulse.net>
 <3F007B12.3070107@pollenation.net>
Message-ID: <m3n0fypmil.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Michael C. Neel wrote:
>> I won't admit how long I used hungarian notation...
>> 
Matt> Shudder. ;-)

When I was learning how to be a Windows programmer in the 3.0 days I
gave hungarian notation a try.  What I found was that I spent way too
much time trying to work out the correct notation for a type that
actually naming the variable.  It never stuck in my mind.  It was much
later that I formed an opinion about the notation.

Funny story:

Around 1990 on one contract I worked on a system of around 200k lines
of Pascal.  It made heavy use of hungarian notation.  People were
constantly complaining that the use of notation and abbreviations was
not consistent.

After a while I became so sick of tripping over the inconsistent
naming that I wrote a program to parse the entire code base and build
a global symbol table.  The program gave you the ability to
interactively browse all of the symbols and show you where they were
used.  You could then globally rename symbols.

After getting the program working I went through and renamed hundreds
of symbols in a matter of hours.  The result was still hungarian
notation, but at least most of the abbreviations were consistent.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul  1 11:07:34 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 21:07:34 +1000
Subject: [albatross-users] Renaming application and execution context classes
In-Reply-To: <C0FC22C08B82074A88B50061764157775C8A@johnson.mediapulse.net>
References: <C0FC22C08B82074A88B50061764157775C8A@johnson.mediapulse.net>
Message-ID: <m3isqmpksp.fsf@ferret.object-craft.com.au>

> Is it possible to combine the classes into one (from the user's
> point of view)?  In an ideal world
> 
> 	app = albatross.App(blah, blah, session_store=albatross.FILE)
> 
> it what I'd like best.
> 
> As for StudyCaps I personally find them much more readable for class
> names, and they point out "hey I'm a class not a function", and use
> lowercase/_'s for var/function names.  This also jives with (most)
> the standard library, and I think it's more important to go along
> with the norm than have your module be different.
> 
> I won't admit how long I used hungarian notation...

Andrew and I thought about something like the above but could not come
up with anything that was not extremely hacky and ugly.

I like the idea of being able to control individual feature sets like
that...

In a sense Albatross is already like that though.  If you look in at
the bottom of albatross/app.py you can hopefully see what I mean.
Take for instance the ModularSessionApp class:

class ModularSessionApp(PickleSignMixin,
                        Application,
                        CachingTemplateLoaderMixin,
                        PageModuleMixin,
                        SessionServerAppMixin):
    def __init__(self, base_url, module_path, template_path, start_page, secret,
                 session_appid, session_server = 'localhost', server_port = 34343, session_age = 1800):
        Application.__init__(self, base_url)
        PickleSignMixin.__init__(self, secret)
        CachingTemplateLoaderMixin.__init__(self, template_path)
        PageModuleMixin.__init__(self, module_path, start_page)
        SessionServerAppMixin.__init__(self, session_appid, session_server, server_port, session_age)

    def create_context(self):
        return SessionAppContext(self)

If we did away with all of the prepackaged classes and changed the
interface to all of the mixins to just accept **kw then the above
could look like this:

class ModularSessionApp(PickleSignMixin,
                        Application,
                        CachingTemplateLoaderMixin,
                        PageModuleMixin,
                        SessionServerAppMixin):
    def __init__(self, **kw):
        Application.__init__(self, **kw)
        PickleSignMixin.__init__(self, **kw)
        CachingTemplateLoaderMixin.__init__(self, **kw)
        PageModuleMixin.__init__(self, **kw)
        SessionServerAppMixin.__init__(self, **kw)

In theory the create_context() method could be removed from the
existing classes.

App Class                       Context Class
SimpleApp                       SimpleAppContext
ModularApp                      SimpleAppContext
RandomModularApp                SimpleAppContext
SimpleSessionApp                SessionAppContext
ModularSessionApp               SessionAppContext
RandomModularSessionApp         SessionAppContext
SimpleSessionFileApp            SessionFileAppContext
ModularSessionFileApp           SessionFileAppContext
RandomModularSessionFileApp     SessionFileAppContext

If we consider SimpleAppContext to be the default execution context
for an application, then all application classes that inherit from
SessionFileAppMixin create a SessionFileAppContext, while all that
inherit from SessionServerAppMixin create SessionAppContext.

Taking the above a little further we could probably allow you to do
something like this:

class MyApp(albatross.AppBase,
            albatross.PickleSignMixin,
            albatross.Application,
            albatross.CachingTemplateLoaderMixin,
            albatross.PageModuleMixin,
            albatross.SessionServerAppMixin):
    pass

Then in Albatross we would define the AppBase class to be something
like this:

class AppBase:
    def __init__(self, **kw):
        for klass in self.__class__.__bases__:
            if klass != AppBase:
                klass.__init__(self, **kw)

All of this is of course untested :-)

One big advantage to all of this is that it allows you to change
application classes very easily.

Going back to yet another suggestion by Cameron Blackwood:

  http://www.object-craft.com.au/pipermail/albatross-users/2002-November/000166.html

This would solve his problem in varying argument lists to application
class constructors.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul  1 11:25:38 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 01 Jul 2003 12:25:38 +0100
Subject: hungarian notation (was Re: [albatross-users] Renaming application
 and execution context classes)
In-Reply-To: <m3n0fypmil.fsf@ferret.object-craft.com.au>
References: <C0FC22C08B82074A88B50061764157775C8A@johnson.mediapulse.net>	<3F007B12.3070107@pollenation.net> <m3n0fypmil.fsf@ferret.object-craft.com.au>
Message-ID: <3F016FB2.3010102@pollenation.net>

Many moons ago, I used to work with OS/2 and its APIs use hungarian 
notation everywhere so we did too. It seemed like a good idea at the 
time but as we got more "used" to it we realised that it caused more 
hastle than benefit.

The amusing thing was that we *all* sat down to come up with our coding 
standards that mandated the use of hungarian notation but over time we 
*individually*, secretly stopped using it in our code. I can't remember 
who admitted that they had stopped using it first but there were sighes 
of relief from everyone :-). I've never used it since.

- Matt

Dave Cole wrote:

>>>>>>"Matt" == Matt Goodall <matt@pollenation.net> writes:
>>>>>>            
>>>>>>
>
>Matt> Michael C. Neel wrote:
>  
>
>>>I won't admit how long I used hungarian notation...
>>>
>>>      
>>>
>Matt> Shudder. ;-)
>
>When I was learning how to be a Windows programmer in the 3.0 days I
>gave hungarian notation a try.  What I found was that I spent way too
>much time trying to work out the correct notation for a type that
>actually naming the variable.  It never stuck in my mind.  It was much
>later that I formed an opinion about the notation.
>
>Funny story:
>
>Around 1990 on one contract I worked on a system of around 200k lines
>of Pascal.  It made heavy use of hungarian notation.  People were
>constantly complaining that the use of notation and abbreviations was
>not consistent.
>
>After a while I became so sick of tripping over the inconsistent
>naming that I wrote a program to parse the entire code base and build
>a global symbol table.  The program gave you the ability to
>interactively browse all of the symbols and show you where they were
>used.  You could then globally rename symbols.
>
>After getting the program working I went through and renamed hundreds
>of symbols in a matter of hours.  The result was still hungarian
>notation, but at least most of the abbreviations were consistent.
>
>- Dave
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From korg at darkqueen.org  Tue Jul  1 11:23:02 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Tue, 01 Jul 2003 21:23:02 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #177 - 13 msgs
In-Reply-To: Your message of "Tue, 01 Jul 2003 20:19:01 +1000."
 <20030701101901.28516.20369.Mailman@numbat.object-craft.com.au>
Message-ID: <200307011123.h61BN2w19204@firewall.darkqueen.org>

From: "Eric S. Johansson" <esj@harvee.org>:
  | 
  | Cameron Blackwood wrote:
  | > 
  | > Im with you there, man. :) 
  | > 
  | > Long descriptive names and '_'s up the whazoo for me :)
  | > 
  | > I want going to comment because I didnt write albatross, so I didnt
  | > feel that I could comment on coding style, but now you mention it...
  | > 
  | > Now, if we cant use generic classes with initiators like:
  | >    Context(storage=Albatross.file)

This would be my preferred solution actually :) 

or even Application(contest=Albatross.store_on_client) 
:)

  | >  2| SimpleAppContext                Context_stored_client
  | >  2| SessionAppContext               Context_stored_server_process
  | >  2| SessionFileAppContext           Context_stored_server_file
  | 
  | I like this series of proposed names because it's something that makes sens
  | e.

*nod* instantly readable! Thats gotta be a Good Thing(tm)

  | 
  | > And the same but longer for the Apps (although again, Id prefer
  | > the app not to need to know how the context was stored, but maybe
  | > thats just me).
  | 
  | there is a couple things about albatross that has plagued me and excessive 
  | visibility of what should be internals is one of them.

*nod* Im agree 100%. Trying to nut out the difference between all the
Application types etc is something I hit again and again.



  | trying to understand the gestalt without much success.  Goodness knows I've
  |  
  | written my fair share of train-by-telepathy documentation so there is no bl

*haha*

  | 
  | one of the gestalt points is communicating information from the python side
  |  to 
  | the HTML side and back.  I've been mucking with it for about a week and I a
  | m no 
  | closer to having a useful knowledge of information transfer issues than I w
  | as 
  | before I started.  The black magic surrounding context.local really isn't c
  | lear 
  | nor is the usage of attributes.  I've been using Python for the better part

I just tend to put everyting in context.local.variable_name
and then  ctx.add_session_vars('variable_name')  and magically it's
useable from my template. Maybe its just laziness on my part, but it
just kinda works.

  | 
  | In any case, this gestalt has driven me to the point of admitting failure. 
  |  I've 
  | surveyed a fair number of template/CGI tools and I think albatross is one o
  | f the 
  | better overall and has the best potential for being small and really effect
  | ive. 

Yeah, I agree. Albatross is so tauntingly almost perfect ;).
(Im not saying I could do better, but there just seem to be a couple
of things that poke me in the eye everytime I think I fully understand
it :)

  |   However, I have been unable to learn enough to do I need to do and I cann
  | ot 
  | ask any more from the community.  You have been more than generous with you
  | r 
  | time  already and for that I am grateful.


Actually, one thing that I think Albatross __REALLY__ __REALLY__ needs
are some simple example/skeleton code fragments. Maybe this should be
a list effort, not something we all wait for object-craft.com.au to
do.

What kind of demo would help you get your stuff done? I was thinking
along the lines of...


1) An demo page that takes input and dumps it to some python code and 
   displays the results in various ways. (Kind of already there, but one
   that tests the various template tags.)
   
2) An example app that lets you:
     Login,
        user,
          edit user.details
          change user.picture
          edit user.story
   
3) A more complex example with a custom al-session-daemon that holds a
   number of 'users' and 'chess games' and (using the newly 
   renamed Context_stored_server_process  :) lets you:
     Login,
        user,
          list/create game
          join game
          play game
            make move
            send message
   (this one is more complex because you have the whole state saving and 
   parallel input problem)

This is top of my head stuff, but if we as a list could come up with a
set of 'these examples would be great', then Id sure be up for trying
to code some of them. (I mean, it'd be a good learning exercise ;)


  | 
  | If you can accept a suggestion, I would suggest creating an very simple 
  | user-friendly layer.  It would reflect the simple nature of CGI environment
  | .  It 
  | would consist of a couple of functions/methods.  The first would be used fo
  | r 
  | generating output.  It would associate a dictionary with an albatross HTML 
  | template.  The keys of the dictionary would map to variable names in the 
  | template.  The matching value to a given key would be substituted into the 
  | variables.  Obviously, nested lists, dictionaries etc. would be permitted s
  | o 
  | that for loops etc. in the templates could be used.
  | 
  | the second function would extract information back from CGI forms.  I haven
  | 't 
  | thought long on the form it would take but I suspect it would also be some 
  | form 
  | of dictionary based structure based on an albatross HTML template.
  | 

Ive been abusing cgi forms for ages :) Im sure we could would work out
a few good examples here. :)

Ive been avoiding templates :) so Im not sure I could do it alone. :-/


  | obviously these ideas aren't fully formed but they are a reflection of my 
  | confidence that that the whole process can be simplified even further than 
  | it 

*nod* I have a hope that it could be simplified, but the objectcraft
people are smart, so maybe it cant.

  | has been to date.  I also believe it should be possible to create a layer o
  | ver 
  | the current albatross API to generate this kind of simple interface.

*nod* although a set of good useable examples would take away the
need, maybe?

cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
$100 invested at 7% interest for 100 years will become $100,000, at
which time it will be worth absolutely nothing.
		-- Lazarus Long, "Time Enough for Love"

From djc at object-craft.com.au  Tue Jul  1 11:26:41 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 21:26:41 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14 msgs
In-Reply-To: <200307010109.h6119m713416@firewall.darkqueen.org>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
Message-ID: <m3el1apjwu.fsf@ferret.object-craft.com.au>

>   | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-generic.html
> 
>         if user:
>             ctx.locals._user = user
>             ctx.add_session_vars('_user')
>                                   ^
>                              remove the '_' ??

The '_' prefix is there to prevent the browser ever being able to
provide _user as an input field in a request.  It effectively protects
the object from external modification.  From:

    http://www.object-craft.com.au/projects/albatross/albatross/tag-input.html

        When merging browser requests the application object places
        the browser supplied value back into the execution context
        value referenced by the name attribute. The application
        request merging will not merge variable names prefixed by
        underscore. Use this to protect application values from
        browser modification.

>   | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-image.html
> 
> Oooooh, nifty!!
> 
> So, having said I feel stoopid for commenting on the docs...  I notice 
> that the page renders as:
> 
>   <al-input type="image" nextpage="m" srcicons/right.gif" whitespace>
>                                       ^
>                                  missing " again.

That is really bizarre.  Another latex2html bug.  It comes up fine in
the PDF documentation.  The actual file used in the document is
doc/doctest/tags-input-image (from the Albatross top level directory).

It looks like this:

    >>> import albatross
    >>> class Ctx(albatross.SimpleContext):
    ...     def input_add(self, *args):
    ...         print args
    ... 
    >>> ctx = Ctx('.')
    >>> albatross.Template(ctx, '<magic>', '''
    ... <al-input type="image" nextpage="m" src="/icons/right.gif" whitespace>
    ... ''').to_html(ctx)
    ('image', 'nextpage,m', None, 0)
    >>> ctx.flush_content()
    <input type="image" src="/icons/right.gif" name="nextpage,m">

The files in the doc/doctest directory are constructed so that we can
run 'make test' in the doc directory and use all of the documentation
fragments as unit tests.

>   | http://www.object-craft.com.au/projects/albatross/albatross/tag-input-file.html
> 
>    "a flag indicating whether or not the list (5.2.2.4) attribute was
>    present."
> 
> is directly followed by:
> 
>    "The request merging allows the user to submit more than one file in a
>    file input field. To simplify application code the Request always
>    returns a list of FileField objects for file inputs"
> 
> So, 'request merging' allows submitting more than one file in a single
> al-imput, but the list attribute allows for multiple al-inputs of the
> same name? ?? 
> 
> It seems unclear.

I blame Andrew :-)  (*ducks*)

>   | On a personal note, I find StudleyCaps quite disturbing.  In 1994 after bei
>   | ng a 
> [...]
>   | again, this is a long-winded way of saying that ordinary_words 
>   | separated_by_the_underscore_character is preferable to me and that I subscr
>   | ibe 
> 
> Im with you there, man. :) 
> 
> Long descriptive names and '_'s up the whazoo for me :)
> 
> I want going to comment because I didnt write albatross, so I didnt
> feel that I could comment on coding style, but now you mention it...

I follow the BDFL.

  http://www.python.org/doc/essays/styleguide.html

        Class Names

        Almost without exception, class names use the CapWords
        convention. Classes for internal use have a leading underscore
        in addition.

Ben (another guy here) says (paraphrased) that when speaking a
(spoken) language you strive to speak like a native, or without an
accent.  The same should apply to programming languages.  If you
follow the BDFL you are less likely to code Python with an accent.

> And the same but longer for the Apps (although again, Id prefer
> the app not to need to know how the context was stored, but maybe
> thats just me).

The application and execution context cooperate to deal with the
session.  The execution context is where the session data is created
manipulated, but the application is usually responsible to loading
and saving the session.  This allows the application to maintain a
long lived connection with a server if necessary.  Remember that an
execution context lives only long enough to process a single browser
request.

Hope that does not make things less clear...

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul  1 11:34:39 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 21:34:39 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14 msgs
In-Reply-To: <3F00F6ED.5050009@harvee.org>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
 <3F00F6ED.5050009@harvee.org>
Message-ID: <m3adbypjjk.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> The documentation looks like it is very good if you have already
Eric> absorbed the gestalt of albatross.  Unfortunately, I've been
Eric> reading the documentation and trying to understand the gestalt
Eric> without much success.  Goodness knows I've written my fair share
Eric> of train-by-telepathy documentation so there is no blame by me
Eric> on that point, just observation.

I am largely responsible for lack of clarity in the documentation.
The criticism above is something I am all too aware of.
train-by-telepathy is a good description :-)

Eric> In any case, this gestalt has driven me to the point of
Eric> admitting failure.  I've surveyed a fair number of template/CGI
Eric> tools and I think albatross is one of the better overall and has
Eric> the best potential for being small and really
Eric> effective. However, I have been unable to learn enough to do I
Eric> need to do and I cannot ask any more from the community.  You
Eric> have been more than generous with your time already and for that
Eric> I am grateful.

It is responses like this that tell us where our weaknesses are.  We
need at more small examples that are carefully explained.  I raised
this with Andrew a week or so ago and asked if he thought we should
remove the popview example.  He said that the popview example was good
because it actually did something real.

Eric> on the documentation front, take a look at CherryPY and pso.
Eric> while I'm not happy with either of them for different reasons,
Eric> they both have good documentation teaching people how to use
Eric> their system.

I will have a look.  Thanks for the tip.

Eric> If you can accept a suggestion, I would suggest creating an very
Eric> simple user-friendly layer.  It would reflect the simple nature
Eric> of CGI environment.  It would consist of a couple of
Eric> functions/methods.  The first would be used for generating
Eric> output.  It would associate a dictionary with an albatross HTML
Eric> template.  The keys of the dictionary would map to variable
Eric> names in the template.  The matching value to a given key would
Eric> be substituted into the variables.  Obviously, nested lists,
Eric> dictionaries etc. would be permitted so that for loops etc. in
Eric> the templates could be used.

Eric> the second function would extract information back from CGI
Eric> forms.  I haven't thought long on the form it would take but I
Eric> suspect it would also be some form of dictionary based structure
Eric> based on an albatross HTML template.

Eric> obviously these ideas aren't fully formed but they are a
Eric> reflection of my confidence that that the whole process can be
Eric> simplified even further than it has been to date.  I also
Eric> believe it should be possible to create a layer over the current
Eric> albatross API to generate this kind of simple interface.

Albatross is supposed to eliminate the need for most of what you
describe above.  Maybe that should be made much clearer in the
documentation (along with some examples showing why).

- Dave

-- 
http://www.object-craft.com.au


From andrewm at object-craft.com.au  Tue Jul  1 11:35:41 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 01 Jul 2003 21:35:41 +1000
Subject: [albatross-users] @#*(&$@!! Waste of a day... :(
In-Reply-To: Message from "Cameron Blackwood" <korg@darkqueen.org>
 of "Tue, 01 Jul 2003 18:43:07 +1000." <200307010843.h618h7K17558@firewall.darkqueen.org>
References: <200307010843.h618h7K17558@firewall.darkqueen.org>
Message-ID: <20030701113541.1BCE13C332@coffee.object-craft.com.au>

>http://www.object-craft.com.au/projects/albatross/albatross/inst-pre.html
>
>Says: 
>   "2.1 Prerequisites
>
>      * Python 2.1 or later."
>
>But...
>
>
>cgiapp.py:26 reads:
>
>        if isinstance(field, list):
>
>
>Which breaks on list....

Damn. And that's not the only one. The booleans True and False are also
missing, and there's also another test failure that I haven't looked
into yet. The fixes are easy enough - I guess that's what a -pre is for
(although I thought we'd tested against 2.1 - 1.5.2 was the target up
until this version). Sorry for wasting your time.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From djc at object-craft.com.au  Tue Jul  1 11:39:27 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 21:39:27 +1000
Subject: [albatross-users] al-input type="image":  internal bug/obscured traceback??
In-Reply-To: <200307010542.h615g5N15836@firewall.darkqueen.org>
References: <200307010542.h615g5N15836@firewall.darkqueen.org>
Message-ID: <m365mmpjbk.fsf@ferret.object-craft.com.au>

> but when I click on the image, I get:
> ----------- vv -----------
> Template traceback (most recent call last):
> 
> Traceback (most recent call last):
>   File "/usr/lib/python2.1/site-packages/albatross/app.py", line 221, in run
>     self.merge_request(ctx)
>   File "/usr/lib/python2.1/site-packages/albatross/app.py", line 301, in merge_request
>     ctx.merge_request()
>   File "/usr/lib/python2.1/site-packages/albatross/context.py", line 279, in merge_request
>     text = self.request.field_value('__albform__')
>   File "/usr/lib/python2.1/site-packages/albatross/cgiapp.py", line 26, in field_value
>     if isinstance(field, list):
> TypeError: isinstance() arg 2 must be a class or type
> ----------- ^^ -----------
> 
> Which doesnt look like it's in my code at all. :-/ (or if it is, how
> to I get a real traceback or some more useful feedback?)
> 
> 
> Infact clicking on the submit button also generates the above error.
> Hmmm?
> 
> Its probably a problem in my actual code, but it's kinda hard to find
> it with the above traceback (which makes it look like an albatross
> internal problem). ??

As you quite rightly point out in a subsequent message this is a
problem in Albatross.  Looks like a 2.2ism sneaked in there.

Andrew and I discussed which level of Python we should target now that
1.5.2 is a distant memory and we settled on 2.1.

We need to fix this for the 1.10pre2 release that I hope to get out
before the end of the week.

If the 1.10pre2 goes well we will rerelease it as 1.10 and announce
it outside of this list.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul  1 11:51:24 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 01 Jul 2003 12:51:24 +0100
Subject: example code (was Re: [albatross-users] Re: Albatross-users digest,
 Vol 1 #177 - 13 msgs)
In-Reply-To: <200307011123.h61BN2w19204@firewall.darkqueen.org>
References: <200307011123.h61BN2w19204@firewall.darkqueen.org>
Message-ID: <3F0175BC.7040102@pollenation.net>

Cameron Blackwood wrote:

>Actually, one thing that I think Albatross __REALLY__ __REALLY__ needs
>are some simple example/skeleton code fragments. Maybe this should be
>a list effort, not something we all wait for object-craft.com.au to
>do.
>
I would be more than happy to help out with this. I also think that 
HOWTOs and an FAQ (built from posts to the mailing list) would be really 
useful. A wiki seems like a good place for this sort of documentation as 
the list memers could then contribute too ... leaving Dave time to 
update the official documentation ;-).

Dave, Andrew - any chance of getting a public wiki on the object craft 
web server?

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From djc at object-craft.com.au  Tue Jul  1 11:46:27 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 21:46:27 +1000
Subject: example code (was Re: [albatross-users] Re: Albatross-users digest, Vol 1 #177 - 13 msgs)
In-Reply-To: <3F0175BC.7040102@pollenation.net>
References: <200307011123.h61BN2w19204@firewall.darkqueen.org>
 <3F0175BC.7040102@pollenation.net>
Message-ID: <m31xxapizw.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Cameron Blackwood wrote:
>> Actually, one thing that I think Albatross __REALLY__ __REALLY__
>> needs are some simple example/skeleton code fragments. Maybe this
>> should be a list effort, not something we all wait for
>> object-craft.com.au to do.
>> 
Matt> I would be more than happy to help out with this. I also think
Matt> that HOWTOs and an FAQ (built from posts to the mailing list)
Matt> would be really useful. A wiki seems like a good place for this
Matt> sort of documentation as the list memers could then contribute
Matt> too ... leaving Dave time to update the official documentation
Matt> ;-).

Matt> Dave, Andrew - any chance of getting a public wiki on the object
Matt> craft web server?

I think that is an excellent idea.  Why didn't we think of that...

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Tue Jul  1 12:34:58 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Tue, 01 Jul 2003 08:34:58 -0400
Subject: example code (was Re: [albatross-users] Re: Albatross-users digest,
 Vol 1 #177 - 13 msgs)
In-Reply-To: <m31xxapizw.fsf@ferret.object-craft.com.au>
References: <200307011123.h61BN2w19204@firewall.darkqueen.org>	<3F0175BC.7040102@pollenation.net> <m31xxapizw.fsf@ferret.object-craft.com.au>
Message-ID: <3F017FF2.5020806@harvee.org>

Dave Cole wrote:

>>>>>>"Matt" == Matt Goodall <matt@pollenation.net> writes:
> 
> 
> Matt> Cameron Blackwood wrote:
> 
>>>Actually, one thing that I think Albatross __REALLY__ __REALLY__
>>>needs are some simple example/skeleton code fragments. Maybe this
>>>should be a list effort, not something we all wait for
>>>object-craft.com.au to do.
>>>
> 
> Matt> I would be more than happy to help out with this. I also think
> Matt> that HOWTOs and an FAQ (built from posts to the mailing list)
> Matt> would be really useful. A wiki seems like a good place for this
> Matt> sort of documentation as the list memers could then contribute
> Matt> too ... leaving Dave time to update the official documentation
> Matt> ;-).
> 
> Matt> Dave, Andrew - any chance of getting a public wiki on the object
> Matt> craft web server?
> 
> I think that is an excellent idea.  Why didn't we think of that...

having acquired a fair degree of scar tissue in this area over the past year, go 
for simple (install, use) and try to find something with a revision history 
capability in case you get the occasional fine witted moron who decides to pee 
in the public pool.

I believe the current version of moinmoin satisfies this requirement.  I've used 
twiki and while it is simple to use, its registration requirements for users, 
configuration, and set up complexity reveal its perlish nature.

---eric



From djc at object-craft.com.au  Tue Jul  1 13:22:52 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 01 Jul 2003 23:22:52 +1000
Subject: example code (was Re: [albatross-users] Re: Albatross-users digest, Vol 1 #177 - 13 msgs)
In-Reply-To: <3F017FF2.5020806@harvee.org>
References: <200307011123.h61BN2w19204@firewall.darkqueen.org>
 <3F0175BC.7040102@pollenation.net>
 <m31xxapizw.fsf@ferret.object-craft.com.au>
 <3F017FF2.5020806@harvee.org>
Message-ID: <m31xxanzyr.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

>> I think that is an excellent idea.  Why didn't we think of that...

Eric> having acquired a fair degree of scar tissue in this area over
Eric> the past year, go for simple (install, use) and try to find
Eric> something with a revision history capability in case you get the
Eric> occasional fine witted moron who decides to pee in the public
Eric> pool.

Eric> I believe the current version of moinmoin satisfies this
Eric> requirement.  I've used twiki and while it is simple to use, its
Eric> registration requirements for users, configuration, and set up
Eric> complexity reveal its perlish nature.

OK.  I have a pristine wiki up at the following URL:

        http://www.object-craft.com.au/projects/albatross/wiki

Go for it.  I suppose the best thing that Andrew and I can do is stand
well back.  Maybe we should restrict ourselves to adding clarification
and corrections when they are necessary.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul  1 15:03:41 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 01 Jul 2003 16:03:41 +0100
Subject: example code (was Re: [albatross-users] Re: Albatross-users digest,
 Vol 1 #177 - 13 msgs)
In-Reply-To: <m31xxanzyr.fsf@ferret.object-craft.com.au>
References: <200307011123.h61BN2w19204@firewall.darkqueen.org>	<3F0175BC.7040102@pollenation.net>	<m31xxapizw.fsf@ferret.object-craft.com.au>	<3F017FF2.5020806@harvee.org> <m31xxanzyr.fsf@ferret.object-craft.com.au>
Message-ID: <3F01A2CD.4040402@pollenation.net>

Do we want to come up with a top-level "menu"? Perhaps HowTo, 
FrequentlyAskedQuestions, CaseStudies, Comments, Gotchas aka 
CommonMistakes, ExampleCode, HintsAndTips etc. Perhaps someone knows a 
similar wiki that they find well structured that we could use as a model?

... or shall we just see what happens?

Cheers, Matt

Dave Cole wrote:

>OK.  I have a pristine wiki up at the following URL:
>
>        http://www.object-craft.com.au/projects/albatross/wiki
>
>Go for it.  I suppose the best thing that Andrew and I can do is stand
>well back.  Maybe we should restrict ourselves to adding clarification
>and corrections when they are necessary.
>
>- Dave
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From tchur at optushome.com.au  Tue Jul  1 15:41:05 2003
From: tchur at optushome.com.au (Tim Churches)
Date: 02 Jul 2003 01:41:05 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14
 msgs
In-Reply-To: <m3adbypjjk.fsf@ferret.object-craft.com.au>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
 <3F00F6ED.5050009@harvee.org>  <m3adbypjjk.fsf@ferret.object-craft.com.au>
Message-ID: <1057074065.1169.12.camel@emilio>

--=-pqyBVPd5zMmGwy36rsLh
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Tue, 2003-07-01 at 21:34, Dave Cole wrote:
> It is responses like this that tell us where our weaknesses are.  We
> need at more small examples that are carefully explained.  I raised
> this with Andrew a week or so ago and asked if he thought we should
> remove the popview example.  He said that the popview example was good
> because it actually did something real.

I think that Albatross needs an explicit pattern library, in the
documentation if possible (or perhaps patterns can be migrataed from the
wiki into the documentation with each new release) - like the Python
Cookbook edited by Alex Martelli, which is very, very useful. The manual
is a bit too phenomenological at the moment: here is a tag, here is a
context, here is a hippopotamus - but not enough examples of how they
relate to one another when solving real-life problems. Such examples are
essential for developing the gestalt to which Eric refers.

These patterns need to go from the basic through to the more complex.

I also think that things like popview are useful in order to demonstrate
what's possible to people who are assessing Albatross - but that these
examples should (also) be hosted directly on an Object Craft Web site so
they don't need to install Albatross to try them out.

--=20

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint =3D 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0



--=-pqyBVPd5zMmGwy36rsLh
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQA/AauReJFGqer5k9ARAnYdAKCNPM0sq7But3UkXBriYcikH7LRdgCeM2+T
Rjn+AnSxJuHXJL7XGGdzvus=
=GXpm
-----END PGP SIGNATURE-----

--=-pqyBVPd5zMmGwy36rsLh--


From neel at mediapulse.com  Tue Jul  1 15:44:24 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Tue, 1 Jul 2003 11:44:24 -0400
Subject: [albatross-users] @#*(&$@!! Waste of a day... :(
Message-ID: <C0FC22C08B82074A88B50061764157776B9592@johnson.mediapulse.net>

As a tip, I've always done:

if type([]) =3D=3D type(mylist)

to avoid these version change problems.  not sure if that would help in
this case.

Mike


-------------------
Python 2.1.1 (#1, Aug 30 2001, 17:36:05)
[GCC 2.96 20000731 (Mandrake Linux 8.1 2.96-0.61mdk)] on linux-i386
Type "copyright", "credits" or "license" for more information.
>>> type(list)
<type 'builtin_function_or_method'>

because a function isnt the same as:

Python 2.2.2 (#2, Feb  5 2003, 10:40:08)
[GCC 3.2.1 (Mandrake Linux 9.1 3.2.1-5mdk)] on linux-i386
Type "help", "copyright", "credits" or "license" for more information.
>>> type(list)
<type 'list'>



From tchur at optushome.com.au  Tue Jul  1 15:45:28 2003
From: tchur at optushome.com.au (Tim Churches)
Date: 02 Jul 2003 01:45:28 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14
 msgs
In-Reply-To: <m3el1apjwu.fsf@ferret.object-craft.com.au>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
 <m3el1apjwu.fsf@ferret.object-craft.com.au>
Message-ID: <1057074328.1169.16.camel@emilio>

--=-ENvrPnEy05Cg7DJtKpMo
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Tue, 2003-07-01 at 21:26, Dave Cole wrote:

> Ben (another guy here) says (paraphrased) that when speaking a
> (spoken) language you strive to speak like a native, or without an
> accent.  The same should apply to programming languages.  If you
> follow the BDFL you are less likely to code Python with an accent.

Or rather, to code Python with the intended Dutch accent...

> The application and execution context cooperate to deal with the
> session.  The execution context is where the session data is created
> manipulated, but the application is usually responsible to loading
> and saving the session.  This allows the application to maintain a
> long lived connection with a server if necessary.  Remember that an
> execution context lives only long enough to process a single browser
> request.

That para should be near the front of the manual if it isn't already.

--=20

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint =3D 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0



--=-ENvrPnEy05Cg7DJtKpMo
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQA/AayYeJFGqer5k9ARArNRAJ98VQkv2ziNzRCYgfuchfdwvgWI4gCfSKqw
jc37IlGaZdqngU/wkpYaGJI=
=ek4z
-----END PGP SIGNATURE-----

--=-ENvrPnEy05Cg7DJtKpMo--


From neel at mediapulse.com  Tue Jul  1 16:13:06 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Tue, 1 Jul 2003 12:13:06 -0400
Subject: [albatross-users] Renaming application and execution context classes
Message-ID: <C0FC22C08B82074A88B50061764157775C8F@johnson.mediapulse.net>

I think there is something too this.  Add to what you have below a top
level albatross function that returns the **kw you need, taking one
argument and doing a preset of classes.  So you could write:

app =3D albatross.App(albatross.AppBase,
		albatross.PickleSignMixin,
            albatross.Application,
            albatross.CachingTemplateLoaderMixin,
            albatross.PageModuleMixin,
            albatross.SessionServerAppMixin)

- or -

app =3D albatross.App(albatross.init_app(albatross.SERVER))

- and even could do -

initapp =3D albatross.init_app(albatross.SERVER)
initapp.pagemodule =3D RandomPageModuleMixin

app =3D =3D albatross.App(initapp)

and yes, this comes from my win32 days =3Dp  long live h_wnd!

Thanks,
Mike

-----Original Message-----
From: Dave Cole [mailto:djc@object-craft.com.au]=20
Sent: Tuesday, July 01, 2003 7:08 AM
To: Michael C. Neel
Cc: albatross-users; korg@darkqueen.org
Subject: Re: [albatross-users] Renaming application and execution
context classes



> Is it possible to combine the classes into one (from the user's
> point of view)?  In an ideal world
>=20
> 	app =3D albatross.App(blah, blah, session_store=3Dalbatross.FILE)
>=20
> it what I'd like best.
>=20
> As for StudyCaps I personally find them much more readable for class
> names, and they point out "hey I'm a class not a function", and use
> lowercase/_'s for var/function names.  This also jives with (most)
> the standard library, and I think it's more important to go along
> with the norm than have your module be different.
>=20
> I won't admit how long I used hungarian notation...

Andrew and I thought about something like the above but could not come
up with anything that was not extremely hacky and ugly.

I like the idea of being able to control individual feature sets like
that...

In a sense Albatross is already like that though.  If you look in at
the bottom of albatross/app.py you can hopefully see what I mean.
Take for instance the ModularSessionApp class:

class ModularSessionApp(PickleSignMixin,
                        Application,
                        CachingTemplateLoaderMixin,
                        PageModuleMixin,
                        SessionServerAppMixin):
    def __init__(self, base_url, module_path, template_path, start_page,
secret,
                 session_appid, session_server =3D 'localhost',
server_port =3D 34343, session_age =3D 1800):
        Application.__init__(self, base_url)
        PickleSignMixin.__init__(self, secret)
        CachingTemplateLoaderMixin.__init__(self, template_path)
        PageModuleMixin.__init__(self, module_path, start_page)
        SessionServerAppMixin.__init__(self, session_appid,
session_server, server_port, session_age)

    def create_context(self):
        return SessionAppContext(self)

If we did away with all of the prepackaged classes and changed the
interface to all of the mixins to just accept **kw then the above
could look like this:

class ModularSessionApp(PickleSignMixin,
                        Application,
                        CachingTemplateLoaderMixin,
                        PageModuleMixin,
                        SessionServerAppMixin):
    def __init__(self, **kw):
        Application.__init__(self, **kw)
        PickleSignMixin.__init__(self, **kw)
        CachingTemplateLoaderMixin.__init__(self, **kw)
        PageModuleMixin.__init__(self, **kw)
        SessionServerAppMixin.__init__(self, **kw)

In theory the create_context() method could be removed from the
existing classes.

App Class                       Context Class
SimpleApp                       SimpleAppContext
ModularApp                      SimpleAppContext
RandomModularApp                SimpleAppContext
SimpleSessionApp                SessionAppContext
ModularSessionApp               SessionAppContext
RandomModularSessionApp         SessionAppContext
SimpleSessionFileApp            SessionFileAppContext
ModularSessionFileApp           SessionFileAppContext
RandomModularSessionFileApp     SessionFileAppContext

If we consider SimpleAppContext to be the default execution context
for an application, then all application classes that inherit from
SessionFileAppMixin create a SessionFileAppContext, while all that
inherit from SessionServerAppMixin create SessionAppContext.

Taking the above a little further we could probably allow you to do
something like this:

class MyApp(albatross.AppBase,
            albatross.PickleSignMixin,
            albatross.Application,
            albatross.CachingTemplateLoaderMixin,
            albatross.PageModuleMixin,
            albatross.SessionServerAppMixin):
    pass

Then in Albatross we would define the AppBase class to be something
like this:

class AppBase:
    def __init__(self, **kw):
        for klass in self.__class__.__bases__:
            if klass !=3D AppBase:
                klass.__init__(self, **kw)

All of this is of course untested :-)

One big advantage to all of this is that it allows you to change
application classes very easily.

Going back to yet another suggestion by Cameron Blackwood:

=20
http://www.object-craft.com.au/pipermail/albatross-users/2002-November/0
00166.html

This would solve his problem in varying argument lists to application
class constructors.

- Dave

--=20
http://www.object-craft.com.au


From neel at mediapulse.com  Tue Jul  1 17:29:38 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Tue, 1 Jul 2003 13:29:38 -0400
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #177 - 13 msgs
Message-ID: <C0FC22C08B82074A88B50061764157775C90@johnson.mediapulse.net>

---------
Actually, one thing that I think Albatross __REALLY__ __REALLY__ needs
are some simple example/skeleton code fragments. Maybe this should be
a list effort, not something we all wait for object-craft.com.au to
do.
---------

I agree here, we can probably do this and keep off some burden.  I have
used albtross in 3 major live deployed appliactions now, and a few
others that are more light weight.  I even have our Sys Adm setup with
some default files so when I say to setup a new site with Albatross,
these are generic enough to get the site running and the templates being
parsed so that the designers can begin to place in the design and pages.

I'll try to get some time to make a mod_python/Albatross example, and
post here when I have more.

Mike


From esj at harvee.org  Tue Jul  1 19:40:58 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Tue, 01 Jul 2003 15:40:58 -0400
Subject: [albatross-users] cognitive speed bumps
Message-ID: <3F01E3CA.5030404@harvee.org>

Since a few folks have urged me to try and document where I missed with 
albatross, I figured I would start with my mental model of this process and go 
from there.  Hopefully it will not be much of a mental goat trail.  I also 
apologize for any speech recognition airs I've missed.  It's sometimes quite 
difficult to catch what's wrong with something you've said by reading.

Internally, I've partitioned the problem of Web programming into a few events:

    * initial setup
    * page creation based on "current known values"
    * retrieval of form data and translation into a set of "current known values"

In the traditional CGI environment, the order seems to be retrieval of form 
data, detection of initial conditions leading to initial set up, then creation 
of "current known values" followed by page creation and display.  It's not very 
intuitive but it is logical once you understand the underlying mechanisms of the 
CGI environment.

However when you introduce a tool to separate presentation information from 
generation of data, that model changes.  My expectation has been somewhat to 
very different from what the various tools smiths have generated.  This is no 
surprise but merely reflects the wide range of mental models folks have for this 
process.

In my expectation, I was picturing templates as named entities which could be 
operated upon by some engine.  The engine would be supplied a template and a set 
of named data elements.  The engine would then apply the value for each named 
data element to each occurrence of the name within the template.  The end result 
of the engines work would be copied/transferred to either an aggregation object 
or storage until all of the components were operated on and ready for display.

Originally I envisioned that the presentation control flow would be part of an 
object encapsulating their representation of a web page.  Unfortunately, that 
becomes rather messy and I can see the wisdom of how albatross put the 
presentation control flow into the HTML template itself.

Data representation on both sides of the fence (CGI, template) should be normal 
python data types such as strings, integers, floats and, objects.  Objects are 
especially important because it may be convenient for an application to 
encapsulate all the data either as methods or instance variables.  In any case, 
it should be possible to run the object and retrieve the appropriate data.

on retrieving data from forms, I think of the data as being logically associated 
by name or by type.  But what should be most important is that *all* of the data 
should be available and that user filters can be applied to reduce or refine the 
data.

the only improvement one could make over the dictionary method of presenting 
data ala cgi.fieldstorage would be if one could logically group the data in the 
form and have a standard tool/method to retrieve just that named set of data and 
place it in a dictionary like structure.

Yes, I like dictionaries and I like even better the ability to use 
__get/setitem__ and related methods to create data transformations.

Now, having explained (hopefully) where my mental model was at, I banged my 
shins rather badly on the attribute based model of albatross.  One failing of 
the documentation is a complete description of the data model used for bridging 
between CGI and template.  When it comes right down to it it doesn't matter a 
piss hole in the snow what model you use as long as it is flexible and 
documented in a way a novice can understand.  Once you know what you have to 
connect to, you can create impedance matching code to translate between your 
object models and the template system object models.

One test for how good the system is comes from the ability to wiggle every 
little widget or knob through a variety of techniques ranging from static data 
assignment through more dynamic data sources such as dictionaries or method 
generated data sets without requiring the user to learn more than a few 
interfaces.  One could argue that the bridge interface between data and 
templates should be the same for all levels of data complexity.

another place I tripped on regarding albatross was the discovery that I could 
not get all of the checkbox data (on and off) back from albatross environment 
without going through a significant amount of convolution which I was willing to 
do but was unsuccessful at getting it to work.

minor points were some of the techniques for determining if a button had been 
pressed or the value of a field by dynamic name.  I'm still not sure the best 
way to handle hidden variables or should I use state preservation techniques? 
I'm sure there is a whole list of techniques we could document for various 
frequently needed circumstances but the question one should always consider is 
that if you are building a large list of tricks, are you solving the problem the 
wrong way?  Good design doesn't need tricks to do the right job.  It should 
follow naturally from the architecture.  It may be necessary to train everyone 
on the new architecture especially if it is based on a very different set of 
techniques from the normal world of experience.  But it should be simple to 
convey and easy to learn enough to get started quickly and still support the 
developer ability to master more complex techniques as the complexity of the 
task rises.

---eric







From matt at pollenation.net  Tue Jul  1 20:01:53 2003
From: matt at pollenation.net (Matt Goodall)
Date: 01 Jul 2003 21:01:53 +0100
Subject: [albatross-users] A great use for the simple http server
Message-ID: <1057089713.26359.13.camel@localhost>

Hi,

I'm not trying to blow my own trumpet, but the simple http server I
wrote and posted in this list would make running the sample apps *much*
easier to run.

For instance, running the popview5 example would be a matter of changing
to the directory and running "al-httd popview.App 8000". No installation
necessary and no need to change base_url, template_path etc.

Cheers, Matt
-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Tue Jul  1 20:55:55 2003
From: matt at pollenation.net (Matt Goodall)
Date: 01 Jul 2003 21:55:55 +0100
Subject: [albatross-users] First post!
Message-ID: <1057092954.32426.54.camel@localhost>

Oh, this isn't /. then? ;-)

I suggested the Wiki so it's only fair that I get the ball rolling.
There is now a hello world example on the wiki. Yeah I know, it's a
completely useless application but it's a start and it's something to
cut and paste to get going with.

One thing that I think could be useful is to add to the
RequestsForContent page. I tend to lose things in the noise of the
mailing list so that page could be a todo list for people to pick items
from later when they have more time or more knowledge.

Dave/Andrew - could you enable the "Mail me my account data" option in
MoinMoin please so that I can get my id wherever I am. Thanks.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net
t: 0113 2252500

From esj at harvee.org  Tue Jul  1 21:34:20 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Tue, 01 Jul 2003 17:34:20 -0400
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14
 msgs
In-Reply-To: <3F015656.2040606@pollenation.net>
References: <200307010109.h6119m713416@firewall.darkqueen.org> <3F00F6ED.5050009@harvee.org> <3F015656.2040606@pollenation.net>
Message-ID: <3F01FE5C.4040400@harvee.org>

Matt Goodall wrote:

> Hi Eric,
> 
> I realise you have posted quite a bit to the list recently (which is 
> definitely not a problem!) and got some answers but could you quickly 
> and clearly summarize the problems you are still facing with Albatross. 
> Perhaps we could then address each issue again and see if we can help you.
> 
> Cheers, Matt
> 
thank you Matt, if the rest of you folks can take another long-winded 
description, here's a big chunk of the problem I'm trying to code up.

If you go to:
harvee.org/cgi-bin/correct.cgi

Login with username: albatross password: albatross

you'll see what I'm trying reimplement.  The reason for the reimplementation is 
that I will be doing something similar with the spam dumpster and I also need to 
create something totally unrelated which is the user configuration editing tool.

by the way, don't worry about pushing on the buttons because I've filtered out 
any real action from albatross.

The vast majority of the web page is quite simple.  The "refresh display" link 
just refers to the CGI so that you can refresh the list of currently trapped 
spam below.  The number of messages contained in the spamtrap is obviously 
dynamically created at p. construction time.  The button "spam on top/bottom" is 
simultaneously state information and user information.  I set the value of the 
button to display the sort order for the list.  One of my first attempts at 
converting to albatross forced me to change the message and just set the text to 
spam on top/bottom and have the button become "click to reverse sorting".  I 
then used a hidden variable to store the state sort of direction.

I would like to figure out how to store the sort state direction information so 
that it would persist even though someone refreshes the display.

On the top and bottom of the spamtrap display are to train buttons.  We've gone 
around the barn a few times on the vocabulary for the buttons and I'm probably 
going to make it "process spamtrap" although there is a vocal minority that 
prefers "lather, rinse, repeat".

Now we come to the messy part, the spamtrap display.  This is what prompted my 
question about displaying two-dimensional data.  Internally, I need to look at 
all messages in the spamtrap, read the messages, extract all fields and correct 
for errors created by spammers, sort the list, truncate to the display size, and 
build the display list.

The easy forms of the display list are either tuples of all the display elements 
or dictionaries of the same.  I've implemented an object model version but it's 
not significantly better than the straight iterative form.

as you can see, each line consists of:

    * a checkbox which is checked or unchecked automatically
    * a background color set according to score
    * destination mailbox
    * from address
    * subject

the subject field is also a link which takes you to another CGI which displays 
the contents of the message.  This is useful when trying to evaluate whether or 
not a message truly is spam.

You will also notice that as you check and uncheck each line, the color changes 
from yellow to red and back again.

the last piece is the periodic refresh.  it's a JavaScript clock that counts 
down and reloads the page.  It's not perfect because tries to repost data from 
the last time a button was pressed.  Obviously, I need to figure out what's the 
right way to do the refresh.  Also, I would like to make the countdown time 
dynamic based on the users configuration at some point.

as I said in my other piece, what has consistently tripped me up is the 
difference between the Python data model and the albatross data model.  Bridging 
the gap is essential in enabling new users to be successful with albatross.

---eric



From tchur at optushome.com.au  Tue Jul  1 21:55:15 2003
From: tchur at optushome.com.au (Tim Churches)
Date: 02 Jul 2003 07:55:15 +1000
Subject: [albatross-users] A great use for the simple http server
In-Reply-To: <1057089713.26359.13.camel@localhost>
References: <1057089713.26359.13.camel@localhost>
Message-ID: <1057096515.1169.73.camel@emilio>

--=-EJn+ORrGrvzYlipqR4Xq
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Wed, 2003-07-02 at 06:01, Matt Goodall wrote:
> Hi,
>=20
> I'm not trying to blow my own trumpet, but the simple http server I
> wrote and posted in this list would make running the sample apps *much*
> easier to run.
>=20
> For instance, running the popview5 example would be a matter of changing
> to the directory and running "al-httd popview.App 8000". No installation
> necessary and no need to change base_url, template_path etc.

I agree - and ease of trying out examples is paramount to hook new
users.


--=20

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint =3D 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0



--=-EJn+ORrGrvzYlipqR4Xq
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQA/AgNDeJFGqer5k9ARAhwyAJ96JEn3nevEATpu8OF6ZQ+KXN37nwCfa6xs
jOxv6x+u3PNkmc342w5sxVM=
=4ZL1
-----END PGP SIGNATURE-----

--=-EJn+ORrGrvzYlipqR4Xq--


From esj at harvee.org  Tue Jul  1 22:45:30 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Tue, 01 Jul 2003 18:45:30 -0400
Subject: [albatross-users] impedance mismatch 1
Message-ID: <3F020F0A.9010907@harvee.org>

I'm reading the documentation and here is my commentary based on where I feel I 
get that slack-jawed look on my face.

http://www.object-craft.com.au/projects/albatross/albatross/tug-simple3.html

section 3.3 "introducing albatross tags"

after the body of code shown, there is a paragraph starting with "the following 
lines construct a sorted list..."

The description of locals leaves me puzzled.  Its behavior is counter to 
everything I have learned in Python to date.  It's this magic data lint trap 
that doesn't behave like a class, a dictionary, a list or anything else I'm 
familiar with.

But I guess I can sort of deal with the dissonance by just assuming that 
context.locals can magically absorb data and associate it with a statically 
supplied name.

quite frankly, most of the discussion about local and global name spaces just 
goes in one ear and out the other because it just doesn't seem important to the 
problem on hand of figuring out how to transfer data from the local context to 
the template context.

Now comes the templates and I get really confused.

expr=name.value()

brings an exclamation of wth?? to my lips.  Why isn't it just name?  what 
happened to the ctx.locals?  Eventually I just shrug and figure it's for some 
hopefully documented reason.  But the next line really brings wrinkles to my brow

expr="environ[name.value()]"

after all if name needs a .value() function why isn't it

expr="environ[name.value()].value()"

seriously grump inspiring.

I think the rest of the explanation with separating out presentation logic and 
the use of the al-exec tag is pretty good.  the only rough spot is comes with 
the description of the "locals" data lint trap.

---eric




From matt at pollenation.net  Wed Jul  2 00:18:57 2003
From: matt at pollenation.net (Matt Goodall)
Date: 02 Jul 2003 01:18:57 +0100
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <3F020F0A.9010907@harvee.org>
References: <3F020F0A.9010907@harvee.org>
Message-ID: <1057105136.30064.125.camel@localhost>

Oh boy, I hope I get this right. I **really** hope it helps and doesn't
just confuse you more. See the comments below.

Cheers, Matt

On Tue, 2003-07-01 at 23:45, Eric S. Johansson wrote:
> The description of locals leaves me puzzled.  Its behavior is counter to 
> everything I have learned in Python to date.  It's this magic data lint trap 
> that doesn't behave like a class, a dictionary, a list or anything else I'm 
> familiar with.

OK, I think I understand what you mean about locals now and where your
confusion comes from. locals is simply an empty class definition:

    class Vars:
        pass
    ...
    ctx.locals = Vars()

Python allows you to set any attribute on any object at any time:

    v = Vars()
    v.my_name = 'Matt'

This is perfectly valid and fairly normal in Python. After all, you
don't have to define instance attributes, you just say "self.my_name =
'Matt'".

This reminds me that I have a note in a text file somewhere that says
something like, "could/should ctx.locals behave like a dictionary?".

I think ctx.locals could be made to behave like a dictionary. An initial
(untested) implementation of Vars to support this might be:

    class Vars:
        def __setitem__(self, name, value):
            setattr(self, name, value)
        def __getitem__(self, name):
            return getattr(self, name)
        etc

That means you can treat a Vars instance as an object or a dictionary,
depending on your preference:

    >>> v = Vars()
    >>> v.a = 'a'
    >>> v['b'] = 'b'
    >>> vars(v)
    {'a': 'a', 'b': 'b'}
    >>> v.a
    'a'
    >>> v['a']
    'a'
    >>> v.b
    'b'
    >>> v['b']
    'b'

There are other methods that would need implementing for Vars to be
truly dictionary-like but I'm sure you get the idea.

Eric, would making ctx.locals behave like a dictionary help your
understanding? Dave/Andrew, are there any technical reasons why this
would not work?

> 
> But I guess I can sort of deal with the dissonance by just assuming that 
> context.locals can magically absorb data and associate it with a statically 
> supplied name.

Hopefully, it's not so magic now ;-).

> 
> quite frankly, most of the discussion about local and global name spaces just 
> goes in one ear and out the other because it just doesn't seem important to the 
> problem on hand of figuring out how to transfer data from the local context to 
> the template context.

I must admit, I need to find out about the global name space!

> 
> Now comes the templates and I get really confused.
> 
> expr=name.value()
> 
> brings an exclamation of wth?? to my lips.  Why isn't it just name?  what 
> happened to the ctx.locals?  Eventually I just shrug and figure it's for some 
> hopefully documented reason.  But the next line really brings wrinkles to my brow
> 
> expr="environ[name.value()]"
> 
> after all if name needs a .value() function why isn't it
> 
> expr="environ[name.value()].value()"
> 
> seriously grump inspiring.

OK, here goes ...

What happened to the ctx.locals
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In a template the scope *is* ctx.locals, i.e. any reference to a
variable called 'my_name' actually means ctx.locals.my_name. You publish
object for the template by storing them in ctx.locals. Earlier, you
talked about the need to "associate a dictionary with an albatross HTML
template"; that's what ctx.locals does.

expr="name.value()" ... Why isn't it just name?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Here's the important bit in the documentation from section 5.3.7,
http://www.object-craft.com.au/projects/albatross/albatross/tag-for.html.

        "The tag uses an instance of the ListIterator identified in the
        local namespace by the iter attribute to iterate over the
        sequence defined by the expression in the expr attribute."

If you look at ListIterator's methods it will probably make more sense
now. It is ListIterator that makes the paging possible and the index()
can be used for things like colouring alternate rows, i.e.
"name.index()%2 and 'red' or 'green'".

By the way, Dave has only just updated this bit of documentation based
on feedback on the list so don't feel bad about missing the importance
of ListIterator. It was documented but it was not as clear as it could
have been.

why isn't it environ[name.value()].value()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
So, why isn't the .value() needed on the end? Hopefully you now
understand that:

      * "name" is a ListIterator created by the <al-for> tag to iterate
        over the items in "keys"
      * "name.value()" is the current item from "keys"
      * "environ" when inside the template refers to the same dictionary
        as "ctx.locals.environ" in the python code.

So all "environ[name.value()]" is actually doing is getting the value of
the item in environ that has a key name.value(). The <al-value> tag has
no need to create anything like the ListIterator as all it does is print
some text to the page.

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From gnb at itga.com.au  Wed Jul  2 00:44:13 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Wed, 02 Jul 2003 10:44:13 +1000
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: Your message of 02 Jul 2003 01:18:57 +0100.
Message-ID: <200307020044.KAA27327@lightning.itga.com.au>

One wrinkle on all this:

>       * "environ" when inside the template refers to the same dictionary
>         as "ctx.locals.environ" in the python code.

... and if ctx.locals.environ doesn't exist, then it looks for environ in the
global context of the _page display function_.  So you can do something like:

----
my_heading = 'Super Application'

class StartPage:
	# ...
	def page_display(self, ctx):
		ctx.run_template('start.ht')
----
then, inside start.ht, something like
	<h1><al-value expr="my_heading"/></h1>
and it will work.  

But the trick is it uses the global namespace of the function that calls
run_template(). If you use page modules, it has to be in the page module.  If
you do what I do and have a library with a base class for Albatross pages that
has a run_template function, then you will get the global namespace of the
library module, not the application, unless you override the page_display in
your concrete page class.

(Which is probably as clear as mud....)




From matt at pollenation.net  Wed Jul  2 00:54:35 2003
From: matt at pollenation.net (Matt Goodall)
Date: 02 Jul 2003 01:54:35 +0100
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <200307020044.KAA27327@lightning.itga.com.au>
References: <200307020044.KAA27327@lightning.itga.com.au>
Message-ID: <1057107274.11964.2.camel@localhost>

Thanks for that correction! Now I understand globals as well :)

On Wed, 2003-07-02 at 01:44, Gregory Bond wrote:
> One wrinkle on all this:
> 
> >       * "environ" when inside the template refers to the same dictionary
> >         as "ctx.locals.environ" in the python code.
> 
> ... and if ctx.locals.environ doesn't exist, then it looks for environ in the
> global context of the _page display function_.  So you can do something like:
> 
> ----
> my_heading = 'Super Application'
> 
> class StartPage:
> 	# ...
> 	def page_display(self, ctx):
> 		ctx.run_template('start.ht')
> ----
> then, inside start.ht, something like
> 	<h1><al-value expr="my_heading"/></h1>
> and it will work.  
> 
> But the trick is it uses the global namespace of the function that calls
> run_template(). If you use page modules, it has to be in the page module.  If
> you do what I do and have a library with a base class for Albatross pages that
> has a run_template function, then you will get the global namespace of the
> library module, not the application, unless you override the page_display in
> your concrete page class.
> 
> (Which is probably as clear as mud....)
> 
> 
> 
> _______________________________________________
> Albatross-users mailing list
> Albatross-users@object-craft.com.au
> https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users
-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From sheila at thinkspot.net  Wed Jul  2 03:57:46 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 01 Jul 2003 20:57:46 -0700
Subject: [albatross-users] multiple session support
Message-ID: <50112117.1057093066@SHEILA-LAPTOP>

Hello,

I got a question today from a team member, and I think the answer is "yes"
but I just wanted to verify that with the list.

The question was:

"Multiple sessions/browser windows open is something to keep in mind. Does
the Albatross framework allow multiple sessions for 1 account / from 1
computer?"

At this point, I'm wondering exactly what he means by this?

I can't find anything about multiple sessions in the albatross
documentation and the closest I found in the mailing list archives is this:
<http://object-craft.com.au/pipermail/albatross-users/2003-April/000340.htm
l>

which isn't really exactly what he is asking about, but I think it touches
on it.

Does anyone have an idea what the purpose or desireability of multiple
sessions for one account from one computer would be? And would it be
possible to implement in Albatross?

(The guy who asked is mostly a PHP and Perl programmer, so I'm not sure if
that has anything to do with where he's coming from or not?)

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org




From esj at harvee.org  Wed Jul  2 12:50:56 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Wed, 02 Jul 2003 08:50:56 -0400
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <1057107274.11964.2.camel@localhost>
References: <200307020044.KAA27327@lightning.itga.com.au> <1057107274.11964.2.camel@localhost>
Message-ID: <3F02D530.6060503@harvee.org>

Matt Goodall wrote:
...
> On Wed, 2003-07-02 at 01:44, Gregory Bond wrote:
 >..

fantastic explanations!  This brings to mind a few things:

1) we now have the start of a wiki section called "puzzled about?"

I'll try to take these explanations and put them into wiki pages so you can see 
what I mean.

2) I think the documentation should have a section called "fundamental concepts" 
in which it illustrates some of these fundamental bits of knowledge necessary to 
understand the environment.  One picture, forming in my head, which may or may 
not be useful to anyone would show the relationship between python code space, 
context.locals, a template, and the final result.


python code    template   end result
   .
   .
   .

context.locals.xyzzy = "plugh"
         |
         +==> <b><al-value expr="xyzzy"> </b>
                  |
                  +===> <b>plugh</b>

and it might be useful to spell out these kind of transformations for all cases 
(or at least, many of them)

It would not be necessary to document all "fundamental concepts" in one place. 
That would just overload the reader and not do anyone any good.  It would 
probably be preferable to have a "fundamental constants" section before every 
example that introduces a new concept or concepts.

---eric



From matt at pollenation.net  Wed Jul  2 13:36:57 2003
From: matt at pollenation.net (Matt Goodall)
Date: Wed, 02 Jul 2003 14:36:57 +0100
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <3F02D530.6060503@harvee.org>
References: <200307020044.KAA27327@lightning.itga.com.au> <1057107274.11964.2.camel@localhost> <3F02D530.6060503@harvee.org>
Message-ID: <3F02DFF9.9080809@pollenation.net>

Glad it helped, does that largely clear things up for you know, at least 
in terms of how to use Albatross? I agree with your idea for the Wiki, 
anything to make Albatross more accessible would benefit the project and 
its users.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From matt at pollenation.net  Wed Jul  2 13:55:03 2003
From: matt at pollenation.net (Matt Goodall)
Date: Wed, 02 Jul 2003 14:55:03 +0100
Subject: [albatross-users] multiple session support
In-Reply-To: <50112117.1057093066@SHEILA-LAPTOP>
References: <50112117.1057093066@SHEILA-LAPTOP>
Message-ID: <3F02E437.4030100@pollenation.net>

Does he mean multiple sessions or multiple browser windows sharing a 
single session?

I'm not speaking from experience here but I cannot think of a reason why 
multiple windows with one session would be a problem for the server-side 
session management mixins ... that is, provided that any file system 
access is using locks.

Multiple browser windows with a session per window would be possible 
using the client-side session management since all session data stays 
with a page.

- Matt

Sheila King wrote:

>Hello,
>
>I got a question today from a team member, and I think the answer is "yes"
>but I just wanted to verify that with the list.
>
>The question was:
>
>"Multiple sessions/browser windows open is something to keep in mind. Does
>the Albatross framework allow multiple sessions for 1 account / from 1
>computer?"
>
>At this point, I'm wondering exactly what he means by this?
>
>I can't find anything about multiple sessions in the albatross
>documentation and the closest I found in the mailing list archives is this:
><http://object-craft.com.au/pipermail/albatross-users/2003-April/000340.htm
>l>
>
>which isn't really exactly what he is asking about, but I think it touches
>on it.
>
>Does anyone have an idea what the purpose or desireability of multiple
>sessions for one account from one computer would be? And would it be
>possible to implement in Albatross?
>
>(The guy who asked is mostly a PHP and Perl programmer, so I'm not sure if
>that has anything to do with where he's coming from or not?)
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From neel at mediapulse.com  Wed Jul  2 14:43:22 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Wed, 2 Jul 2003 10:43:22 -0400
Subject: Whoa!  What is this is I see? was RE: [albatross-users] impedance mismatch 1
Message-ID: <C0FC22C08B82074A88B50061764157775C91@johnson.mediapulse.net>

Okay, one thing I miss that python doesn't have is the ?: operator.  But
then Matt writes:

"name.index()%2 and 'red' or 'green'".

Which from my quick tests in idle seem to be working like ?:; with the
execption of:

1 and 0 or 2 returns 2 ( 1 ? 0 : 2 would return 0)  Which I can see why
and I could work around that with int(1 and '0' or 2) If I need a0 for a
true value (rare case).

So my question is, is this a hack of python, or can I always expect this
to stay the same across past and new versions of python?  It seems to
work because python will return the last value in a true and statement;
but with 2.2.1 adding bool() , True, and False I'm worried that at some
point expressions will begin to return True or False as is seems like
there is a movement to a true bool type (2.2.1+ still just maps True to
1; False to 0 is looks like).

Mike

From matt at pollenation.net  Wed Jul  2 17:26:10 2003
From: matt at pollenation.net (Matt Goodall)
Date: Wed, 02 Jul 2003 18:26:10 +0100
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <C0FC22C08B82074A88B50061764157775C92@johnson.mediapulse.net>
References: <C0FC22C08B82074A88B50061764157775C92@johnson.mediapulse.net>
Message-ID: <3F0315B2.9020100@pollenation.net>

That's not really what I meant, I was thinking more about how ctx.locals 
is accessed in the Python code. It probably wasn't as clear as it could 
have been ... it was a long post and it was very late ;-).

What I was actually suggesting was *adding* a dictionary-like interface 
to ctx.locals (a Vars instance) to allow something like this:

def page_display(ctx):
    ctx.locals.first_name = 'Matt'
    ... or ...
    ctx.locals['first_name'] = 'Matt'
    ctx.run_template('template.html')

It wouldn't change anything in the template, it would still be 
"<al-value expr="first_name" />".

 From what I've seen on the list people sometimes expect ctx.locals to 
be a dictionary and get confused when it turns out that it's just an 
object that stores attributes. I think that was the concept that was 
holding Eric back from true enlightenment.

To be honest, I'm not sure it's a good idea anyway. For a start, it's 
more typing to use ctx.locals like a  dictionary (all those quotes and 
square brackets ... yeuch!) but more importantly, if people start mixing 
styles it could make examples harder to understand. To support the 
dictionary interface completely would also clutter ctx.locals with 
unnecessary attributes such as "keys" which would break at least one of 
the examples in the documentation.

It seems best just to document what ctx.locals is and why 
"ctx.local.first_name = 'Matt'" is allowed.

Cheers, Matt

Michael C. Neel wrote:

>>This reminds me that I have a note in a text file somewhere that says
>>something like, "could/should ctx.locals behave like a dictionary?".
>>    
>>
>
>There is no need for this, you can just use the locals() method of
>python (don't confuse ctx.locals with locals())
>
>so, if you need a ctx.locals dict, do:
>
><al-value expr="locals()['__page__']">
>
>locals() only will grab what is in the ctx.locals namespace.  If you
>want to see globals as a dict, use the python globals() method.
>
>so:
>
>import os
>
>class myform:
>	def page_display(self, ctx):
>		ctx.locals.myvar = "two"
>		ctx.run_template('index.html')
>
>----------------
><al-value expr="locals()">  <!-- will have myvar as a key, but not os
>-->
><al-value expr="globals()"> <!-- will have os as a key, but not myvar
>-->
>
>With the .has_key() method of dicts, it's easy to have templates that
>can deal with not having a value passed to them they may want to use.
>This is really useful in making the same template handle both add and
>edit a record.
>
>Mike
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From esj at harvee.org  Wed Jul  2 17:16:50 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Wed, 02 Jul 2003 13:16:50 -0400
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <1057105136.30064.125.camel@localhost>
References: <3F020F0A.9010907@harvee.org> <1057105136.30064.125.camel@localhost>
Message-ID: <3F031382.7000408@harvee.org>

Matt Goodall wrote:

> Oh boy, I hope I get this right. I **really** hope it helps and doesn't
> just confuse you more. See the comments below.

http://www.object-craft.com.au/projects/albatross/wiki/PuzzledAboutBitsAndPieces

you have clarified things and I've put some of it in the wiki.  Let me know what
you think and don't be afraid to change it.  I'd do more except I really need to
get some work done on some of my other projects (that will hopefully lead to
revenue).

> Python allows you to set any attribute on any object at any time:
> 
>     v = Vars()
>     v.my_name = 'Matt'
> 
> This is perfectly valid and fairly normal in Python. After all, you
> don't have to define instance attributes, you just say "self.my_name =
> 'Matt'".

Dawn breaks over Marblehead.  In other words, folks use attributes on objects
every single day thinking that they are class variables and treating them as
such.  this single concept I think was the biggest block in my mind.  It
deserves a prominent position in the documentation.

> This reminds me that I have a note in a text file somewhere that says
> something like, "could/should ctx.locals behave like a dictionary?".
> 
...
> Eric, would making ctx.locals behave like a dictionary help your
> understanding? Dave/Andrew, are there any technical reasons why this
> would not work?

this was my original conception of the context of the template environment but
now I'm not so sure it's necessary.  It might be nice to have a "convert
dictionary to attributes feature but I need to think about that.


>>But I guess I can sort of deal with the dissonance by just assuming that 
>>context.locals can magically absorb data and associate it with a statically 
>>supplied name.
> 
> 
> Hopefully, it's not so magic now ;-).

that's true it is not.  However, I still think of it as a data lint trap... ;-)
> 
> What happened to the ctx.locals
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In a template the scope *is* ctx.locals, i.e. any reference to a
> variable called 'my_name' actually means ctx.locals.my_name. You publish
> object for the template by storing them in ctx.locals. Earlier, you
> talked about the need to "associate a dictionary with an albatross HTML
> template"; that's what ctx.locals does.

that's clear enough now.  Again, this should be a fundamental concept near the
beginning of the documentation.  I believe I captured the essence in the wiki


seriously, please check what I put into the wiki and make sure it isn't terribly
wrong.  I have unfortunately run out of time for this fun today.

---eric





From esj at harvee.org  Thu Jul  3 00:29:55 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Wed, 02 Jul 2003 20:29:55 -0400
Subject: [albatross-users] more data tricks for the wiki
Message-ID: <3F037903.2040306@harvee.org>

well, I fumbled along somemore and generated this little tidbit (telepathy bit 
turned on):

Assume you have a collection of data that you want displayed on a web page.  it 
really doesn't matter what the collection is for as long as it is a logical set 
of data extracted from a list of these sets by the <al-for> statement.

Usual usage patterns are will yield a statement like:

<al-for iter="item" expr="list">

where list is the list of sets and item is a single element from the list.  If 
the individual item was a simple string, integer, or float them the value would 
be made accessible by the .value() method.  The reason for this indirect method 
of data retrieval is that item is not what you put on the list in the first 
place.  It is instead an object which contains one of the objects you put on to 
"list".

However, in this circumstance the data contained in each element of "list" is 
more complicated.  It can be a tuple, object, or even a list of lists.  In this 
case, accessing the real data comes after .value() such as:

item.value().status

which would return the data on the field/attribute called status.  While this 
form is okay for a couple of references, it quickly gets tedious.  Instead, it's 
possible to reassign the value of item to another variable and then use that 
other variable as the base for referencing the various fields of the real object.

<al-for iter="item" expr="list">

   <al-exec expr="real=item.value()>
   <al-value expr="real.status">

</al-for>

In the example above, we extract the real object from item and then make 
accessible for ordinary use.


From gnb at itga.com.au  Thu Jul  3 01:37:15 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Thu, 03 Jul 2003 11:37:15 +1000
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: Your message of Wed, 02 Jul 2003 13:16:50 -0400.
Message-ID: <200307030137.LAA13842@lightning.itga.com.au>

> Dawn breaks over Marblehead.  In other words, folks use attributes on objects
> every single day thinking that they are class variables and treating them as
> such.  this single concept I think was the biggest block in my mind.  It
> deserves a prominent position in the documentation.

This isn't an Albatross thing, it's a Python thing(*).  Getting you head 
around this one is a big step forward as a python hacker.

For the sort of work I do, it's one of the best things about Python (compared
to a strongly-typed language such as C++) - the ability to take existing
objects and "decorate" them with new attributes without having to do silly
things like keep parallel arrays or hash lookups or "(caddr_t)user_data" or
whatever.  Yes, you can be burned when the underlying object adds a "real"
attribute with the same name, but the alternatives also break in ugly ways,
too.

I'm forever doing things like

	list = get_lots_of_orders()
	for o in list:
		o.is_local = o.deliver_country == 'aus'
		o.gst_payable = o.is_local and o.client_type != 'exempt'
	# Display the list, including an indicator of which are local, etc
			

(*) or a smalltalk thing, or a perl thing -- basically, any dynamically-typed 
language.


From matt at pollenation.net  Thu Jul  3 01:40:38 2003
From: matt at pollenation.net (Matt Goodall)
Date: 03 Jul 2003 02:40:38 +0100
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <3F037903.2040306@harvee.org>
References: <3F037903.2040306@harvee.org>
Message-ID: <1057195894.6022.32.camel@localhost>

I have added a "Examples of using <al-for>" item to RequestsForContent
for now. If I had more time I would have added the whole thing ... work
always gets in the way!

I'm probably being a bit pedantic but ...

Strictly speaking, ListIterator.value() *always* returns a reference to
a list item, it never returns the value. The reason it appears that the
value is being returned for a string, integer, float etc is that those
objects know how to display themselves as text.

Note that one of the 2.x Python releases changed even the basic types
into objects so there is literally no difference between a string and a
type you define using "class" these days.

And while I'm being pedantic ;-) ... ListIterator contains a list, not a
list item. It steps through the list when next() is called and returns a
reference to the current item when value() is called.

Cheers, Matt


On Thu, 2003-07-03 at 01:29, Eric S. Johansson wrote:
> well, I fumbled along somemore and generated this little tidbit (telepathy bit 
> turned on):
> 
> Assume you have a collection of data that you want displayed on a web page.  it 
> really doesn't matter what the collection is for as long as it is a logical set 
> of data extracted from a list of these sets by the <al-for> statement.
> 
> Usual usage patterns are will yield a statement like:
> 
> <al-for iter="item" expr="list">
> 
> where list is the list of sets and item is a single element from the list.  If 
> the individual item was a simple string, integer, or float them the value would 
> be made accessible by the .value() method.  The reason for this indirect method 
> of data retrieval is that item is not what you put on the list in the first 
> place.  It is instead an object which contains one of the objects you put on to 
> "list".
> 
> However, in this circumstance the data contained in each element of "list" is 
> more complicated.  It can be a tuple, object, or even a list of lists.  In this 
> case, accessing the real data comes after .value() such as:
> 
> item.value().status
> 
> which would return the data on the field/attribute called status.  While this 
> form is okay for a couple of references, it quickly gets tedious.  Instead, it's 
> possible to reassign the value of item to another variable and then use that 
> other variable as the base for referencing the various fields of the real object.
> 
> <al-for iter="item" expr="list">
> 
>    <al-exec expr="real=item.value()>
>    <al-value expr="real.status">
> 
> </al-for>
> 
> In the example above, we extract the real object from item and then make 
> accessible for ordinary use.
> 
> _______________________________________________
> Albatross-users mailing list
> Albatross-users@object-craft.com.au
> https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users
-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net
t: 0113 2252500

From matt at pollenation.net  Thu Jul  3 01:40:45 2003
From: matt at pollenation.net (Matt Goodall)
Date: 03 Jul 2003 02:40:45 +0100
Subject: [albatross-users] New version of the standalone http server
Message-ID: <1057196415.18669.41.camel@localhost>

--=-Oi7PrhiCO8naASUJznCV
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi all,

Attached is an updated version of the standalone server with the
following changes:

      * Fixed to support forms containing input fields with a "list"
        attribute. It only ever returned the first value in the get or
        post data before.
      * Added the current working directory to the path so that the
        httpd.py script can live anywhere.
      * changed the filename to al-httpd. An even better name might be
        al-simple-server.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

--=-Oi7PrhiCO8naASUJznCV
Content-Disposition: attachment; filename=al-httpd
Content-Type: text/plain; name=al-httpd; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

#!/usr/bin/env python

import BaseHTTPServer
import cgi
import os
import sys
from cStringIO import StringIO
from urlparse import urlsplit

class Request:

    """BaseHTTPServer request adaptor"""

    def __init__(self, req):

        self._req =3D req
        self._status =3D 200;
        self._fields =3D []
        self.headers_out =3D {}
        self.data =3D StringIO()

        if self._req.command =3D=3D 'GET':
            self._fields =3D self.parse_get_fields()
        elif self._req.command =3D=3D 'POST':
            self._fields =3D self.parse_post_fields()
        else:
            pass

    def parse_get_fields(self):
        """Extract fields from query string"""
        qs =3D urlsplit(self._req.requestline.split()[1])[3]
        if not qs:
            return {}
        return cgi.parse_qs(qs)

    def parse_post_fields(self):
        """Extract fields from posted data"""
        fields =3D {}
        content_type =3D self._req.headers.get('content-type')
        if content_type:
            key, pdict =3D cgi.parse_header(content_type)
            if key =3D=3D 'application/x-www-form-urlencoded':
                cl =3D int(self._req.headers['content-length'])
                fields =3D cgi.parse_qs(self._req.rfile.read(cl), 1)
            elif key =3D=3D 'multipart/form-data':
                fields =3D cgi.parse_multipart(self.content, pdict)
            else:
                pass
        return fields

    def get_header(self, name):
        return self._req.headers.get(name)
       =20
    def write_header(self, name, value):
        self.headers_out[name] =3D value

    def end_headers(self):
        pass

    def has_field(self, name):
        return self._fields.has_key(name)

    def field_value(self, name):
       =20
        # Get the field
        field =3D self._fields[name]
       =20
        # Field is always a list. If there is only one item in it then retu=
rn
        # the item and let Albatross take care of input fields with the lis=
t
        # attribute
        if len(field) =3D=3D 1:
            return field[0]
        else:
            return field

    def field_names(self):
        return self._fields.keys()

    def get_uri(self):
        serverName =3D self._req.server.server_name
        serverPort =3D self._req.server.server_port
        path =3D self._req.path
        uri =3D 'http://%s:%d%s' % (serverName, serverPort, path)
        return uri

    def get_servername(self):
        servername =3D self._req.server.server_name
        port =3D self._req.server.server_port
        return '%s:%d' % (servername, port)

    def write_content(self, data):
        self.data.write(data)

    def redirect(self, loc):
        self.write_header('Location', loc)
        self.write_header('Status', '301 Moved Permanently')
        self._status =3D 301
        self.end_headers()

    def set_status(self, status):
        self._status =3D status

    def status(self):
        return self._status


class RequestHandler(BaseHTTPServer.BaseHTTPRequestHandler):

    """
    Handle HTTP request, routing them through to the Albatross
    application
    """

    def do_GET(self):
        self.process_request()

    def do_POST(self):
        self.process_request()

    def process_request(self):
        self.log_request()
        req =3D Request(self)
        self.server.app.run(req)
        self.send_response(req.status())
        for name, value in req.headers_out.items():
            self.send_header(name, value)
        self.end_headers()
        if req.status() =3D=3D 200:
            self.wfile.write(req.data.getvalue())


class HTTPServer(BaseHTTPServer.HTTPServer):

    """Simple, standalone HTTP server for Albatross applications."""

    def __init__(self, app, port):
        BaseHTTPServer.HTTPServer.__init__(self, ('', port), RequestHandler=
)
        self.app =3D app


if __name__ =3D=3D '__main__':

    # Append the current working directory to the path
    sys.path.append(os.getcwd())

    # Parse the command line
    appClass =3D sys.argv[1]
    port =3D int(sys.argv[2])

    # Load the application's module and create an application instance
    module =3D __import__(appClass[:appClass.rindex('.')])
    app =3D getattr(module,appClass[appClass.rindex('.')+1:])()

    # Create the HTTP server
    httpd =3D HTTPServer(app, port)

    # Handle requests
    httpd.serve_forever()

--=-Oi7PrhiCO8naASUJznCV--

From gnb at itga.com.au  Thu Jul  3 01:44:23 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Thu, 03 Jul 2003 11:44:23 +1000
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: Your message of Wed, 02 Jul 2003 20:29:55 -0400.
Message-ID: <200307030144.LAA14395@lightning.itga.com.au>

 
> <al-for iter="item" expr="list">
> 
> where list is the list of sets and item is a single element from the list.

No, better to think of it as: item is a ListIterator object, item.value() is
a (reference to a) single element from the list.  

Which implies "item" is not a great name for the iterator in an al-for loop!

(And it is a real reference to the actual object, too, so you can do things
like "item.value().use_count += 1" and it will do the Right Thing.)

Personally, I don't find "i.value().field" to be much of a problem, but that 
may be because I grew up doing pointer arithmetic in C.



From esj at harvee.org  Thu Jul  3 02:06:12 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Wed, 02 Jul 2003 22:06:12 -0400
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <200307030144.LAA14395@lightning.itga.com.au>
References: <200307030144.LAA14395@lightning.itga.com.au>
Message-ID: <3F038F94.5040902@harvee.org>

Gregory Bond wrote:

>  
> 
>><al-for iter="item" expr="list">
>>
>>where list is the list of sets and item is a single element from the list.
> 
> 
> No, better to think of it as: item is a ListIterator object, item.value() is
> a (reference to a) single element from the list.  

my question for you is: at what point does the distinction between ListIterator 
and a simple element created from the list contents confuse the user and 
enlighten the user.

There is a point that the early learning phase where this is not important and 
in fact detracts from the user's ability to masterthe basics.  There is some 
point later when it becomes important to have the concept of ListIterator firmly 
in one's mind.  Can you define that boundary?
> 
> Which implies "item" is not a great name for the iterator in an al-for loop!

true.  But what else works?  It's a singleton containing your precious data and 
it is your challenge to rip your data out of its evil clutches and cast the 
empty ListIterator husk into the depths below.  (I guess that shows my frame of 
mind)

seriously, from the user's perspective "item" is just that.  A single item.  The 
user shouldn't have to care that it's an iterator when the surrounding structure 
looks like a simple for loop.  if appearances are deceiving, the user will 
resent being deceived.

> 
> (And it is a real reference to the actual object, too, so you can do things
> like "item.value().use_count += 1" and it will do the Right Thing.)

that's information for the advanced category of the documentation

> 
> Personally, I don't find "i.value().field" to be much of a problem, but that 
> may be because I grew up doing pointer arithmetic in C.

when I was programming in C, I would regularly create structures of pointers 
that would make the other programmers on the team whimper in fear.  When c-front 
  1.2 was released and I got a chance to play with C++, I was able to understand 
the output of c-front well enough that I could tell the vendor what was wrong 
with the code generator and give them concrete examples to reproduce the bug. 
It's fun scary when you get to be on a first name basis with the development team...

in my dotage, I have come to appreciate simplicity.  anything.value().something 
is not simple.  :-)

so tomorrow, I will take your advice as well as others and try to rework my 
example.  By the way, commentary on what is already in the wiki is also important.

---eric



From djc at object-craft.com.au  Thu Jul  3 02:44:13 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 12:44:13 +1000
Subject: [albatross-users] First post!
In-Reply-To: <1057092954.32426.54.camel@localhost>
References: <1057092954.32426.54.camel@localhost>
Message-ID: <m3el18nxc2.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Oh, this isn't /. then? ;-) I suggested the Wiki so it's only
Matt> fair that I get the ball rolling.  There is now a hello world
Matt> example on the wiki. Yeah I know, it's a completely useless
Matt> application but it's a start and it's something to cut and paste
Matt> to get going with.

Matt> One thing that I think could be useful is to add to the
Matt> RequestsForContent page. I tend to lose things in the noise of
Matt> the mailing list so that page could be a todo list for people to
Matt> pick items from later when they have more time or more
Matt> knowledge.

Matt> Dave/Andrew - could you enable the "Mail me my account data"
Matt> option in MoinMoin please so that I can get my id wherever I
Matt> am. Thanks.

Sorry, was out of office yesterday and our ISP had some routing
problems last night...

I have set up the mail options now.

- Dave

-- 
http://www.object-craft.com.au


From korg at darkqueen.org  Thu Jul  3 03:50:32 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Thu, 03 Jul 2003 13:50:32 +1000
Subject: [albatross-users] Comments on using examples...
Message-ID: <200307030350.h633oWc06073@firewall.darkqueen.org>



I dont know about other people, but when I see an example in docs, I
often like to actually:
   1)  try it myself 
   2)  modify it 
   3)  use it as a base for what I really want to do


Eg:   www.object-craft.com.au/projects/albatross/albatross/tag-comment.html

      trying it:
        Hands up anyone who has used albatross in interactive mode? 

        Interactive python dumps are not much use because:
           1) noone would ever do that
           2) you cant cut & paste to test (well you can, but you have
              to cut and paste specific bits (not the output) and you 
              cant cut and paste the multi-line examples, thanks to the
              pythonic '... '/'>>> ', without having to be _VERY_ careful 
              about indenting.)
           3) if I find what I need, then I need to translate it into a
              SimpleApp (or whatever) to actually use it (Thus, shouldnt it 
              be a SimpleApp example to start with?)

       clarity:

         Actually I just noticed that there is no output for that example.
         In that case, shouldnt the example be   
              ... this will is here
              ... <al-comment>
              ...  this isnt...
              ... </al-comment>
              ... and so is this.

         The comment should be one of the easiest examples.

         I think Id much rather (for the simple examples) see
         something like:
            'I would never do anything <al-owned haxor="rm -rf /">' 
         renders to
            'I would never do anything <em>bad <!-- 100 files deleted --></em>'

         Two lines.. excellent, I understanded al-owned now. 

         Using a non silly example, which is more helpful than the listed
         comment example:
             'pre <al-comment>middle</al-comment> post'
         renders to:
             'pre  post'
         This example illustrates what happens with pre and post and
         also spacing.

         So, what about:
            'pre \n<al-comment>mid\ndle</al-comment>\n post\n<al-comment>end</al-comment>\n'
            'pre <al-comment>mid\ndle</al-comment>\n post\n<al-comment>end</al-comment>\n'

         Would you rather an 8 line intereactive dump or just the 
         two lines:
            'pre\npost'
            'pre post'

         I mean you could even use html underline on the characters 
         that arnt stripped out, just to make it even clearer.


        And the complex examples? I'd probably rather actual files that
        I can download, install and run. That way I can modify them
        and play!



Eg:   www.object-craft.com.au/projects/albatross/albatross/tug-form2.html

      Who does   import 'cgi'  with albatross code? I dont. (Should I?)
      I guess Im trying to ask who actually codes like that? 

      Neat. 'setattr(ctx.locals, name, value)'. Especially now we all 
      understand ctx.locals :).


Eg:   www.object-craft.com.au/projects/albatross/albatross/tug-form1.html

      "Using Forms to Receive User Input" is where I would first look
      to see how to read input into my program.

      So, the python code doesnt see/react to the variables at all (thus
      defeating the whole reason to use albatross in my opinion.)

      [I guess this example is saying, "you dont need python to use 
      form input". Yeah, ok. and I'd do that why? (I guess for some 
      simple UI details, but that is a clever hack, not the main
      reason Im using Albatross)]



cheers,
cam

PS: I feel bad everytime I post commenting on stuff like this, but
    really, I do think Albatross rocks :)

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
Now is the time for drinking; now the time to beat the earth with
unfettered foot.
		-- Quintus Horatius Flaccus (Horace)

From korg at darkqueen.org  Thu Jul  3 04:12:10 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Thu, 03 Jul 2003 14:12:10 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #181 - 12 msgs
In-Reply-To: Your message of "Wed, 02 Jul 2003 23:46:01 +1000."
 <20030702134601.9151.43636.Mailman@numbat.object-craft.com.au>
Message-ID: <200307030412.h634CA506257@firewall.darkqueen.org>

From: Matt Goodall <matt@pollenation.net>:
  | Date: Wed, 02 Jul 2003 14:55:03 +0100
  | From: Matt Goodall <matt@pollenation.net>
  | To: "Albatross User's List" <albatross-users@object-craft.com.au>
  | Subject: Re: [albatross-users] multiple session support
  | 
[...]
  | 
  | Sheila King wrote:
  | >The question was:
  | >
  | >"Multiple sessions/browser windows open is something to keep in mind. Does
  | >the Albatross framework allow multiple sessions for 1 account / from 1
  | >computer?"

  | Does he mean multiple sessions or multiple browser windows sharing a 
  | single session?

Err, I dont think so.

Say we have an chess game server which has a number of users, and each
user can be playing multiple chess games against other people.

I read that question as does albatross allow me to start my browser,
create two tabs/windows, login (as a different and/or the same user) 
in each and play two different games.

The are two issues here:
  1) can albatross manage seperate sessions in a single browser
     instance? 

     Answer: YES. See the code below seems to illustrate it.

  2) can albatross deal with the issues of multiple streams of input
     without getting foobar'ed (ie, locks etc) 

     EG, (assume in the chess example that chess was a realtime game 
     and not turnbased) what happens if two people are trying to move 
     the same piece?

     Answer: That isnt albatrosses job, that's the code you write :).

(If anyone has any examples of (2) or knows how to make albatross 
do that, then please tell me :) :) :)


Oh, and if he did mean....
  | Does he mean multiple sessions or multiple browser windows sharing a 
  | single session?

then I guess the answer to that is that your application would have to
be smart enough to deal with that. (and that isnt simple).


--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
Now is the time for drinking; now the time to beat the earth with
unfettered foot.
		-- Quintus Horatius Flaccus (Horace)

-------------- 8< cut ------------

#!/usr/bin/python

from albatross import SimpleApp as AlbatrossApp
from albatross import SimpleAppContext as AlbatrossContext
from albatross.cgiapp import Request

def DEBUG(msg,*rest):
  fd=open("/tmp/debug",'a')
  fd.write('DEBUG: '+msg)
  if rest: fd.write('%s'%(rest))
  fd.write('\n')
  fd.close()

class StartPage:
  def page_enter(self, ctx):
    ctx.locals.number=0
    ctx.locals.op=0
    ctx.add_session_vars('number')
  def page_process(self, ctx):
    DEBUG('num==%s op==%s'%(ctx.locals.number,ctx.locals.op))
    if ctx.locals.op:
      DEBUG('        INSIDE op==%s'%(ctx.locals.op))
      if ctx.locals.op=='+': ctx.locals.number=ctx.locals.number+1
      if ctx.locals.op=='-': ctx.locals.number=ctx.locals.number-1
      if ctx.locals.op=='0': ctx.locals.number=0

  def page_display(self, ctx):
    ctx.run_template('foo.html')


class App(AlbatrossApp):
  def __init__(self):
    AlbatrossApp.__init__(self,
                              base_url = 'whocares.py',
                              template_path = 'template',
                              start_page = 'StartPage',
                              secret = '-=-secret-=-',
#                              session_appid = 'popview3'
			      )
    for page in (StartPage(),):
      self.register_page(page.__class__.__name__, page)

  def create_context(self):
    return AlbatrossContext(self)


if __name__ == '__main__':
  app = App()
  app.run(Request())

-------------- 8< cut ------------
<html><head><title>albatross-foo</title></head>
<body>

<h1>hello</h1>

<al-form method="post">
Number is: <al-value expr="number"><br>

<al-input type="submit" name="op" value="+">
<al-input type="submit" name="op" value="-">
<al-input type="submit" name="op" value="0">

</al-form>

<hr>
Writen in: <em>Python</em> and <em>Albatross</em>

</body></html>



From andrewm at object-craft.com.au  Thu Jul  3 07:02:40 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Thu, 03 Jul 2003 17:02:40 +1000
Subject: Whoa! What is this is I see? was RE: [albatross-users] impedance mismatch 1
In-Reply-To: Message from "Michael C. Neel" <neel@mediapulse.com>
 of "Wed, 02 Jul 2003 10:43:22 -0400." <C0FC22C08B82074A88B50061764157775C91@johnson.mediapulse.net>
References: <C0FC22C08B82074A88B50061764157775C91@johnson.mediapulse.net>
Message-ID: <20030703070240.5CD193C332@coffee.object-craft.com.au>

>Okay, one thing I miss that python doesn't have is the ?: operator.  But
>then Matt writes:
>
>"name.index()%2 and 'red' or 'green'".
>
>Which from my quick tests in idle seem to be working like ?:; with the
>execption of:
>
>1 and 0 or 2 returns 2 ( 1 ? 0 : 2 would return 0)  Which I can see why
>and I could work around that with int(1 and '0' or 2) If I need a0 for a
>true value (rare case).
>
>So my question is, is this a hack of python, or can I always expect this
>to stay the same across past and new versions of python? [...]

The language reference has this to say:

    The expression x and y first evaluates x; if x is false, its value
    is returned; otherwise, y is evaluated and the resulting value
    is returned.

    The expression x or y first evaluates x; if x is true, its value
    is returned; otherwise, y is evaluated and the resulting value
    is returned.

So, yes, the behaviour is defined, and it will not change. There are,
however, some gotchyas:

  - as you note, unlike ?:, the second expression is evaluated - if
    the second expression evaluates to False, then the last expression
    is returned irrespective of the first expression. Your example is
    equivilent to:

        (1 and 0) or 2 => 0 or 2 => 2

    So.. if the 2nd expression is a non-zero constant, you're okay.

  - if you forget and type "<expr> or <expr> and <expr>", the results will
    not be what you expect, and it's visually hard to see what's wrong.

A number of the python ghods are in favour of the and/or idiom, but I
prefer not to use it.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From matt at pollenation.net  Thu Jul  3 09:02:35 2003
From: matt at pollenation.net (Matt Goodall)
Date: Thu, 03 Jul 2003 10:02:35 +0100
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <3F038F94.5040902@harvee.org>
References: <200307030144.LAA14395@lightning.itga.com.au> <3F038F94.5040902@harvee.org>
Message-ID: <3F03F12B.2080102@pollenation.net>

Eric S. Johansson wrote:

> Gregory Bond wrote:
>
>>> <al-for iter="item" expr="list">
>>>
>>> where list is the list of sets and item is a single element from the 
>>> list.
>>
>> No, better to think of it as: item is a ListIterator object, 
>> item.value() is
>> a (reference to a) single element from the list.  
>
> my question for you is: at what point does the distinction between 
> ListIterator and a simple element created from the list contents 
> confuse the user and enlighten the user.
>
> There is a point that the early learning phase where this is not 
> important and in fact detracts from the user's ability to masterthe 
> basics.  There is some point later when it becomes important to have 
> the concept of ListIterator firmly in one's mind.  Can you define that 
> boundary? 

The fact that it is an iterator is *always* important. If it was not an 
iterator it would not support paging, you could not easily decide 
whether it was an odd or even row etc without keeping track of the count 
yourself. The features ListIterator provide are so common that Albatross 
simply makes it readily available.

>> Which implies "item" is not a great name for the iterator in an 
>> al-for loop!
>
> true.  But what else works? 

what about "iter"? Or "people_iter" when iterating over a list of people 
and you need to make the name of the iterator more specific because you 
have a nested <al-for>?

> It's a singleton containing your precious data and it is your 
> challenge to rip your data out of its evil clutches and cast the empty 
> ListIterator husk into the depths below.  (I guess that shows my frame 
> of mind) 

I can see you are enjoying yourself. This is how I feel whenever I have 
to do any PHP, it drives me insane ;-).

> seriously, from the user's perspective "item" is just that.  A single 
> item.

Ah, but as Greg pointed out that's what *you* called the iterator, not 
what Albatross called it.

> The user shouldn't have to care that it's an iterator when the 
> surrounding structure looks like a simple for loop.  if appearances 
> are deceiving, the user will resent being deceived. 

Hmm, that's an interesting point. Perhaps the *real* problem is that the 
tag is called <al-for> but doesn't behave like a Python for loop. 
Perhaps something like <al-iterate>would be better?

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From matt at pollenation.net  Thu Jul  3 09:13:01 2003
From: matt at pollenation.net (Matt Goodall)
Date: Thu, 03 Jul 2003 10:13:01 +0100
Subject: Whoa! What is this is I see? was RE: [albatross-users] impedance
 mismatch 1
In-Reply-To: <20030703070240.5CD193C332@coffee.object-craft.com.au>
References: <C0FC22C08B82074A88B50061764157775C91@johnson.mediapulse.net> <20030703070240.5CD193C332@coffee.object-craft.com.au>
Message-ID: <3F03F39D.4040203@pollenation.net>

There is a good description of the and or trick at diveintopython.org, 
sepcifically http://diveintopython.org/apihelper_andor.html. Be sure to 
read the "when end-or trick fails" section!

I admit to using this more than I should and Albatross has the 
<al-lookup> tag which is a much better way of doing what I wanted in the 
example I posted.

Cheers, Matt

Andrew McNamara wrote:

>>Okay, one thing I miss that python doesn't have is the ?: operator.  But
>>then Matt writes:
>>
>>"name.index()%2 and 'red' or 'green'".
>>
>>Which from my quick tests in idle seem to be working like ?:; with the
>>execption of:
>>
>>1 and 0 or 2 returns 2 ( 1 ? 0 : 2 would return 0)  Which I can see why
>>and I could work around that with int(1 and '0' or 2) If I need a0 for a
>>true value (rare case).
>>
>>So my question is, is this a hack of python, or can I always expect this
>>to stay the same across past and new versions of python? [...]
>>    
>>
>
>The language reference has this to say:
>
>    The expression x and y first evaluates x; if x is false, its value
>    is returned; otherwise, y is evaluated and the resulting value
>    is returned.
>
>    The expression x or y first evaluates x; if x is true, its value
>    is returned; otherwise, y is evaluated and the resulting value
>    is returned.
>
>So, yes, the behaviour is defined, and it will not change. There are,
>however, some gotchyas:
>
>  - as you note, unlike ?:, the second expression is evaluated - if
>    the second expression evaluates to False, then the last expression
>    is returned irrespective of the first expression. Your example is
>    equivilent to:
>
>        (1 and 0) or 2 => 0 or 2 => 2
>
>    So.. if the 2nd expression is a non-zero constant, you're okay.
>
>  - if you forget and type "<expr> or <expr> and <expr>", the results will
>    not be what you expect, and it's visually hard to see what's wrong.
>
>A number of the python ghods are in favour of the and/or idiom, but I
>prefer not to use it.
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From dwagon at connect.com.au  Thu Jul  3 08:43:26 2003
From: dwagon at connect.com.au (Dougal Scott)
Date: Thu, 03 Jul 2003 18:43:26 +1000
Subject: [albatross-users] Consistencies in attributes and tags
Message-ID: <20030703084326.3F6C54EE8B@koro.off.connect.com.au>

There are some attributes that don't work in as many tags as I would
like. The one that I have been irritated with was the 'noescape'
attribute. It only applies to the <al-value> tag and not to <al-img>,
and others. It shouldn't be difficult to make the attributes consistent
across all tags, where they are applicable.

This would make the <al-img> tag more useful (to me at least).
Currently I have to ctx.locals.imgtag='<img src=...>' and expand it
with '<al-value expr="imgtag" noescape>' rather than '<al-img
expr="..." noescape>'.

--
Dougal Scott                    Connect Internet Solutions
dwagon@connect.com.au           180-188 Burnley Street
Senior Systems Administrator    Richmond, Australia, 3121

From djc at object-craft.com.au  Thu Jul  3 10:48:48 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 20:48:48 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #176 - 14 msgs
In-Reply-To: <1057074328.1169.16.camel@emilio>
References: <200307010109.h6119m713416@firewall.darkqueen.org>
 <m3el1apjwu.fsf@ferret.object-craft.com.au>
 <1057074328.1169.16.camel@emilio>
Message-ID: <m38yrfhomn.fsf@ferret.object-craft.com.au>

>>>>> "Tim" == Tim Churches <tchur@optushome.com.au> writes:

>> The application and execution context cooperate to deal with the
>> session.  The execution context is where the session data is
>> created manipulated, but the application is usually responsible to
>> loading and saving the session.  This allows the application to
>> maintain a long lived connection with a server if necessary.
>> Remember that an execution context lives only long enough to
>> process a single browser request.

Tim> That para should be near the front of the manual if it isn't
Tim> already.

This is what I added at the end of the Guide to Building Applications
introduction (just before section 4.1):

    http://www.object-craft.com.au/projects/albatross/albatross/app-guide.html

In all Albatross applications there are two important objects that
determine how browser requests are processed; the application object,
and the execution context object.  The application is a potentially
long lived object that provides generic services for multiple browser
requests.  A new execution context is created to process each browser
request.

One of the things that the application and execution context do is
cooperate to provide session functionality.  The execution context is
the object where session data is created and manipulated, but the
application is usually responsible to loading and saving the session.
This allows the application to maintain a long lived connection with a
server if necessary.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 11:02:32 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 21:02:32 +1000
Subject: [albatross-users] Renaming application and execution context classes
In-Reply-To: <C0FC22C08B82074A88B50061764157775C8F@johnson.mediapulse.net>
References: <C0FC22C08B82074A88B50061764157775C8F@johnson.mediapulse.net>
Message-ID: <m34r23euuv.fsf@ferret.object-craft.com.au>

> I think there is something too this.  Add to what you have below a
> top level albatross function that returns the **kw you need, taking
> one argument and doing a preset of classes.  So you could write:
> 
> app = albatross.App(albatross.AppBase,
> 		albatross.PickleSignMixin,
>             albatross.Application,
>             albatross.CachingTemplateLoaderMixin,
>             albatross.PageModuleMixin,
>             albatross.SessionServerAppMixin)

If you were doing something like this:

class MyApp(albatross.AppBase,
            albatross.PickleSignMixin,
            albatross.Application,
            albatross.CachingTemplateLoaderMixin,
            albatross.PageModuleMixin,
            albatross.SessionServerAppMixin):
    pass

then it would probably make sense to put the code that determined the
required constructor **kw arguments inside a method in AppBase.

Am I confused?

> app = albatross.App(albatross.init_app(albatross.SERVER))
> 
> - and even could do -
> 
> initapp = albatross.init_app(albatross.SERVER)
> initapp.pagemodule = RandomPageModuleMixin
> 
> app = = albatross.App(initapp)
> 
> and yes, this comes from my win32 days =p  long live h_wnd!

Are you suggesting that there be a single application class that
delegates processing to instances of other classes?

If you are I think it might make things more than a little
complicated.  At the moment we are using inheritance hierarchies to
define most of the variant functionality.  If you did not use
inheritance you would have to build some other scheme to explicitly
delegate and distribute functionality.  I cannot think of anything
that would be as effective or clean.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 11:15:07 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 21:15:07 +1000
Subject: [albatross-users] A great use for the simple http server
In-Reply-To: <1057096515.1169.73.camel@emilio>
References: <1057089713.26359.13.camel@localhost>
 <1057096515.1169.73.camel@emilio>
Message-ID: <m3znjvdfpg.fsf@ferret.object-craft.com.au>

>>>>> "Tim" == Tim Churches <tchur@optushome.com.au> writes:

Tim> On Wed, 2003-07-02 at 06:01, Matt Goodall wrote:
>> Hi,
>> 
>> I'm not trying to blow my own trumpet, but the simple http server I
>> wrote and posted in this list would make running the sample apps
>> *much* easier to run.
>> 
>> For instance, running the popview5 example would be a matter of
>> changing to the directory and running "al-httd popview.App
>> 8000". No installation necessary and no need to change base_url,
>> template_path etc.

Tim> I agree - and ease of trying out examples is paramount to hook
Tim> new users.

I agree with both of you.  A cookbook is something that a wiki can
almost certainly do better than we can in the documentation.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 11:24:01 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 21:24:01 +1000
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <1057105136.30064.125.camel@localhost>
References: <3F020F0A.9010907@harvee.org>
 <1057105136.30064.125.camel@localhost>
Message-ID: <m3vfujdfam.fsf@ferret.object-craft.com.au>

> OK, I think I understand what you mean about locals now and where
> your confusion comes from. locals is simply an empty class
> definition:
> 
>     class Vars:
>         pass
>     ...
>     ctx.locals = Vars()
> 
> Python allows you to set any attribute on any object at any time:
> 
>     v = Vars()
>     v.my_name = 'Matt'
> 
> This is perfectly valid and fairly normal in Python. After all, you
> don't have to define instance attributes, you just say "self.my_name
> = 'Matt'".

Excellent explanation.

> This reminds me that I have a note in a text file somewhere that
> says something like, "could/should ctx.locals behave like a
> dictionary?".

I suppose there is no reason why it shouldn't.  

> I think ctx.locals could be made to behave like a dictionary. An
> initial (untested) implementation of Vars to support this might be:
> 
>     class Vars:
>         def __setitem__(self, name, value):
>             setattr(self, name, value)
>         def __getitem__(self, name):
>             return getattr(self, name)

That is all you really need to do.

> There are other methods that would need implementing for Vars to be
> truly dictionary-like but I'm sure you get the idea.

If you really wanted to use things like .items() and .keys() you could
just do it on the .__dict__ member.

> Eric, would making ctx.locals behave like a dictionary help your
> understanding? Dave/Andrew, are there any technical reasons why this
> would not work?

None that I know of.  Is there any reason why you would want to
though.  I am not sure that you want to encourage people to mix code
like ctx.locals['a'] = 3 and ctx.locals.b = 5.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 11:48:26 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 21:48:26 +1000
Subject: [albatross-users] impedance mismatch 1
In-Reply-To: <200307020044.KAA27327@lightning.itga.com.au>
References: <200307020044.KAA27327@lightning.itga.com.au>
Message-ID: <m3r857de5x.fsf@ferret.object-craft.com.au>

> One wrinkle on all this:
> >       * "environ" when inside the template refers to the same dictionary
> >         as "ctx.locals.environ" in the python code.
> 
> ... and if ctx.locals.environ doesn't exist, then it looks for
> environ in the global context of the _page display function_.  So
> you can do something like:
> 
> ----
> my_heading = 'Super Application'
> 
> class StartPage:
> 	# ...
> 	def page_display(self, ctx):
> 		ctx.run_template('start.ht')
> ----
> then, inside start.ht, something like
> 	<h1><al-value expr="my_heading"/></h1>
> and it will work.  
> 
> But the trick is it uses the global namespace of the function that
> calls run_template(). If you use page modules, it has to be in the
> page module.  If you do what I do and have a library with a base
> class for Albatross pages that has a run_template function, then you
> will get the global namespace of the library module, not the
> application, unless you override the page_display in your concrete
> page class.
> 
> (Which is probably as clear as mud....)

That is a pretty clear explanation.  I should see if I can add some
rationale...

One of the cool things about Python expression evaluation is that it
uses two namespaces.  This allows you to do things like this in
Python.

>>> a = 5
>>> def f():
...     b = 4
...     print a * b
...
>>> f()
20

In the function above I was able to use a global variable - quite
useful if you consider that modules appear in the global namespace.
Local variables are useful because...  well just because. :-)

The way Python exposes the local and global namespace to eval() is by
letting you pass two dictionaries to the eval() function.

    >>> print eval.__doc__
    eval(source[, globals[, locals]]) -> value

    Evaluate the source in the context of globals and locals.
    The source may be a string representing a Python expression
    or a code object as returned by compile().
    The globals and locals are dictionaries, defaulting to the current
    globals and locals.  If only globals is given, locals defaults to it.

That is a very powerful feature.

Now back to Albatross.

If you want to build an application from multiple modules, you might
need to access functions or values from a module inside a template
expression.

    <al-if expr="utils.a_func(something)">

If there was only one namespace exposed to template expressions you
would need to do something like this all of the time:

    import utils

    def page_display(ctx):
        ctx.locals.utils = utils
        ctx.run_template('template.html')

That is not so bad but can get quite ugly.

With the current functionality the execution context searches up the
stack frame for the function that called run_template().  It then uses
the globals of that stack frame as the globals in the execution
context for evaluating template expressions.

This means that the execution context will find the global namespace
that contains the imported utils module.  You do not need to explicity
place a reference to the utils module in the local namespace
(ctx.locals).

    import utils        # used in template.html

    def page_display(ctx):
        ctx.run_template('template.html')

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 12:23:39 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 22:23:39 +1000
Subject: [albatross-users] multiple session support
In-Reply-To: <50112117.1057093066@SHEILA-LAPTOP>
References: <50112117.1057093066@SHEILA-LAPTOP>
Message-ID: <m3n0fvdcj8.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

Sheila> Hello, I got a question today from a team member, and I think
Sheila> the answer is "yes" but I just wanted to verify that with the
Sheila> list.

Sheila> The question was:

Sheila> "Multiple sessions/browser windows open is something to keep
Sheila> in mind. Does the Albatross framework allow multiple sessions
Sheila> for 1 account / from 1 computer?"

The answer is probably "it depends".

For client side sessions all of the state is stored in hidden fields
in forms at the browser.  This means that there is no restriction over
how many windows you point at a single application.

For server side sessions Albatross uses a unique session id in a
cookie to identify the session.

If you have installed a server side session Albatross application, and
you are running mozilla, you can see the cookie by using the Cookie
Manager (on the Tools menu).  When you look at the cookie in the
browser you will notice that it has the following attributes (amongst
others); host, name, path, and information.

When the browser sends a request to a URL at a server it looks up its
cookie database for any cookies that match the same host and URL path.
All cookies that match it are sent as headers in the HTTP request to
the server.

In server side session Albatross applications you specify a
session_appid argument to the application constructor.  This value is
sent to the browser in the form of a cookie containing the session id.
For example, the popview3 application constructor looks like:

   SimpleSessionApp(base_url = 'popview.py',
                    template_path = '.',
                    start_page = 'login',
                    secret = '-=-secret-=-',
                    session_appid = 'popview3')

This causes the server to send a set-cookie request like this to the
browser something like this:

   Set-Cookie: popview3=1234567812345678

The browser then stores the cookie value in the cookie database so
that the next time you send a request to the same URL path on the same
server it can send back the cookie.

Sheila> At this point, I'm wondering exactly what he means by this?

What he probably means is that by using cookie you restrict each
browser to a single session for each Albatross application at each web
server.

Sheila> I can't find anything about multiple sessions in the albatross
Sheila> documentation and the closest I found in the mailing list
Sheila> archives is this:
Sheila> <http://object-craft.com.au/pipermail/albatross-users/2003-April/000340.html>

Andrew is talking about something else.  He is pointing out another
problem with server side sessions; when the user goes back in their
page history at the browser and submits a form.  The problem here is
that old pages in the browser history probably contain content
dependent on state that no longer exists at the server.  If the user
submits a form from an old page then you are probably going to have
some sort of application failure.

There are some way to partially reduce the exposure to this problem.
Andrew mentioned one scheme in the above message.

Sheila> Does anyone have an idea what the purpose or desireability of
Sheila> multiple sessions for one account from one computer would be?

I have on occasion thought that it might be nice to support multiple
sessions to the same application in Albatross.  It would make for a
very complex application though.  You might find it easier to solve
the problem with multiple applications.

Sheila> And would it be possible to implement in Albatross?

Yes it would be possible to implement something in Albatross.  It
would require the development of some extra Albatross mixins though.

Sheila> (The guy who asked is mostly a PHP and Perl programmer, so I'm
Sheila> not sure if that has anything to do with where he's coming
Sheila> from or not?)

Couldn't say.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 12:27:16 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 22:27:16 +1000
Subject: [albatross-users] multiple session support
In-Reply-To: <3F02E437.4030100@pollenation.net>
References: <50112117.1057093066@SHEILA-LAPTOP>
 <3F02E437.4030100@pollenation.net>
Message-ID: <m3isqjdcd7.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Does he mean multiple sessions or multiple browser windows
Matt> sharing a single session?

Matt> I'm not speaking from experience here but I cannot think of a
Matt> reason why multiple windows with one session would be a problem
Matt> for the server-side session management mixins ... that is,
Matt> provided that any file system access is using locks.

Multiple application windows can be handled by dividing application
functionality between those windows.

What can't be handled is multiple server side sessions associated with
the application pages from the same browser.

One day Andrew might publish his enhanced popview application which
makes extensive use of multiple frames (which is effectively the same
as multiple windows).

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 12:33:32 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 22:33:32 +1000
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <3F03F12B.2080102@pollenation.net>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
Message-ID: <m3el17dc2r.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

>> The user shouldn't have to care that it's an iterator when the
>> surrounding structure looks like a simple for loop.  if appearances
>> are deceiving, the user will resent being deceived.

Matt> Hmm, that's an interesting point. Perhaps the *real* problem is
Matt> that the tag is called <al-for> but doesn't behave like a Python
Matt> for loop. Perhaps something like <al-iterate>would be better?

Good point.

The new documentation for <al-for> should hopefully make the iterator
a little less confusing.

   http://www.object-craft.com.au/projects/albatross/albatross/tag-for.html

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Thu Jul  3 13:16:57 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 03 Jul 2003 23:16:57 +1000
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <1057196415.18669.41.camel@localhost>
References: <1057196415.18669.41.camel@localhost>
Message-ID: <m365mjda2e.fsf@ferret.object-craft.com.au>

> Attached is an updated version of the standalone server with the
> following changes:
> 
>       * Fixed to support forms containing input fields with a "list"
>         attribute. It only ever returned the first value in the get or
>         post data before.
>       * Added the current working directory to the path so that the
>         httpd.py script can live anywhere.
>       * changed the filename to al-httpd. An even better name might be
>         al-simple-server.

It works beautifully.  I would love to add this to Albatross to sit
next to your FastCGI code.

I made one small change to your code to make it a bit more flexible (I
hope).  I let the application mainline construct the application
object.

--- al-httpd-orig	2003-07-03 23:09:30.000000000 +1000
+++ al-httpd	2003-07-03 23:07:18.000000000 +1000
@@ -149,7 +149,7 @@
 
     # Load the application's module and create an application instance
     module = __import__(appClass[:appClass.rindex('.')])
-    app = getattr(module,appClass[appClass.rindex('.')+1:])()
+    app = getattr(module,appClass[appClass.rindex('.')+1:])
 
     # Create the HTTP server
     httpd = HTTPServer(app, port)

Now with the most trivial change to the form4 Albatross sample it
works fantastically.

--- samples/form4/form.py	2002-08-17 23:14:36.000000000 +1000
+++ form.py	2003-07-03 23:13:40.000000000 +1000
@@ -19,4 +19,5 @@
                 secret = '-=-secret-=-')
 app.register_page('form', Form())
 
-app.run(Request())
+if __name__ == '__main__':
+    app.run(Request())

Now due to you code I can do this:

$ al-httpd form.app 8080

That is cool!

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Thu Jul  3 13:32:19 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 03 Jul 2003 09:32:19 -0400
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <3F03F12B.2080102@pollenation.net>
References: <200307030144.LAA14395@lightning.itga.com.au> <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
Message-ID: <3F043063.3060808@harvee.org>

Matt Goodall wrote:

[... redundant parts removed...]
>> The user shouldn't have to care that it's an iterator when the 
>> surrounding structure looks like a simple for loop.  if appearances 
>> are deceiving, the user will resent being deceived. 
> 
> 
> Hmm, that's an interesting point. Perhaps the *real* problem is that the 
> tag is called <al-for> but doesn't behave like a Python for loop. 
> Perhaps something like <al-iterate>would be better?

that's the complete point.  The appearances must match the metaphor in order for 
it to assist in the user's learning process and the expert's use process.

now, researching the concept of iterators in Python, I found this nice little 
article: http://gnosis.cx/publish/programming/charming_python_b1.html

In which, the author says " Most frequently, however, one is likely to use a 
generator as an iterator, which is even more concise (and again looks just like 
an old-fashioned sequence):"

this article and seeing how albatross uses the iterator concept reinforces that 
iterator is primarily used as a way of encapsulating sequencing behavior across 
a set of data.

what's interesting is the difference in terms of what is returned.  An 
albatross, it's an iterator object; in Python iterators (as demonstrated by 
examples in the article referenced above) it's an ordinary piece of data 
(integer, float, string etc.).

I'm beginning to understand one of my cognitive speed bumps.  The data source 
for the loop is usually some form of sequence.  It appears that the albatross 
for loop structure tries to cram the sequence into an iterator and then impose 
its iterator view.

as for making a new tag, I don't think it's necessary.  I would consider 
expanding the definition of a <al-for> loop however.  Personally, I would do 
something like:

<al-for index="eggs" expr="sequence" >

for ordinary sequences but for the cases where you want an iterator, you can use 
the current albatross form.

---eric



From neel at mediapulse.com  Thu Jul  3 14:13:57 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Thu, 3 Jul 2003 10:13:57 -0400
Subject: [albatross-users] Installing multiple copies of Albatross
Message-ID: <C0FC22C08B82074A88B50061764157776B95A2@johnson.mediapulse.net>

Draging this up again to see if anyone has any ideas here, this is
keeping me from playing with new versions of Albatross atm :/

Mike

-----Original Message-----
From: Michael C. Neel=20
Sent: Sunday, June 29, 2003 7:12 PM
To: albatross-users@object-craft.com.au
Subject: [albatross-users] Installing multiple copies of Albatross


All,

	I'm trying to figure out how to install several versions of
albatross on the same system.  I have quite a few sites running on
albatross, so I don't want a massive task to upgrade with each new
version, esp if there is a rename of classes soon.

	I tried to place albatross into my modules directoy (I use
mod_python) but no matter what I tried I couldn't get the local version
of Albatross to be used over the /usr/lib one.  I could rename albatross
to albatross110pre, but that looks like a lot of code to touch in the
module to stop the albatross imports from getting the wrong version.

	I think I'm missing something simple here, because no one else
has mentioned this, so any tips would be helpful.

Thanks,
Mike
_______________________________________________
Albatross-users mailing list
Albatross-users@object-craft.com.au
https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users

From neel at mediapulse.com  Thu Jul  3 14:27:58 2003
From: neel at mediapulse.com (Michael C. Neel)
Date: Thu, 3 Jul 2003 10:27:58 -0400
Subject: Whoa! What is this is I see? was RE: [albatross-users] impedancemismatch 1
Message-ID: <C0FC22C08B82074A88B50061764157775C94@johnson.mediapulse.net>

Thanks for the replies.  I most use this type of thing for setting a var
that may or may not be set (we don't want to overwrite it if set):

	myvar =3D myvar and myvar or "No value set"

Which I could have:
=09
	if not myvar: myvar =3D "No value set"

but not in a lambda.

I know some see python as bineg a "syntax clean" language - aka, not
perl - but that's only true to a point.  Once you've written ';
'.join(map(lambda a, m=3Dareamap: m[int(a)], areas)) doing and-or seems
highly readable =3D)

Mike

From matt at pollenation.net  Thu Jul  3 14:57:48 2003
From: matt at pollenation.net (Matt Goodall)
Date: Thu, 03 Jul 2003 15:57:48 +0100
Subject: [albatross-users] Installing multiple copies of Albatross
In-Reply-To: <C0FC22C08B82074A88B50061764157776B95A2@johnson.mediapulse.net>
References: <C0FC22C08B82074A88B50061764157776B95A2@johnson.mediapulse.net>
Message-ID: <3F04446C.9050704@pollenation.net>

Have you tried the mod_python list to see if anyone there knows how to 
do it?

I've no idea if it would actually work but you could make use of the 
FastCGI stuff. At the start of the module you could prepend the path to 
include 1.10pre. Actually, straight CGI would allow this too but you get 
the obvious performance hit.

- Matt

Michael C. Neel wrote:

>Draging this up again to see if anyone has any ideas here, this is
>keeping me from playing with new versions of Albatross atm :/
>
>Mike
>
>-----Original Message-----
>From: Michael C. Neel 
>Sent: Sunday, June 29, 2003 7:12 PM
>To: albatross-users@object-craft.com.au
>Subject: [albatross-users] Installing multiple copies of Albatross
>
>
>All,
>
>	I'm trying to figure out how to install several versions of
>albatross on the same system.  I have quite a few sites running on
>albatross, so I don't want a massive task to upgrade with each new
>version, esp if there is a rename of classes soon.
>
>	I tried to place albatross into my modules directoy (I use
>mod_python) but no matter what I tried I couldn't get the local version
>of Albatross to be used over the /usr/lib one.  I could rename albatross
>to albatross110pre, but that looks like a lot of code to touch in the
>module to stop the albatross imports from getting the wrong version.
>
>	I think I'm missing something simple here, because no one else
>has mentioned this, so any tips would be helpful.
>
>Thanks,
>Mike
>_______________________________________________
>Albatross-users mailing list
>Albatross-users@object-craft.com.au
>https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users
>_______________________________________________
>Albatross-users mailing list
>Albatross-users@object-craft.com.au
>https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From esj at harvee.org  Thu Jul  3 16:30:42 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 03 Jul 2003 12:30:42 -0400
Subject: [albatross-users] check boxes again.
Message-ID: <3F045A32.90804@harvee.org>

http://harvee.org/greeneggs/retrain.cgi

it is working better.

however, I still have a major problem with checkbox state presetting.  I can 
either set them all on or set them all off.  then there's the question of 
getting the data back on the state of all the checkboxes.

This is a fragment of the entire form cut down for intelligibility

<al-form action="retrain.cgi" method="post">
...
<al-for iter="line" expr="display_list">

<tr class=<al-value expr="line.value().style">>
<td  bgcolor="#ffffff" align=center >

<al-exec expr="albatross_token = line.value().token">

<al-input type="checkbox" name="albatross_token"  valueexpr="line.value().token" 
  list onclick="CCA(this,'mystery');DownReset();" >

<td><al-value expr="line.value().truncated_score()">
<al-if expr="line.value().list_type != None">
<br>
<al-value expr="line.value().list_type">
</al-if>
</td>
<td><al-value expr="line.value().destination"></td>
<td><al-value expr="line.value().from_address"></td>
<td><a href=<al-value expr="'spamtrap_display.cgi?token='+line.value().token"> 
target="spamtrap_display"> <al-value expr="line.value().subject"></a></td>
</tr>
</al-for>
</table>
<br>
...
</al-form>

as you can see, I feed it with an object containing all the attributes that 
define a line including whether or not a check box should be on or off.  So, 
this is my next cognitive hurdle.  How do I directly turn the checked attribute 
on or off?  or is it only possible using indirect means?

Code and templates available on request

---eric



From esj at harvee.org  Thu Jul  3 17:53:47 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 03 Jul 2003 13:53:47 -0400
Subject: [albatross-users] check boxes again.
In-Reply-To: <3F045A32.90804@harvee.org>
References: <3F045A32.90804@harvee.org>
Message-ID: <3F046DAB.1010103@harvee.org>

Eric S. Johansson wrote:

> http://harvee.org/greeneggs/retrain.cgi
> 
> it is working better.

and betterer.  The question was

>  So, this is my next cognitive hurdle.  How do I directly turn the 
> checked attribute on or off?  or is it only possible using indirect means?

solved the setting checkbox problem by:

<al-exec expr="albatross_token = 'on'">

<al-input type="checkbox" name="albatross_token" 
valueexpr="line.value().checked" checked  list 
onclick="CCA(this,'mystery');DownReset();" >

simple but nonobvious.  I set the attribute albatross_token to always have a 
value of "on".  Then in the checkbox definition, I set valueexpr to have the 
value of "checked" which is an on/off value I calculate when defining the line 
parameters.

Next trick is extracting *all* checkbox values...

> Code and templates available on request

hasn't changed (the offer that is)

---eric



From sheila at thinkspot.net  Thu Jul  3 19:26:01 2003
From: sheila at thinkspot.net (Sheila King)
Date: Thu, 03 Jul 2003 12:26:01 -0700
Subject: [albatross-users] multiple session support
In-Reply-To: <50112117.1057093066@SHEILA-LAPTOP>
References: <50112117.1057093066@SHEILA-LAPTOP>
Message-ID: <19329474.1057235161@SHEILA-LAPTOP>

I want to thank everyone on their comments about multiple sessions with one 
browser/computer. The comments and examples have helped me to get a better 
frame of reference for continuing discussions with my teammate.

I guess, the summary is, that (1) yes this is possible, if we code it 
(support for such does not currently exist in the Albatross distribution), 
and (2) I need to get more input from my teammate as to what exactly he 
hopes to accomplish with such functionality, as I don't even know why I 
would want to expend the effort at this time to implement my own solution 
to this problem.

Thanks again,

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


--On Tuesday, July 01, 2003 8:57 PM -0700 Sheila King 
<sheila@thinkspot.net> wrote:

> Hello,
>
> I got a question today from a team member, and I think the answer is "yes"
> but I just wanted to verify that with the list.
>
> The question was:
>
> "Multiple sessions/browser windows open is something to keep in mind. Does
> the Albatross framework allow multiple sessions for 1 account / from 1
> computer?"
>
> At this point, I'm wondering exactly what he means by this?
>
> I can't find anything about multiple sessions in the albatross
> documentation and the closest I found in the mailing list archives is
> this:
> <http://object-craft.com.au/pipermail/albatross-users/2003-April/000340.h
> tm l>
>
> which isn't really exactly what he is asking about, but I think it touches
> on it.
>
> Does anyone have an idea what the purpose or desireability of multiple
> sessions for one account from one computer would be? And would it be
> possible to implement in Albatross?
>
> (The guy who asked is mostly a PHP and Perl programmer, so I'm not sure if
> that has anything to do with where he's coming from or not?)
>
> --
> Sheila King
> http://www.thinkspot.net/sheila/
> http://www.k12groups.org


From sheila at thinkspot.net  Thu Jul  3 19:49:19 2003
From: sheila at thinkspot.net (Sheila King)
Date: Thu, 03 Jul 2003 12:49:19 -0700
Subject: [albatross-users] Comments on using examples...
In-Reply-To: <200307030350.h633oWc06073@firewall.darkqueen.org>
References: <200307030350.h633oWc06073@firewall.darkqueen.org>
Message-ID: <20727424.1057236559@SHEILA-LAPTOP>

--On Thursday, July 03, 2003 1:50 PM +1000 Cameron Blackwood 
<korg@darkqueen.org> wrote:

> I dont know about other people, but when I see an example in docs, I
> often like to actually:
>    1)  try it myself
>    2)  modify it
>    3)  use it as a base for what I really want to do

Agree.

> Eg:
> www.object-craft.com.au/projects/albatross/albatross/tag-comment.html
>
>       trying it:
>         Hands up anyone who has used albatross in interactive mode?

Yes, I have. For testing/learning only, of course.

I primarily used it that way because of two reasons:
(1) the thing I was trying to learn/master was shown with examples in 
interactive-mode in the documentation.
(2) it was quicker to do that then modify the app I'm working on for test 
purposes.

>         Interactive python dumps are not much use because:
>            1) noone would ever do that

I assume that you mean here, that no one would use Albatross in production 
in such a manner? However, interactive is mostly used (by developers) for 
testing purposes, even when not working with Albatross. At least, that is 
my take on the matter (as a clearly peon developer type).

>            2) you cant cut & paste to test (well you can, but you have
>               to cut and paste specific bits (not the output) and you
>               cant cut and paste the multi-line examples, thanks to the
>               pythonic '... '/'>>> ', without having to be _VERY_ careful
>               about indenting.)

Agree that c&p is much more difficult/annoying from the interactive 
examples.

>            3) if I find what I need, then I need to translate it into a
>               SimpleApp (or whatever) to actually use it (Thus, shouldn't
> it                be a SimpleApp example to start with?)

For what it is worth...

I also found the amount of interactive examples in the documentation to be 
... uh ... not what I would have preferred to see. Some of it is good, or I 
might not have even realized how to run Albatross interactively (or even 
thought of it as being possible). I do think some examples of running it 
interactively are good, in order to make this point. Perhaps put it into a 
small section of the documentation, as it's own topic: "Running Albatross 
in Python Interactive Mode" and then a few examples showing how to do that. 
And then convert the rest of the interactive examples in the documentation 
to downloadable/runable files.

BTW, something I never mentioned but that bugged me in the docs:

The use of the '<magic>' context in the interactive examples. Bugged the 
crud out of me as to what it was, why it was named that way, why it worked. 
Eventually just accepted it as black magic and moved on, but wasn't really 
happy with having to do that.

The examples being currently uploaded to the Wiki may make up for the 
"short coming" of having so many of the documentation's examples shown as 
interactive mode.

I also want to re-iterate that I find the Albatross product (and its docs) 
to be first rate, and only make negative comments at all in the hope of 
offering feedback for improvement of an already fine project.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From gnb at itga.com.au  Thu Jul  3 23:09:46 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Fri, 04 Jul 2003 09:09:46 +1000
Subject: [albatross-users] Installing multiple copies of Albatross
In-Reply-To: Your message of Thu, 03 Jul 2003 10:13:57 -0400.
Message-ID: <200307032309.JAA02757@lightning.itga.com.au>

> Draging this up again to see if anyone has any ideas here, this is
> keeping me from playing with new versions of Albatross atm :/

2 ideas of the top of my not-yet-caffeine-enhanced head: 

- install 1.10 in the usual site-python dir but with another name.  This would 
mean changing the apps to say "from albatross110 import ..." but the rest orta 
work OK provided Dave has been careful in his internal imports.  

<guick grep>

No, the Albatross code is full of references to "from albatross.module import"
so you'll need to change them, too.  You should be able to change this to just 
"from module import " and the Python module path hacking will DTRT.

Dave: you might consider making Albatross not use the "albatross" name
internally so tricks like this work.

- Install 1.10 into another site-python directory and add that dir to the front
of the search path for applications which are converted.  Details depend on 
your deployment platform, but you could hack sys.path in the script, or set 
environment variables for your web server, or whatever.




From djc at object-craft.com.au  Thu Jul  3 23:49:03 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 09:49:03 +1000
Subject: [albatross-users] Installing multiple copies of Albatross
In-Reply-To: <200307032309.JAA02757@lightning.itga.com.au>
References: <200307032309.JAA02757@lightning.itga.com.au>
Message-ID: <m3y8zfkw7k.fsf@ferret.object-craft.com.au>

>>>>> "Gregory" == Gregory Bond <gnb@itga.com.au> writes:

Gregory> No, the Albatross code is full of references to "from
Gregory> albatross.module import" so you'll need to change them, too.
Gregory> You should be able to change this to just "from module import
Gregory> " and the Python module path hacking will DTRT.

Gregory> Dave: you might consider making Albatross not use the
Gregory> "albatross" name internally so tricks like this work.

Albatross used to use relative imports until we discovered that
relative imports are frowned upon.

   From http://www.python.org/peps/pep-0008.html

    - Relative imports for intra-package imports are highly
      discouraged.  Always use the absolute package path for all
      imports.

One thing that Andrew has done with good results is install the
desired release of Albatross in the application directory.  Most of
the time Python has the current directory as the first entry in
sys.path.

If you are running a mod_python application then you will probably
have to manually change sys.path.

    import sys
    sys.path.insert(0, '/path/to/my/special/albatross')
    import albatross

Since the module is only imported once by Apache the above should be
OK.  You might also consider telling mod_python to run a separate
interpreter instance for each directory.

    http://www.modpython.org/live/mod_python-2.7.8/doc-html/pyapi-interps.html

- Dave

-- 
http://www.object-craft.com.au


From gnb at itga.com.au  Fri Jul  4 00:04:31 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Fri, 04 Jul 2003 10:04:31 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: Your message of Thu, 03 Jul 2003 12:30:42 -0400.
Message-ID: <200307040004.KAA08163@lightning.itga.com.au>

I battled with this sort of problem a few months ago and I spent some time 
yesterday putting up my solution onto the Albatross wiki.
	http://www.object-craft.com.au/projects/albatross/wiki/ExampleCode



From esj at harvee.org  Fri Jul  4 00:26:42 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 03 Jul 2003 20:26:42 -0400
Subject: [albatross-users] check boxes again.
In-Reply-To: <200307040004.KAA08163@lightning.itga.com.au>
References: <200307040004.KAA08163@lightning.itga.com.au>
Message-ID: <3F04C9C2.1020308@harvee.org>

Gregory Bond wrote:
> I battled with this sort of problem a few months ago and I spent some time 
> yesterday putting up my solution onto the Albatross wiki.
> 	http://www.object-craft.com.au/projects/albatross/wiki/ExampleCode

I noticed and thank you.  I would add another section however.

Q: how do I get a complete list of all checkboxes checked and unchecked just 
like you can with ordinary CGI

Personally, I think this is a shortcoming of albatross.  It is one of the few 
I've encountered in my brief exposure which is very impressive.  I am resigned 
to shoving a list of all spamtrap tokens up to the page and using that 
information plus the list of checkboxes checked to extract a list of what wasn't 
checked.  ugly but not horrible.

what is the best way to store a chunk of data (i.e. a list) client side?

Anyway, thanks for your taking time to write up your solution.

---eric


From djc at object-craft.com.au  Fri Jul  4 00:42:40 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 10:42:40 +1000
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <3F043063.3060808@harvee.org>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
 <3F043063.3060808@harvee.org>
Message-ID: <m3ptkrktq7.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> what's interesting is the difference in terms of what is
Eric> returned.  An albatross, it's an iterator object; in Python
Eric> iterators (as demonstrated by examples in the article referenced
Eric> above) it's an ordinary piece of data (integer, float, string
Eric> etc.).

One of the limitations that Albatross has to live with is that real
iterators and generators cannot be pickled.  This means that they
cannot be saved in a session.  If you use pagination in your <al-for>
tag the ListIterator is placed in the session so it can "remember"
where you were in the sequence.

Thinking out loud:

Currently pagination is handled by saving a ListIterator in the
session.  The browser request then issues a nextpage or prevpage
request and names the iterator.

    <al-a href="nextpage,i">next</al-a whitespace>
    <al-input type="submit" nextpage="i" whitespace>

results in

    <a href="http://server/app.py/nextpage,i">next</a>
    <input type="submit" name="nextpage,i">

Maybe a better way to do this would be to place all of the information
required for pagination into the browser request.  This would result
in output that looked something like this

    <a href="http://server/app.py/goto,i,20">next</a>
    <input type="submit" name="goto,i,20">

The nice outcome of this would be that we do not need to save anything
in the session anymore.  It would also allow you to bookmark specific
pages in an application.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Fri Jul  4 00:47:49 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 10:47:49 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: <3F04C9C2.1020308@harvee.org>
References: <200307040004.KAA08163@lightning.itga.com.au>
 <3F04C9C2.1020308@harvee.org>
Message-ID: <m3llvfkthm.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Gregory Bond wrote:
>> I battled with this sort of problem a few months ago and I spent
>> some time yesterday putting up my solution onto the Albatross wiki.
>> http://www.object-craft.com.au/projects/albatross/wiki/ExampleCode

Eric> I noticed and thank you.  I would add another section however.

Eric> Q: how do I get a complete list of all checkboxes checked and
Eric> unchecked just like you can with ordinary CGI

Eric> Personally, I think this is a shortcoming of albatross.  It is
Eric> one of the few I've encountered in my brief exposure which is
Eric> very impressive.  I am resigned to shoving a list of all
Eric> spamtrap tokens up to the page and using that information plus
Eric> the list of checkboxes checked to extract a list of what wasn't
Eric> checked.  ugly but not horrible.

Eric> what is the best way to store a chunk of data (i.e. a list)
Eric> client side?

Eric> Anyway, thanks for your taking time to write up your solution.

You might be able to get closer what you want by using the alias
attribute.

   http://www.object-craft.com.au/projects/albatross/albatross/tag-input-alias.html

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Fri Jul  4 01:01:48 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 11:01:48 +1000
Subject: [albatross-users] Consistencies in attributes and tags
In-Reply-To: <20030703084326.3F6C54EE8B@koro.off.connect.com.au>
References: <20030703084326.3F6C54EE8B@koro.off.connect.com.au>
Message-ID: <m3adbvksub.fsf@ferret.object-craft.com.au>

> There are some attributes that don't work in as many tags as I would
> like. The one that I have been irritated with was the 'noescape'
> attribute. It only applies to the <al-value> tag and not to
> <al-img>, and others.  It shouldn't be difficult to make the
> attributes consistent across all tags, where they are applicable.
> 
> This would make the <al-img> tag more useful (to me at least).
> Currently I have to ctx.locals.imgtag='<img src=...>' and expand it
> with '<al-value expr="imgtag" noescape>' rather than '<al-img
> expr="..." noescape>'.

I cannot see any problem in adding the noescape attribute to al-img.

We should probably add the ability to disable content escaping
everywhere.  What do you think Andrew?

- Dave

-- 
http://www.object-craft.com.au


From gnb at itga.com.au  Fri Jul  4 01:10:24 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Fri, 04 Jul 2003 11:10:24 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: Your message of Thu, 03 Jul 2003 20:26:42 -0400.
Message-ID: <200307040110.LAA28357@lightning.itga.com.au>

> Q: how do I get a complete list of all checkboxes checked and unchecked just 
> like you can with ordinary CGI

Presumably, you get this list from the same place you get the list of all the 
items you are adding to the form.  (i.e. you need exactly this information to 
display the page in the first place, so it can't be that hard!!)

> what is the best way to store a chunk of data (i.e. a list) client side?

If you add a variable to the session (with ctx.add_session_vars()), then it
will be there for you next time the page_process() function is called. You can
add (almost) any Python datatype to the session, so a vector of strings is
easy.  This can be expensive in bandwith if you use client-side sessions,
depending on how much data you add.  We use file-based server sessions with up
to 100k of session data to do this sort of thing with no problems at all.

Perhaps the model you want is something like this (warning: untested):

--------
def page_enter(self, ctx):
	ctx.locals.all_ids = []
	ctx.add_session_vars('all_ids')

def page_display(self, ctx):
	ctx.locals.spam_ids = []
	ctx.locals.not_spam_ids = []
	for i in self.list_of_emails():
		id = i.message_id()
		ctx.locals.all_ids.append(id)
		if i.is_spam():
			ctx.locals.spam_ids.append(id)
		elif i.is_not_spam():
			ctx.locals.not_spam_ids.append(id)
	ctx.run_template('list.htm')

def page_process(self, ctx):
	# Here, ctx.locals.all_ids is a vector of all the IDs
	# ctx.locals.spam_ids is a vector of all the IDs for which the "is spam" checkbox was selected
	# likewise, ctx.locals.not_spam_ids
	for id in ctx.locals.all_ids:
		if id in ctx.locals.spam_ids:
			mark_id_spam(id)
		elif id in ctx.locals.not_spam_ids:
			mark_id_not_spam(id)
		else:
			mark_id_unknown_status(id)

----------

Now depending on how long each vector might be, the "i in smap_ids" might be
too slow (it's O(n) in the number of items) so you could start by converting
the vector into a dict.  But this is unlikely to be an issue for any sensible
web page - your browser will explode trying to render the form well before this
becomes a bottleneck.

And you can avoid storing the all_ids vector in the session by changing your 
process function to something like this:

def page_process(self, ctx):
	for i in self.list_of_emails():
		id = i.message_id()
		if id in ctx.locals.spam_ids:
			mark_id_spam(id)
		elif id in ctx.locals.not_spam_ids:
			mark_id_not_spam(id)
		else:
			mark_id_unknown_status(id)


i.e. use the same for loop in the process function as you used in the al-for 
template that generated the form.



From andrewm at object-craft.com.au  Fri Jul  4 01:18:46 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Fri, 04 Jul 2003 11:18:46 +1000
Subject: [albatross-users] Consistencies in attributes and tags
In-Reply-To: Message from Dave Cole <djc@object-craft.com.au>
 of "04 Jul 2003 11:01:48 +1000." <m3adbvksub.fsf@ferret.object-craft.com.au>
References: <20030703084326.3F6C54EE8B@koro.off.connect.com.au>  <m3adbvksub.fsf@ferret.object-craft.com.au>
Message-ID: <20030704011847.F0F923C332@coffee.object-craft.com.au>

>
>> There are some attributes that don't work in as many tags as I would
>> like. The one that I have been irritated with was the 'noescape'
>> attribute. It only applies to the <al-value> tag and not to
>> <al-img>, and others.  It shouldn't be difficult to make the
>> attributes consistent across all tags, where they are applicable.
>> 
>> This would make the <al-img> tag more useful (to me at least).
>> Currently I have to ctx.locals.imgtag='<img src=...>' and expand it
>> with '<al-value expr="imgtag" noescape>' rather than '<al-img
>> expr="..." noescape>'.
>
>I cannot see any problem in adding the noescape attribute to al-img.
>
>We should probably add the ability to disable content escaping
>everywhere.  What do you think Andrew?

Sounds reasonable.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From djc at object-craft.com.au  Fri Jul  4 01:48:34 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 11:48:34 +1000
Subject: [albatross-users] Consistencies in attributes and tags
In-Reply-To: <20030704011847.F0F923C332@coffee.object-craft.com.au>
References: <20030703084326.3F6C54EE8B@koro.off.connect.com.au>
 <m3adbvksub.fsf@ferret.object-craft.com.au>
 <20030704011847.F0F923C332@coffee.object-craft.com.au>
Message-ID: <m3znjvjc3x.fsf@ferret.object-craft.com.au>

--=-=-=

>>>>> "Andrew" == Andrew McNamara <andrewm@object-craft.com.au> writes:

>>> There are some attributes that don't work in as many tags as I
>>> would like. The one that I have been irritated with was the
>>> 'noescape' attribute. It only applies to the <al-value> tag and
>>> not to <al-img>, and others.  It shouldn't be difficult to make
>>> the attributes consistent across all tags, where they are
>>> applicable.
>>> 
>>> This would make the <al-img> tag more useful (to me at least).
>>> Currently I have to ctx.locals.imgtag='<img src=...>' and expand
>>> it with '<al-value expr="imgtag" noescape>' rather than '<al-img
>>> expr="..." noescape>'.
>>  I cannot see any problem in adding the noescape attribute to
>> al-img.
>> 
>> We should probably add the ability to disable content escaping
>> everywhere.  What do you think Andrew?

Andrew> Sounds reasonable.

OK, here is a patch to enable noescape in <al-input value="...">,
<al-img src="...">, <al-select optionexpr="..."> (values returned by
expression), and <al-textarea>.

It is completely untested - but will be in the next release.

- Dave


--=-=-=
Content-Type: text/x-patch
Content-Disposition: attachment; filename=noescape.patch

Index: albatross/tags.py
===================================================================
RCS file: /usr/local/cvsroot/object-craft/albatross/albatross/tags.py,v
retrieving revision 1.121
diff -u -r1.121 tags.py
--- albatross/tags.py	23 Jun 2003 05:10:57 -0000	1.121
+++ albatross/tags.py	4 Jul 2003 01:47:12 -0000
@@ -142,7 +142,10 @@
         if name:
             ctx.write_content(' name="%s"' % name)
         if value is not None:
-            ctx.write_content(' value="%s"' % escape(value))
+            if self.has_attrib('noescape'):
+                ctx.write_content(' value="%s"' % value)
+            else:
+                ctx.write_content(' value="%s"' % escape(value))
         if not self.has_attrib('disabled'):
             ctx.input_add(self.__itype, name, value, self.has_attrib('list'))
 
@@ -248,7 +251,11 @@
     def to_html(self, ctx):
         ctx.write_content('<img')
         self.write_attribs_except(ctx, 'expr', 'src')
-        ctx.write_content(' src="%s"' % escape(self.eval_attrib(ctx, 'expr')))
+        src = self.eval_attrib(ctx, 'expr')
+        if self.has_attrib('noescape'):
+            ctx.write_content(' src="%s"' % src)
+        else:
+            ctx.write_content(' src="%s"' % escape(src))
         ctx.write_content('>')
 
 # Using the albatross <FORM> tag allows Albatross to record the
@@ -369,7 +376,10 @@
         if type(item) is type(()):
             ctx.write_content(' value="%s"' % escape(value))
         ctx.write_content('>')
-        ctx.write_content(escape(text))
+        if self.has_attrib('noescape'):
+            ctx.write_content(text)
+        else:
+            ctx.write_content(escape(text))
         ctx.write_content('</option>')
         ctx.write_content('\n')
 
@@ -395,6 +405,8 @@
             value = ctx.get_value(name)
             if not value:
                 ctx.write_content('')
+            elif self.has_attrib('noescape'):
+                ctx.write_content(value)
             else:
                 ctx.write_content(escape(value))
         else:

--=-=-=


-- 
http://www.object-craft.com.au

--=-=-=--


From djc at object-craft.com.au  Fri Jul  4 02:45:04 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 12:45:04 +1000
Subject: [albatross-users] Standalone HTTP Albatross server
In-Reply-To: <1057006507.14170.12.camel@localhost>
References: <1056331020.29342.79.camel@localhost>
 <m365mpxi2u.fsf@ferret.object-craft.com.au>
 <1057006507.14170.12.camel@localhost>
Message-ID: <m3r857j9hr.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> On Sun, 2003-06-29 at 12:02, Dave Cole wrote:
>> Andrew, do you think we should put this into Albatross now?

Matt> I just remembered that there is a possible bug in Albatross
Matt> 1.10pre1 that breaks this. I can send a patch against 1.10pre1
Matt> if it would help?

Matt> See here for more:
Matt> http://object-craft.com.au/pipermail/albatross-users/2003-June/000460.html

Yup.  I am supposed to fix that this weekend so I can release
1.10pre2...  You are more than welcome to beat me to it...

- Dave

-- 
http://www.object-craft.com.au


From dougal.scott at member.sage-au.org.au  Fri Jul  4 06:50:11 2003
From: dougal.scott at member.sage-au.org.au (dougal.scott@member.sage-au.org.au)
Date: Fri, 04 Jul 2003 16:50:11 +1000
Subject: [albatross-users] Newbie question
Message-ID: <20030704065011.7E8784EE3E@koro.off.connect.com.au>

In order to better learn Albatross I'm converting an existing cgi. This
is a fairly simple cgi which gets some info from the user and then
calls itself a few times with different options to generate some
embedded images.

I'm having problems converting this to Albatross. How do I repeatedly
pass a different context to another module.

I can do it the bad way by generating an expression that is an img tag
such as  "<img src=.../genimg?imgtype=1&var1=foo&var2-bar>". This
involves going through the locals and generating the arguments from
that - and adding some other ones, in this case imgtype.

However this is a hack and I would prefer to do it neatly. Also it
means that the context is accepted from the URL, where I would prefer
it not to be, so people can't play with the options.

This is where either Albatross or my limited understanding of Albatross
breaks down. Albatross seems to be based around the concept of a page -
and one invocation of a module per page. What I need is multiple
modules being invoked to generate a single page.

Can anyone point me in the right direction?

--
Dougal Scott                    Connect Internet Solutions
dwagon@connect.com.au           180-188 Burnley Street
Senior Systems Administrator    Richmond, Australia, 3121

From gnb at itga.com.au  Fri Jul  4 08:09:00 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Fri, 04 Jul 2003 18:09:00 +1000
Subject: [albatross-users] Newbie question
In-Reply-To: Your message of Fri, 04 Jul 2003 16:50:11 +1000.
Message-ID: <200307040809.SAA18580@lightning.itga.com.au>

> In order to better learn Albatross I'm converting an existing cgi. This
> is a fairly simple cgi which gets some info from the user and then
> calls itself a few times with different options to generate some
> embedded images.

Hmmm... AFAICT this is not the sort of application where Albatross shines, so
I'm not real surprised you are running into some conceptual issues.  About the
best thing that springs to mind is to have the main app generate the images
into a temp directory before spitting out the HTML, then reference them with a
call to a simple (non-Albatross) cgi that simply returns the cached image.

Another approach might be to use a Random app which is much more
tolerant of unusual page orders.   Have a "page" called "images" in the
trandom app that doesn't update the session state, but creates and
returns the requested image based on session info, then refer to them
from the "main" page with something like "..../application?image=1"
(Random apps are not as well documented as the non-Random versions, and
they have quite a different design.)



From matt at pollenation.net  Fri Jul  4 09:20:02 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 04 Jul 2003 10:20:02 +0100
Subject: [albatross-users] Standalone HTTP Albatross server
In-Reply-To: <m3r857j9hr.fsf@ferret.object-craft.com.au>
References: <1056331020.29342.79.camel@localhost>	<m365mpxi2u.fsf@ferret.object-craft.com.au>	<1057006507.14170.12.camel@localhost> <m3r857j9hr.fsf@ferret.object-craft.com.au>
Message-ID: <3F0546C2.2010606@pollenation.net>

This is a multi-part message in MIME format.
--------------060906010806030007050909
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Dave Cole wrote:

>Yup.  I am supposed to fix that this weekend so I can release
>1.10pre2...  You are more than welcome to beat me to it...
>
Patch attached, it should apply to the 1.10pre release using "cat 
base_url.patch | patch --strip 0".

Cheers, Matt

p.s. how do you avoid the need for --strip?

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com


--------------060906010806030007050909
Content-Type: text/plain;
 name="base_url.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="base_url.patch"

diff -ru ./albatross/app.py ../albatross-1.10pre1-patched/albatross/app.py
--- ./albatross/app.py	2003-06-07 06:16:51.000000000 +0100
+++ ../albatross-1.10pre1-patched/albatross/app.py	2003-07-04 10:10:34.000000000 +0100
@@ -206,7 +206,11 @@
     def __init__(self, base_url):
         ResourceMixin.__init__(self)
         apply(self.register_tagclasses, tags.tags)
-        self.__base_url = base_url
+        # Use a base_url of '/' if not provided
+        if base_url:
+            self.__base_url = base_url
+        else:
+            self.__base_url = '/'
 
     def run(self, req):
         '''Process a single browser request
diff -ru ./albatross/randompage.py ../albatross-1.10pre1-patched/albatross/randompage.py
--- ./albatross/randompage.py	2003-01-27 10:23:27.000000000 +0000
+++ ../albatross-1.10pre1-patched/albatross/randompage.py	2003-07-04 09:56:27.000000000 +0100
@@ -21,7 +21,7 @@
         try:
             base_path = urlparse.urlparse(self.base_url())[2]
             uri_path = urlparse.urlparse(uri)[2]
-            page = string.split(uri_path, base_path + '/', 1)[1]
+            page = string.split(uri_path, base_path, 1)[1]
         except IndexError:
             pass
         if not page:

--------------060906010806030007050909--


From matt at pollenation.net  Fri Jul  4 09:52:20 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 04 Jul 2003 10:52:20 +0100
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <m365mjda2e.fsf@ferret.object-craft.com.au>
References: <1057196415.18669.41.camel@localhost> <m365mjda2e.fsf@ferret.object-craft.com.au>
Message-ID: <3F054E54.5000308@pollenation.net>

Please feel free to include it in Albatross, that would obviously 
benefit me. I think the only thing that is missing is the file upload 
support but I still hope to fix that.

I don't really have a strong feeling on your change but it seems that 
making the application's module create an app object may make it less 
flexible although your solution is definitely the simplest and, as you 
point out, involves only a trivial change to the code.

An even better solution might be to split al-httpd into a module, say 
albatross.server.httpd, leaving only the __main__ code in al-httpd 
itself. Applications could then create an instance of the http server 
manually *if necessary* or use al-httpd in the same way as now. I see 
the mod_python vs cgi vs fcgi vs httpd as a deployment decision and 
nothing much to do with the application code itself or in other words 
... who hows how someone might want to deploy an application based on 
Albatross on their server.

I'm not sure if I have explained that too well but what I can imagine 
doing is having a series of startup scripts for the application 
depending on how I'm using it (i.e. dev vs prod). Sometime like this 
(completely untested):

-- standalone.py --
from albatross.server import httpd
import myapp

app = myapp.App(base_url = '/', base_dir = '/home/matt/src/myapp')

httpd = httpd.HTTPServer(app, 8000)
httpd.serve_forever()
--

-- app.py --
from albatross import cgiapp
import myapp

app = myapp.App(base_url = '/dev/myapp/app.py', base_dir = 
'/home/matt/src/myapp')
app.run(cgiapp.Request())
--

etc

What are your thoughts on this? I will split the al-httpd script up 
today, as detailed above, if you think it's a good approach, or if you 
want to keep it simple, I'm happy with that too.

Cheers, Matt

Dave Cole wrote:

>>Attached is an updated version of the standalone server with the
>>following changes:
>>
>>      * Fixed to support forms containing input fields with a "list"
>>        attribute. It only ever returned the first value in the get or
>>        post data before.
>>      * Added the current working directory to the path so that the
>>        httpd.py script can live anywhere.
>>      * changed the filename to al-httpd. An even better name might be
>>        al-simple-server.
>>    
>>
>
>It works beautifully.  I would love to add this to Albatross to sit
>next to your FastCGI code.
>
>I made one small change to your code to make it a bit more flexible (I
>hope).  I let the application mainline construct the application
>object.
>
>--- al-httpd-orig	2003-07-03 23:09:30.000000000 +1000
>+++ al-httpd	2003-07-03 23:07:18.000000000 +1000
>@@ -149,7 +149,7 @@
> 
>     # Load the application's module and create an application instance
>     module = __import__(appClass[:appClass.rindex('.')])
>-    app = getattr(module,appClass[appClass.rindex('.')+1:])()
>+    app = getattr(module,appClass[appClass.rindex('.')+1:])
> 
>     # Create the HTTP server
>     httpd = HTTPServer(app, port)
>
>Now with the most trivial change to the form4 Albatross sample it
>works fantastically.
>
>--- samples/form4/form.py	2002-08-17 23:14:36.000000000 +1000
>+++ form.py	2003-07-03 23:13:40.000000000 +1000
>@@ -19,4 +19,5 @@
>                 secret = '-=-secret-=-')
> app.register_page('form', Form())
> 
>-app.run(Request())
>+if __name__ == '__main__':
>+    app.run(Request())
>
>Now due to you code I can do this:
>
>$ al-httpd form.app 8080
>
>That is cool!
>
>- Dave
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com
t: 0113 2252500



From andrewm at object-craft.com.au  Fri Jul  4 10:52:09 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Fri, 04 Jul 2003 20:52:09 +1000
Subject: [albatross-users] Newbie question
In-Reply-To: Message from dougal.scott@member.sage-au.org.au
 of "Fri, 04 Jul 2003 16:50:11 +1000." <20030704065011.7E8784EE3E@koro.off.connect.com.au>
References: <20030704065011.7E8784EE3E@koro.off.connect.com.au>
Message-ID: <20030704105209.793393C332@coffee.object-craft.com.au>

>I can do it the bad way by generating an expression that is an img tag
>such as  "<img src=.../genimg?imgtype=1&var1=foo&var2-bar>". This
>involves going through the locals and generating the arguments from
>that - and adding some other ones, in this case imgtype.

That suggests to me that you're using the hidden field session mixin -
if you use a server side session mixin, the session will be available
when the image is loaded (just be aware that the browser will probably
try to load multiple images at the same time, resulting in multiple
processes accessing the one session, which can cause a mess).

A similar problem occurs in applications using frames - it can be done,
although more care than usual is needed (all frames dispatch through one
albatross page, with a hidden field in each frame signalling to the app 
which is being loaded).

The alternative you suggest is okay - I've used it in apps. If security
is a problem, you can use md5 hashes to create a hard to guess expression.

Greg Bond's suggestion is also a good one which I've used in Albatross
apps before: the main application generates all the images and places
them where they can be fetched as htdocs (rather than cgi). Again, md5
hashes can be used to obscure filenames. Some mechanism will be needed
to clean the files up after use (older than 5 minutes, etc). This option
will result in less load on the web server.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From esj at harvee.org  Fri Jul  4 12:17:56 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Fri, 04 Jul 2003 08:17:56 -0400
Subject: [albatross-users] check boxes again.
In-Reply-To: <m3llvfkthm.fsf@ferret.object-craft.com.au>
References: <200307040004.KAA08163@lightning.itga.com.au>	<3F04C9C2.1020308@harvee.org> <m3llvfkthm.fsf@ferret.object-craft.com.au>
Message-ID: <3F057074.7070205@harvee.org>

Dave Cole wrote:

> You might be able to get closer what you want by using the alias
> attribute.
> 
>    http://www.object-craft.com.au/projects/albatross/albatross/tag-input-alias.html

I'm probably being entirely too dim again but it looks like alias is useful for 
unique names for the checkbox.  I'm not sure how it answers my questions about 
getting checkbox is in the right state.  If you look at:

http://harvee.org/greeneggs/retrain.cgi

you will see that I've managed to accomplish this goal using the following 
technique.

...
<al-if expr="line.value().checked=='on'">
     <al-exec expr="albatross_token = line.value().token">
<al-else>
     <al-exec expr="albatross_token = 'off'">
</al-if>
<al-input type="checkbox" name="albatross_token"  valueexpr="line.value().token" 
checked  list onclick="CCA(this,'mystery');DownReset();" >
...

and on submit, I get a list of tokens that have been checked.  If I follow 
suggestions given another messages, I should be able to store a list of all 
tokens as a hidden or session variable and then by simple process of 
illumination, figure out what wasn't checked.  Like I said, ugly, tedious, but 
not horrible.  I can live with it.  After all, the goal is to get the job done 
so I can get out the next release of this antispam system.

I still need to solve how to dynamically shift the auto approval and auto trash 
limits as one trains the Bayesian portion of the system.  I keep trying to 
reduce the administrator's workload and I've put far more work into the Bayesian 
side than the sender pays side because Bayesian filters are that much more 
flaky/touchy-feely/squishy etc.

but seriously Dave, please consider adding some form of option to allow for 
direct setting of checked or not checked (extension of checked attribute 
perhaps) and for listing all checkboxes.

---eric

onward and upward




From djc at object-craft.com.au  Fri Jul  4 13:11:04 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 23:11:04 +1000
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <3F054E54.5000308@pollenation.net>
References: <1057196415.18669.41.camel@localhost>
 <m365mjda2e.fsf@ferret.object-craft.com.au>
 <3F054E54.5000308@pollenation.net>
Message-ID: <m37k6ycu8n.fsf@ferret.object-craft.com.au>

> Please feel free to include it in Albatross, that would obviously
> benefit me. I think the only thing that is missing is the file
> upload support but I still hope to fix that.
> 
> I don't really have a strong feeling on your change but it seems
> that making the application's module create an app object may make
> it less flexible although your solution is definitely the simplest
> and, as you point out, involves only a trivial change to the code.

Your original the harness expected the mainline to define an
application class constructor that takes zero arguments.  I figured
that since a CGI or mod_python app has to create an application object
in the mainline anyway, why not just reference that.

In other words, all mainlines do something like this:

app = AppClass(args_if_any)

> An even better solution might be to split al-httpd into a module,
> say albatross.server.httpd, leaving only the __main__ code in
> al-httpd itself.

LOL.  That is what I did with your first version before I had
understood your al-httpd intention :-).

> Applications could then create an instance of the http server
> manually *if necessary* or use al-httpd in the same way as now.  I
> see the mod_python vs cgi vs fcgi vs httpd as a deployment decision
> and nothing much to do with the application code itself or in other
> words ... who hows how someone might want to deploy an application
> based on Albatross on their server.

Exactly!

The al-httpd program and httpdapp module will be an excellent addition
to Albatross.

> I'm not sure if I have explained that too well but what I can
> imagine doing is having a series of startup scripts for the
> application depending on how I'm using it (i.e. dev vs
> prod). Sometime like this (completely untested):
> 
> -- standalone.py --
> from albatross.server import httpd
> import myapp
> 
> app = myapp.App(base_url = '/', base_dir = '/home/matt/src/myapp')
> 
> httpd = httpd.HTTPServer(app, 8000)
> httpd.serve_forever()
> --
> 
> -- app.py --
> from albatross import cgiapp
> import myapp
> 
> app = myapp.App(base_url = '/dev/myapp/app.py', base_dir =
> '/home/matt/src/myapp')
> app.run(cgiapp.Request())
> --

> What are your thoughts on this? I will split the al-httpd script up
> today, as detailed above, if you think it's a good approach, or if
> you want to keep it simple, I'm happy with that too.

I agree 100% that it should be split into an importable module and a
program to launch applications.

Andrew, do you have any excellent ideas about naming of the program
and module?

Off the top of my head I would like to rename the current
session-server directory to tools and place the new al-httpd program
in there.  I also think that the importable module should probably be
called httpdapp and should live side by side with fcgiapp, apacheapp,
and cgiapp.

Matt, this is different to what you have said above, but I think more
consistent with the current layout.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Fri Jul  4 13:32:45 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 04 Jul 2003 14:32:45 +0100
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <m37k6ycu8n.fsf@ferret.object-craft.com.au>
References: <1057196415.18669.41.camel@localhost>	<m365mjda2e.fsf@ferret.object-craft.com.au>	<3F054E54.5000308@pollenation.net> <m37k6ycu8n.fsf@ferret.object-craft.com.au>
Message-ID: <3F0581FD.3060205@pollenation.net>

Dave Cole wrote:

>>What are your thoughts on this? I will split the al-httpd script up
>>today, as detailed above, if you think it's a good approach, or if
>>you want to keep it simple, I'm happy with that too.
>>    
>>
>
>I agree 100% that it should be split into an importable module and a
>program to launch applications.
>
>Andrew, do you have any excellent ideas about naming of the program
>and module?
>
>Off the top of my head I would like to rename the current
>session-server directory to tools and place the new al-httpd program
>in there.  I also think that the importable module should probably be
>called httpdapp and should live side by side with fcgiapp, apacheapp,
>and cgiapp.
>
>Matt, this is different to what you have said above, but I think more
>consistent with the current layout.
>
>- Dave
>
>  
>
Cool, I think this is the best solution overall - the simplicity of 
al-httpd with the flexibility of the module. If you're happy to make 
these changes then go ahead but just say if you want me to do anything 
to help.

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From djc at object-craft.com.au  Fri Jul  4 13:28:08 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 04 Jul 2003 23:28:08 +1000
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <3F0581FD.3060205@pollenation.net>
References: <1057196415.18669.41.camel@localhost>
 <m365mjda2e.fsf@ferret.object-craft.com.au>
 <3F054E54.5000308@pollenation.net>
 <m37k6ycu8n.fsf@ferret.object-craft.com.au>
 <3F0581FD.3060205@pollenation.net>
Message-ID: <m3k7aybevr.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Cool, I think this is the best solution overall - the simplicity
Matt> of al-httpd with the flexibility of the module. If you're happy
Matt> to make these changes then go ahead but just say if you want me
Matt> to do anything to help.

I will wait to see what Andrew says.  He always pokes holes in my
stoopid ideas.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sat Jul  5 01:07:49 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Fri, 04 Jul 2003 21:07:49 -0400
Subject: [albatross-users] check boxes again.
In-Reply-To: <200307040110.LAA28357@lightning.itga.com.au>
References: <200307040110.LAA28357@lightning.itga.com.au>
Message-ID: <3F0624E5.7080104@harvee.org>

Gregory Bond wrote:

>>Q: how do I get a complete list of all checkboxes checked and unchecked just 
>>like you can with ordinary CGI
> 
> 
> Presumably, you get this list from the same place you get the list of all the 
> items you are adding to the form.  (i.e. you need exactly this information to 
> display the page in the first place, so it can't be that hard!!)

I do.  When I build the display list, I start out first by scanning the entire 
set of files in the spamtrap.  Then, I pass each file (containing a single 
e-mail message) to an object which extracts necessary information from the 
headers and if any errors occur, fill in with appropriate defaults.  This object 
defines" a line" in the display.

Each of these line objects are placed on a dictionary.  In the process of 
putting line objects in a dictionary, I build a another list (AKA sort list) 
containing token and some value to sort on.  After sorting, the top N entries 
are are extracted from the sort list and the token values are used to extract 
line objects from the dictionary.  These line objects are then placed on a list 
which is given to locals for use in the display template.  Simple as 3.14159.

> 
> 
>>what is the best way to store a chunk of data (i.e. a list) client side?
> 
> 
> If you add a variable to the session (with ctx.add_session_vars()), then it
> will be there for you next time the page_process() function is called. You can
> add (almost) any Python datatype to the session, so a vector of strings is
> easy.  This can be expensive in bandwith if you use client-side sessions,
> depending on how much data you add.  We use file-based server sessions with up
> to 100k of session data to do this sort of thing with no problems at all.

this is good to know.  It shouldn't be bad.  I only expect to see 20 or 30 20 
byte entries.

> 
> Perhaps the model you want is something like this (warning: untested):

thanks for the clue.  not only did it teach me more about when to use 
page_display and page_process, but it also gave me a clue on how to handle the data.

here's my working test fragment.  All_tokens is a complete list of all tokens 
entered into the display list.  Albatross_tokens is a list return by the 
checkbox list of checkboxes checked:

             for token in context.locals.all_tokens:
                 #syslog.syslog("working on %s"% token)
                 if token in context.locals.albatross_token:
                     syslog.syslog("token is spam %s"% token)
                 else:
                     syslog.syslog("token is NOT spam %s"% token)

cool thing is I get a usable data set and since the data set is expected to be 
in the sub 100 entry range, it shouldn't suck too bad.  The worst bit is 
processing every single piece of e-mail in the spamtrap for headers etc. I 
shudder to think how bad as it going to be when processing the dumpster (where 
spamtrap entries go to die) according to age.

tomorrow, I'm going to complete moving over the Bayesian discriminator 
correction code from ad hoc CGI to albatross and declare victory.

Many thanks to all you who have helped.  I could not have done without you and I 
hope my commentary helped improve documentation.  If there's interest, I will 
continue the cognitive disconnect series on subsequent chapters.

---eric




From korg at darkqueen.org  Sat Jul  5 01:57:23 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Sat, 05 Jul 2003 11:57:23 +1000
Subject: [albatross-users] Re: Albatross-users digest, Vol 1 #186 - 12 msgs
In-Reply-To: Your message of "Fri, 04 Jul 2003 16:51:01 +1000."
 <20030704065101.28833.86489.Mailman@numbat.object-craft.com.au>
Message-ID: <200307050157.h651vOo28275@firewall.darkqueen.org>

From: dougal.scott@member.sage-au.org.au:
  | Date: Fri, 04 Jul 2003 16:50:11 +1000
  | Subject: [albatross-users] Newbie question
  | 
  | 
  | In order to better learn Albatross I'm converting an existing cgi. This
  | is a fairly simple cgi which gets some info from the user and then
  | calls itself a few times with different options to generate some
  | embedded images.
  | 
  | I'm having problems converting this to Albatross. How do I repeatedly
  | pass a different context to another module.
  | 
  | I can do it the bad way by generating an expression that is an img tag
  | such as  "<img src=.../genimg?imgtype=1&var1=foo&var2-bar>". This
  | involves going through the locals and generating the arguments from
  | that - and adding some other ones, in this case imgtype.
  | 
  | However this is a hack and I would prefer to do it neatly. Also it
  | means that the context is accepted from the URL, where I would prefer
  | it not to be, so people can't play with the options.
  | 
  | This is where either Albatross or my limited understanding of Albatross
  | breaks down. Albatross seems to be based around the concept of a page -
  | and one invocation of a module per page. What I need is multiple
  | modules being invoked to generate a single page.
  | 
  | Can anyone point me in the right direction?



I uses something like this [below].... (it may need hacking to get going, but 
you should be able modify the dict in application to get it starting..

(Oh and modify the urls that are coded into the template).

It uses the store_server_side_file (SimpleSessionFileApp) to store 
context, and uses a URL hack to flag image requests.

The function:

      data=image_generator.map_plot()

returns a png. You could pass ctx.locals.whatever to it if you need
to or do the processing in the page_display() etc.

Often I subclass all my albatross pages
from another class so that image checking is less work.

This works for me :). 


PLEASE tell me if Im doing something wrong or abusive of albatross ;)


I think it is fine because the template page loads first (and processes 
any ctx/form changes and then it requests the image, so the image page 
has the latest context as generated from the last template page). To 
be really save, I guess you should skip any processing in page_process() 
if it is an image.

Actually, Ive been meaning to ask, is there a way to flush data out of
a request and NOT store/process the request? Ie, if it is an image,
then just write the data and dont do anything with the context (like
save or modify it?).


  | Dougal Scott                    Connect Internet Solutions
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
Don't go to bed with no price on your head.
		-- Baretta



------------------- 8< cut -----------------------
#!/usr/bin/python

import os,string,re
import image_gen

from albatross        import SimpleSessionFileApp  as Albatross_Application
from albatross        import SessionFileAppContext as Albatross_Context
from albatross.cgiapp import Request               as Albatross_Request

# I used to use this on the url before the click data context
# COORDS=re.compile("(?P<x>[0-9][0-9]*),(?P<y>[0-9][0-9]*)")

image_generator=image_gen.load('test')


PageList={}

def addpage(p):
  PageList[ p.__class__.__name__ ]=p
def DEBUG(msg):
  open("/tmp/debug",'a').write('%s\n'%(msg))



class Start:
  def page_enter(self, ctx):
    DEBUG("page entered!!")
    ctx.locals.username=None
    ctx.locals.passwd=None
    ctx.locals.square='morons'
    ctx.locals.pos=[0,0]
    ctx.locals.click=None
    ctx.locals.zoom=None
    ctx.add_session_vars('username','passwd','square','pos','click')

  def  page_process(s,ctx):
    DEBUG('b:%s'%(ctx.get_uri()))
    DEBUG('b:%s'%(ctx.base_url()))
    DEBUG('c:%s'%(ctx.current_url()))
    if ctx.locals.zoom:
      DEBUG("setting zoom = %s"%(ctx.locals.zoom))
      image_generator.set_zoom(ctx.locals.zoom)
    if ctx.locals.click:
      ctx.locals.pos=[ ctx.locals.pos[0]+ctx.locals.click.x,
                       ctx.locals.pos[1]+ctx.locals.click.y ]
      ctx.locals.click=None
    if ctx.locals.square=='tree':
      DEBUG('tree!!!')
    DEBUG('%s'%(ctx.locals))

  def page_display(self, ctx):
    page=os.environ['REQUEST_URI'][len(os.environ['SCRIPT_NAME'])+1:]
    DEBUG("page is %s"%(page))
    if page=='map_image':
      data=image_generator.map_plot()
      ctx.request.write_header('Content-Type','image/png') 
      ctx.request.write_header('Content-Length',str(len(data)))
      ctx.request.write_header('Pragma', 'no-cache')
      DEBUG("image squre %s"%(ctx.locals.square))
      ctx.write_content(data)
    else:
      DEBUG("page squre %s"%(ctx.locals.square))
      ctx.run_template('template_map.html')

#class Image:

addpage(Start())
#addpage(Image())


#############################################################
# Albatross objects
#############################################################


class request(Albatross_Request):
  def __init__(s):
    Albatross_Request.__init__(s)

class context(Albatross_Context):
  def __init__(s,app):
    Albatross_Context.__init__(s,app)


class application(Albatross_Application):
  def __init__(s):
    dict= {  'base_url':'20circle',
             'template_path':'/home/korg/projects/20circle/data/web_template/',
             'start_page':'Start',
             'secret':'-=-secret-=-',
             'session_appid':'dumb',          # server file stored
             'session_dir':'/tmp/session/',   # server file stored
             }
    Albatross_Application.__init__(s,**dict)
    if PageList:
      for n,v in PageList.items(): s.register_page( n, v )

  def create_context(s):
      ctx=context(s)
      ctx.load_template('macros.html').to_html(ctx)
      return ctx


def main():
  a = application()
  a.run(request())

if __name__=="__main__":  main()



------------------- 8< cut -----------------------

<html>
<head>
  <title>Map screen</title>
</head>

<body>


<h1>Welcome</h1>
<al-form method="post">

<hr>
<h2>Games Details:</h2>

pos: <al-value expr="username"><br>
pos: <al-value expr="passwd"><br>
pos: <al-value expr="square"><br>
pos: <al-value expr="pos"><br>

<!-- albatross 1.01 used this and url regexes to get coords -->
<!--a href="http://10.0.0.254/cgi-bin/20circle"-->
<!--al-img ismap expr="'http://10.0.0.254/cgi-bin/20circle/map_image'"-->
<!--/a-->

<!-- albatross 1.10pre1 uses this and click data type-->
<al-input type="image" name="click" src="http://10.0.0.254/cgi-bin/20circle/map_image" whitespace>

<al-input type="submit" name="square" value="square">
<al-input type="submit" name="square" value="circle">
<al-input type="submit" name="square" value="tree">

<al-input type="submit" name="zoom" value="3.">
<al-input type="submit" name="zoom" value="2.">
<al-input type="submit" name="zoom" value="1.">
<al-input type="submit" name="zoom" value=".5">

<hr>

</al-form>

<hr>

<al-expand name="copyright"/>

</body>
</html>




From djc at object-craft.com.au  Sat Jul  5 02:53:01 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 12:53:01 +1000
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <m3k7aybevr.fsf@ferret.object-craft.com.au>
References: <1057196415.18669.41.camel@localhost>
 <m365mjda2e.fsf@ferret.object-craft.com.au>
 <3F054E54.5000308@pollenation.net>
 <m37k6ycu8n.fsf@ferret.object-craft.com.au>
 <3F0581FD.3060205@pollenation.net>
 <m3k7aybevr.fsf@ferret.object-craft.com.au>
Message-ID: <m3vfuh7khe.fsf@ferret.object-craft.com.au>

>>>>> "Dave" == Dave Cole <djc@object-craft.com.au> writes:

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:
Matt> Cool, I think this is the best solution overall - the simplicity
Matt> of al-httpd with the flexibility of the module. If you're happy
Matt> to make these changes then go ahead but just say if you want me
Matt> to do anything to help.

Dave> I will wait to see what Andrew says.  He always pokes holes in
Dave> my stoopid ideas.

Andrew didn't want me to rename the session-server directory so I have
added a new standalone-server directory.

In preparing the source I noticed some formatting differences between
Albatross and your code.  I went to PEP8 and had a read.  Albatross is
still not PEP8 clean :-(.  I might do some minor reformatting.

While placing the code into Albatross I realised that there is no
copyright on the code.  What do you want to do?

* Include a Matt Goodall BSD style copyright notice.

* Assign the copyright to us so we can just apply the same BSD style
  notice as is on the rest of the code.

Since the code is not ours you need to decide what to do here.

Anyway, I have split the mainline out so it now looks like this:

- - al-httpd - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
#!/usr/bin/env python
import os
import sys
from albatross import httpdapp

if __name__ == '__main__':
    # Append the current working directory to the path.
    sys.path.append(os.getcwd())

    # Parse the command line.
    try:
        app_module, app_object = sys.argv[1].split('.')
        port = int(sys.argv[2])
    except:
        sys.exit('usage: %s main-module.app-object port'
                 % os.path.basename(sys.argv[0]))

    # Load the application's module and get application instance.
    module = __import__(app_module)
    app = getattr(module, app_object)

    # Create the HTTP server and serve requests.
    httpd = httpdapp.HTTPServer(app, port)
    httpd.serve_forever()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 06:18:30 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 16:18:30 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: <3F0624E5.7080104@harvee.org>
References: <200307040110.LAA28357@lightning.itga.com.au>
 <3F0624E5.7080104@harvee.org>
Message-ID: <m3k7ax7ayx.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Many thanks to all you who have helped.  I could not have done
Eric> without you and I hope my commentary helped improve
Eric> documentation.  If there's interest, I will continue the
Eric> cognitive disconnect series on subsequent chapters.

Any comment on how the documentation is lacking or could be improved
is always welcome.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 06:34:08 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 16:34:08 +1000
Subject: [albatross-users] al-for and Template frustration and confusion
In-Reply-To: <1056929696.5434.427.camel@localhost>
References: <49126450.1056069096@SHEILA-LAPTOP>
 <20030620075929.E5F743C332@coffee.object-craft.com.au>
 <9948525.1056116333@SHEILA-LAPTOP>
 <m3fzltv3fq.fsf@ferret.object-craft.com.au>
 <981458372.1056843004@goddess> <1056877957.5434.265.camel@localhost>
 <m3llvlmd7l.fsf@ferret.object-craft.com.au>
 <1056929696.5434.427.camel@localhost>
Message-ID: <m3brw97a8v.fsf@ferret.object-craft.com.au>

> Introduction
> ~~~~~~~~~~~~
> I think the following are either missing or not specifically listed:
>  - simple and flexible to use (as opposed to the framework itself)

Not sure I should add a comment like that :-)

>  - lightweight enough for cgi

Added.

>  - form population and parsing

Added.

> Front Matter
> ~~~~~~~~~~~~
> Add a "See Also" link to www.fastcgi.com and maybe python.org

Added.
 
> 3.5.3 Multiple Argument Macros
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> This could probably do with a small example anyway.

Good point.  Will see what I can do.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 07:50:06 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 17:50:06 +1000
Subject: [albatross-users] Standalone HTTP Albatross server
In-Reply-To: <3F0546C2.2010606@pollenation.net>
References: <1056331020.29342.79.camel@localhost>
 <m365mpxi2u.fsf@ferret.object-craft.com.au>
 <1057006507.14170.12.camel@localhost>
 <m3r857j9hr.fsf@ferret.object-craft.com.au>
 <3F0546C2.2010606@pollenation.net>
Message-ID: <m3fzll5s5t.fsf@ferret.object-craft.com.au>

> Dave Cole wrote:
> >Yup.  I am supposed to fix that this weekend so I can release
> >1.10pre2...  You are more than welcome to beat me to it...
> >
> Patch attached, it should apply to the 1.10pre release using "cat
> base_url.patch | patch --strip 0".
> 
> Cheers, Matt
> 
> p.s. how do you avoid the need for --strip?

Applied the patch.  Thanks.

What I usually do is make sure that my version of the source and the
original version share the same parent directory.  I can then create a
patch from the parent directory like this:

$ diff -ru albatross-orig albatross > albatross.patch

Then the patch can be applied by doing this from inside the unpacked
albatross directory, regardless of whether it is called
albatross-orig, albatross-1.10pre1, or albatross.

$ patch -sp1 < albatross.patch

Someone else might be able to come up with better instructions.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 08:14:48 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 18:14:48 +1000
Subject: [albatross-users] Comments on using examples...
In-Reply-To: <20727424.1057236559@SHEILA-LAPTOP>
References: <200307030350.h633oWc06073@firewall.darkqueen.org>
 <20727424.1057236559@SHEILA-LAPTOP>
Message-ID: <m3brw95r0n.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

Sheila> --On Thursday, July 03, 2003 1:50 PM +1000 Cameron Blackwood
Sheila> <korg@darkqueen.org> wrote:

>> I dont know about other people, but when I see an example in docs,
>> I often like to actually: 1) try it myself 2) modify it 3) use it
>> as a base for what I really want to do

Sheila> Agree.

>> Eg:
>> www.object-craft.com.au/projects/albatross/albatross/tag-comment.html
>> 
>> trying it: Hands up anyone who has used albatross in interactive
>> mode?

Point taken.  The interactive examples are a bit of a pain in the
freckle to copy and paste,

Would it be better if there was a module or class provided that made
interactive playing easier?

The albatross.Application class traps exceptions and tries to build a
tag traceback.  It might be nice if something like that was available
for playing around outside of applications.

Sheila> I primarily used it that way because of two reasons: (1) the
Sheila> thing I was trying to learn/master was shown with examples in
Sheila> interactive-mode in the documentation.  (2) it was quicker to
Sheila> do that then modify the app I'm working on for test purposes.

>> Interactive python dumps are not much use because: 1) noone would
>> ever do that

Interactive mode is a useful tool for testing ideas.

I think that with the new al-httpd program from Matt Goodall will make
some types of testing much easier.

Sheila> I assume that you mean here, that no one would use Albatross
Sheila> in production in such a manner? However, interactive is mostly
Sheila> used (by developers) for testing purposes, even when not
Sheila> working with Albatross. At least, that is my take on the
Sheila> matter (as a clearly peon developer type).

I am always firing up an interactive interpreter to check my
understanding of various things in Python.

>> 2) you cant cut & paste to test (well you can, but you have to cut
>> and paste specific bits (not the output) and you cant cut and paste
>> the multi-line examples, thanks to the pythonic '... '/'>>> ',
>> without having to be _VERY_ careful about indenting.)

Sheila> Agree that c&p is much more difficult/annoying from the
Sheila> interactive examples.

The examples are formatted like that so that a testing program can run
the samples and ensure that the correct output is produced.  This has
the benefit of ensuring that the documentation is always correct.  It
also provides more unit tests for the toolkit.

We could probably change the testing program so that it used two
files; a .py file and a .out file.  This would allow us to change
something like doctest/tags-if1 from

- - doctest/tags-if1 - - - - - - - - - - - - - - - - - - - - - - - -
>>> import albatross
>>> ctx = albatross.SimpleContext('.')
>>> ctx.locals.a = 10
>>> albatross.Template(ctx, '<magic>', '''
... <al-if expr="a < 15">
...  a (<al-value expr="a">) is less than 15.
... <al-else>
...  a (<al-value expr="a">) is greater than or equal to 15.
... </al-if>
... ''').to_html(ctx)
>>> ctx.flush_content()
a (10) is less than 15.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

to this:

- - doctest/tags-if1.py - - - - - - - - - - - - - - - - - - - - - -
import albatross

# Create an execution context for evaluating templates and place a
# value into the local namespace.
ctx = albatross.SimpleContext('.')
ctx.locals.a = 10

# Evaluate an inline template.
albatross.Template(ctx, '<magic>', '''
<al-if expr="a < 15">
 a (<al-value expr="a">) is less than 15.
<al-else>
 a (<al-value expr="a">) is greater than or equal to 15.
</al-if>
''').to_html(ctx)

# Display the template output.
ctx.flush_content()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -


- - doctest/tags-if1.out - - - - - - - - - - - - - - - - - - - - - -
a (10) is less than 15.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Oh, I think that the doctext samples should be documented too.

Sheila> I also found the amount of interactive examples in the
Sheila> documentation to be ... uh ... not what I would have preferred
Sheila> to see.  Some of it is good, or I might not have even realized
Sheila> how to run Albatross interactively (or even thought of it as
Sheila> being possible).  I do think some examples of running it
Sheila> interactively are good, in order to make this point.  Perhaps
Sheila> put it into a small section of the documentation, as it's own
Sheila> topic: "Running Albatross in Python Interactive Mode" and then
Sheila> a few examples showing how to do that. And then convert the
Sheila> rest of the interactive examples in the documentation to
Sheila> downloadable/runable files.

Good idea.

Sheila> BTW, something I never mentioned but that bugged me in the
Sheila> docs:

Sheila> The use of the '<magic>' context in the interactive
Sheila> examples. Bugged the crud out of me as to what it was, why it
Sheila> was named that way, why it worked. Eventually just accepted it
Sheila> as black magic and moved on, but wasn't really happy with
Sheila> having to do that.

It was a stupid joke.  There is nothing magic about it - it is just a
fake filename.  When you run Python interactively and case an
exception Python calls interactive input '<stdin>'.

>>> 1/0
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
ZeroDivisionError: integer division or modulo by zero

Sheila> The examples being currently uploaded to the Wiki may make up
Sheila> for the "short coming" of having so many of the
Sheila> documentation's examples shown as interactive mode.

Sheila> I also want to re-iterate that I find the Albatross product
Sheila> (and its docs) to be first rate, and only make negative
Sheila> comments at all in the hope of offering feedback for
Sheila> improvement of an already fine project.

Thanks for the feedback.  If people think that the above changes to
the doctest samples make sense then I will change them.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 09:00:23 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 19:00:23 +1000
Subject: [albatross-users] Proposed changes to Albatross placed in wiki
Message-ID: <m3vfuhxs9k.fsf@ferret.object-craft.com.au>

Rather than have all of these proposals fade into obscurity I have
created a new page where they can be catalogued and discussed.

   http://www.object-craft.com.au/projects/albatross/wiki/ProposedChanges

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Sat Jul  5 09:27:44 2003
From: matt at pollenation.net (Matt Goodall)
Date: 05 Jul 2003 10:27:44 +0100
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <m3vfuh7khe.fsf@ferret.object-craft.com.au>
References: <1057196415.18669.41.camel@localhost>
 <m365mjda2e.fsf@ferret.object-craft.com.au>
 <3F054E54.5000308@pollenation.net>
 <m37k6ycu8n.fsf@ferret.object-craft.com.au>
 <3F0581FD.3060205@pollenation.net>
 <m3k7aybevr.fsf@ferret.object-craft.com.au>
 <m3vfuh7khe.fsf@ferret.object-craft.com.au>
Message-ID: <1057397238.8364.11.camel@localhost>

On Sat, 2003-07-05 at 03:53, Dave Cole wrote:
> In preparing the source I noticed some formatting differences between
> Albatross and your code.  I went to PEP8 and had a read.  Albatross is
> still not PEP8 clean :-(.  I might do some minor reformatting.

I must read that PEP myself.

> While placing the code into Albatross I realised that there is no
> copyright on the code.  What do you want to do?
> 
> * Include a Matt Goodall BSD style copyright notice.
> 
> * Assign the copyright to us so we can just apply the same BSD style
>   notice as is on the rest of the code.
> 
> Since the code is not ours you need to decide what to do here.

I hereby assign the copyright to you to be included in Albatross under
the same licence as the rest of the project. Seems easier that way ;-).
By the way the LICENSE file says (c) 2002.

It would be nice if my name and email address appeared in the header,
similar to the header you added to the fcgi module.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sat Jul  5 09:39:23 2003
From: matt at pollenation.net (Matt Goodall)
Date: 05 Jul 2003 10:39:23 +0100
Subject: [albatross-users] al-for and Template frustration and confusion
In-Reply-To: <m3brw97a8v.fsf@ferret.object-craft.com.au>
References: <49126450.1056069096@SHEILA-LAPTOP>
 <20030620075929.E5F743C332@coffee.object-craft.com.au>
 <9948525.1056116333@SHEILA-LAPTOP>
 <m3fzltv3fq.fsf@ferret.object-craft.com.au> <981458372.1056843004@goddess>
 <1056877957.5434.265.camel@localhost>
 <m3llvlmd7l.fsf@ferret.object-craft.com.au>
 <1056929696.5434.427.camel@localhost>
 <m3brw97a8v.fsf@ferret.object-craft.com.au>
Message-ID: <1057397962.30521.24.camel@localhost>

On Sat, 2003-07-05 at 07:34, Dave Cole wrote:
> > Introduction
> > ~~~~~~~~~~~~
> > I think the following are either missing or not specifically listed:
> >  - simple and flexible to use (as opposed to the framework itself)
> 
> Not sure I should add a comment like that :-)

Hah, I'm not sure I worded that very well :). Probably one of my late
night sessions again.

I /think/ what I was trying to highlight is that the intro talks of an
"application framework which is flexible and extensible". What I was
trying (very, very badly) to say is that the application code itself is
very simple and flexible ... as opposed to the framework code. Is that
better? Nah, probably best to just ignore me on this ;-).

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sat Jul  5 09:51:01 2003
From: matt at pollenation.net (Matt Goodall)
Date: 05 Jul 2003 10:51:01 +0100
Subject: [albatross-users] Comments on using examples...
In-Reply-To: <m3brw95r0n.fsf@ferret.object-craft.com.au>
References: <200307030350.h633oWc06073@firewall.darkqueen.org>
 <20727424.1057236559@SHEILA-LAPTOP>
 <m3brw95r0n.fsf@ferret.object-craft.com.au>
Message-ID: <1057398661.8364.35.camel@localhost>

On Sat, 2003-07-05 at 09:14, Dave Cole wrote:
> Thanks for the feedback.  If people think that the above changes to
> the doctest samples make sense then I will change them.

Yeah, I think what Sheila says makes a lot of sense. I don't think I
have ever copy and pasted code from the documentation but the
interactive style code certainly makes it harder for me to read.

It might be a bit stupid sometimes but at first I prefer to write *all*
code myself using the documentation as a reference. I tend to find I
understand things better that way. That's the reason I dislike most IDE
let-us-do-it-all-for-you tools too.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From djc at object-craft.com.au  Sat Jul  5 10:27:57 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 20:27:57 +1000
Subject: [albatross-users] New version of the standalone http server
In-Reply-To: <1057397238.8364.11.camel@localhost>
References: <1057196415.18669.41.camel@localhost>
 <m365mjda2e.fsf@ferret.object-craft.com.au>
 <3F054E54.5000308@pollenation.net>
 <m37k6ycu8n.fsf@ferret.object-craft.com.au>
 <3F0581FD.3060205@pollenation.net>
 <m3k7aybevr.fsf@ferret.object-craft.com.au>
 <m3vfuh7khe.fsf@ferret.object-craft.com.au>
 <1057397238.8364.11.camel@localhost>
Message-ID: <m3ptkpxo7m.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> On Sat, 2003-07-05 at 03:53, Dave Cole wrote:
>> In preparing the source I noticed some formatting differences
>> between Albatross and your code.  I went to PEP8 and had a read.
>> Albatross is still not PEP8 clean :-(.  I might do some minor
>> reformatting.

Matt> I must read that PEP myself.

>> While placing the code into Albatross I realised that there is no
>> copyright on the code.  What do you want to do?
>> 
>> * Include a Matt Goodall BSD style copyright notice.
>> 
>> * Assign the copyright to us so we can just apply the same BSD
>> style notice as is on the rest of the code.
>> 
>> Since the code is not ours you need to decide what to do here.

Matt> I hereby assign the copyright to you to be included in Albatross
Matt> under the same licence as the rest of the project. Seems easier
Matt> that way ;-).  By the way the LICENSE file says (c) 2002.

Matt> It would be nice if my name and email address appeared in the
Matt> header, similar to the header you added to the fcgi module.

Thanks.  That makes it easy.

I have added this to the top of the files:

#
# Copyright 2003 by Object Craft P/L, Melbourne, Australia.
#
# LICENCE - see LICENCE file distributed with this software for details.
#
# AUTHOR(S)
#       Matt Goodall <matt AT pollenation DOT net>

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 10:41:40 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 20:41:40 +1000
Subject: [albatross-users] al-for and Template frustration and confusion
In-Reply-To: <m3brw97a8v.fsf@ferret.object-craft.com.au>
References: <49126450.1056069096@SHEILA-LAPTOP>
 <20030620075929.E5F743C332@coffee.object-craft.com.au>
 <9948525.1056116333@SHEILA-LAPTOP>
 <m3fzltv3fq.fsf@ferret.object-craft.com.au>
 <981458372.1056843004@goddess> <1056877957.5434.265.camel@localhost>
 <m3llvlmd7l.fsf@ferret.object-craft.com.au>
 <1056929696.5434.427.camel@localhost>
 <m3brw97a8v.fsf@ferret.object-craft.com.au>
Message-ID: <m3llvdxnkr.fsf@ferret.object-craft.com.au>

>>>>> "Dave" == Dave Cole <djc@object-craft.com.au> writes:

>> 3.5.3 Multiple Argument Macros ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> This could probably do with a small example anyway.

Dave> Good point.  Will see what I can do.

Added a small example.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul  5 11:11:34 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 05 Jul 2003 21:11:34 +1000
Subject: [albatross-users] Albatross 1.10 pre2 released
In-Reply-To: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
Message-ID: <m3znjttehl.fsf@ferret.object-craft.com.au>

I have just packaged up the 1.10pre2 release of Albatross.  This
resolves the problems encountered in the 1.10pre1 release and includes
some very useful extra functionality.

The changes since 1.10pre1 are:

* Standalone BaseHTTPServer support.

* mod_python Request now works with mod_python 2 and 3.

* Python 2.2 features have been reverted to Python 2.1 equivalents.

* Reintroduced TemplateLoadError as subclass of ApplicationError.

* Nested <al-form> tags now raise ApplicationError.

* Added noescape to <al-input>, <al-img>, <al-select>, and
  <al-textarea>.

* Implemented lazy attribute expression compilation in all tags.

* Now using Python-2.3 for building documentation.

* Minor reformatting to be closer to PEP-8.

We are now including a summary of changes between releases in the back
of the manual.

    http://www.object-craft.com.au/projects/albatross/albatross/changes.html

There still is no documentation of the FastCGI and Standalone
BaseHTTPServer support.  I will try to add some documentation for
these additions before the 1.10 final release.

Our intention is to make this the last pre release of 1.10.  This time
next week we will release the real 1.10.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sat Jul  5 12:40:18 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Sat, 05 Jul 2003 08:40:18 -0400
Subject: [albatross-users] Proposed changes to Albatross placed in wiki
In-Reply-To: <m3vfuhxs9k.fsf@ferret.object-craft.com.au>
References: <m3vfuhxs9k.fsf@ferret.object-craft.com.au>
Message-ID: <3F06C732.3070902@harvee.org>

Dave Cole wrote:

> Rather than have all of these proposals fade into obscurity I have
> created a new page where they can be catalogued and discussed.
> 
>    http://www.object-craft.com.au/projects/albatross/wiki/ProposedChanges

I've updated ClassRenaming to include the more descriptive names floated by 
Cameron.  Personally I find them more useful (and speakable) in contrast to the 
first set of proposed changes.  Since I don't know enough about the current 
classes to flush out the set, I'm hoping somebody else will

---eric



From matt at pollenation.net  Sat Jul  5 20:43:15 2003
From: matt at pollenation.net (Matt Goodall)
Date: 05 Jul 2003 21:43:15 +0100
Subject: [albatross-users] [wiki] simple example to display a list and allow sorting and
 filtering
Message-ID: <1057437185.6158.4.camel@localhost>

I just added a new example,
http://www.object-craft.com.au/projects/albatross/wiki/Data_20Presentation, that demonstrates getting a list of people from a datastore (of sorts), displaying the list, sorting the list by clicking on a table header and a simple list filter.

Could whoever added, "data presentation from a database or whatever
source comes to your mind with some form processing for sorting columns
and filtering data", to the RequestsForContent page have a look through
and see if it makes sense.

Cheers, Matt
-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sat Jul  5 23:32:41 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 00:32:41 +0100
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <m3ptkrktq7.fsf@ferret.object-craft.com.au>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
 <3F043063.3060808@harvee.org>  <m3ptkrktq7.fsf@ferret.object-craft.com.au>
Message-ID: <1057447961.6158.23.camel@localhost>

On Fri, 2003-07-04 at 01:42, Dave Cole wrote:
> >>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:
> 
> Eric> what's interesting is the difference in terms of what is
> Eric> returned.  An albatross, it's an iterator object; in Python
> Eric> iterators (as demonstrated by examples in the article referenced
> Eric> above) it's an ordinary piece of data (integer, float, string
> Eric> etc.).
> 
> One of the limitations that Albatross has to live with is that real
> iterators and generators cannot be pickled.  This means that they
> cannot be saved in a session.  If you use pagination in your <al-for>
> tag the ListIterator is placed in the session so it can "remember"
> where you were in the sequence.

Hmm, so the ListIterator is stored in the session and the ListIterator
contains a reference to the sequence being iterated over. Does that mean
that the whole sequence will be pickled, i.e. the entire result set from
a database query perhaps?

> Thinking out loud:
> 
> Currently pagination is handled by saving a ListIterator in the
> session.  The browser request then issues a nextpage or prevpage
> request and names the iterator.
> 
>     <al-a href="nextpage,i">next</al-a whitespace>
>     <al-input type="submit" nextpage="i" whitespace>
> 
> results in
> 
>     <a href="http://server/app.py/nextpage,i">next</a>
>     <input type="submit" name="nextpage,i">
> 
> Maybe a better way to do this would be to place all of the information
> required for pagination into the browser request.  This would result
> in output that looked something like this
> 
>     <a href="http://server/app.py/goto,i,20">next</a>
>     <input type="submit" name="goto,i,20">
> 
> The nice outcome of this would be that we do not need to save anything
> in the session anymore.  It would also allow you to bookmark specific
> pages in an application.

I was thinking of implementing this just the other day although my
reason for wanting purely request based paging was that clicking the
next page link and then using the browser's refresh button makes the
ListIterator jump forward another page.

My only thought so far on this matter is to add a scope attribute to the
<al-for> tag whose default value is "session", "request" being the other
choice.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sun Jul  6 00:02:18 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 01:02:18 +0100
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <1057447961.6158.23.camel@localhost>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
 <3F043063.3060808@harvee.org>  <m3ptkrktq7.fsf@ferret.object-craft.com.au>
 <1057447961.6158.23.camel@localhost>
Message-ID: <1057449738.28875.28.camel@localhost>

On Sun, 2003-07-06 at 00:32, Matt Goodall wrote:
> On Fri, 2003-07-04 at 01:42, Dave Cole wrote:
> > >>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:
> > 
> > Eric> what's interesting is the difference in terms of what is
> > Eric> returned.  An albatross, it's an iterator object; in Python
> > Eric> iterators (as demonstrated by examples in the article referenced
> > Eric> above) it's an ordinary piece of data (integer, float, string
> > Eric> etc.).
> > 
> > One of the limitations that Albatross has to live with is that real
> > iterators and generators cannot be pickled.  This means that they
> > cannot be saved in a session.  If you use pagination in your <al-for>
> > tag the ListIterator is placed in the session so it can "remember"
> > where you were in the sequence.
> 
> Hmm, so the ListIterator is stored in the session and the ListIterator
> contains a reference to the sequence being iterated over. Does that mean
> that the whole sequence will be pickled, i.e. the entire result set from
> a database query perhaps?

The answer is ... no. Phew ;-).

ListIterator has a __getstate__ method so only the start and pagesize
are pickled. 

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sun Jul  6 00:17:52 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 01:17:52 +0100
Subject: [albatross-users] check boxes again.
In-Reply-To: <3F04C9C2.1020308@harvee.org>
References: <200307040004.KAA08163@lightning.itga.com.au>
 <3F04C9C2.1020308@harvee.org>
Message-ID: <1057450672.19240.37.camel@localhost>

On Fri, 2003-07-04 at 01:26, Eric S. Johansson wrote:

> Q: how do I get a complete list of all checkboxes checked and unchecked just 
> like you can with ordinary CGI

Are you sure CGI does this? As far as I was aware browsers only send a
request parameter for a checkbox when it is checked. If the browser
doesn't send a request parameter how can CGI even know there were
unchecked checkboxes on the page? Perhaps the browser behaves
differently when the form data is posted?

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sun Jul  6 00:49:03 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 01:49:03 +0100
Subject: [albatross-users] check boxes again.
In-Reply-To: <3F057074.7070205@harvee.org>
References: <200307040004.KAA08163@lightning.itga.com.au>
 <3F04C9C2.1020308@harvee.org> <m3llvfkthm.fsf@ferret.object-craft.com.au>
 <3F057074.7070205@harvee.org>
Message-ID: <1057452543.19240.53.camel@localhost>

On Fri, 2003-07-04 at 13:17, Eric S. Johansson wrote:
> but seriously Dave, please consider adding some form of option to allow for 
> direct setting of checked or not checked (extension of checked attribute 
> perhaps) and for listing all checkboxes.

Isn't this exactly the sort of thing the NameRecorderMixin could do? It
is already told about every input that is added to a form so it could
potentially track the values of inputs with the list attribute, pickle
that into the hidden form form field and use it again when merge_request
is called.

There are two problems I can see with this though:

      * For large forms, the amount of hidden data would increase
        significantly.
      * The state of the checkboxes is passed to page_process as a list
        of the checked values. That would probably have to change to a
        dictionary to pass checked and unchecked in the same structure.
        That's a nasty change to make :(.

Just my thoughts ... I may have completely misunderstood the
NameRecorderMixin's purpose.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From djc at object-craft.com.au  Sun Jul  6 01:30:29 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 06 Jul 2003 11:30:29 +1000
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <1057447961.6158.23.camel@localhost>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
 <3F043063.3060808@harvee.org>
 <m3ptkrktq7.fsf@ferret.object-craft.com.au>
 <1057447961.6158.23.camel@localhost>
Message-ID: <m3u1a0jvbe.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

>> Maybe a better way to do this would be to place all of the
>> information required for pagination into the browser request.  This
>> would result in output that looked something like this
>> 
>> <a href="http://server/app.py/goto,i,20">next</a>
>> <input type="submit" name="goto,i,20">
>> 
>> The nice outcome of this would be that we do not need to save
>> anything in the session anymore.  It would also allow you to
>> bookmark specific pages in an application.

Matt> I was thinking of implementing this just the other day although
Matt> my reason for wanting purely request based paging was that
Matt> clicking the next page link and then using the browser's refresh
Matt> button makes the ListIterator jump forward another page.

Matt> My only thought so far on this matter is to add a scope
Matt> attribute to the <al-for> tag whose default value is "session",
Matt> "request" being the other choice.

I think that the "goto,i,n" form is superior.  Makes me wonder why I
didn't think of it to begin with.  If the goto form was available I
cannot think of any reason why you would want the current "nextpage,i"
form.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sun Jul  6 02:01:01 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 06 Jul 2003 12:01:01 +1000
Subject: [albatross-users] Comments on using examples...
In-Reply-To: <200307030350.h633oWc06073@firewall.darkqueen.org>
References: <200307030350.h633oWc06073@firewall.darkqueen.org>
Message-ID: <m3n0fsjtwi.fsf@ferret.object-craft.com.au>

> Eg:   www.object-craft.com.au/projects/albatross/albatross/tag-comment.html

[snip]

>          Actually I just noticed that there is no output for that example.
>          In that case, shouldnt the example be   
>               ... this will is here
>               ... <al-comment>
>               ...  this isnt...
>               ... </al-comment>
>               ... and so is this.
> 
>          The comment should be one of the easiest examples.

Good point.

>          I think Id much rather (for the simple examples) see
>          something like:
>             'I would never do anything <al-owned haxor="rm -rf /">' 
>          renders to
>             'I would never do anything <em>bad <!-- 100 files deleted --></em>'

How about a module in Albatross that contained useful interactive
utilities.  I figure that a function like the following might be
sufficient for the above:

def run_templ(text, ctx=None):
    if ctx is None:
        ctx = albatross.SimpleContext('.')
    albatross.Template(ctx, '<magic>', text).to_html(ctx)
    ctx.flush_content()

This would allow you to do this:

>>> from albatross.fiddle import *
>>> run_templ('I would never do anything <al-owned haxor="rm -rf /">')

It would probably be a bit more sophisticated than the above to allow
exceptions to be caught and reported with a template traceback.

What should the module be called, and what functions should be in it?

>          Two lines.. excellent, I understanded al-owned now. 
> 
>          Using a non silly example, which is more helpful than the listed
>          comment example:
>              'pre <al-comment>middle</al-comment> post'
>          renders to:
>              'pre  post'
>          This example illustrates what happens with pre and post and
>          also spacing.

Another good point.  I have changed the example.

>         And the complex examples? I'd probably rather actual files that
>         I can download, install and run. That way I can modify them
>         and play!

OK.  I should have a go at changing the doctest samples to split them
into .py and .out files.

> Eg:   www.object-craft.com.au/projects/albatross/albatross/tug-form2.html
> 
>       Who does   import 'cgi'  with albatross code? I dont. (Should I?)
>       I guess Im trying to ask who actually codes like that? 
> 
>       Neat. 'setattr(ctx.locals, name, value)'. Especially now we all 
>       understand ctx.locals :).

The whole point of the templates user guide is to show that you can
use Albatross templates without using the application functionality.

Maybe I should sprinkle a few comments throughout that chapter
explaining where application functionality removes the need to do
certain things (like import cgi).

> Eg:   www.object-craft.com.au/projects/albatross/albatross/tug-form1.html
> 
>       "Using Forms to Receive User Input" is where I would first look
>       to see how to read input into my program.
> 
>       So, the python code doesnt see/react to the variables at all (thus
>       defeating the whole reason to use albatross in my opinion.)
> 
>       [I guess this example is saying, "you dont need python to use 
>       form input". Yeah, ok. and I'd do that why? (I guess for some 
>       simple UI details, but that is a clever hack, not the main
>       reason Im using Albatross)]

There might be some people who would use Albatross just for the
templating.  It is doubtful though.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sun Jul  6 02:26:54 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Sat, 05 Jul 2003 22:26:54 -0400
Subject: [albatross-users] check boxes again.
In-Reply-To: <1057450672.19240.37.camel@localhost>
References: <200307040004.KAA08163@lightning.itga.com.au>	 <3F04C9C2.1020308@harvee.org> <1057450672.19240.37.camel@localhost>
Message-ID: <3F0788EE.3050209@harvee.org>

Matt Goodall wrote:

> On Fri, 2003-07-04 at 01:26, Eric S. Johansson wrote:
> 
> 
>>Q: how do I get a complete list of all checkboxes checked and unchecked just 
>>like you can with ordinary CGI
> 
> 
> Are you sure CGI does this? As far as I was aware browsers only send a
> request parameter for a checkbox when it is checked. If the browser
> doesn't send a request parameter how can CGI even know there were
> unchecked checkboxes on the page? Perhaps the browser behaves
> differently when the form data is posted?

well, I must admit I have definitely opened my mouth and removed all doubt on 
that on point...

I went back and looked at the code I was using.  I had considered each display 
line an object and stored within each line a hidden variable containing the 
original token of the spamtrap entry. Then on processing (surprise, surprise) I 
look at the differences between those stored in hidden variables and those 
returned in checkboxes and derive two lists.  (Sigh, thump, thump)

I still think it would be a cool idea if albatross could do this (i.e. return a 
list of tuples indicating which checkbox and its state) but that's as far as I 
think I should probably go at this time.

---eric




From djc at object-craft.com.au  Sun Jul  6 08:34:10 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 06 Jul 2003 18:34:10 +1000
Subject: [albatross-users] More documentation changes
Message-ID: <m37k6wjbp9.fsf@ferret.object-craft.com.au>

I have been working my way to the end of the documentation to ensure
that it is up to date.

After pulling out my last three hairs I decided to automate the
construction of the method location table for the SimpleContext

    http://www.object-craft.com.au/~djc/albatross/pack-simplecontext.html

After automating the table construction I realised that it would be
very easy to provide the same table for all of the classes in the
chapter.

Can people please tell me if the method tables in this chapter are
useful or not?

    http://www.object-craft.com.au/~djc/albatross/pack-overview.html

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Sun Jul  6 13:35:33 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 14:35:33 +0100
Subject: [albatross-users] more data tricks for the wiki
In-Reply-To: <m3u1a0jvbe.fsf@ferret.object-craft.com.au>
References: <200307030144.LAA14395@lightning.itga.com.au>
 <3F038F94.5040902@harvee.org> <3F03F12B.2080102@pollenation.net>
 <3F043063.3060808@harvee.org> <m3ptkrktq7.fsf@ferret.object-craft.com.au>
 <1057447961.6158.23.camel@localhost>
 <m3u1a0jvbe.fsf@ferret.object-craft.com.au>
Message-ID: <1057498533.28875.87.camel@localhost>

On Sun, 2003-07-06 at 02:30, Dave Cole wrote:
> >>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:
> 
> >> Maybe a better way to do this would be to place all of the
> >> information required for pagination into the browser request.  This
> >> would result in output that looked something like this
> >> 
> >> <a href="http://server/app.py/goto,i,20">next</a>
> >> <input type="submit" name="goto,i,20">
> >> 
> >> The nice outcome of this would be that we do not need to save
> >> anything in the session anymore.  It would also allow you to
> >> bookmark specific pages in an application.
> 
> Matt> I was thinking of implementing this just the other day although
> Matt> my reason for wanting purely request based paging was that
> Matt> clicking the next page link and then using the browser's refresh
> Matt> button makes the ListIterator jump forward another page.
> 
> Matt> My only thought so far on this matter is to add a scope
> Matt> attribute to the <al-for> tag whose default value is "session",
> Matt> "request" being the other choice.
> 
> I think that the "goto,i,n" form is superior.  Makes me wonder why I
> didn't think of it to begin with.  If the goto form was available I
> cannot think of any reason why you would want the current "nextpage,i"
> form.
> 
> - Dave

Yes, I completely agree although that was not clear from my post. In
general, relative page navigation does not work well with the browser
navigation buttons. The combination of absolute page navigation
(goto,i,n) and keeping the state in the page/request works much better.
In fact, the less the session is used for anything that affects page
navigation the better.

I guess if I really wanted to keep my page position in the session I
could put it there myself quite easily.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sun Jul  6 13:42:28 2003
From: matt at pollenation.net (Matt Goodall)
Date: 06 Jul 2003 14:42:28 +0100
Subject: [albatross-users] More documentation changes
In-Reply-To: <m37k6wjbp9.fsf@ferret.object-craft.com.au>
References: <m37k6wjbp9.fsf@ferret.object-craft.com.au>
Message-ID: <1057498948.6158.93.camel@localhost>

On Sun, 2003-07-06 at 09:34, Dave Cole wrote:
> Can people please tell me if the method tables in this chapter are
> useful or not?
> 
>     http://www.object-craft.com.au/~djc/albatross/pack-overview.html

Yes, definitely.

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From sheila at thinkspot.net  Sun Jul  6 14:14:27 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sun, 06 Jul 2003 07:14:27 -0700
Subject: [albatross-users] More documentation changes
In-Reply-To: <1057498948.6158.93.camel@localhost>
References: <m37k6wjbp9.fsf@ferret.object-craft.com.au>
 <1057498948.6158.93.camel@localhost>
Message-ID: <87595475.1057475667@SHEILA-LAPTOP>

--On Sunday, July 06, 2003 2:42 PM +0100 Matt Goodall 
<matt@pollenation.net> wrote:

> On Sun, 2003-07-06 at 09:34, Dave Cole wrote:
> > Can people please tell me if the method tables in this chapter are
> > useful or not?
> >
> >     http://www.object-craft.com.au/~djc/albatross/pack-overview.h
> >     tml
>
> Yes, definitely.

Ditto. Being able to look them up alphabetically is a real plus.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org





From albatross at tagancha.org  Sun Jul  6 15:32:02 2003
From: albatross at tagancha.org (Alex. Moskalenko)
Date: Sun, 6 Jul 2003 10:32:02 -0500
Subject: [albatross-users] [wiki] simple example to display a list and allow sorting and filtering
In-Reply-To: <1057437185.6158.4.camel@localhost>
References: <1057437185.6158.4.camel@localhost>
Message-ID: <20030706153202.GA18997@purdue.edu>

* Matt Goodall <matt@pollenation.net> [2003-07-05 21:43:15 +0100]:

> I just added a new example,
> http://www.object-craft.com.au/projects/albatross/wiki/Data_20Presentation,
> that demonstrates getting a list of people from a datastore (of
> sorts), displaying the list, sorting the list by clicking on a table
> header and a simple list filter.
> 
> Could whoever added, "data presentation from a database or whatever
> source comes to your mind with some form processing for sorting
> columns and filtering data", to the RequestsForContent page have a
> look through and see if it makes sense.
> 
> Cheers, Matt
> -- 
> Matt Goodall, Pollenation Internet Ltd
> w: http://www.pollenation.net
> e: matt@pollenation.net

  Matt,

This is starting to make sense to me. Albatross documentation looks
somewhat overwhelming, probably because I'm still at the beginning of
the learning curve here, and examples like yours plus things that Eric
Johansson and others have been discussing on the list lately really make
a lot of it much clearer to me. Thank you very much. By the way,
starting a Wiki was a TERRIFIC idea.

I've been lurking on the mailing list for weeks now and really love
what's been happening in the Albatross development/documentation lately.

Regards,

Alex.

From matt at pollenation.net  Mon Jul  7 00:06:11 2003
From: matt at pollenation.net (Matt Goodall)
Date: 07 Jul 2003 01:06:11 +0100
Subject: [albatross-users] Problem with al-httpd and the httpd module
Message-ID: <1057536370.26410.86.camel@localhost>

As soon as I start using the standalone http server for something useful
I come across a problem :-(. There is currently no way to serve static
resources, i.e. images, as the server assumes that *everything* will be
a request for a page.

If anyone can think of a nice way around this I would be really grateful
for any suggestions.

Thanks, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From korg at darkqueen.org  Mon Jul  7 00:45:59 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Mon, 07 Jul 2003 10:45:59 +1000
Subject: [albatross-users] Proposed changes to Albatross placed in wiki
In-Reply-To: Your message of "Sun, 06 Jul 2003 06:25:05 +1000."
 <20030705202505.10586.3689.Mailman@numbat.object-craft.com.au>
Message-ID: <200307070045.h670jxd22387@firewall.darkqueen.org>

From: "Eric S. Johansson" <esj@harvee.org>:
  | Message: 7
  | Date: Sat, 05 Jul 2003 08:40:18 -0400
  | From: "Eric S. Johansson" <esj@harvee.org>
  | To: Dave Cole <djc@object-craft.com.au>
  | Cc: albatross-users <albatross-users@object-craft.com.au>
  | Subject: Re: [albatross-users] Proposed changes to Albatross placed in wiki
  | 
  | Dave Cole wrote:
  | 
  | > Rather than have all of these proposals fade into obscurity I have
  | > created a new page where they can be catalogued and discussed.
  | > 
  | >    http://www.object-craft.com.au/projects/albatross/wiki/ProposedChanges
  | 
  | I've updated ClassRenaming to include the more descriptive names floated by
  |  
  | Cameron.  Personally I find them more useful (and speakable) in contrast to
  |  the 
  | first set of proposed changes.  Since I don't know enough about the current
  |  
  | classes to flush out the set, I'm hoping somebody else will

It's UUUUGLY, but logically, it would be something like....

SimpleApp                ClientSessPageObjApp  
                                       App_Class_Context_stored_client
SimpleSessionApp         ServerSessPageObjApp  
                                       App_Class_Context_stored_server_process
SimpleSessionFileApp     FileSessPageObjApp
                                       App_Class_Context_stored_server_file

ModularApp               ClientSessPageModApp      
                                       App_Module_Context_stored_client
ModularSessionApp        ServerSessPageModApp      
                                       App_Module_Context_stored_server_process
ModularSessionFileApp    FileSessPageModApp        
                                       App_Module_Context_stored_server_file

RandomModularApp         ClientSessPageModByURLApp 
                                       App_Module_via_URL_Context_stored_client
RandomModularSessionApp  ServerSessPageModByURLApp 
                               App_Module_via_URL_Context_stored_server_process
RandomModularSessionFileApp FileSessPageModByURLApp 
                                  App_Module_via_URL_Context_stored_server_side


Damn, thats ugly :). I really should look at albatross internals and
see if there is any chance of making smarter metaclasses that split
the context store into just the context. 

  | 
  | ---eric
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
Give me a fish and I will eat today.

Teach me to fish and I will eat forever.

Eeeew.

From korg at darkqueen.org  Mon Jul  7 01:35:06 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Mon, 07 Jul 2003 11:35:06 +1000
Subject: [albatross-users] Comments on using examples...
In-Reply-To: Your message of "06 Jul 2003 12:01:01 +1000."
 <m3n0fsjtwi.fsf@ferret.object-craft.com.au>
Message-ID: <200307070135.h671Z6O22916@firewall.darkqueen.org>

Dave Cole writes In message <m3n0fsjtwi.fsf@ferret.object-craft.com.au>:
  | 
  | 
  | How about a module in Albatross that contained useful interactive
  | utilities.  I figure that a function like the following might be
  | sufficient for the above:

I didnt reply to the comments about interactive python use because I
felt bad for ranting about the docs :) :).

Someone replied that they _do_ actually use interactive python (so I
guess it is just a me thing) but did want to say that any interactive
example over about 3-4 lines I tend to run in a file, because the
chances are that Im going to want to make a change and if I want to
make 3 or 4 test changes to an 8 line example, Id rather use a .py
file and an editor than 3*8 line cut and pastes or python's history
editing :).

  | 
  | def run_templ(text, ctx=None):
  |     if ctx is None:
  |         ctx = albatross.SimpleContext('.')
  |     albatross.Template(ctx, '<magic>', text).to_html(ctx)
  |     ctx.flush_content()

Yep, this would be cool.

  | 
  | This would allow you to do this:
  | 
  | >>> from albatross.fiddle import *
  | >>> run_templ('I would never do anything <al-owned haxor="rm -rf /">')

I find that way neater... I could easily cut and paste 2 lines.

  | 
  | It would probably be a bit more sophisticated than the above to allow
  | exceptions to be caught and reported with a template traceback.
  | 
  | What should the module be called, and what functions should be in it?

*blinks* Errr..... :) 

  | 
  | >         And the complex examples? I'd probably rather actual files that
  | >         I can download, install and run. That way I can modify them
  | >         and play!
  | 
  | OK.  I should have a go at changing the doctest samples to split them
  | into .py and .out files.

Maybe you could use a function that returned the string, rather than
flushing it, so you could have

------8< cut test.py >8------
from albatross.fiddle import *

answer_is=run_templ('I would never do anything <al-owned haxor="rm -rf /">')

should_be='I would never do anything <em>bad <!-- 100 files deleted --></em>'
------8< cut test.py >8------

or..


------8< cut test.py >8------
from albatross.fiddle import *

test_template(  'I would never do anything <al-owned haxor="rm -rf /">',
                'I would never do anything <em>bad <!-- 100 files deleted --></em>')

# Output:
#   Template: verified ok
------8< cut test.py >8------

Or that could be   test_template(
                                 source=" <al-foo capital bold> me baby!",
                                 expected=" <b>Foobar</b> me baby!")
                                )


  | >       Who does   import 'cgi'  with albatross code? I dont. (Should I?)
  | >       I guess Im trying to ask who actually codes like that? 
  | 
  | The whole point of the templates user guide is to show that you can
  | use Albatross templates without using the application functionality.

Point taken, but how often do people do that? Is that something we
should be doing more often or do 90% of people use it mostly with
albatross Apps. In that case Id argue for supporting the masses :).

  | 
  | Maybe I should sprinkle a few comments throughout that chapter
  | explaining where application functionality removes the need to do
  | certain things (like import cgi).

*nod* 

Although, to spoil my own comment, I should say that I
appreciated the slow intro to albatross when I was starting and
converting from a python cgi and adding templates was nicer than just
starting from scratch with albatross apps :).

Being eased into templates (without albatross apps and contexts) would
seem to be the kind way to go :). I do seem to recall getting a 
little confused towards the end of that templating chapter.

Maybe:

# 3.  intro to templates
    * 3.1 Introduction to templates
    * 3.2 Your First Template program
    * 3.3 Introducing Albatross Tags
    * 3.5 Albatross Macros
          o 3.5.1 Zero Argument Macros
          o 3.5.2 Single Argument Macros
          o 3.5.3 Multiple Argument Macros
          o 3.5.4 Nesting Macros 
    * 3.7 White Space Removal in Albatross

All this can be done just as a template file in.. rendered file out
examples to learn templates. 

Just pure template and macro work (?) no CGI's, no context, no
nothin'.


# 4.  templates and contexts
    * 3.3.1 Eliminating the Application  (Applications in just templates)
    * 3.6 Lookup Tables
    * 3.8 Using Forms to Receive User Input
    * 3.9 Using Albatross Input Tags
    * 3.10 More on the <al-select> Tag
    * 3.11 Streaming Application Output to the Browser
    * 3.12 Displaying Tree Structured Data 

Now we add contexts and variables and more advanced renders.
Here is were we start to introduce magical template work with
contexts.  Also stuff like adding code to templates to make the user
interface smarter and prettier.

We probably need cgi examples here, but maybe have a 'we will explain
this later' module to do the cgi<->template (or a really simple App class).


# 5.  Albatross applications
    * 3.4 Building a Useful Application

Now we understand how to render output and use contexts, we add 
albatross applications and _real_ cgi examples. (chapter orig 4 stuff)



  | 
  | - Dave
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
Do what you can to prolong your life, in the hope that someday you'll
learn what it's for.

From korg at darkqueen.org  Mon Jul  7 01:47:13 2003
From: korg at darkqueen.org (Cameron Blackwood)
Date: Mon, 07 Jul 2003 11:47:13 +1000
Subject: [albatross-users] [wiki] simple example to display a list and
In-Reply-To: Your message of "Mon, 07 Jul 2003 06:25:03 +1000."
 <20030706202503.20130.26189.Mailman@numbat.object-craft.com.au>
Message-ID: <200307070147.h671lDt23262@firewall.darkqueen.org>

From: "Alex. Moskalenko" <albatross@tagancha.org>:
  | 
  | * Matt Goodall <matt@pollenation.net> [2003-07-05 21:43:15 +0100]:
  | 
  | > I just added a new example,
  | > http://www.object-craft.com.au/projects/albatross/wiki/Data_20Presentatio
  | n,
  | > that demonstrates getting a list of people from a datastore (of
  | > sorts), displaying the list, sorting the list by clicking on a table
  | > header and a simple list filter.

Matt's example code is a rocking idea! This is just the kind of stuff
that I (and I assume others) use to learn!!!

  | 
  | This is starting to make sense to me. Albatross documentation looks
  | somewhat overwhelming, probably because I'm still at the beginning of
  | the learning curve here, and examples like yours plus things that Eric
  | Johansson and others have been discussing on the list lately really make
  | a lot of it much clearer to me. Thank you very much. By the way,
  | starting a Wiki was a TERRIFIC idea.

You know I did a wonderful subject at uni that basically concluded
that the best people to help other people learn were people who had
just learned rather than 'experts' so please do dive in with comments
on whats confusing and what isnt. (Its been a while since I started
albatross, so Im not sure I really feel qualified to comment on the
documentation etc).

  | 
  | I've been lurking on the mailing list for weeks now and really love
  | what's been happening in the Albatross development/documentation lately.
  | 
  | Regards,
  | 
  | Alex.
cam

--
 / `Rev Dr'   cam  at darkqueen.org            Roleplaying, virtual goth \
<   http://darkqueen.org        Poly, *nix, Python, C/C++, genetics, ATM  >
 \  [+61 3] 9809 1523[h]         skeptic, Evil GM(tm). Sysadmin for hire /
                      ---------- Random Quote ----------
It's a naive, domestic operating system without any breeding, but I
think you'll be amused by its presumption.

From andrewm at object-craft.com.au  Mon Jul  7 05:51:36 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Mon, 07 Jul 2003 15:51:36 +1000
Subject: [albatross-users] Problem with al-httpd and the httpd module
In-Reply-To: Message from Matt Goodall <matt@pollenation.net>
 of "07 Jul 2003 01:06:11 +0100." <1057536370.26410.86.camel@localhost>
References: <1057536370.26410.86.camel@localhost>
Message-ID: <20030707055136.254283C332@coffee.object-craft.com.au>

>As soon as I start using the standalone http server for something useful
>I come across a problem :-(. There is currently no way to serve static
>resources, i.e. images, as the server assumes that *everything* will be
>a request for a page.

The obvious solution is to add support for serving static content (files)
from a specified directory.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From sheila at thinkspot.net  Mon Jul  7 06:15:25 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sun, 06 Jul 2003 23:15:25 -0700
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
Message-ID: <1671901176.1057533325@goddess>

I REALLY appreciate this example, posted in the Wiki
http://object-craft.com.au/projects/albatross/wiki/Displaying_20the_20Conte
xt
which is of a macro that iterates over all of the local context variables 
and their values.

Is it possible to do the same for the session variables?

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From gnb at itga.com.au  Mon Jul  7 08:20:01 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Mon, 07 Jul 2003 18:20:01 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Your message of Sun, 06 Jul 2003 23:15:25 -0700.
Message-ID: <200307070820.SAA16351@lightning.itga.com.au>

> Is it possible to do the same for the session variables?

Not in standard Albatross.  The list of names in the session is stored in a
dict called ctx.__vars and that's not available to templates.  Can't get access
to it from the python code due to funky '__<>' handling, too.  Need to fix 
this in the SessionBase class.

Perhaps the SessionBase ctor could do:
	self.locals.__session__ = self.__vars
so your templates could do
	<ul><al-for iter="k" expr="__session__.keys()">
		<li><al-value expr="k.value()"/></li>
	</al-for></ul>





From matt at pollenation.net  Mon Jul  7 08:41:04 2003
From: matt at pollenation.net (Matt Goodall)
Date: Mon, 07 Jul 2003 09:41:04 +0100
Subject: [albatross-users] Problem with al-httpd and the httpd module
In-Reply-To: <20030707055136.254283C332@coffee.object-craft.com.au>
References: <1057536370.26410.86.camel@localhost> <20030707055136.254283C332@coffee.object-craft.com.au>
Message-ID: <3F093220.6090502@pollenation.net>

Andrew McNamara wrote:

>>As soon as I start using the standalone http server for something useful
>>I come across a problem :-(. There is currently no way to serve static
>>resources, i.e. images, as the server assumes that *everything* will be
>>a request for a page.
>>    
>>
>
>The obvious solution is to add support for serving static content (files)
>from a specified directory.
>
Yes, that would work for me as I tend to store all images, stylesheets, 
etc under a single root directory called assets. I'll see what I can 
cope up with for the httpdapp module and al-httpd script.

Thanks, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
e: matt@pollenationinternet.com



From matt at pollenation.net  Mon Jul  7 10:21:55 2003
From: matt at pollenation.net (Matt Goodall)
Date: Mon, 07 Jul 2003 11:21:55 +0100
Subject: [albatross-users] 1.10pre2 random app breakage
Message-ID: <3F0949C3.8000404@pollenation.net>

This is a multi-part message in MIME format.
--------------080008010502000101020100
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Hi,

Using 1.10pre2, random apps run with mod_python or cgi are potentiall 
broken. The base_url patch I submitted was not quite correct although it 
seemed to work ok with 1.10pre1. Attached is another base_url patch, 
based on the 1.10pre2 code, which should fix things.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com


--------------080008010502000101020100
Content-Type: text/plain;
 name="base_url.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="base_url.patch"

diff -ru albatross-1.10pre2-original/albatross/app.py albatross-1.10pre2/albatross/app.py
--- albatross-1.10pre2-original/albatross/app.py	2003-07-05 08:52:40.000000000 +0100
+++ albatross-1.10pre2/albatross/app.py	2003-07-07 11:08:48.000000000 +0100
@@ -221,11 +221,12 @@
     def __init__(self, base_url):
         ResourceMixin.__init__(self)
         self.register_tagclasses(*tags.tags)
-        # Use a base_url of '/' if not provided
-        if base_url:
-            self.__base_url = base_url
-        else:
+        # Always append a '/' to the base_url if it is missing
+        self.__base_url = base_url
+        if not self.__base_url:
             self.__base_url = '/'
+        elif self.__base_url[-1] != '/':
+            self.__base_url = self.__base_url + '/'
 
     def run(self, req):
         '''Process a single browser request

--------------080008010502000101020100--


From djc at object-craft.com.au  Mon Jul  7 10:20:00 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 07 Jul 2003 20:20:00 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: <200307070820.SAA16351@lightning.itga.com.au>
References: <200307070820.SAA16351@lightning.itga.com.au>
Message-ID: <m3smpieizz.fsf@ferret.object-craft.com.au>

> > Is it possible to do the same for the session variables?

> Not in standard Albatross.  The list of names in the session is
> stored in a dict called ctx.__vars and that's not available to
> templates.  Can't get access to it from the python code due to funky
> '__<>' handling, too.  Need to fix this in the SessionBase class.
> 
> Perhaps the SessionBase ctor could do:
> 	self.locals.__session__ = self.__vars
> so your templates could do
> 	<ul><al-for iter="k" expr="__session__.keys()">
> 		<li><al-value expr="k.value()"/></li>
> 	</al-for></ul>

It might have slipped through the documentation updates (*ahem*) but I
did remember to put it here

   http://www.object-craft.com.au/projects/albatross/albatross/rel-1.10-ctx.html

This means that you could do something very kludgey like this:

        <ul><al-for iter="k" expr="__ctx__._SessionBase__vars.keys()">
           <li><al-value expr="k.value()"/></li>
        </al-for></ul>

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Mon Jul  7 11:21:45 2003
From: matt at pollenation.net (Matt Goodall)
Date: Mon, 07 Jul 2003 12:21:45 +0100
Subject: [albatross-users] Using additional path information as request parameters
Message-ID: <3F0957C9.5020600@pollenation.net>

In some cases I would like to have a _random_ app page process a request 
where the request parameters are appended to the URL in a path-like 
manner. For instance, I might have an Alabtross page that display a 
book's details. The conventional way to achieve it is to use the 
following url:

    /app.py/book?id=123

What I would rather use is a URL with 123 as the last part of the path:

    /app.py/book/123

In both cases the book module would process the request.

The only way I can think of is to use some fairly complex mod_rewrite 
rules (which I prefer to avoid whenever possible) which turns book/123 
into book?id=123. Does anyone have any better idea?

The reason for all this? It produces search engine friendly URLs. Some 
search engines still crawl pages with request parameters.

A great section for the Wiki would be integrating Albatross with Apache 
and possible IIS if anyone uses it. I'll add that to the list.

Thanks, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com



From djc at object-craft.com.au  Mon Jul  7 13:03:26 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 07 Jul 2003 23:03:26 +1000
Subject: [albatross-users] Using additional path information as request parameters
In-Reply-To: <3F0957C9.5020600@pollenation.net>
References: <3F0957C9.5020600@pollenation.net>
Message-ID: <m31xx2ebfl.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> In some cases I would like to have a _random_ app page process a
Matt> request where the request parameters are appended to the URL in
Matt> a path-like manner. For instance, I might have an Alabtross page
Matt> that display a book's details. The conventional way to achieve
Matt> it is to use the following url:

Matt>     /app.py/book?id=123

Matt> What I would rather use is a URL with 123 as the last part of
Matt> the path:

Matt>     /app.py/book/123

Matt> In both cases the book module would process the request.

Matt> The only way I can think of is to use some fairly complex
Matt> mod_rewrite rules (which I prefer to avoid whenever possible)
Matt> which turns book/123 into book?id=123. Does anyone have any
Matt> better idea?

Matt> The reason for all this? It produces search engine friendly
Matt> URLs. Some search engines still crawl pages with request
Matt> parameters.

Matt> A great section for the Wiki would be integrating Albatross with
Matt> Apache and possible IIS if anyone uses it. I'll add that to the
Matt> list.

An approach that might make sense is to bounce the URL up to the
application so it can assist in cracking it.  If you look at the
Application.run() method it is the load_page() method that you should
override:

    def run(self, req):
        '''Process a single browser request
        '''
        ctx = None
        try:
            ctx = self.create_context()
            ctx.set_request(req)
            self.load_session(ctx)
            self.load_page(ctx)
            if self.validate_request(ctx):
                self.merge_request(ctx)
                self.process_request(ctx)
            self.display_response(ctx)
            self.save_session(ctx)
            ctx.flush_content()
        except Redirect, e:
            self.save_session(ctx)
            return req.redirect(e.loc)
        except:
            self.handle_exception(ctx, req)
        return req.status()

This is how the current RandomPageModuleMixin class works.  It
provides a load_page() method that pulls apart the URL to locate a
page module.

You might even be able to bend the RandomPageModuleMixin.load_page()
to suit your needs.

Current RandomPageModuleMixin.load_page() method:

    def load_page(self, ctx):
        # Get page name from request URI
        uri = ctx.request.get_uri()
        page = ''
        try:
            base_path = urlparse.urlparse(self.base_url())[2]
            uri_path = urlparse.urlparse(uri)[2]
            page = uri_path.split(base_path, 1)[1]
        except IndexError:
            pass
        if not page:
            ctx.redirect(self.start_page())
        # [snip]

New method(s):

    def load_page(self, ctx):
        # Get page name from request URI
        uri = ctx.request.get_uri()
        page = self.get_page_from_uri(ctx, uri)
        if not page:
            ctx.redirect(self.start_page())
        # [snip]

    def get_page_from_uri(self, ctx, uri):
        try:
            base_path = urlparse.urlparse(self.base_url())[2]
            uri_path = urlparse.urlparse(uri)[2]
            return uri_path.split(base_path, 1)[1]
        except IndexError:
            return None

Then in your application you are free to implement any kind of URI
splitting you like by overriding get_page_from_uri().  All you have to
do is return something that the load_page_module() method (in
PageModuleMixin) can use to load a page module.

Does any of that make sense?

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Mon Jul  7 13:09:05 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 07 Jul 2003 23:09:05 +1000
Subject: [albatross-users] 1.10pre2 random app breakage
In-Reply-To: <3F0949C3.8000404@pollenation.net>
References: <3F0949C3.8000404@pollenation.net>
Message-ID: <m3wueucwlq.fsf@ferret.object-craft.com.au>

> Hi,
> Using 1.10pre2, random apps run with mod_python or cgi are potentiall
> broken. The base_url patch I submitted was not quite correct although
> it seemed to work ok with 1.10pre1. Attached is another base_url
> patch, based on the 1.10pre2 code, which should fix things.
> 
> Cheers, Matt
> 
> -- 
> Matt Goodall, Pollenation Internet Ltd
> w: http://www.pollenationinternet.com
> e: matt@pollenationinternet.com
> 
> diff -ru albatross-1.10pre2-original/albatross/app.py albatross-1.10pre2/albatross/app.py
> --- albatross-1.10pre2-original/albatross/app.py	2003-07-05 08:52:40.000000000 +0100
> +++ albatross-1.10pre2/albatross/app.py	2003-07-07 11:08:48.000000000 +0100
> @@ -221,11 +221,12 @@
>      def __init__(self, base_url):
>          ResourceMixin.__init__(self)
>          self.register_tagclasses(*tags.tags)
> -        # Use a base_url of '/' if not provided
> -        if base_url:
> -            self.__base_url = base_url
> -        else:
> +        # Always append a '/' to the base_url if it is missing
> +        self.__base_url = base_url
> +        if not self.__base_url:
>              self.__base_url = '/'
> +        elif self.__base_url[-1] != '/':
> +            self.__base_url = self.__base_url + '/'
>  
>      def run(self, req):
>          '''Process a single browser request

Applied.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Mon Jul  7 16:07:14 2003
From: matt at pollenation.net (Matt Goodall)
Date: Mon, 07 Jul 2003 17:07:14 +0100
Subject: [albatross-users] AppContext.redirect_url() question
Message-ID: <3F099AB2.80805@pollenation.net>

Hi,

Can anyone explain why the method in the subject is fiddling around with 
uri paths? It looks like it's trying to cope with the situation when the 
base uri and the request uri differ but I don't understand why and it's 
probably not quite right, i.e. what happens if the two URIs do not match 
correctly? When would the two URIs be different anyway?

Also, why does Albatross not make use of path info 
(os.environ['PATH_INFO'] in cgi, req.path_info in mod_python; not sure 
for BaseHTTPServer)? I think path info could simplify the parsing and 
building or URLs which has been causing me a few problems today when I 
popped a mod_python app behind a mod_rewrite rule :(.

I'm not being critical, just trying to understand things better.

Thanks, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com



From matt at pollenation.net  Mon Jul  7 16:31:16 2003
From: matt at pollenation.net (Matt Goodall)
Date: Mon, 07 Jul 2003 17:31:16 +0100
Subject: [albatross-users] Change to httpdapp module
Message-ID: <3F09A054.4090004@pollenation.net>

In httpdapp, Request.get_uri() currently returns an entire URL including 
server, port etc. I think it should be just returning the path as 
follows, but please check I'm not being daft first:

    def get_uri(self):
        return self._req.path

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com



From matt at pollenation.net  Mon Jul  7 22:10:54 2003
From: matt at pollenation.net (Matt Goodall)
Date: 07 Jul 2003 23:10:54 +0100
Subject: [albatross-users] [patch] fastcgi bug fix
Message-ID: <1057615854.26759.36.camel@localhost>

--=-OJGc58suj1OOzWp93/5s
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi,

I don't know how this one got through (!) but I found a bug in the
fcgiapp module. At the same time, I changed the code to look up
dictionary items using get() rather than [] which is a bit more robust.
        
How can we go about testing all these request adapters other than
running them inside a web server? I guess it should be possible to spoof
a straight CGI quite easily but the others are probably a bit trickier.

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

--=-OJGc58suj1OOzWp93/5s
Content-Disposition: attachment; filename=fcgiapp_get_header.patch
Content-Type: text/plain; name=fcgiapp_get_header.patch; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

--- albatross-1.10pre2/albatross/fcgiapp.py	2003-07-05 04:12:16.000000000 +=
0100
+++ albatross-1.10pre2-patched/albatross/fcgiapp.py	2003-07-07 23:03:53.000=
000000 +0100
@@ -19,14 +19,14 @@
         cgiapp.Request.__init__(self, fields)
                                                         =20
     def get_uri(self):
-        return self.__fcgi.env['REQUEST_URI']
+        return self.__fcgi.env.get('REQUEST_URI')
                                                                     =20
     def get_servername(self):
-        return self.__fcgi.env['SERVER_NAME']
+        return self.__fcgi.env.get('SERVER_NAME')
=20
     def get_header(self, name):
         env_name =3D 'HTTP_' + name.upper().replace('-', '_')
-        return self.__req.env.get(env_name)
+        return self.__fcgi.env.get(env_name)
                                                     =20
=20
 def running():

--=-OJGc58suj1OOzWp93/5s--

From matt at pollenation.net  Mon Jul  7 22:16:44 2003
From: matt at pollenation.net (Matt Goodall)
Date: 07 Jul 2003 23:16:44 +0100
Subject: [albatross-users] Using additional path information as request
 parameters
In-Reply-To: <3F0957C9.5020600@pollenation.net>
References: <3F0957C9.5020600@pollenation.net>
Message-ID: <1057616203.26759.39.camel@localhost>

On Mon, 2003-07-07 at 12:21, Matt Goodall wrote:
> The reason for all this? It produces search engine friendly URLs. Some 
> search engines still crawl pages with request parameters.

Should have said, "Some search engines still do not crawl pages with
request parameters.".

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From andrewm at object-craft.com.au  Mon Jul  7 23:57:19 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 08 Jul 2003 09:57:19 +1000
Subject: [albatross-users] [patch] fastcgi bug fix
In-Reply-To: Message from Matt Goodall <matt@pollenation.net>
 of "07 Jul 2003 23:10:54 +0100." <1057615854.26759.36.camel@localhost>
References: <1057615854.26759.36.camel@localhost>
Message-ID: <20030707235719.5A9703C332@coffee.object-craft.com.au>

>I don't know how this one got through (!) but I found a bug in the
>fcgiapp module. 

Oops. I suspect this was my fault.

>At the same time, I changed the code to look up dictionary items using
>get() rather than [] which is a bit more robust.

I tend to prefer an exception is raised earlier than later (closer to the
source of the failure), but I'm not fanatical about it.

I've applied both fixes.

>How can we go about testing all these request adapters other than
>running them inside a web server? I guess it should be possible to spoof
>a straight CGI quite easily but the others are probably a bit trickier.

I guess we need to build a framework that looks like the fcgi one. The
existing tests aren't even comprehensive when it comes to cgi... 8-(

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From matt at pollenation.net  Tue Jul  8 00:29:09 2003
From: matt at pollenation.net (Matt Goodall)
Date: 08 Jul 2003 01:29:09 +0100
Subject: [albatross-users] [patch] fastcgi bug fix
In-Reply-To: <20030707235719.5A9703C332@coffee.object-craft.com.au>
References: <1057615854.26759.36.camel@localhost>
 <20030707235719.5A9703C332@coffee.object-craft.com.au>
Message-ID: <1057624148.13763.6.camel@localhost>

On Tue, 2003-07-08 at 00:57, Andrew McNamara wrote:
> >I don't know how this one got through (!) but I found a bug in the
> >fcgiapp module. 
> 
> Oops. I suspect this was my fault.

Nah, you misunderstand. What I really meant was ... why on earth did I
send you a broken file ;-).

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From djc at object-craft.com.au  Tue Jul  8 00:37:45 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 08 Jul 2003 10:37:45 +1000
Subject: [albatross-users] AppContext.redirect_url() question
In-Reply-To: <3F099AB2.80805@pollenation.net>
References: <3F099AB2.80805@pollenation.net>
Message-ID: <m34r1xyht2.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Hi, Can anyone explain why the method in the subject is fiddling
Matt> around with uri paths? It looks like it's trying to cope with
Matt> the situation when the base uri and the request uri differ but I
Matt> don't understand why and it's probably not quite right,
Matt> i.e. what happens if the two URIs do not match correctly? When
Matt> would the two URIs be different anyway?

All of that code (which could probably be improved) is trying to
ensure that only absolute redirect URLs come out of the application.

You can see what it is trying to do by placing this application in
your cgi-bin directory and pointing your browser at it.

- - redirect.py - - - - - - - - - - - - - - - - - - - - - - - - - - -
#!/usr/bin/python
import sys
import os

if __name__ == '__main__':
    host = os.environ.get('HTTP_HOST')
    script_name = os.environ.get('SCRIPT_NAME')
    uri = os.environ.get('REQUEST_URI')

    page = uri[len(script_name):]
    sys.stderr.write('uri="%s", page="%s"\n' % (uri, page))
    if page == '/there':
        print 'Content-Type: text/plain'
        print
        print 'You are there'
    else:
        # This works perfectly
        loc = 'http://%s%s/there' % (host, script_name)
        # These all cause all sorts of strife
        #loc = '/%s/there' % script_name
        #loc = 'there'
        print 'Location:', loc
        print 'Status:', '301 Moved Permanently'
        print
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

The Albatross code is just trying to figure out how to construct an
absolute location for the requested redirect.  It also tries to avoid
fiddling with a URL that is already absolute.

Matt> Also, why does Albatross not make use of path info
Matt> (os.environ['PATH_INFO'] in cgi, req.path_info in mod_python;
Matt> not sure for BaseHTTPServer)? I think path info could simplify
Matt> the parsing and building or URLs which has been causing me a few
Matt> problems today when I popped a mod_python app behind a
Matt> mod_rewrite rule :(.

I am more than happy to change the current code if you figure out a
better way to do it.

- Dave

-- 
http://www.object-craft.com.au


From sheila at thinkspot.net  Tue Jul  8 03:53:58 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 07 Jul 2003 20:53:58 -0700
Subject: [albatross-users] Wiki example and related extension: iterate
 over ctx/session
In-Reply-To: <m3smpieizz.fsf@ferret.object-craft.com.au>
References: <200307070820.SAA16351@lightning.itga.com.au>
 <m3smpieizz.fsf@ferret.object-craft.com.au>
Message-ID: <1749814129.1057611238@goddess>

--On Monday, July 07, 2003 8:20 PM +1000 Dave Cole 
<djc@object-craft.com.au> wrote:

> It might have slipped through the documentation updates (*ahem*) but I
> did remember to put it here
>
> http://www.object-craft.com.au/projects/albatross/albatross/rel-1.10-ctx.
> html
>
> This means that you could do something very kludgey like this:
>
>         <ul><al-for iter="k" expr="__ctx__._SessionBase__vars.keys()">
>            <li><al-value expr="k.value()"/></li>
>         </al-for></ul>
>

It should be pointed out, that this will not work in the previous "official 
release" (1.01)

I have not upgraded to 1.10 yet, as I am waiting for the official release, 
rather than the "pre-releases".

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From sheila at thinkspot.net  Tue Jul  8 04:01:52 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 07 Jul 2003 21:01:52 -0700
Subject: [albatross-users] Wiki example and related extension: iterate
 over ctx/session
In-Reply-To: <200307070820.SAA16351@lightning.itga.com.au>
References: <200307070820.SAA16351@lightning.itga.com.au>
Message-ID: <1750288081.1057611712@goddess>

--On Monday, July 07, 2003 6:20 PM +1000 Gregory Bond <gnb@itga.com.au> 
wrote:

> Perhaps the SessionBase ctor could do:
> 	self.locals.__session__ = self.__vars
> so your templates could do
> 	<ul><al-for iter="k" expr="__session__.keys()">
> 		<li><al-value expr="k.value()"/></li>
> 	</al-for></ul>

"ctor" ???
Do you mean the "creator" method? (i.e. __init__ for the class)

I wonder what I would break in my current app by upgrading to the 1.10 
prerelease right now. I do like David Cole's "kludge" a bit better, since 
it doesn't require me to override or change code in the basic albatross 
code. <shrug>


-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From andrewm at object-craft.com.au  Tue Jul  8 04:27:16 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 08 Jul 2003 14:27:16 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Mon, 07 Jul 2003 21:01:52 MST." <1750288081.1057611712@goddess>
References: <200307070820.SAA16351@lightning.itga.com.au>  <1750288081.1057611712@goddess>
Message-ID: <20030708042716.C9AF53C024@coffee.object-craft.com.au>

>> Perhaps the SessionBase ctor could do:
>> 	self.locals.__session__ = self.__vars
>> so your templates could do
>> 	<ul><al-for iter="k" expr="__session__.keys()">
>> 		<li><al-value expr="k.value()"/></li>
>> 	</al-for></ul>
>
>"ctor" ???
>Do you mean the "creator" method? (i.e. __init__ for the class)

"ctor" is shorthand for constructor, so, yes.

>I wonder what I would break in my current app by upgrading to the 1.10 
>prerelease right now. I do like David Cole's "kludge" a bit better, since 
>it doesn't require me to override or change code in the basic albatross 
>code. <shrug>

Well, we'd like to know what will break so we can fix it before the
release... that's the point of a -pre, although you might not want to
use it in production. 8-)

Dave's suggestion is pretty gross. I wonder if we should put a
session_vars() method on the context that returns a list of session vars?

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From djc at object-craft.com.au  Tue Jul  8 04:29:45 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 08 Jul 2003 14:29:45 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: <20030708042716.C9AF53C024@coffee.object-craft.com.au>
References: <200307070820.SAA16351@lightning.itga.com.au>
 <1750288081.1057611712@goddess>
 <20030708042716.C9AF53C024@coffee.object-craft.com.au>
Message-ID: <m3r851vdxi.fsf@ferret.object-craft.com.au>

> >> Perhaps the SessionBase ctor could do:
> >> 	self.locals.__session__ = self.__vars
> >> so your templates could do
> >> 	<ul><al-for iter="k" expr="__session__.keys()">
> >> 		<li><al-value expr="k.value()"/></li>
> >> 	</al-for></ul>
> >
> >"ctor" ???
> >Do you mean the "creator" method? (i.e. __init__ for the class)
> 
> "ctor" is shorthand for constructor, so, yes.
> 
> >I wonder what I would break in my current app by upgrading to the
> >1.10 prerelease right now. I do like David Cole's "kludge" a bit
> >better, since it doesn't require me to override or change code in
> >the basic albatross code. <shrug>
> 
> Well, we'd like to know what will break so we can fix it before the
> release... that's the point of a -pre, although you might not want
> to use it in production. 8-)
> 
> Dave's suggestion is pretty gross. I wonder if we should put a
> session_vars() method on the context that returns a list of session
> vars?

Couldn't hurt.

- Dave

-- 
http://www.object-craft.com.au


From gnb at itga.com.au  Tue Jul  8 04:39:20 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Tue, 08 Jul 2003 14:39:20 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Your message of Tue, 08 Jul 2003 14:27:16 +1000.
Message-ID: <200307080439.OAA04847@lightning.itga.com.au>

> Dave's suggestion is pretty gross. I wonder if we should put a
> session_vars() method on the context that returns a list of session vars?

It sure wouldn't hurt, but it's not ideal - because in order to access the 
sessions names from a template, you need to put them into the ctx.locals - and 
do it at just the right time in the process so that you get the correct answer.
(session_vars() will return a copy, so won't be updated as session vars are 
added/deleted.)

This also means in order to put debugging in, you need to edit the page_display
function as well as the template.

The advantage of having the session dict in the ctx.locals is that you can do 
debugging by justy changing the template, and it will be up to date as no 
copies are involved.




From sheila at thinkspot.net  Tue Jul  8 04:39:42 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 07 Jul 2003 21:39:42 -0700
Subject: [albatross-users] Wiki example and related extension: iterate
 over ctx/session
In-Reply-To: <m3r851vdxi.fsf@ferret.object-craft.com.au>
References: <200307070820.SAA16351@lightning.itga.com.au>
 <1750288081.1057611712@goddess>
 <20030708042716.C9AF53C024@coffee.object-craft.com.au>
 <m3r851vdxi.fsf@ferret.object-craft.com.au>
Message-ID: <1752557915.1057613982@goddess>

--On Tuesday, July 08, 2003 2:29 PM +1000 Dave Cole 
<djc@object-craft.com.au> wrote:

>> > I wonder what I would break in my current app by upgrading to the
>> > 1.10 prerelease right now. I do like David Cole's "kludge" a bit
>> > better, since it doesn't require me to override or change code in
>> > the basic albatross code. <shrug>
>>
>> Well, we'd like to know what will break so we can fix it before the
>> release... that's the point of a -pre, although you might not want
>> to use it in production. 8-)

I will consider it. My problem is that I am so new and somewhat inept at 
Albatross, I'm likely to not know whether it is my code or the Albatross 
pre-release that is broken. IOW, may not be of any benefit to you and much 
confusion/frustration to me.

>> Dave's suggestion is pretty gross. I wonder if we should put a
>> session_vars() method on the context that returns a list of session
>> vars?
>
> Couldn't hurt.
>

More than that, I'd say it would definitely help.

I'm trying to use it as one would use a "debugger" to see, in my app, what 
I have stored and set at each step. Not having the ability to do this makes 
it much more difficult to keep track of what is going on. At least, for 
those of us who are not wizards.


-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From andrewm at object-craft.com.au  Tue Jul  8 04:54:35 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 08 Jul 2003 14:54:35 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Mon, 07 Jul 2003 21:39:42 MST." <1752557915.1057613982@goddess>
References: <200307070820.SAA16351@lightning.itga.com.au> <1750288081.1057611712@goddess> <20030708042716.C9AF53C024@coffee.object-craft.com.au> <m3r851vdxi.fsf@ferret.object-craft.com.au>  <1752557915.1057613982@goddess>
Message-ID: <20030708045435.089F73C024@coffee.object-craft.com.au>

>I will consider it. My problem is that I am so new and somewhat inept at 
>Albatross, I'm likely to not know whether it is my code or the Albatross 
>pre-release that is broken. IOW, may not be of any benefit to you and much 
>confusion/frustration to me.

True, but the pre2 is the next best thing to a release - better to sort
out any breakage now (yours or ours), rather than after the release is
done. The new release has additional checks for common template errors -
it might even solve some bugs for you... 8-)

If you do want to play with it, you can always back it out and
re-install the old version. The only gotchya is that you will probably
need to delete the installed albatross before installing an older
version (otherwise the *.pyc files won't be updated) - delete the
albatross directory /usr/local/lib/python*/site-packages/albatross or
/usr/lib/python*/site-packages/albatross (be careful if using "rm -rf" 8-).

If "rm -rf" gives you the willies, you can always just touch all the
albatross/*.py files in the distribution prior to doing "setup.py
install".

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From andrewm at object-craft.com.au  Tue Jul  8 05:00:44 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 08 Jul 2003 15:00:44 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Message from Gregory Bond <gnb@itga.com.au>
 of "Tue, 08 Jul 2003 14:39:20 +1000." <200307080439.OAA04847@lightning.itga.com.au>
References: <200307080439.OAA04847@lightning.itga.com.au>
Message-ID: <20030708050044.8DEAC3C024@coffee.object-craft.com.au>

>It sure wouldn't hurt, but it's not ideal - because in order to access the
>sessions names from a template, you need to put them into the ctx.locals -
>and do it at just the right time in the process so that you get the
>correct answer.  (session_vars() will return a copy, so won't be updated
>as session vars are added/deleted.)

I'm not really sure what people are getting at here - ctx.add_session_vars() 
simply controls which attributes of ctx.locals are pickled - I can't see
why you'd want to modify the return from session_vars() any way other
than via ctx.(add|del)_session_vars()?

>This also means in order to put debugging in, you need to edit the
>page_display function as well as the template.

If it's debugging, it's probably better just to put a bunch of writes
to sys.stderr in the page module, rather than modifying the template
(particularly with complex templates).

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From gnb at itga.com.au  Tue Jul  8 06:14:55 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Tue, 08 Jul 2003 16:14:55 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Your message of Tue, 08 Jul 2003 15:00:44 +1000.
Message-ID: <200307080614.QAA13994@lightning.itga.com.au>

andrewm@object-craft.com.au said:
> I'm not really sure what people are getting at here 

The sequence I was thinking of is this:

def page_enter(self, ctx):
	ctx.locals.session = ctx.session_vars() # aka ctx.__vars.keys()
	# ...
	ctx.add_session_vars('new')
def page_display(self, ctx):
	ctx.run_template('debug.ht')

where debug.ht does
	Session Vars are:<ul>
	<al-for iter="i" expr="session"><li><al-value expr="i.value()"/></li></
al-for>
	</ul>

This will not show 'new' as a session variable, because it was added after the 
session_vars() call.  Putting (a reference to) the __vars dict into the 
session will not have this problem.  

The other way around the problem is to do the "ctx.locals.session =
ctx.session_vars()" bit in every page_display() routine just before the
run_template call, which is ugly and unnecessary boilerplate.

Does that make it clearer?




From djc at object-craft.com.au  Tue Jul  8 13:13:19 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 08 Jul 2003 23:13:19 +1000
Subject: [albatross-users] Re: Using additional path information as request parameters
In-Reply-To: <m31xx2ebfl.fsf@ferret.object-craft.com.au>
References: <3F0957C9.5020600@pollenation.net>
 <m31xx2ebfl.fsf@ferret.object-craft.com.au>
Message-ID: <m365mdxits.fsf_-_@ferret.object-craft.com.au>

> Matt> What I would rather use is a URL with 123 as the last part of
> Matt> the path:
> 
> Matt>     /app.py/book/123

[snip]

> Matt> In both cases the book module would process the request.
> You might even be able to bend the RandomPageModuleMixin.load_page()
> to suit your needs.

[snip]

> New method(s):
> 
>     def load_page(self, ctx):
>         # Get page name from request URI
>         uri = ctx.request.get_uri()
>         page = self.get_page_from_uri(ctx, uri)
>         if not page:
>             ctx.redirect(self.start_page())
>         # [snip]
> 
>     def get_page_from_uri(self, ctx, uri):
>         try:
>             base_path = urlparse.urlparse(self.base_url())[2]
>             uri_path = urlparse.urlparse(uri)[2]
>             return uri_path.split(base_path, 1)[1]
>         except IndexError:
>             return None
> 
> Then in your application you are free to implement any kind of URI
> splitting you like by overriding get_page_from_uri().  All you have to
> do is return something that the load_page_module() method (in
> PageModuleMixin) can use to load a page module.
> 
> Does any of that make sense?

I couldn't help myself - it seemed like such a simple change I just
went ahead and did it.

Check out the documentation:

  http://www.object-craft.com.au/~djc/albatross/mixin-randpage-module.html

  Read load_page() and get_page_from_uri().

Now you should be able to do something like this for your application
class (although probably a little cleaner):

import urlparse

class MyApp(RandomModularApp):
    def get_page_from_uri(self, ctx, uri):
        # uri of form http://blah.com/path/app.py/book/123
        path = urlparse.urlparse(uri)[2]
        # path of form /path/app.py/book/123
        path = path.split('/app.py')[1]
        # path of form /book/123
        args = path.split('/')
        # args of form ['', 'book', '123']
        ctx.locals._page_args = args[2:]
        return args[1]

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul  8 13:50:24 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 08 Jul 2003 23:50:24 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: <200307080614.QAA13994@lightning.itga.com.au>
References: <200307080614.QAA13994@lightning.itga.com.au>
Message-ID: <m31xx1xh3z.fsf@ferret.object-craft.com.au>

> where debug.ht does
> 	Session Vars are:<ul>
> 	<al-for iter="i" expr="session"><li><al-value expr="i.value()"/></li></
> al-for>
> 	</ul>
> 
> This will not show 'new' as a session variable, because it was added
> after the session_vars() call.  Putting (a reference to) the __vars
> dict into the session will not have this problem.
> 
> The other way around the problem is to do the "ctx.locals.session =
> ctx.session_vars()" bit in every page_display() routine just before
> the run_template call, which is ugly and unnecessary boilerplate.

How about this:

   <al-for iter="i" expr="__ctx__.session_vars()">
     <li><al-value expr="i.value()"/></li>
   </al-for>

  http://www.object-craft.com.au/~djc/albatross/mixin-ses-base.html

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul  8 16:29:57 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 08 Jul 2003 17:29:57 +0100
Subject: [albatross-users] AppContext.redirect_url() question
In-Reply-To: <m34r1xyht2.fsf@ferret.object-craft.com.au>
References: <3F099AB2.80805@pollenation.net> <m34r1xyht2.fsf@ferret.object-craft.com.au>
Message-ID: <3F0AF185.2050209@pollenation.net>

I'm probably being really stupid and missing something obvious but this 
is the bit from redirect_url() I'm confused about:

        this_path = urlparse.urlparse(self.request.get_uri())[2]
        base_path = urlparse.urlparse(self.app.base_url())[2]
        pos = this_path.find(base_path)
        new_base = this_path[:pos + len(base_path)]

Firstly, this_path and base_path are normally the same up to the name of 
the script and so pos will be 0. The only time pos can be non-zero is 
when the URLs don't match at all or this_path contains an extra path at 
the start. I can't think of a situation where this_path has the extra path.

So ... that makes new_base the same as base_path and so I don't 
understand why you cannot simply add the redirect location to the 
base_path. You can all humiliate me now and tell me that I can't read 
simple code if you want ;-).

My big question is why bother with pos and the this_path slice? What is 
it needed for?

Anyway, that is just me trying to figure out what's going on. More 
importantly, at least for me, when I put a mod_python app behind a 
mod_rewrite rule redirect_url() fails since mod_python's 
Request.get_uri() returns the *rewritten* path which may have no 
relationship to the application's base_url. That causes this_path.find() 
to return -1 and after that things get a little messy ;-).

I have been playing around with using PATH_INFO (in CGI speak) as an 
alternative to decoding the URLs. A set of patches follow, I would be 
interested in your thoughts.

Cheers, Matt

Dave Cole wrote:

>>>>>>"Matt" == Matt Goodall <matt@pollenation.net> writes:
>>>>>>            
>>>>>>
>
>Matt> Hi, Can anyone explain why the method in the subject is fiddling
>Matt> around with uri paths? It looks like it's trying to cope with
>Matt> the situation when the base uri and the request uri differ but I
>Matt> don't understand why and it's probably not quite right,
>Matt> i.e. what happens if the two URIs do not match correctly? When
>Matt> would the two URIs be different anyway?
>
>All of that code (which could probably be improved) is trying to
>ensure that only absolute redirect URLs come out of the application.
>
>You can see what it is trying to do by placing this application in
>your cgi-bin directory and pointing your browser at it.
>
>- - redirect.py - - - - - - - - - - - - - - - - - - - - - - - - - - -
>#!/usr/bin/python
>import sys
>import os
>
>if __name__ == '__main__':
>    host = os.environ.get('HTTP_HOST')
>    script_name = os.environ.get('SCRIPT_NAME')
>    uri = os.environ.get('REQUEST_URI')
>
>    page = uri[len(script_name):]
>    sys.stderr.write('uri="%s", page="%s"\n' % (uri, page))
>    if page == '/there':
>        print 'Content-Type: text/plain'
>        print
>        print 'You are there'
>    else:
>        # This works perfectly
>        loc = 'http://%s%s/there' % (host, script_name)
>        # These all cause all sorts of strife
>        #loc = '/%s/there' % script_name
>        #loc = 'there'
>        print 'Location:', loc
>        print 'Status:', '301 Moved Permanently'
>        print
>- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
>The Albatross code is just trying to figure out how to construct an
>absolute location for the requested redirect.  It also tries to avoid
>fiddling with a URL that is already absolute.
>
>Matt> Also, why does Albatross not make use of path info
>Matt> (os.environ['PATH_INFO'] in cgi, req.path_info in mod_python;
>Matt> not sure for BaseHTTPServer)? I think path info could simplify
>Matt> the parsing and building or URLs which has been causing me a few
>Matt> problems today when I popped a mod_python app behind a
>Matt> mod_rewrite rule :(.
>
>I am more than happy to change the current code if you figure out a
>better way to do it.
>
>- Dave
>
>  
>

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com
t: +44 (0)113 2252500



From matt at pollenation.net  Tue Jul  8 17:03:19 2003
From: matt at pollenation.net (Matt Goodall)
Date: Tue, 08 Jul 2003 18:03:19 +0100
Subject: [albatross-users] [patch] get_uri -> get_path_info changes
Message-ID: <3F0AF957.6050402@pollenation.net>

This is a multi-part message in MIME format.
--------------080504040305000105010102
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

WARNING: This email contains experimental patches which may be harmful 
to small children and production systems.

Attached are a series of patches that are sitting on my machine, some of 
which have already been accepted and applied:

    01.base_url.patch - a patch to fix my previous patch. Already
    applied by Dave I think.

    02.fcgiapp_get_header.patch - erm, see above ;-). Already applied by
    Andrew

    03.httpdapp.patch - simplified version of Request.get_uri().

    04.get_path_info.patch - introduces get_path_info() methods into the
    request adapters.

    05.use_path_info.patch - switch from using Request.get_uri() to
    Request.get_path_info(). This affects how the random page loader
    works and how the current and redirect urls are calculated.

All patches are based on 1.10pre2 and were applied in the order above.

I feel that I am getting a little out of my depth with these changes in 
that I'm not entirely sure what effect they may have on the rest of the 
framework. What I can say is that the popview5 and form4 samples still 
work and that a simple application (two linked pages, one with basic 
form processing) written as RandomModularSessionApp and SimpleSessionApp 
is working in cgi, fcgi, mod_python and httpdapp mode. These simple apps 
also work behind a mod_rewrite rule which previously failed for me and 
is the cause of my experimentation.

I would really, really appreciate some feedback on the path_info changes 
since the mod_rewrite stuff could be quite important to me.

I think I'd better do some real work tomorrow :(.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com


--------------080504040305000105010102
Content-Type: text/plain;
 name="01.base_url.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="01.base_url.patch"

diff -ru albatross-1.10pre2-original/albatross/app.py albatross-1.10pre2/albatross/app.py
--- albatross-1.10pre2-original/albatross/app.py	2003-07-05 08:52:40.000000000 +0100
+++ albatross-1.10pre2/albatross/app.py	2003-07-07 11:08:48.000000000 +0100
@@ -221,11 +221,12 @@
     def __init__(self, base_url):
         ResourceMixin.__init__(self)
         self.register_tagclasses(*tags.tags)
-        # Use a base_url of '/' if not provided
-        if base_url:
-            self.__base_url = base_url
-        else:
+        # Always append a '/' to the base_url if it is missing
+        self.__base_url = base_url
+        if not self.__base_url:
             self.__base_url = '/'
+        elif self.__base_url[-1] != '/':
+            self.__base_url = self.__base_url + '/'
 
     def run(self, req):
         '''Process a single browser request

--------------080504040305000105010102
Content-Type: text/plain;
 name="02.fcgiapp_get_header.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="02.fcgiapp_get_header.patch"

--- albatross-1.10pre2/albatross/fcgiapp.py	2003-07-05 04:12:16.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/fcgiapp.py	2003-07-07 23:03:53.000000000 +0100
@@ -19,14 +19,14 @@
         cgiapp.Request.__init__(self, fields)
                                                          
     def get_uri(self):
-        return self.__fcgi.env['REQUEST_URI']
+        return self.__fcgi.env.get('REQUEST_URI')
                                                                      
     def get_servername(self):
-        return self.__fcgi.env['SERVER_NAME']
+        return self.__fcgi.env.get('SERVER_NAME')
 
     def get_header(self, name):
         env_name = 'HTTP_' + name.upper().replace('-', '_')
-        return self.__req.env.get(env_name)
+        return self.__fcgi.env.get(env_name)
                                                      
 
 def running():

--------------080504040305000105010102
Content-Type: text/plain;
 name="03.httpdapp.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="03.httpdapp.patch"

--- albatross-1.10pre2/albatross/httpdapp.py	2003-07-05 11:29:01.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/httpdapp.py	2003-07-08 00:18:58.000000000 +0100
@@ -82,11 +82,7 @@
         return self._fields.keys()
 
     def get_uri(self):
-        server_name = self._req.server.server_name
-        server_port = self._req.server.server_port
-        path = self._req.path
-        uri = 'http://%s:%d%s' % (server_name, server_port, path)
-        return uri
+        return self._req.path
 
     def get_servername(self):
         server_name = self._req.server.server_name

--------------080504040305000105010102
Content-Type: text/plain;
 name="04.get_path_info.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="04.get_path_info.patch"

--- albatross-1.10pre2/albatross/apacheapp.py	2003-07-05 04:12:16.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/apacheapp.py	2003-07-07 21:41:50.000000000 +0100
@@ -17,6 +17,9 @@
     def get_uri(self):
         return self.__req.uri
 
+    def get_path_info(self):
+        return self.__req.path_info
+
     def get_servername(self):
         try:
             return self.__req.hostname                          # mod_python 3
--- albatross-1.10pre2/albatross/cgiapp.py	2003-07-05 04:49:13.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/cgiapp.py	2003-07-08 00:08:40.000000000 +0100
@@ -60,6 +60,9 @@
     def get_uri(self):
         return os.environ.get('REQUEST_URI')
 
+    def get_path_info(self):
+        return os.environ.get('PATH_INFO')
+
     def get_servername(self):
         return os.environ.get('SERVER_NAME')
 
--- albatross-1.10pre2/albatross/fcgiapp.py	2003-07-08 00:07:02.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/fcgiapp.py	2003-07-07 23:18:15.000000000 +0100
@@ -21,6 +21,9 @@
     def get_uri(self):
         return self.__fcgi.env.get('REQUEST_URI')
                                                                      
+    def get_path_info(self):
+        return self.__fcgi.env.get('PATH_INFO')
+                                                                     
     def get_servername(self):
         return self.__fcgi.env.get('SERVER_NAME')
 
--- albatross-1.10pre2/albatross/httpdapp.py	2003-07-08 00:21:08.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/httpdapp.py	2003-07-08 00:20:38.000000000 +0100
@@ -84,6 +84,9 @@
     def get_uri(self):
         return self._req.path
 
+    def get_path_info(self):
+        return self._req.path
+
     def get_servername(self):
         server_name = self._req.server.server_name
         port = self._req.server.server_port

--------------080504040305000105010102
Content-Type: text/plain;
 name="05.use_path_info.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="05.use_path_info.patch"

Only in albatross-1.10pre2-patched/albatross: .randompage.py.swp
diff -ru albatross-1.10pre2.2/albatross/app.py albatross-1.10pre2-patched/albatross/app.py
--- albatross-1.10pre2.2/albatross/app.py	2003-07-08 15:30:20.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/app.py	2003-07-08 12:17:35.000000000 +0100
@@ -191,19 +191,22 @@
         return self.app.base_url()
 
     def current_url(self):
-        uri = self.request.get_uri()
-        scheme, netloc, path = urlparse.urlparse(uri)[:3]
-        return path
+        if not self.request.get_path_info():
+            this_path = ''
+        else:
+            this_path = urlparse.urlparse(self.request.get_path_info())[2]
+        base_path = urlparse.urlparse(self.app.base_url())[2]
+        if base_path and base_path[-1] == '/':
+            if this_path and this_path[0] == '/':
+                this_path = this_path[1:]
+        return base_path + this_path
 
     def redirect_url(self, loc):
-        this_path = urlparse.urlparse(self.request.get_uri())[2]
         base_path = urlparse.urlparse(self.app.base_url())[2]
-        pos = this_path.find(base_path)
-        new_base = this_path[:pos + len(base_path)]
-        if new_base[-1] != '/':
-            new_base = new_base + '/'
+        if base_path[-1] != '/':
+            base_path = base_path + '/'
         return urlparse.urlunparse(('http', self.request.get_servername(),
-                                    new_base + loc, '', '', ''))
+                                    base_path + loc, '', '', ''))
 
     def redirect(self, loc):
         scheme, netloc, path = urlparse.urlparse(loc)[:3]
diff -ru albatross-1.10pre2.2/albatross/randompage.py albatross-1.10pre2-patched/albatross/randompage.py
--- albatross-1.10pre2.2/albatross/randompage.py	2003-07-05 08:52:40.000000000 +0100
+++ albatross-1.10pre2-patched/albatross/randompage.py	2003-07-08 15:30:56.000000000 +0100
@@ -18,14 +18,9 @@
 
     def load_page(self, ctx):
         # Get page name from request URI
-        uri = ctx.request.get_uri()
-        page = ''
-        try:
-            base_path = urlparse.urlparse(self.base_url())[2]
-            uri_path = urlparse.urlparse(uri)[2]
-            page = uri_path.split(base_path, 1)[1]
-        except IndexError:
-            pass
+        page = ctx.request.get_path_info()
+        if page and page[0] == '/':
+            page = page[1:]
         if not page:
             ctx.redirect(self.start_page())
 

--------------080504040305000105010102--


From gnb at itga.com.au  Tue Jul  8 23:22:42 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Wed, 09 Jul 2003 09:22:42 +1000
Subject: [albatross-users] Wiki example and related extension: iterate over ctx/session
In-Reply-To: Your message of 08 Jul 2003 23:50:24 +1000.
Message-ID: <200307082322.JAA28934@lightning.itga.com.au>

djc@object-craft.com.au said:
> How about this:
>    <al-for iter="i" expr="__ctx__.session_vars()">
>      <li><al-value expr="i.value()"/></li>
>    </al-for> 

Kool.  Works for me!



From sheila at thinkspot.net  Wed Jul  9 04:41:15 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 08 Jul 2003 21:41:15 -0700
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
Message-ID: <312402881.1057700475@SHEILA-LAPTOP>

OK, so I decided to try an upgrade to the latest pre-release.

Downloaded it and put it on my server.
Note: I do not have, and am not able to have root perms on the 
community server where my account is. However, I have installed my 
own local copy of Python, which I have aliased as "pythonx", and when 
I previously installed albatross 1.01 the following install command 
worked just fine:

$ pythonx setup.py install

and installed everything properly.

OK, so I uploaded the tarball of 1.10pre2 to my home directory and 
unzipped it; everything looked fine.

I gave the command:

$ pythonx setup.py install

All of the output looked fine.

I cd to the /test directory.

Unfortunately, there is a problem with running the test programs as 
directed. I entered

$ make

and got the following error output:

PYTHONPATH=.. python all.py
Traceback (innermost last):
  File "all.py", line 14, in ?
    import unittest
ImportError: No module named unittest
make: *** [unittests] Error 1

So, instead I tried this:

$ pythonx all.py

which seemed to work mostly fine. I did get one error result:

======================================================================
ERROR: runTest (sessions.basic.ServerBasicCase)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "sessions/basic.py", line 26, in runTest
    ctx.load_session()
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/session.py", 
line 50, in load_session
    sesid = self.app.new_session()
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/session.py", 
line 149, in new_session
    self._server_connect()
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/session.py", 
line 104, in _server_connect
    raise ServerError('could not connect to session server')
ServerError: could not connect to session server

----------------------------------------------------------------------
Ran 92 tests in 0.289s

FAILED (errors=1)


Which, I guess is expected, as I am not running the session server ???

Anyhow, the next thing I did was try to go to my application, which I 
had tested this afternoon before upgrading to the latest albatross.

I get the following error output when I try to go to the base URL for 
my application:

Template traceback (most recent call last):

Traceback (most recent call last):
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py", line 
238, in run
    self.load_page(ctx)
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/randompage.py
", line 35, in load_page
    self.load_page_module(ctx, page)
  File 
"/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py", line 
372, in load_page_module
    file, filepath, desc = imp.find_module(name, [dirname])
ImportError: No module named login

Which is really weird, because the login.py file hasn't moved. The 
app worked the last time I tried it, and the only thing that has 
changed, is that I have upgraded albatross.

Thoughts? Suggestions?



From andrewm at object-craft.com.au  Wed Jul  9 05:16:21 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 09 Jul 2003 15:16:21 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Tue, 08 Jul 2003 21:41:15 MST." <312402881.1057700475@SHEILA-LAPTOP>
References: <312402881.1057700475@SHEILA-LAPTOP>
Message-ID: <20030709051621.9388C3C024@coffee.object-craft.com.au>

>Unfortunately, there is a problem with running the test programs as 
>directed. I entered
>
>$ make
>
>and got the following error output:
>
>PYTHONPATH=.. python all.py
>Traceback (innermost last):
>  File "all.py", line 14, in ?
>    import unittest
>ImportError: No module named unittest
>make: *** [unittests] Error 1
>
>So, instead I tried this:
>
>$ pythonx all.py

Yes - the makefile ran the wrong python - your solution is the correct
one.  You could also have used "PYTHONPATH=.. pythonx all.py", which
would allow you to run the tests without installing albatross.

>which seemed to work mostly fine. I did get one error result:
[...]
>ServerError: could not connect to session server
>
>----------------------------------------------------------------------
>Ran 92 tests in 0.289s
>
>FAILED (errors=1)
>
>Which, I guess is expected, as I am not running the session server ???

That's right - if you started the session server (doesn't need to be
root), this test would have passed, but there's not really a lot of
point (and your hosts might get upset with random processes listening
on random ports).

>Anyhow, the next thing I did was try to go to my application, which I 
>had tested this afternoon before upgrading to the latest albatross.
>
>I get the following error output when I try to go to the base URL for 
>my application:
>
>Template traceback (most recent call last):
>
>Traceback (most recent call last):
>  File 
>"/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py", line 
>238, in run
>    self.load_page(ctx)
>  File 
>"/big/dom/xdomain/lib/python2.2/site-packages/albatross/randompage.py
>", line 35, in load_page
>    self.load_page_module(ctx, page)
>  File 
>"/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py", line 
>372, in load_page_module
>    file, filepath, desc = imp.find_module(name, [dirname])
>ImportError: No module named login
>
>Which is really weird, because the login.py file hasn't moved. The 
>app worked the last time I tried it, and the only thing that has 
>changed, is that I have upgraded albatross.

That's a curly one. I'd suggest putting some debug in before the above
line in albatross. Write os.getcwd() and dirname to sys.stderr. If you
don't have access to the web server's stderr output, you could write
the debug to a file in /tmp, etc.

"dirname" is derived from the module name ("login") and the module_path
you specified when creating the app object.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From sheila at thinkspot.net  Wed Jul  9 05:41:11 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 08 Jul 2003 22:41:11 -0700
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709051621.9388C3C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
Message-ID: <315999543.1057704071@SHEILA-LAPTOP>

--On Wednesday, July 09, 2003 3:16 PM +1000 Andrew McNamara 
<andrewm@object-craft.com.au> wrote:

> > Anyhow, the next thing I did was try to go to my application,
> > which I  had tested this afternoon before upgrading to the latest
> > albatross.
> >
> > I get the following error output when I try to go to the base URL
> > for  my application:
> >
> > Template traceback (most recent call last):
> >
> > Traceback (most recent call last):
> >  File
> > "/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py",
> > line  238, in run
> >    self.load_page(ctx)
> >  File
> > "/big/dom/xdomain/lib/python2.2/site-packages/albatross/randompag
> > e.py ", line 35, in load_page
> >    self.load_page_module(ctx, page)
> >  File
> > "/big/dom/xdomain/lib/python2.2/site-packages/albatross/app.py",
> > line  372, in load_page_module
> >    file, filepath, desc = imp.find_module(name, [dirname])
> > ImportError: No module named login
> >
> > Which is really weird, because the login.py file hasn't moved.
> > The  app worked the last time I tried it, and the only thing that
> > has  changed, is that I have upgraded albatross.
>


> That's a curly one. I'd suggest putting some debug in before the
> above line in albatross. Write os.getcwd() and dirname to
> sys.stderr. If you don't have access to the web server's stderr
> output, you could write the debug to a file in /tmp, etc.
>
> "dirname" is derived from the module name ("login") and the
> module_path you specified when creating the app object.
>

OK, good suggestions. I have done this, and the output was as follows:

os.getcwd() = /big/dom/xdomain/cgi-bin/qa_auth
dirname = /


cwd looks correct, as that is the directory wherein the cgi file is 
stored for the main app file.

I am using the SessionFileAppMixin and there is no place to specify 
an alternate directory for the template files (or at least there 
wasn't in the previous version).

My class definition and __init__ is as follows:

class qaApp(PickleSignMixin,
              Application,
              TemplateLoaderMixin,
              RandomPageModuleMixin,
              SessionFileAppMixin):

    def __init__(self, base_url = 'mgr.py',
                   page_path = '.',
                   start_page = 'login',
                   secret = '-=-secret-=-',
                   session_appid = 'questadmin',
                   session_dir = './sessions'):
        Application.__init__(self, base_url)
        PickleSignMixin.__init__(self, secret)
        TemplateLoaderMixin.__init__(self, page_path)
        RandomPageModuleMixin.__init__(self, page_path, start_page)
        SessionFileAppMixin.__init__(self, session_appid, session_dir)

Ah, actually I do see, there is a "page_path" and it is correct. The 
pages are all in the same directory with the main app file.

K, so nothing looks wrong to me here. :/

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org




From andrewm at object-craft.com.au  Wed Jul  9 06:03:47 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 09 Jul 2003 16:03:47 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Tue, 08 Jul 2003 22:41:11 MST." <315999543.1057704071@SHEILA-LAPTOP>
References: <312402881.1057700475@SHEILA-LAPTOP> <20030709051621.9388C3C024@coffee.object-craft.com.au>  <315999543.1057704071@SHEILA-LAPTOP>
Message-ID: <20030709060347.283353C024@coffee.object-craft.com.au>

>OK, good suggestions. I have done this, and the output was as follows:
>
>os.getcwd() = /big/dom/xdomain/cgi-bin/qa_auth
>dirname = /

Okay - this means that load_page_module has been called with name="/login" - 
this means the page name extraction code in RandomPageModuleMixin.load_page() 
is buggy. Dave will look at this shortly, I hope.

Dave - we probably also need to add a check to ensure load_page_module
is never called with an absolute path - it's a security problem.

>    def __init__(self, base_url = 'mgr.py',
>                   page_path = '.',
>                   start_page = 'login',
>                   secret = '-=-secret-=-',
>                   session_appid = 'questadmin',
>                   session_dir = './sessions'):
>        Application.__init__(self, base_url)
>        PickleSignMixin.__init__(self, secret)
>        TemplateLoaderMixin.__init__(self, page_path)
>        RandomPageModuleMixin.__init__(self, page_path, start_page)
>        SessionFileAppMixin.__init__(self, session_appid, session_dir)
>
>Ah, actually I do see, there is a "page_path" and it is correct. The 
>pages are all in the same directory with the main app file.

As a workaround, can you try setting base_url to 'mgr.py/' (note the
addition of a slash to the end of the string)?

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From sheila at thinkspot.net  Wed Jul  9 06:29:09 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 08 Jul 2003 23:29:09 -0700
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709060347.283353C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
Message-ID: <318876900.1057706949@SHEILA-LAPTOP>

--On Wednesday, July 09, 2003 4:03 PM +1000 Andrew McNamara 
<andrewm@object-craft.com.au> wrote:

>
> As a workaround, can you try setting base_url to 'mgr.py/' (note the
> addition of a slash to the end of the string)?

OK, well this resulted in a very weird error.

Instead of sending me to

http://www.example.com/cgi-bin/qa_auth/mgr.py/login

which is where it *should* send me, it sent me to the following:

http://www.example.com/cgi-b/login

which resulted in a 404 error from my web site's Apache server (i.e. 
the usual 404 page for my site, and not the Bad URL error from 
albatross)

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org



From andrewm at object-craft.com.au  Wed Jul  9 06:42:14 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 09 Jul 2003 16:42:14 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Tue, 08 Jul 2003 23:29:09 MST." <318876900.1057706949@SHEILA-LAPTOP>
References: <312402881.1057700475@SHEILA-LAPTOP> <20030709051621.9388C3C024@coffee.object-craft.com.au> <315999543.1057704071@SHEILA-LAPTOP> <20030709060347.283353C024@coffee.object-craft.com.au>  <318876900.1057706949@SHEILA-LAPTOP>
Message-ID: <20030709064214.A35E43C024@coffee.object-craft.com.au>

>> As a workaround, can you try setting base_url to 'mgr.py/' (note the
>> addition of a slash to the end of the string)?
>
>OK, well this resulted in a very weird error.
>
>Instead of sending me to
>
>http://www.example.com/cgi-bin/qa_auth/mgr.py/login
>
>which is where it *should* send me, it sent me to the following:
>
>http://www.example.com/cgi-b/login
>
>which resulted in a 404 error from my web site's Apache server (i.e. 
>the usual 404 page for my site, and not the Bad URL error from 
>albatross)

Ha. So much for that idea. The cause of the problem is Matt Goodall's
patch to allow random page applications to work with his standalone
server. Until we work out a solution that allows both deployment methods
to work, it's probably enough just to revert this change - randompage.py,
line 26:

            page = uri_path.split(base_path, 1)[1]

should be changed to read:

            page = uri_path.split(base_path + '/', 1)[1]

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From sheila at thinkspot.net  Wed Jul  9 07:03:05 2003
From: sheila at thinkspot.net (Sheila King)
Date: Wed, 09 Jul 2003 00:03:05 -0700
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709064214.A35E43C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <318876900.1057706949@SHEILA-LAPTOP>
 <20030709064214.A35E43C024@coffee.object-craft.com.au>
Message-ID: <320913679.1057708985@SHEILA-LAPTOP>

--On Wednesday, July 09, 2003 4:42 PM +1000 Andrew McNamara 
<andrewm@object-craft.com.au> wrote:

> >> As a workaround, can you try setting base_url to 'mgr.py/' (note
> >> the addition of a slash to the end of the string)?
> >
> > OK, well this resulted in a very weird error.
> >
> > Instead of sending me to
> >
> > http://www.example.com/cgi-bin/qa_auth/mgr.py/login
> >
> > which is where it *should* send me, it sent me to the following:
> >
> > http://www.example.com/cgi-b/login
> >
> > which resulted in a 404 error from my web site's Apache server
> > (i.e.  the usual 404 page for my site, and not the Bad URL error
> > from  albatross)
>
> Ha. So much for that idea. The cause of the problem is Matt
> Goodall's patch to allow random page applications to work with his
> standalone server. Until we work out a solution that allows both
> deployment methods to work, it's probably enough just to revert
> this change - randompage.py, line 26:
>
>             page = uri_path.split(base_path, 1)[1]
>
> should be changed to read:
>
>             page = uri_path.split(base_path + '/', 1)[1]

OK, well that gets me past that error. Now I have to deal with the 
fact that I've apparently got a problem due to setting session vars 
before they exist in the local name space (which is an error now, and 
didn't used to be).

Also, there is now something weird with the way the headers are being 
written, and the headers are appearing in my web page.

Thanks for your help so far. I don't think I would have figured out 
the path problems with the page module on my own.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org






From andrewm at object-craft.com.au  Wed Jul  9 08:00:10 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 09 Jul 2003 18:00:10 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Wed, 09 Jul 2003 00:03:05 MST." <320913679.1057708985@SHEILA-LAPTOP>
References: <312402881.1057700475@SHEILA-LAPTOP> <20030709051621.9388C3C024@coffee.object-craft.com.au> <315999543.1057704071@SHEILA-LAPTOP> <20030709060347.283353C024@coffee.object-craft.com.au> <318876900.1057706949@SHEILA-LAPTOP> <20030709064214.A35E43C024@coffee.object-craft.com.au>  <320913679.1057708985@SHEILA-LAPTOP>
Message-ID: <20030709080010.EC71C3C024@coffee.object-craft.com.au>

>OK, well that gets me past that error. Now I have to deal with the 
>fact that I've apparently got a problem due to setting session vars 
>before they exist in the local name space (which is an error now, and 
>didn't used to be).

Ah - yes. I'm sure this change seemed like a good idea at the time... 8-)

I think very strange things happened later in the application's life if
you forgot to set the session var. Rather than have albatross explode
in obscure ways, we're enforcing the requirement that variables be set
before they're registered.

>Also, there is now something weird with the way the headers are being 
>written, and the headers are appearing in my web page.

The header code had to have a revamp to allow arbitary headers to be set
by the application (content encodings and things like that). Are you
able to strip the application down to the simplest form that exhibits
this problem and forward the result to me?

>Thanks for your help so far. I don't think I would have figured out 
>the path problems with the page module on my own.

Sticking debug into albatross itself is a useful technique (albatross
should have more hooks for debugging). There's nothing sacred about the
module - we intend people to mess around inside it.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From matt at pollenation.net  Wed Jul  9 09:03:54 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 10:03:54 +0100
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709064214.A35E43C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <318876900.1057706949@SHEILA-LAPTOP>
 <20030709064214.A35E43C024@coffee.object-craft.com.au>
Message-ID: <1057741433.27471.50.camel@localhost>

On Wed, 2003-07-09 at 07:42, Andrew McNamara wrote:

> Ha. So much for that idea. The cause of the problem is Matt Goodall's
> patch to allow random page applications to work with his standalone
> server.

Eek, sorry!

The 2nd patch for base_url will not fix this either as I never even
considered that base_url would not be an absolute path on the server
:-(.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Wed Jul  9 09:06:41 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 10:06:41 +0100
Subject: [albatross-users] AppContext.redirect_url() question
In-Reply-To: <3F0AF185.2050209@pollenation.net>
References: <3F099AB2.80805@pollenation.net>
 <m34r1xyht2.fsf@ferret.object-craft.com.au>
 <3F0AF185.2050209@pollenation.net>
Message-ID: <1057741600.29259.54.camel@localhost>

On Tue, 2003-07-08 at 17:29, Matt Goodall wrote:
> I'm probably being really stupid and missing something obvious but this 
> is the bit from redirect_url() I'm confused about:
> 
>         this_path = urlparse.urlparse(self.request.get_uri())[2]
>         base_path = urlparse.urlparse(self.app.base_url())[2]
>         pos = this_path.find(base_path)
>         new_base = this_path[:pos + len(base_path)]
> 
> Firstly, this_path and base_path are normally the same up to the name of 
> the script and so pos will be 0. The only time pos can be non-zero is 
> when the URLs don't match at all or this_path contains an extra path at 
> the start. I can't think of a situation where this_path has the extra path.
> 
> So ... that makes new_base the same as base_path and so I don't 
> understand why you cannot simply add the redirect location to the 
> base_path. You can all humiliate me now and tell me that I can't read 
> simple code if you want ;-).
> 
> My big question is why bother with pos and the this_path slice? What is 
> it needed for?

OK, now I know after reading Sheila's "Upgrade problems" thread. It's
for when base_url is not absolute.

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From andrewm at object-craft.com.au  Wed Jul  9 09:22:45 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 09 Jul 2003 19:22:45 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: Message from Matt Goodall <matt@pollenation.net>
 of "09 Jul 2003 10:03:54 +0100." <1057741433.27471.50.camel@localhost>
References: <312402881.1057700475@SHEILA-LAPTOP> <20030709051621.9388C3C024@coffee.object-craft.com.au> <315999543.1057704071@SHEILA-LAPTOP> <20030709060347.283353C024@coffee.object-craft.com.au> <318876900.1057706949@SHEILA-LAPTOP> <20030709064214.A35E43C024@coffee.object-craft.com.au>  <1057741433.27471.50.camel@localhost>
Message-ID: <20030709092245.79B793C024@coffee.object-craft.com.au>

>> Ha. So much for that idea. The cause of the problem is Matt Goodall's
>> patch to allow random page applications to work with his standalone
>> server.
>
>Eek, sorry!
>
>The 2nd patch for base_url will not fix this either as I never even
>considered that base_url would not be an absolute path on the server
>:-(.

Ah well - we applied it without noticing either. Add some comments next
time, Dave... (if in doubt, blame Dave)... 8-)

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From matt at pollenation.net  Wed Jul  9 09:44:09 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 10:44:09 +0100
Subject: [albatross-users] [patch] get_uri -> get_path_info changes
In-Reply-To: <3F0AF957.6050402@pollenation.net>
References: <3F0AF957.6050402@pollenation.net>
Message-ID: <1057743848.28628.75.camel@localhost>

In light of the problem Sheila found, below are some comments on the
patches.

On Tue, 2003-07-08 at 18:03, Matt Goodall wrote:

>     01.base_url.patch - a patch to fix my previous patch. Already
>     applied by Dave I think.

This was clearly the wrong solution as the original base_url patch did
not consider that the base_url might not be absolute.

>     02.fcgiapp_get_header.patch - erm, see above ;-). Already applied by
>     Andrew

Necessary and already applied.

>     03.httpdapp.patch - simplified version of Request.get_uri().

Not necessary but still an improvement. It's kind of irrelevant though
as it doesn't work with the old base_url stuff.

>     04.get_path_info.patch - introduces get_path_info() methods into the
>     request adapters.

Not much point with 05 but it may be useful anyway. Dave/Andrew, it's up
to you.

>     05.use_path_info.patch - switch from using Request.get_uri() to
>     Request.get_path_info(). This affects how the random page loader
>     works and how the current and redirect urls are calculated.

Completely breaks redirect_url when the base_url is not absolute. Do not
apply.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From djc at object-craft.com.au  Wed Jul  9 11:45:01 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 09 Jul 2003 21:45:01 +1000
Subject: [albatross-users] AppContext.redirect_url() question
References: <3F099AB2.80805@pollenation.net>
 <m34r1xyht2.fsf@ferret.object-craft.com.au>
 <3F0AF185.2050209@pollenation.net>
Message-ID: <m3u19vrkjm.fsf@ferret.object-craft.com.au>

> I'm probably being really stupid and missing something obvious but
> this is the bit from redirect_url() I'm confused about:

Not stupid, just not able to read my mind...

>         this_path = urlparse.urlparse(self.request.get_uri())[2]
>         base_path = urlparse.urlparse(self.app.base_url())[2]
>         pos = this_path.find(base_path)
>         new_base = this_path[:pos + len(base_path)]

The idea here is that the URI for a request might be something like
this:

        http://blah.com/path/to/app/app.py/page/here

When you call the application constructor you might have specified a
base_url of 'app.py'.  The toolkit needs to be able to find the
page/here part of the request URI.

The first step is to isolate the path component of both request URI
and the base_url.  That is done by the first two lines.

    this_path = urlparse.urlparse(self.request.get_uri())[2]
    base_path = urlparse.urlparse(self.app.base_url())[2]

For the example above this leads to the following paths

    this_path path/to/app/app.py/page/here and 
    base_path app.py

Then the following lines get executed:

    pos = this_path.find(base_path)
    new_base = this_path[:pos + len(base_path)]

So new_base is

    new_base = path/to/app/app.py

> Firstly, this_path and base_path are normally the same up to the
> name of the script and so pos will be 0. The only time pos can be
> non-zero is when the URLs don't match at all or this_path contains
> an extra path at the start. I can't think of a situation where
> this_path has the extra path.

You already answered this in a later email :-)

I had to go out tonight so did not get to finish this message until
now.

> Anyway, that is just me trying to figure out what's going on. More
> importantly, at least for me, when I put a mod_python app behind a
> mod_rewrite rule redirect_url() fails since mod_python's
> Request.get_uri() returns the *rewritten* path which may have no
> relationship to the application's base_url. That causes
> this_path.find() to return -1 and after that things get a little
> messy ;-).

I have been thinking that there should be enough information in the
request to remove the need to specify base_url.  The base_url is only
there for the following reasons:

* To allow the application to generate URLs to itself.

* To allow the application to work out the above (split base from
  page).

> I have been playing around with using PATH_INFO (in CGI speak) as an
> alternative to decoding the URLs. A set of patches follow, I would
> be interested in your thoughts.

I will have a look.  I would be interested to know if you think we can
eliminate the need to specify a base_url to the application
constructor.  This would probably make URI parsing a lot more easy.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Wed Jul  9 13:04:38 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 09 Jul 2003 23:04:38 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709060347.283353C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
Message-ID: <m3ptkjrgux.fsf@ferret.object-craft.com.au>

>>>>> "Andrew" == Andrew McNamara <andrewm@object-craft.com.au> writes:

>> OK, good suggestions. I have done this, and the output was as
>> follows:
>> 
>> os.getcwd() = /big/dom/xdomain/cgi-bin/qa_auth dirname = /

Andrew> Okay - this means that load_page_module has been called with
Andrew> name="/login" - this means the page name extraction code in
Andrew> RandomPageModuleMixin.load_page() is buggy. Dave will look at
Andrew> this shortly, I hope.

I have been wondering if it is possible for the application to
discover the base_url from the web server.  It certainly is using CGI,
so probably is with FastCGI too, but I am not so sure about
mod_python.  I am certain the BaseHTTPServer cannot supply enough
information.

With CGI the SCRIPT_NAME environment tells you the base_url.

Just played a bit With mod_python.  The path_info does not help me as
much as it could.  Maybe I have missed something in my Apache
configuration.  I have allowed .htaccess files by doing this:

<Directory /var/www/alsamp/>
    AllowOverride FileInfo Indexes
    Order allow,deny
    Allow from all
</Directory>

Then in the /var/www/alsamp/simple directory I have added this
.htacess file:

DirectoryIndex simple.py
SetHandler python-program
PythonHandler simple

Now the funny thing is that when I look at the Request.path_info
information there is one path element missing...

uri         /alsamp/simple/something/missing
path_info   /missing

I was sure that the 'something/' would appear in the path_info.

If there was some way to find the missing path element then there
would be no need for the base_url argument to the application
constructor.  This would allow the toolkit to be certain about the
part of the URI that corresponded to the application location, and so
it would be able to find the path to the right of that.

Andrew> Dave - we probably also need to add a check to ensure
Andrew> load_page_module is never called with an absolute path - it's
Andrew> a security problem.

I think we need to prevent any attempt to climb directories.  Didn't
you have some code for that somewhere?

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Wed Jul  9 13:35:26 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 09 Jul 2003 23:35:26 +1000
Subject: [albatross-users] Upgrade problems to albatross 1.10pre2
In-Reply-To: <20030709092245.79B793C024@coffee.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <318876900.1057706949@SHEILA-LAPTOP>
 <20030709064214.A35E43C024@coffee.object-craft.com.au>
 <1057741433.27471.50.camel@localhost>
 <20030709092245.79B793C024@coffee.object-craft.com.au>
Message-ID: <m3llv7rffl.fsf@ferret.object-craft.com.au>

>>>>> "Andrew" == Andrew McNamara <andrewm@object-craft.com.au> writes:

>>> Ha. So much for that idea. The cause of the problem is Matt
>>> Goodall's patch to allow random page applications to work with his
>>> standalone server.
>>  Eek, sorry!
>> 
>> The 2nd patch for base_url will not fix this either as I never even
>> considered that base_url would not be an absolute path on the
>> server :-(.

Andrew> Ah well - we applied it without noticing either. Add some
Andrew> comments next time, Dave... (if in doubt, blame Dave)... 8-)

Always a good move - you are seldom wrong...

I am completely stumped trying to get a path_info that is not missing
the first directory element.  Try out the following files with
mod_python and tell me what I am doing wrong.

- Dave

- - /var/www/alsamp/simple/.htaccess - - - - - - - - - - - - - - - -
DirectoryIndex app.py
SetHandler python-program
PythonHandler app
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

- - /var/www/alsamp/simple/app.py - - - - - - - - - - - - - - - - -
from mod_python import apache
from albatross import SimpleContext
 
def handler(req):
    req.content_type = 'text/html'
    req.send_http_header()
 
    ctx = SimpleContext('.')
    ctx.locals.req = req
    templ = ctx.load_template('/var/www/alsamp/simple/req.html')
    ctx.push_content_trap()
    templ.to_html(ctx)
    req.write(ctx.pop_content_trap())
 
    return apache.OK
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

- - /var/www/alsamp/simple/req.html - - - - - - - - - - - - - - - -
<html>
 <head>
  <title>The mod_python environment</title>
 </head>
 <body>
  <b>Request members</b><br>
  <al-exec expr="r = req._req; n = dir(r)">
  <table>
   <al-for iter="i" expr="n">
    <tr>
     <td><al-value expr="i.value()"></td>
     <td><al-value expr="getattr(r, i.value())"></td>
    <tr>
   </al-for>
  </table>
  <b>get_config()</b><br>
  <al-exec expr="t = req.get_config(); n = t.keys(); n.sort()">
  <table>
   <al-for iter="i" expr="n">
    <tr>
     <td><al-value expr="i.value()"></td>
     <td><al-value expr="t[i.value()]"></td>
    <tr>
   </al-for>
  </table>
  <b>get_dirs()</b><br>
  <al-exec expr="t = req.get_dirs(); n = t.keys(); n.sort()">
  <table>
   <al-for iter="i" expr="n">
    <tr>
     <td><al-value expr="i.value()"></td>
     <td><al-value expr="t[i.value()]"></td>
    <tr>
   </al-for>
  </table>
 </body>
</html>
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Wed Jul  9 14:16:51 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 15:16:51 +0100
Subject: [albatross-users] [patch] Another attempt at "fixing" base_url
Message-ID: <1057760211.28628.269.camel@localhost>

--=-9GXKMzKrz+DiGLRJckg4
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Attached is another patch, against stock 1.10pre2, which attempts to
make the base_url + httpdapp combination work without breaking
everything else ;-).

This time, with the knowledge that a relative base_url is perfectly
valid, I have taken the approach of "fixing up" the base_url where it is
used to determine the page and build redirect urls. I hope this approach
is a little more successful than the previous one!

I have tested this for a RandomModularSessionApp in cgiapp, fcgiapp and
apacheapp mode with an absolute and relative base_url. It also works for
httpdapp.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

--=-9GXKMzKrz+DiGLRJckg4
Content-Disposition: attachment; filename=base_url.patch
Content-Type: text/x-patch; name=base_url.patch; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

diff -ru albatross-1.10pre2/albatross/app.py albatross-1.10pre2-matt/albatr=
oss/app.py
--- albatross-1.10pre2/albatross/app.py	2003-07-05 08:52:40.000000000 +0100
+++ albatross-1.10pre2-matt/albatross/app.py	2003-07-09 15:11:13.000000000 =
+0100
@@ -198,6 +198,9 @@
     def redirect_url(self, loc):
         this_path =3D urlparse.urlparse(self.request.get_uri())[2]
         base_path =3D urlparse.urlparse(self.app.base_url())[2]
+        # If no base_path then the application is mounted at the root
+        if not base_path:
+            base_path =3D '/'
         pos =3D this_path.find(base_path)
         new_base =3D this_path[:pos + len(base_path)]
         if new_base[-1] !=3D '/':
@@ -221,11 +224,7 @@
     def __init__(self, base_url):
         ResourceMixin.__init__(self)
         self.register_tagclasses(*tags.tags)
-        # Use a base_url of '/' if not provided
-        if base_url:
-            self.__base_url =3D base_url
-        else:
-            self.__base_url =3D '/'
+        self.__base_url =3D base_url
=20
     def run(self, req):
         '''Process a single browser request
diff -ru albatross-1.10pre2/albatross/randompage.py albatross-1.10pre2-matt=
/albatross/randompage.py
--- albatross-1.10pre2/albatross/randompage.py	2003-07-05 08:52:40.00000000=
0 +0100
+++ albatross-1.10pre2-matt/albatross/randompage.py	2003-07-09 13:14:32.000=
000000 +0100
@@ -23,6 +23,12 @@
         try:
             base_path =3D urlparse.urlparse(self.base_url())[2]
             uri_path =3D urlparse.urlparse(uri)[2]
+            # Fixup base_path to end with a '/' so that it can be used as
+            # a root path.
+            if not base_path:
+                base_path =3D '/'
+            elif base_path[-1] !=3D '/':
+                base_path =3D  base_path + '/'
             page =3D uri_path.split(base_path, 1)[1]
         except IndexError:
             pass

--=-9GXKMzKrz+DiGLRJckg4--

From matt at pollenation.net  Wed Jul  9 15:50:54 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 16:50:54 +0100
Subject: path_info (was Re: [albatross-users] Upgrade problems to albatross
 1.10pre2)
In-Reply-To: <m3llv7rffl.fsf@ferret.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <318876900.1057706949@SHEILA-LAPTOP>
 <20030709064214.A35E43C024@coffee.object-craft.com.au>
 <1057741433.27471.50.camel@localhost>
 <20030709092245.79B793C024@coffee.object-craft.com.au>
 <m3llv7rffl.fsf@ferret.object-craft.com.au>
Message-ID: <1057765854.28628.282.camel@localhost>

On Wed, 2003-07-09 at 14:35, Dave Cole wrote:

> I am completely stumped trying to get a path_info that is not missing
> the first directory element.  Try out the following files with
> mod_python and tell me what I am doing wrong.

I get the same missing path element when relying on the DirectoryIndex
but it works when I add app.py to the URL:

	/~matt/simple/1/2/3
	filename: /home/matt/public_html/simple/1
	path_info: /2/3

	/~matt/simple/app.py/1/2/3
	filename: /home/matt/public_html/simple/app.py
	path:info: /1/2/3

Very strange, it's as if mod_python just assumes that the first bit is
the script name.

I'll ask on the mod_python list. We might have an answer by the time
Australia wakes up.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Wed Jul  9 16:13:16 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 17:13:16 +0100
Subject: base_url discovery (was Re: [albatross-users] Upgrade problems to
 albatross 1.10pre2)
In-Reply-To: <m3ptkjrgux.fsf@ferret.object-craft.com.au>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <m3ptkjrgux.fsf@ferret.object-craft.com.au>
Message-ID: <1057767195.28628.311.camel@localhost>

On Wed, 2003-07-09 at 14:04, Dave Cole wrote:
> I have been wondering if it is possible for the application to
> discover the base_url from the web server.  It certainly is using CGI,
> so probably is with FastCGI too,

As you pointed out, neither of these should be a problem.

> but I am not so sure about mod_python.

The problem with this is the one I hit the other day when using
mod_rewrite. The uri that Albatross sees is the rewritten uri which can
be completely different to what is typed into the browser url bar.

Having said that, mod_python has req.the_request which is the http
request header, i.e. 'GET /~matt/simple/app.py/1/2/3 HTTP/1.1'. We may
be able to make use of that but I haven't looked into it anymore.

> I am certain the BaseHTTPServer cannot supply enough
> information.

No, but it probably doesn't need to as the application is always mounted
at '/'.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Wed Jul  9 16:35:26 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 17:35:26 +0100
Subject: [albatross-users] missing path element in path_info
Message-ID: <1057768526.27471.335.camel@localhost>

Hi,

Over in the world of Albatross
(http://www.object-craft.com.au/projects/albatross/) we've been playing
with path_info and have come across something that we don't quite
understand.

Here's some example code to demonstrate:

----- .htaccess -----
DirectoryIndex app.py
SetHandler python-program
PythonHandler app
-----

----- app.py -----
from mod_python import apache
                                                                                def handler(req):
    req.content_type = 'text/plain'
    req.send_http_header()

    req.write('the_request: %s\n' % req.the_request)
    req.write('unparsed_uri: %s\n' % req.unparsed_uri)
    req.write('uri: %s\n' % req.uri)
    req.write('filename: %s\n' % req.filename)
    req.write('path_info: %s\n' % req.path_info)

    return apache.OK
-----

If I put those files in /~matt/test on the web server then this is what
I see for each URL:

        = URL /~matt/test/1/2/3 =
        the_request: GET /~matt/test/app.py/1/2/3 HTTP/1.1
        unparsed_uri: /~matt/test/app.py/1/2/3
        uri: /~matt/test/app.py/1/2/3
        filename: /home/matt/public_html/test/app.py
        path_info: /1/2/3
        
        = /~matt/test/app.py/1/2/3 =
        the_request: GET /~matt/test/app.py/1/2/3 HTTP/1.1
        unparsed_uri: /~matt/test/app.py/1/2/3
        uri: /~matt/test/app.py/1/2/3
        filename: /home/matt/public_html/test/app.py
        path_info: /1/2/3
        
        = /~matt/test/1/2/3 =
        the_request: GET /~matt/test/1/2/3 HTTP/1.1
        unparsed_uri: /~matt/test/1/2/3
        uri: /~matt/test/1/2/3
        filename: /home/matt/public_html/test/1
        path_info: /2/3
        
The first two are as expected but the last one drops app.py completely
and thinks that the first element of path_info is part of the filename.

Can anyone explain this?

Server details:

	Apache 1.3.27
	mod_python 2.7.8
	Linux 2.4.x, Gentoo and Debian (I think).

Thanks in advance.

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Wed Jul  9 16:42:32 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 17:42:32 +0100
Subject: [albatross-users] request debugging results
Message-ID: <1057768952.27471.342.camel@localhost>

I just added some debugging to the cgi, fcgi and mod_python request
adapters to dump everything out. I thought the results might be useful
with the current discussions about base_url and path_info.

- Matt


***** cgiapp
PATH_INFO: /home
SCRIPT_FILENAME:
/home/matt/public_html/albatross/path_info/connector.cgi
SCRIPT_NAME: /~matt/albatross/path_info/connector.cgi
HTTP_HOST: mattgoodall.dyndns.org
QUERY_STRING:
PATH_TRANSLATED: /home/httpd/htdocs/home
REQUEST_URI: /~matt/albatross/path_info/connector.cgi/home
SERVER_NAME: mattgoodall.dyndns.org
SERVER_PORT: 80
DOCUMENT_ROOT: /home/httpd/htdocs/

***** fcgiapp
PATH_INFO: /home
SCRIPT_FILENAME:
/home/matt/public_html/albatross/path_info/connector.fcgi
SCRIPT_NAME: /~matt/albatross/path_info/connector.fcgi
HTTP_HOST: mattgoodall.dyndns.org
QUERY_STRING:
PATH_TRANSLATED: /home/httpd/htdocs/home
REQUEST_URI: /~matt/albatross/path_info/connector.fcgi/home
SERVER_NAME: mattgoodall.dyndns.org
SERVER_PORT: 80
DOCUMENT_ROOT: /home/httpd/htdocs/

***** apacheapp
req.the_request: GET /~matt/albatross/path_info/connector.py/home
HTTP/1.1
req.unparsed_uri: /~matt/albatross/path_info/connector.py/home
req.uri: /~matt/albatross/path_info/connector.py/home
req.filename: /home/matt/public_html/albatross/path_info/connector.py
req.path_info: /home
req.args: None
req.server.server_hostname: mattgoodall.dyndns.org
req.server.port: 80

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Wed Jul  9 21:35:32 2003
From: matt at pollenation.net (Matt Goodall)
Date: 09 Jul 2003 22:35:32 +0100
Subject: [albatross-users] missing path element in path_info
In-Reply-To: <1057768526.27471.335.camel@localhost>
References: <1057768526.27471.335.camel@localhost>
Message-ID: <1057786532.28628.527.camel@localhost>

OK, so I got the example output completely wrong (boy, what a day I've
had!) but once I corrected it I got a response from the mod_python list
which I have summarised below.

re: the missing path element in path_info ...

        "In the last one, the reason "1" is not part of path_info:
        path_info is something apache figures out. I think the rule is
        that it sees that "test" is a directory, and the next thing
        after it isn't, thus, it must refer to a file (whether or not
        the file exists doesn't matter). Everything after a file is
        path_info."

re: is there a foolproof way to deduce the correct base url and
path_info ...

        "... it is impossible or very difficult to do, because the
        transformation from "~matt/test" to
        "/home/matt/public_html/test" is not "reversible" and is hidden
        somewhere in apache config, yet this mapping is necessary to
        produce the results above."

If you're interested in the full thread see
http://www.modpython.org/pipermail/mod_python/2003-July/003420.html.

Cheers, Matt


On Wed, 2003-07-09 at 17:35, Matt Goodall wrote:
> Hi,
> 
> Over in the world of Albatross
> (http://www.object-craft.com.au/projects/albatross/) we've been playing
> with path_info and have come across something that we don't quite
> understand.
> 
> Here's some example code to demonstrate:
> 
> ----- .htaccess -----
> DirectoryIndex app.py
> SetHandler python-program
> PythonHandler app
> -----
> 
> ----- app.py -----
> from mod_python import apache
>                                                                                 def handler(req):
>     req.content_type = 'text/plain'
>     req.send_http_header()
> 
>     req.write('the_request: %s\n' % req.the_request)
>     req.write('unparsed_uri: %s\n' % req.unparsed_uri)
>     req.write('uri: %s\n' % req.uri)
>     req.write('filename: %s\n' % req.filename)
>     req.write('path_info: %s\n' % req.path_info)
> 
>     return apache.OK
> -----
> 
> If I put those files in /~matt/test on the web server then this is what
> I see for each URL:
> 
>         = URL /~matt/test/1/2/3 =
>         the_request: GET /~matt/test/app.py/1/2/3 HTTP/1.1
>         unparsed_uri: /~matt/test/app.py/1/2/3
>         uri: /~matt/test/app.py/1/2/3
>         filename: /home/matt/public_html/test/app.py
>         path_info: /1/2/3
>         
>         = /~matt/test/app.py/1/2/3 =
>         the_request: GET /~matt/test/app.py/1/2/3 HTTP/1.1
>         unparsed_uri: /~matt/test/app.py/1/2/3
>         uri: /~matt/test/app.py/1/2/3
>         filename: /home/matt/public_html/test/app.py
>         path_info: /1/2/3
>         
>         = /~matt/test/1/2/3 =
>         the_request: GET /~matt/test/1/2/3 HTTP/1.1
>         unparsed_uri: /~matt/test/1/2/3
>         uri: /~matt/test/1/2/3
>         filename: /home/matt/public_html/test/1
>         path_info: /2/3
>         
> The first two are as expected but the last one drops app.py completely
> and thinks that the first element of path_info is part of the filename.
> 
> Can anyone explain this?
> 
> Server details:
> 
> 	Apache 1.3.27
> 	mod_python 2.7.8
> 	Linux 2.4.x, Gentoo and Debian (I think).
> 
> Thanks in advance.
> 
> - Matt
-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net
t: 0113 2252500

From rbw at demec.ufpe.br  Wed Jul  9 17:59:35 2003
From: rbw at demec.ufpe.br (Ramiro Brito Willmersdorf)
Date: Wed, 9 Jul 2003 14:59:35 -0300
Subject: [albatross-users] Encrypting passwords in a session
Message-ID: <20030709175935.GA17731@demec.ufpe.br>

Hi, 

I just built a small application. I think albatross is great,
because once you get the hang of it, things flow very smoothly.

I need one last thing though, which I'm sure someone must have
needed before. I'm storing passwords in session variables,
and I'm using the file system to store passwords. This
seems to be the glaring security hole (not that there
isnt' any others :). Surely someone has solved a similar
problem before? I tried googling for something without
luck.

Most schemes I came up with involved storing an encryption
key as a session variable together with the encrypted password.
Duh! The next best thing was to generate a random key,
encrypt the password with it, store the encrypted password in
the server and the key (unique for each session) in the
user browser with a cookie (which I was trying hard to avoid,
since I want my application to work with the simplest text 
browsers.)

As cryptography algoriths are things best left to professionals,
I'm asking for suggestions here.

Many thanks for any input,
-- 
Ramiro Brito Willmersdorf            rbw@demec.ufpe.br  
GPG key: http://www.demec.ufpe.br/~rbw/GPG/gpg_key.txt

From esj at harvee.org  Thu Jul 10 01:03:34 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Wed, 09 Jul 2003 21:03:34 -0400
Subject: [albatross-users] Encrypting passwords in a session
In-Reply-To: <20030709175935.GA17731@demec.ufpe.br>
References: <20030709175935.GA17731@demec.ufpe.br>
Message-ID: <3F0CBB66.7060102@harvee.org>

Ramiro Brito Willmersdorf wrote:

> 
> Most schemes I came up with involved storing an encryption
...

> As cryptography algoriths are things best left to professionals,
> I'm asking for suggestions here.

good: http://www.unixpapa.com/auth/
better: http://cookies.lcs.mit.edu/

It seems to me that this kind of problems is something that is very common and 
destined to be reimplemented (poorly).  it might make sense to create a class 
called Simple_authenticated_application (derived from SimpleApp) which gives one 
the ability to be directed to a login template and do all the fun things like 
timeouts etc.  I'll think about the class overnight and post suggestion in the 
morning.

---eric



From esj at harvee.org  Thu Jul 10 13:00:13 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 10 Jul 2003 09:00:13 -0400
Subject: [albatross-users] Encrypting passwords in a session
In-Reply-To: <3F0CBB66.7060102@harvee.org>
References: <20030709175935.GA17731@demec.ufpe.br> <3F0CBB66.7060102@harvee.org>
Message-ID: <3F0D635D.6070807@harvee.org>

Eric S. Johansson wrote:
 > I'll think about
> the class overnight and post suggestion in the morning.

Simple_authenticated_application would need to additional arguments.  First is 
the login screen URL, second is the lifetime of an authentication token. 
Default values should be available for both.

It also seems that we wouldn't need much in the way of additional:

logout_user(): invalidates login token for this user and removes token from HTML 
output stream.
who_is_user(): yields username of user.  (Note: Login process should also set 
remote_user environment variable)
login_duration(): (would be nice) yields time since user login

all of the checking for a valid login, redirection to login screen etc. would 
all happen invisibly to the user as part of the new class which means the 
application code really doesn't need to change.

using the design for authentication tokens shown in section 4 of 
http://cookies.lcs.mit.edu/pubs/webauth:tr.pdf, we don't need to have any 
persistent server side session state if we're willing to accept some of the 
trade-offs (i.e. difficulty in terminating authentication authorization before 
expiration)

make sense?

---eric



From esj at harvee.org  Thu Jul 10 20:31:16 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Thu, 10 Jul 2003 16:31:16 -0400
Subject: [albatross-users] HTML editors
Message-ID: <3F0DCD14.5090906@harvee.org>

are there any good HTML editors that can be expanded to include albatross tags 
(and any sub tags).

I normally use something like Mozilla composer or amaya to create the rough HTML 
outline that I need.  Then I go and hack it to bloody bits in Emacs and I'm 
almost at the cusp of creating albatross specific macros in my speech 
recognition environment but I figured I should ask before working hard.

---eric



From matt at pollenation.net  Thu Jul 10 23:02:23 2003
From: matt at pollenation.net (Matt Goodall)
Date: 11 Jul 2003 00:02:23 +0100
Subject: [albatross-users] Problem with al-httpd and the httpd module
In-Reply-To: <20030710222658.GA1841@rantac.com>
References: <1057536370.26410.86.camel@localhost>
 <20030707055136.254283C332@coffee.object-craft.com.au>
 <3F093220.6090502@pollenation.net>  <20030710222658.GA1841@rantac.com>
Message-ID: <1057878142.8421.86.camel@localhost>

On Thu, 2003-07-10 at 23:26, Osvaldo Santana Neto wrote:
> On Mon, Jul 07, 2003 at 09:41:04AM +0100, Matt Goodall wrote:
> > Andrew McNamara wrote:
> > 
> > >>As soon as I start using the standalone http server for something useful
> > >>I come across a problem :-(. There is currently no way to serve static
> > >>resources, i.e. images, as the server assumes that *everything* will be
> > >>a request for a page.
> > >>   
> > >>
> > >
> > >The obvious solution is to add support for serving static content (files)
> > >from a specified directory.
> > >
> > Yes, that would work for me as I tend to store all images, stylesheets, 
> > etc under a single root directory called assets. I'll see what I can 
> > cope up with for the httpdapp module and al-httpd script.
> > 
> > Thanks, Matt
> 
> I've the same problem. How did you resolved this problem?

Sorry, but I haven't had the chance to do anything with it yet. 

Dave Cole has since added some code that could also help. See his post,
http://www.object-craft.com.au/pipermail/albatross-users/2003-July/000682.html. I was wondering about using this to provide a "page" to serve the static content. The get_page_from_url() is not in 1.10pre2 but I guess it will not be too long before another release.

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From osvaldo at rantac.com  Thu Jul 10 22:26:58 2003
From: osvaldo at rantac.com (Osvaldo Santana Neto)
Date: Thu, 10 Jul 2003 19:26:58 -0300
Subject: [albatross-users] Problem with al-httpd and the httpd module
In-Reply-To: <3F093220.6090502@pollenation.net>
References: <1057536370.26410.86.camel@localhost> <20030707055136.254283C332@coffee.object-craft.com.au> <3F093220.6090502@pollenation.net>
Message-ID: <20030710222658.GA1841@rantac.com>

On Mon, Jul 07, 2003 at 09:41:04AM +0100, Matt Goodall wrote:
> Andrew McNamara wrote:
> 
> >>As soon as I start using the standalone http server for something useful
> >>I come across a problem :-(. There is currently no way to serve static
> >>resources, i.e. images, as the server assumes that *everything* will be
> >>a request for a page.
> >>   
> >>
> >
> >The obvious solution is to add support for serving static content (files)
> >from a specified directory.
> >
> Yes, that would work for me as I tend to store all images, stylesheets, 
> etc under a single root directory called assets. I'll see what I can 
> cope up with for the httpdapp module and al-httpd script.
> 
> Thanks, Matt

I've the same problem. How did you resolved this problem?


-- 
Osvaldo Santana Neto
osvaldo@rantac.com - ICQ: 11287184

From sheila at thinkspot.net  Fri Jul 11 07:38:57 2003
From: sheila at thinkspot.net (Sheila King)
Date: Fri, 11 Jul 2003 00:38:57 -0700
Subject: [albatross-users] Writing Headers
Message-ID: <144878163.1057883937@SHEILA-LAPTOP>

OK, I'm sufficiently confused about the new method of handling 
headers to remark upon it.

In the application that I am working on (which I started with 
Albatross 1.01), I needed to send custom headers of my own. Since 
header handling has changed in 1.10, I need to fix my app.

But I'm going over the new documentation and getting a bit confused.

(P.S. you need to fix the copyright date in a lot of your files...as 
it is only copyright 2001.)

Anyhow, I'm trying to fix this up, and the documentation seems to 
have some flaws/omissions or something.

In section 7 (documentation for Mixins)
http://object-craft.com.au/projects/albatross/albatross/mixin-ref.html

there is no mention of the new ResponseMixin (this is a new Mixin 
class, right?)
Also, when I try to look up in the documentation index the 
"write_headers()" method, I go to "w" and
http://object-craft.com.au/projects/albatross/albatross/genindex.html
#letter-w

the "write_headers" method is listed there, but it is listed for all 
the AppContexts from which it was removed? But it is not listed for 
the new ResponseMixin class.

When I search the new .pdf document for "set_header" it only appears 
in the schematic diagrams for the pre-rolled classes (Chapter 8)

Same with "write_header" except for one mention in Chapter 5 for the 
Fake Application harness. Otherwise only mentioned in the schematic 
diagrams of Chapter 8.

I sure would like to see some examples of setting headers so I would 
have a better idea how I ought to use this. At least, what I mean is, 
the documentation would be more complete with such an example. I'm 
sure I will figure it out eventually by reviewing the source code and 
experimenting, but it would be nicer if I didn't have to.

Anyhow, I will put this off until tomorrow. Am too tired to mess with 
it any more right now.

Good night,

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From matt at pollenation.net  Fri Jul 11 09:03:16 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 11 Jul 2003 10:03:16 +0100
Subject: [albatross-users] Writing Headers
In-Reply-To: <144878163.1057883937@SHEILA-LAPTOP>
References: <144878163.1057883937@SHEILA-LAPTOP>
Message-ID: <3F0E7D54.9020006@pollenation.net>

You are right, this stuff is handled by the ResponseMixin which is 
included by the AppContext class so it should be there for all of the 
standard context classes. Since there is no documentation for this at 
the moment you are best off looking at the code for more detail, it is 
in app.py and is very simple.

As it happens I played around with this just the other day. I probably 
don't understand it all but I have written a "page" that sends an image 
back to the browser which will probably get you going.

    def page_display(ctx):
        f = None
        try:
            f = file('/home/matt/public_html/test/assets/images/%s' % 
ctx.locals.file)
            img_data = f.read()
        finally:
            if f:
                f.close()
        ctx.set_header('content-type', 'image/png')
        ctx.send_content(img_data)

The above code is only an experiment but it should demonstrate the 
basics. Obviously, if you are returning non-HTML then you will need to 
set the "content-type" to the correct MIME type, not hard code it like I 
did (have a look at the mimetypes module).

If you are returning HTML and just need to set your own headers then I 
suspect you only need to call ctx.set_header() and leave the rest to 
Albatross, i.e.

    def page_display(ctx):
        ctx.set_header('blah-blah', '123')
        ctx.run_template('something.html')

You don't seem to need to worry about calling write_headers(), it's 
presumably handled for you by the context and request adapter.

Hope this helps.

Cheers, Matt

Sheila King wrote:

> OK, I'm sufficiently confused about the new method of handling headers 
> to remark upon it.
>
> In the application that I am working on (which I started with 
> Albatross 1.01), I needed to send custom headers of my own. Since 
> header handling has changed in 1.10, I need to fix my app.
>
> But I'm going over the new documentation and getting a bit confused.
>
> (P.S. you need to fix the copyright date in a lot of your files...as 
> it is only copyright 2001.)
>
> Anyhow, I'm trying to fix this up, and the documentation seems to have 
> some flaws/omissions or something.
>
> In section 7 (documentation for Mixins)
> http://object-craft.com.au/projects/albatross/albatross/mixin-ref.html
>
> there is no mention of the new ResponseMixin (this is a new Mixin 
> class, right?)
> Also, when I try to look up in the documentation index the 
> "write_headers()" method, I go to "w" and
> http://object-craft.com.au/projects/albatross/albatross/genindex.html
> #letter-w
>
> the "write_headers" method is listed there, but it is listed for all 
> the AppContexts from which it was removed? But it is not listed for 
> the new ResponseMixin class.
>
> When I search the new .pdf document for "set_header" it only appears 
> in the schematic diagrams for the pre-rolled classes (Chapter 8)
>
> Same with "write_header" except for one mention in Chapter 5 for the 
> Fake Application harness. Otherwise only mentioned in the schematic 
> diagrams of Chapter 8.
>
> I sure would like to see some examples of setting headers so I would 
> have a better idea how I ought to use this. At least, what I mean is, 
> the documentation would be more complete with such an example. I'm 
> sure I will figure it out eventually by reviewing the source code and 
> experimenting, but it would be nicer if I didn't have to.
>
> Anyhow, I will put this off until tomorrow. Am too tired to mess with 
> it any more right now.
>
> Good night,
>

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com
t: +44 (0)113 2252500



From matt at pollenation.net  Fri Jul 11 14:07:26 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 11 Jul 2003 15:07:26 +0100
Subject: [albatross-users] [Fwd: Welcome to the "Pyweb" mailing list]
Message-ID: <3F0EC49E.2000208@pollenation.net>

Just in case you don't subscribe to the comp.lang.python.announce 
newsgroup ...

- Matt

-------- Original Message --------
Subject: 	Welcome to the "Pyweb" mailing list
Date: 	Fri, 11 Jul 2003 09:52:57 -0400
From: 	pyweb-request@amk.ca
To: 	matt@pollenation.net



Welcome to the Pyweb@amk.ca mailing list!

To post to this list, send your email to:

  pyweb@amk.ca

General information about the mailing list is at:

  http://www.amk.ca/mailman/listinfo/pyweb

If you ever want to unsubscribe or change your options (eg, switch to
or from digest mode, change your password, etc.), visit your
subscription page at:

  http://www.amk.ca/mailman/options/pyweb/matt%40pollenation.net

You can also make such adjustments via email by sending a message to:

  Pyweb-request@amk.ca

with the word `help' in the subject or body (don't include the
quotes), and you will get back a message with instructions.

You must know your password to change your options (including changing
the password, itself) or to unsubscribe.  It is:

  mu5hr00m

Normally, Mailman will remind you of your amk.ca mailing list
passwords once every month, although you can disable this if you
prefer.  This reminder will also include instructions on how to
unsubscribe or change your account options.  There is also a button on
your options page that will email your current password to you.


-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenationinternet.com
t: +44 (0)113 2252500



From matt at pollenation.net  Fri Jul 11 14:10:52 2003
From: matt at pollenation.net (Matt Goodall)
Date: Fri, 11 Jul 2003 15:10:52 +0100
Subject: [albatross-users] [Fwd: Welcome to the "Pyweb" mailing list]
In-Reply-To: <3F0EC49E.2000208@pollenation.net>
References: <3F0EC49E.2000208@pollenation.net>
Message-ID: <3F0EC56C.3060104@pollenation.net>

Jeez, I don't believe I just did that, I think I'm losing it! I'll just 
pop off to change my password and then I'll send the correct link.

- Matt


From matt at pollenationinternet.com  Fri Jul 11 14:13:40 2003
From: matt at pollenationinternet.com (Matt Goodall)
Date: Fri, 11 Jul 2003 15:13:40 +0100
Subject: [albatross-users] [Fwd: Mailing list for Web framework discussion]
Message-ID: <3F0EC614.90703@pollenationinternet.com>

Just in case you don't subscribe to the comp.lang.python.announce 
newsgroup ...

- Matt

-------- Original Message --------
Subject: Mailing list for Web framework discussion
Date: Fri, 11 Jul 2003 08:46:20 -0400
From: A.M. Kuchling <amk@amk.ca>
Organization: None
Newsgroups: comp.lang.python.announce

I've created a mailing list called pyweb for discussion of Python Web
frameworks.  The initial topic of the list will likely be a discussion
of what to use as an example web application, continuing the current
python-list thread.

It's a Mailman list, so just point your web browser at:
      http://www.amk.ca/mailman/listinfo/pyweb/
and follow the usual steps.

--amk




From djc at object-craft.com.au  Sat Jul 12 03:16:45 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 13:16:45 +1000
Subject: [albatross-users] Writing Headers
In-Reply-To: <3F0E7D54.9020006@pollenation.net>
References: <144878163.1057883937@SHEILA-LAPTOP>
 <3F0E7D54.9020006@pollenation.net>
Message-ID: <m3smpcl9ia.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> You are right, this stuff is handled by the ResponseMixin which
Matt> is included by the AppContext class so it should be there for
Matt> all of the standard context classes. Since there is no
Matt> documentation for this at the moment you are best off looking at
Matt> the code for more detail, it is in app.py and is very simple.

Ooops - I completely forgot about documenting ResponseMixin.  I will
do that today.

The ResponseMixin was introduced to simplify the process of sending
different types of content to the browser.  Matt is 100% on the money
with his examples.

Matt>     def page_display(ctx):
Matt>         f = None
Matt>         try:
Matt>             f = file('/home/matt/public_html/test/assets/images/%s' %
Matt>                      ctx.locals.file)
Matt>             img_data = f.read()
Matt>         finally:
Matt>             if f:
Matt>                 f.close()
Matt>         ctx.set_header('content-type', 'image/png')
Matt>         ctx.send_content(img_data)

Matt> The above code is only an experiment but it should demonstrate
Matt> the basics.

It is more than an experiment.  It is a perfect example of how to
manipulate the headers.

One subtle thing we did in relation to this was to make sure that the
session is saved before the output is flushed to the browser.  This
allows you to do things like include non-inline dynamically generated
images inside pages served to the browser.

Matt> If you are returning HTML and just need to set your own headers
Matt> then I suspect you only need to call ctx.set_header() and leave
Matt> the rest to Albatross, i.e.
Matt> 
Matt>     def page_display(ctx):
Matt>         ctx.set_header('blah-blah', '123')
Matt>         ctx.run_template('something.html')
Matt> 
Matt> You don't seem to need to worry about calling write_headers(), it's
Matt> presumably handled for you by the context and request adapter.

Yes.  Excellent summary.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 03:18:51 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 13:18:51 +1000
Subject: [albatross-users] Problem with al-httpd and the httpd module
In-Reply-To: <1057878142.8421.86.camel@localhost>
References: <1057536370.26410.86.camel@localhost>
 <20030707055136.254283C332@coffee.object-craft.com.au>
 <3F093220.6090502@pollenation.net> <20030710222658.GA1841@rantac.com>
 <1057878142.8421.86.camel@localhost>
Message-ID: <m3of00l9es.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Dave Cole has since added some code that could also help. See
Matt> his post,
Matt> http://www.object-craft.com.au/pipermail/albatross-users/2003-July/000682.html. I
Matt> was wondering about using this to provide a "page" to serve the
Matt> static content. The get_page_from_url() is not in 1.10pre2 but I
Matt> guess it will not be too long before another release.

I am going to try to get a pre3 release out this weekend.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 03:25:23 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 13:25:23 +1000
Subject: [albatross-users] HTML editors
In-Reply-To: <3F0DCD14.5090906@harvee.org>
References: <3F0DCD14.5090906@harvee.org>
Message-ID: <m3k7aol93w.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> are there any good HTML editors that can be expanded to include
Eric> albatross tags (and any sub tags).

I just installed the latest version of bluefish to see if it was
better than the releases I tried around a year ago.  They seem to have
made some real progress.

        http://bluefish.openoffice.nl/

Eric> I normally use something like Mozilla composer or amaya to
Eric> create the rough HTML outline that I need.  Then I go and hack
Eric> it to bloody bits in Emacs and I'm almost at the cusp of
Eric> creating albatross specific macros in my speech recognition
Eric> environment but I figured I should ask before working hard.

Have a look at bluefish.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 03:45:44 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 13:45:44 +1000
Subject: [albatross-users] Encrypting passwords in a session
In-Reply-To: <3F0D635D.6070807@harvee.org>
References: <20030709175935.GA17731@demec.ufpe.br>
 <3F0CBB66.7060102@harvee.org> <3F0D635D.6070807@harvee.org>
Message-ID: <m3fzlcl85z.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Eric S. Johansson wrote:
>> I'll think about the class overnight and post suggestion in the
>> morning.

That would be good.

Eric> Simple_authenticated_application would need to additional
Eric> arguments.  First is the login screen URL, second is the
Eric> lifetime of an authentication token. Default values should be
Eric> available for both.

Eric> It also seems that we wouldn't need much in the way of
Eric> additional:

Eric> logout_user(): invalidates login token for this user and removes
Eric> token from HTML output stream.  who_is_user(): yields username
Eric> of user.  (Note: Login process should also set remote_user
Eric> environment variable) login_duration(): (would be nice) yields
Eric> time since user login

Eric> all of the checking for a valid login, redirection to login
Eric> screen etc. would all happen invisibly to the user as part of
Eric> the new class which means the application code really doesn't
Eric> need to change.

I am looking forward to seeing the sample.  Adding something like this
to the documentation would be a very good thing.

Eric> using the design for authentication tokens shown in section 4 of
Eric> http://cookies.lcs.mit.edu/pubs/webauth:tr.pdf, we don't need to
Eric> have any persistent server side session state if we're willing
Eric> to accept some of the trade-offs (i.e. difficulty in terminating
Eric> authentication authorization before expiration)

Thanks for the URL.  I have some more reading to do...

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 03:58:13 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 13:58:13 +1000
Subject: base_url discovery (was Re: [albatross-users] Upgrade problems to albatross 1.10pre2)
In-Reply-To: <1057767195.28628.311.camel@localhost>
References: <312402881.1057700475@SHEILA-LAPTOP>
 <20030709051621.9388C3C024@coffee.object-craft.com.au>
 <315999543.1057704071@SHEILA-LAPTOP>
 <20030709060347.283353C024@coffee.object-craft.com.au>
 <m3ptkjrgux.fsf@ferret.object-craft.com.au>
 <1057767195.28628.311.camel@localhost>
Message-ID: <m3brw0l7l6.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> On Wed, 2003-07-09 at 14:04, Dave Cole wrote:
>> I have been wondering if it is possible for the application to
>> discover the base_url from the web server.  It certainly is using
>> CGI, so probably is with FastCGI too,

Matt> As you pointed out, neither of these should be a problem.

>> but I am not so sure about mod_python.

Matt> The problem with this is the one I hit the other day when using
Matt> mod_rewrite. The uri that Albatross sees is the rewritten uri
Matt> which can be completely different to what is typed into the
Matt> browser url bar.

That probably means that it will be very hard to remove base_url.
Into the too hard basket.

Matt> Having said that, mod_python has req.the_request which is the
Matt> http request header, i.e. 'GET /~matt/simple/app.py/1/2/3
Matt> HTTP/1.1'. We may be able to make use of that but I haven't
Matt> looked into it anymore.

>> I am certain the BaseHTTPServer cannot supply enough information.

Matt> No, but it probably doesn't need to as the application is always
Matt> mounted at '/'.

True.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sat Jul 12 04:01:46 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Sat, 12 Jul 2003 00:01:46 -0400
Subject: [albatross-users] HTML editors
In-Reply-To: <m3k7aol93w.fsf@ferret.object-craft.com.au>
References: <3F0DCD14.5090906@harvee.org> <m3k7aol93w.fsf@ferret.object-craft.com.au>
Message-ID: <3F0F882A.9010308@harvee.org>

Dave Cole wrote:

>>>>>>"Eric" == Eric S Johansson <esj@harvee.org> writes:
> 
> 
> Eric> are there any good HTML editors that can be expanded to include
> Eric> albatross tags (and any sub tags).
> 
> I just installed the latest version of bluefish to see if it was
> better than the releases I tried around a year ago.  They seem to have
> made some real progress.
> 
>         http://bluefish.openoffice.nl/
> 

I know about bluefish and have tried the 0.9 version that came with Red Hat 8 
and I'm quite aware of its improvements.  they still don't do syntax 
highlighting correctly (just use don't in the body of your text).  I also 
haven't figured out how to add albatross tags.  I will dig further.

---eric



From djc at object-craft.com.au  Sat Jul 12 04:08:04 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 14:08:04 +1000
Subject: [albatross-users] Change to httpdapp module
In-Reply-To: <3F09A054.4090004@pollenation.net>
References: <3F09A054.4090004@pollenation.net>
Message-ID: <m34r1sl74r.fsf@ferret.object-craft.com.au>

Matt> In httpdapp, Request.get_uri() currently returns an entire URL
Matt> including server, port etc. I think it should be just returning
Matt> the path as follows, but please check I'm not being daft first:
Matt> 
Matt>     def get_uri(self):
Matt>         return self._req.path

I think the above is the correct thing to return.  I will change the
module.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 04:42:59 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 14:42:59 +1000
Subject: [albatross-users] [patch] get_uri -> get_path_info changes
In-Reply-To: <1057743848.28628.75.camel@localhost>
References: <3F0AF957.6050402@pollenation.net>
 <1057743848.28628.75.camel@localhost>
Message-ID: <m3wueojqy4.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> In light of the problem Sheila found, below are some comments on
Matt> the patches.

Matt> On Tue, 2003-07-08 at 18:03, Matt Goodall wrote:

>> 01.base_url.patch - a patch to fix my previous patch. Already
>> applied by Dave I think.

Matt> This was clearly the wrong solution as the original base_url
Matt> patch did not consider that the base_url might not be absolute.

Now back to this in Application.__init__():

        # Use a base_url of '/' if not provided
        if base_url:
            self.__base_url = base_url
        else:
            self.__base_url = '/'

>> 03.httpdapp.patch - simplified version of Request.get_uri().

Matt> Not necessary but still an improvement. It's kind of irrelevant
Matt> though as it doesn't work with the old base_url stuff.

Applied (already said this in other email).

>> 04.get_path_info.patch - introduces get_path_info() methods into
>> the request adapters.

Matt> Not much point with 05 but it may be useful anyway. Dave/Andrew,
Matt> it's up to you.

Applied.

>> 05.use_path_info.patch - switch from using Request.get_uri() to
>> Request.get_path_info(). This affects how the random page loader
>> works and how the current and redirect urls are calculated.

Matt> Completely breaks redirect_url when the base_url is not
Matt> absolute. Do not apply.

Not applied.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 05:04:31 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 15:04:31 +1000
Subject: [albatross-users] [patch] Another attempt at "fixing" base_url
In-Reply-To: <1057760211.28628.269.camel@localhost>
References: <1057760211.28628.269.camel@localhost>
Message-ID: <m3r84wjpy8.fsf@ferret.object-craft.com.au>

> Attached is another patch, against stock 1.10pre2, which attempts to
> make the base_url + httpdapp combination work without breaking
> everything else ;-).

The pain that goes into a few lines of code.  I am sorry that you have
suffered so much.  It really highlights a large shortcoming in the
current code - comments.

> This time, with the knowledge that a relative base_url is perfectly
> valid, I have taken the approach of "fixing up" the base_url where
> it is used to determine the page and build redirect urls. I hope
> this approach is a little more successful than the previous one!

I have been looking at the patch and think it can be simplified.  See
comments below.

> diff -ru albatross-1.10pre2/albatross/app.py albatross-1.10pre2-matt/albatross/app.py
> --- albatross-1.10pre2/albatross/app.py	2003-07-05 08:52:40.000000000 +0100
> +++ albatross-1.10pre2-matt/albatross/app.py	2003-07-09 15:11:13.000000000 +0100
> @@ -198,6 +198,9 @@
>      def redirect_url(self, loc):
>          this_path = urlparse.urlparse(self.request.get_uri())[2]
>          base_path = urlparse.urlparse(self.app.base_url())[2]
> +        # If no base_path then the application is mounted at the root
> +        if not base_path:
> +            base_path = '/'
>          pos = this_path.find(base_path)
>          new_base = this_path[:pos + len(base_path)]
>          if new_base[-1] != '/':
> @@ -221,11 +224,7 @@
>      def __init__(self, base_url):
>          ResourceMixin.__init__(self)
>          self.register_tagclasses(*tags.tags)
> -        # Use a base_url of '/' if not provided
> -        if base_url:
> -            self.__base_url = base_url
> -        else:
> -            self.__base_url = '/'
> +        self.__base_url = base_url
>  
>      def run(self, req):
>          '''Process a single browser request

If you leave the base_url code untouched in the Application
constructor then there is no need for the change in redirect_url().
This means that we can leave this file unpatched.

The effect of this non-change also changes the patch to randompage.py

> diff -ru albatross-1.10pre2/albatross/randompage.py albatross-1.10pre2-matt/albatross/randompage.py
> --- albatross-1.10pre2/albatross/randompage.py	2003-07-05 08:52:40.000000000 +0100
> +++ albatross-1.10pre2-matt/albatross/randompage.py	2003-07-09 13:14:32.000000000 +0100
> @@ -23,6 +23,12 @@
>          try:
>              base_path = urlparse.urlparse(self.base_url())[2]
>              uri_path = urlparse.urlparse(uri)[2]
> +            # Fixup base_path to end with a '/' so that it can be used as
> +            # a root path.
> +            if not base_path:
> +                base_path = '/'
> +            elif base_path[-1] != '/':
> +                base_path =  base_path + '/'
>              page = uri_path.split(base_path, 1)[1]
>          except IndexError:
>              pass

The first check is no longer necessary if we do not change base_url
logic in the Application constructor.  This reduces the patch to this:

         try:
             base_path = urlparse.urlparse(self.base_url())[2]
             uri_path = urlparse.urlparse(uri)[2]
+            # Fixup base_path to end with a '/' so that it can be used as
+            # a root path.
+            if not base_path.endswith('/'):
+                base_path =  base_path + '/'
             page = uri_path.split(base_path, 1)[1]
         except IndexError:
             pass

Can you see any reason why this is not equivalent to your patch?  Am I
missing something else?

- Dave

-- 
http://www.object-craft.com.au


From sheila at thinkspot.net  Sat Jul 12 05:43:09 2003
From: sheila at thinkspot.net (Sheila King)
Date: Fri, 11 Jul 2003 22:43:09 -0700
Subject: [albatross-users] Session Variables not saved
Message-ID: <40971904.1057963389@SHEILA-LAPTOP>

I am still having serious difficulties with my application since the 
upgrade to Albatross 1.10pre2.

I am using RandomApp with Server side file sessions.

I got past the problems with the base_url (thanks to tips from Dave) 
and I got past the header writing issues (I think). Thanks Matt.

But now it appears that my session/authentication code is all fubar.

The thing is, I finally figured out tonight, that even though I am 
(in my code) adding session vars to the session, they do not appear 
to be saved.

I am working on making a stripped down example of my code to show 
this (as I cannot share the app itself) but wanted to toss this 
feeler out there, just in case it "rang a bell" for anyone.

Once I get a stripped down example, I will post it to the list. I 
don't know if I will get it done within the next 1.5 hours. If not, I 
will be off to bed, and it will have to wait until late tomorrow.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From djc at object-craft.com.au  Sat Jul 12 07:06:50 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 17:06:50 +1000
Subject: [albatross-users] Documented ResponseMixin class
Message-ID: <m3k7aojkad.fsf@ferret.object-craft.com.au>

For the curious I have documented the ResponseMixin class.

    http://www.object-craft.com.au/~djc/albatross/mixin-response.html

This will be included in the pre3 release that I hope to get out this
weekend.  The only outstanding task is to confirm my changes to Matt's
latest base_url patch.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 07:11:32 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 17:11:32 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: <1057450672.19240.37.camel@localhost>
References: <200307040004.KAA08163@lightning.itga.com.au>
 <3F04C9C2.1020308@harvee.org> <1057450672.19240.37.camel@localhost>
Message-ID: <m3el0wjk2j.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> On Fri, 2003-07-04 at 01:26, Eric S. Johansson wrote:
>> Q: how do I get a complete list of all checkboxes checked and
>> unchecked just like you can with ordinary CGI

Matt> Are you sure CGI does this? As far as I was aware browsers only
Matt> send a request parameter for a checkbox when it is checked. If
Matt> the browser doesn't send a request parameter how can CGI even
Matt> know there were unchecked checkboxes on the page? Perhaps the
Matt> browser behaves differently when the form data is posted?

The browser never sends checkbox fields that are not checked.  That is
one of the reasons for the NameRecorderMixin.  It records all of the
input fields in a form (in the __albstate__ hidden field) so that the
request merging can do things like clear the value associated with an
unchecked checkbox.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sat Jul 12 07:12:18 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 12 Jul 2003 17:12:18 +1000
Subject: [albatross-users] check boxes again.
In-Reply-To: <m3el0wjk2j.fsf@ferret.object-craft.com.au>
References: <200307040004.KAA08163@lightning.itga.com.au>
 <3F04C9C2.1020308@harvee.org> <1057450672.19240.37.camel@localhost>
 <m3el0wjk2j.fsf@ferret.object-craft.com.au>
Message-ID: <m3adbkjk19.fsf@ferret.object-craft.com.au>

>>>>> "Dave" == Dave Cole <djc@object-craft.com.au> writes:

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:
Matt> On Fri, 2003-07-04 at 01:26, Eric S. Johansson wrote:
>>> Q: how do I get a complete list of all checkboxes checked and
>>> unchecked just like you can with ordinary CGI

Matt> Are you sure CGI does this? As far as I was aware browsers only
Matt> send a request parameter for a checkbox when it is checked. If
Matt> the browser doesn't send a request parameter how can CGI even
Matt> know there were unchecked checkboxes on the page? Perhaps the
Matt> browser behaves differently when the form data is posted?

Dave> The browser never sends checkbox fields that are not checked.
Dave> That is one of the reasons for the NameRecorderMixin.  It
Dave> records all of the input fields in a form (in the __albstate__
                                                        ^^^^^^^^^^^^
Ooops, make that the __albform__ field.

- Dave

-- 
http://www.object-craft.com.au


From sheila at thinkspot.net  Sat Jul 12 08:21:52 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sat, 12 Jul 2003 01:21:52 -0700
Subject: [albatross-users] Session Variables not saved
In-Reply-To: <40971904.1057963389@SHEILA-LAPTOP>
References: <40971904.1057963389@SHEILA-LAPTOP>
Message-ID: <50495048.1057972912@SHEILA-LAPTOP>

--On Friday, July 11, 2003 10:43 PM -0700 Sheila King 
<sheila@thinkspot.net> wrote:

> I am still having serious difficulties with my application since
> the  upgrade to Albatross 1.10pre2.
>
> I am using RandomApp with Server side file sessions.
[snip]
>
> But now it appears that my session/authentication code is all fubar.
>
> The thing is, I finally figured out tonight, that even though I am
> (in my code) adding session vars to the session, they do not appear
> to be saved.
>
> I am working on making a stripped down example of my code to show
> this ...

OK, here is the stripped down example. It's ugly, but I think 
demonstrates my point...

Code is available here:
http://www.mathxy.com/temp/

The .zip file contains all the files without the .txt extensions (if 
you want the whole thing to run for yourself)

Working example is available here:
http://mathxy.com/cgi-bin/test/mgr.py

account number = 123456
email = test@example.com
passwd = test1234

All of the local variables and session vars are displayed to the page.

What is so weird is that the session vars do contain
_last_access
_authstatus

as session variables when the login page is first displayed

But then when you log in and go to the expired page (which is an 
error...it should take you to the Main Account page, but this doesn't 
work right...is my problem I'm trying to solve...)

Notice that the _last_access and _authstatus are no longer session 
variables.

And what is weird is that there is no place in the code where I am 
removing the session or deleting any session variables. So how did 
those variables get removed from the session?

I believe that the reason my authentication code no longer works, is 
because the _last_access variable is no longer stored in the session.

If I'm doing something stupid here, OK...but I don't think so. This 
code (or something darn similar to it) worked fine with the last 
version of Albatross. I am beating my head against a wall here to 
figure out why.

One thing that I thought was possible, is that it is happening in the 
"redirect" as I know that the account gets validated long enough to 
go into the main.py file but when it rechecks for validation there, 
it fails and redirects to the expired page.

Basic process is:

mgr.py <-- main application
login.py <-- start page module
line 15 of login.py is apparently successfully authenticating the 
account
line 20 in login.py is redirecting to main.py
line 5 of main.py is determining that the authentication is not valid
line 6 of main.py is redirecting to expired.py

That's what's happening now. Authentication check should *not* be 
failing in main.py line 5.

:\


-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org





From matt at pollenation.net  Sat Jul 12 10:24:29 2003
From: matt at pollenation.net (Matt Goodall)
Date: 12 Jul 2003 11:24:29 +0100
Subject: [albatross-users] [patch] Another attempt at "fixing" base_url
In-Reply-To: <m3r84wjpy8.fsf@ferret.object-craft.com.au>
References: <1057760211.28628.269.camel@localhost>
 <m3r84wjpy8.fsf@ferret.object-craft.com.au>
Message-ID: <1058005469.22647.62.camel@localhost>

On Sat, 2003-07-12 at 06:04, Dave Cole wrote:
> > Attached is another patch, against stock 1.10pre2, which attempts to
> > make the base_url + httpdapp combination work without breaking
> > everything else ;-).

<snip> Dave's simplified versions of my patches.

> Can you see any reason why this is not equivalent to your patch?  Am I
> missing something else?

Seems ok to me although the whole patches to patches thing is a little
hard to follow ;-).

I'm finding it difficult to picture exactly what's in your source code
right now. I have applied the changes to 1.10pre2 that I think you have
applied here and in your recent messages to the list and it seems to be
working ok.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sat Jul 12 10:59:52 2003
From: matt at pollenation.net (Matt Goodall)
Date: 12 Jul 2003 11:59:52 +0100
Subject: [albatross-users] [patch] Another attempt at "fixing" base_url
In-Reply-To: <m3r84wjpy8.fsf@ferret.object-craft.com.au>
References: <1057760211.28628.269.camel@localhost>
 <m3r84wjpy8.fsf@ferret.object-craft.com.au>
Message-ID: <1058007592.5901.115.camel@localhost>

On Sat, 2003-07-12 at 06:04, Dave Cole wrote:
> > Attached is another patch, against stock 1.10pre2, which attempts to
> > make the base_url + httpdapp combination work without breaking
> > everything else ;-).
> 
> The pain that goes into a few lines of code.  I am sorry that you have
> suffered so much. 

Hey, no problem. It was my choice to contribute to the project. I
thought Albatross looked really useful but saw a couple of "gaps" that I
needed to fill, mostly for my own use. In this case, I didn't consider
one of the possibilities for base_url so didn't test things fully.

Besides, we're not in the open source thing for an easy ride, it's about
community, creating great software (generally better than the commercial
offerings) and seeing others make use and benefit from the code. The
positive comments about my small contributions to the project are more
than enough payback.

/matt steps down from his soapbox ;-)

> It really highlights a large shortcoming in the
> current code - comments.

Yes, a few more comments in the important places would be useful
(redirect_url is probably a good example) but unit tests are probably
more important.

Obviously, there's the interaction with the web server to take into
account too which is often not easy to test. Having said that, I now
have a suite of simple test applications using some (six, so far) of the
standard application classes, each running as cgi, fcgi etc. I'm running
through these apps manually at the moment (very tedious!) but if I can
script up that you are welcome to suck the suite into Albatross as a
final set of tests.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From matt at pollenation.net  Sat Jul 12 21:26:01 2003
From: matt at pollenation.net (Matt Goodall)
Date: 12 Jul 2003 22:26:01 +0100
Subject: [albatross-users] Session Variables not saved
In-Reply-To: <50495048.1057972912@SHEILA-LAPTOP>
References: <40971904.1057963389@SHEILA-LAPTOP>
 <50495048.1057972912@SHEILA-LAPTOP>
Message-ID: <1058045161.15822.129.camel@localhost>

On Sat, 2003-07-12 at 09:21, Sheila King wrote:
> --On Friday, July 11, 2003 10:43 PM -0700 Sheila King 
> <sheila@thinkspot.net> wrote:
> 
> > I am still having serious difficulties with my application since
> > the  upgrade to Albatross 1.10pre2.

I just spent a few minutes stepping through the code debugging this, I'm
getting the same here too.

The problem is that on redirect only the 'Location' and 'Status' headers
are being sent to the browser. The critical one that is missing is
'Set-Cookie', i.e. the browser is never told what its session id is!

This appears to affect all the request adapters. I'm also fairly sure
that I didn't break it this time ;-).

Hope this helps.

Cheers, Matt

p.s. if anyone needs a recommendation for a Python debugger then wingide
works really well.

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From sheila at thinkspot.net  Sat Jul 12 22:22:09 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sat, 12 Jul 2003 15:22:09 -0700
Subject: [albatross-users] Session Variables not saved
In-Reply-To: <1058045161.15822.129.camel@localhost>
References: <40971904.1057963389@SHEILA-LAPTOP>
 <50495048.1057972912@SHEILA-LAPTOP> <1058045161.15822.129.camel@localhost>
Message-ID: <344633526.1058023329@goddess>

--On Saturday, July 12, 2003 10:26 PM +0100 Matt Goodall 
<matt@pollenation.net> wrote:

> On Sat, 2003-07-12 at 09:21, Sheila King wrote:
>> --On Friday, July 11, 2003 10:43 PM -0700 Sheila King
>> <sheila@thinkspot.net> wrote:
>>
>> > I am still having serious difficulties with my application since
>> > the  upgrade to Albatross 1.10pre2.

> The problem is that on redirect only the 'Location' and 'Status' headers
> are being sent to the browser. The critical one that is missing is
> 'Set-Cookie', i.e. the browser is never told what its session id is!

Thank you SO much for that sanity check.

Although I did look at the redirect code for the cgiapp Request class, I 
didn't see what the problem was. Grrr. I *should* have though, as I had 
worked with that code closely before. Guess I was too tired and was just 
trying to figure out exactly where the error was occurring. Have already 
spent more than a full day on this issue alone. :((

Thank goodness for your post. I was about to crank up and make an even more 
stripped down version.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From djc at object-craft.com.au  Sun Jul 13 02:32:52 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 12:32:52 +1000
Subject: [albatross-users] [patch] Another attempt at "fixing" base_url
In-Reply-To: <1058007592.5901.115.camel@localhost>
References: <1057760211.28628.269.camel@localhost>
 <m3r84wjpy8.fsf@ferret.object-craft.com.au>
 <1058007592.5901.115.camel@localhost>
Message-ID: <m3brvz2m23.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Hey, no problem. It was my choice to contribute to the
Matt> project. I thought Albatross looked really useful but saw a
Matt> couple of "gaps" that I needed to fill, mostly for my own
Matt> use. In this case, I didn't consider one of the possibilities
Matt> for base_url so didn't test things fully.

Testing is a hard job.

Matt> Besides, we're not in the open source thing for an easy ride,
Matt> it's about community, creating great software (generally better
Matt> than the commercial offerings) and seeing others make use and
Matt> benefit from the code. The positive comments about my small
Matt> contributions to the project are more than enough payback.

It is a big help.  You see things differently to us in some ways.
This tests our assumptions and makes the code better.

Matt> Yes, a few more comments in the important places would be useful
Matt> (redirect_url is probably a good example) but unit tests are
Matt> probably more important.

In a way we are always playing catchup with documentation.  Comments
are something that I should start to work on too.

Matt> Obviously, there's the interaction with the web server to take
Matt> into account too which is often not easy to test. Having said
Matt> that, I now have a suite of simple test applications using some
Matt> (six, so far) of the standard application classes, each running
Matt> as cgi, fcgi etc. I'm running through these apps manually at the
Matt> moment (very tedious!) but if I can script up that you are
Matt> welcome to suck the suite into Albatross as a final set of
Matt> tests.

That would be excellent.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sun Jul 13 02:35:19 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 12:35:19 +1000
Subject: [albatross-users] Session Variables not saved
In-Reply-To: <344633526.1058023329@goddess>
References: <40971904.1057963389@SHEILA-LAPTOP>
 <50495048.1057972912@SHEILA-LAPTOP>
 <1058045161.15822.129.camel@localhost> <344633526.1058023329@goddess>
Message-ID: <m37k6n2ly0.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

>> The problem is that on redirect only the 'Location' and 'Status'
>> headers are being sent to the browser. The critical one that is
>> missing is 'Set-Cookie', i.e. the browser is never told what its
>> session id is!

Sheila> Thank you SO much for that sanity check.

I looked at the code and could not see anything you were doing wrong -
I assumed I was just missing something.  Turned out it was Albatross
doing something wrong - thanks again Matt.

I am just in the process of packaging up a pre3 release.  This is
hopefully the last one.

Sheila> Although I did look at the redirect code for the cgiapp
Sheila> Request class, I didn't see what the problem was. Grrr. I
Sheila> *should* have though, as I had worked with that code closely
Sheila> before. Guess I was too tired and was just trying to figure
Sheila> out exactly where the error was occurring. Have already spent
Sheila> more than a full day on this issue alone. :((

Ouch.  Sorry about that.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sun Jul 13 02:55:40 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 12:55:40 +1000
Subject: [albatross-users] Albatross 1.10 pre3 released
In-Reply-To: <m3znjttehl.fsf@ferret.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
Message-ID: <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>

I have just packaged up the 1.10pre3 release of Albatross.  This
resolves the problems encountered in the 1.10pre2 release.

The changes since 1.10pre2 are:

* Request classes now have a get_path_info() method.

* Redirect now sends the session id cookie when it is defined.  The
  logic has been moved into ResponseMixin.redirect().

* SessionBase now has a session_vars() method that returns a list of
  the names in the session.

* Random page URI handling is back to 1.10pre1 state.  The base_path
  used to split the requested URI will always end with '/'.

* Random page URI cracking has been moved into the get_page_from_uri()
  method.  This allows application to redefine the cracking method.

* More documentation updates - especially in the prepackaged classes
  chapter.

* Most samples have been changed to allow them to be run using the
  al-httpd program.

There still is no documentation of the FastCGI and Standalone
BaseHTTPServer support.

The pre2 was supposed to be the last pre release of 1.10.  This time
for sure...  If there are no issues reported with this one we will
release the real 1.10 by the end of the week.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sun Jul 13 04:48:58 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Sun, 13 Jul 2003 00:48:58 -0400
Subject: [albatross-users] crafting get arg urls
Message-ID: <3F10E4BA.90802@harvee.org>

I'm hitting another cognitive brick wall.  It's probably because I'm up too late 
but here is what's messing with my head.

Need to create a get style URL for the spamtrap entry display.  was using a 
single argument and manually generated the anchor record just fine.  Tried 
adding an argument and things got confused with the & and all.

here's what I'm generating for albatross:
<al-a 
href="'spamtrap_display.cgi?repository=dumpster&token='+line.value().token" 
target="spamtrap_display">

here is what I'm getting:

<a href="'spamtrap_display.cgi?repository=dumpster&token='+line.value().token" 
target="spamtrap_display">

what I wanted:

<a href="spamtrap_display.cgi?repository=dumpster&token=1234567890123456" 
target="spamtrap_display">

now I'm probably missing an exec somewhere and if I need to grope around, I'll 
figure it out sooner or later.

The real problem is getting clarity on quoting mixed in with variable execution. 
  I realize as I'm working on these problems that I don't know or can't find 
easily a simple statement on how to retrieve values and substitute them into 
strings simply and easily as the case above demonstrates.

it's a real simple thing and common to virtually every aspect of albatross.  The 
answer to this problem will probably occurred to me in the morning but for now, 
it's irritating the hell out of me.

---eric



From sheila at thinkspot.net  Sun Jul 13 05:57:48 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sat, 12 Jul 2003 22:57:48 -0700
Subject: [albatross-users] Redirect problems continue...
Message-ID: <9733165.1058050668@SHEILA-LAPTOP>

OK, thanks for the really quick release of patches for the issues with 
redirect and not saving the cookie etc...

I've installed the pre3 release and tried again with the stripped down 
application code I shared last night. As a reminder, code is available here:
http://mathxy.com/temp/

Unfortunately, the code, when I attempt to run it, now generates a 500 
Internal Server error. Here is the error entry from the logs:

%% [Sun Jul 13 01:37:59 2003] GET /cgi-bin/test/mgr.py HTTP/1.1
%% 500 /big/dom/xexample/cgi-bin/test/mgr.py
%request
Accept: 
application/x-shockwave-flash,text/xml,application/xml,application/xhtml+xm
l,text/html;q=0.9,text/pl 
ain;q=0.8,video/x-mng,image/png,image/jpeg,image/gif;q=0.2,*/*;q=0.1
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Accept-Encoding: gzip,deflate,compress;q=0.9
Accept-Language: en-us,en;q=0.5
Cache-Control: max-age=0
Connection: keep-alive
Host: example.com
Keep-Alive: 300
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3) 
Gecko/20030312
%response
%stderr
Traceback (most recent call last):
File "mgr.py", line 60, in ?
app.run(qaRequest())
File "/big/dom/xexample/lib/python2.2/site-packages/albatross/app.py", line 
253, in run
return ctx.redirect(e.loc)
File "/big/dom/xexample/lib/python2.2/site-packages/albatross/app.py", line 
217, in redirect
raise Redirect(loc)
albatross.app.Redirect: <albatross.app.Redirect instance at 0x82716e4>


I'm looking at the app.py file, and I don't see what the problem is. The 
code from app.py involved in raising the redirect exception appears to me 
to be unchanged between 1.01 and 1.10pre3.

???


-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From djc at object-craft.com.au  Sun Jul 13 06:08:23 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 16:08:23 +1000
Subject: [albatross-users] Redirect problems continue...
In-Reply-To: <9733165.1058050668@SHEILA-LAPTOP>
References: <9733165.1058050668@SHEILA-LAPTOP>
Message-ID: <m3he5rhsbs.fsf@ferret.object-craft.com.au>

> OK, thanks for the really quick release of patches for the issues
> with redirect and not saving the cookie etc...
> 
> I've installed the pre3 release and tried again with the stripped
> down application code I shared last night. As a reminder, code is
> available here: http://mathxy.com/temp/
> 
> Unfortunately, the code, when I attempt to run it, now generates a
> 500 Internal Server error. Here is the error entry from the logs:
> 
> Traceback (most recent call last):
> File "mgr.py", line 60, in ?
> app.run(qaRequest())
> File "/big/dom/xexample/lib/python2.2/site-packages/albatross/app.py",
> line 253, in run
> return ctx.redirect(e.loc)
> File "/big/dom/xexample/lib/python2.2/site-packages/albatross/app.py",
> line 217, in redirect
> raise Redirect(loc)
> albatross.app.Redirect: <albatross.app.Redirect instance at 0x82716e4>

> I'm looking at the app.py file, and I don't see what the problem
> is. The code from app.py involved in raising the redirect exception
> appears to me to be unchanged between 1.01 and 1.10pre3.

I can't believe I did that...

Hold on a second.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sun Jul 13 06:52:45 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 16:52:45 +1000
Subject: [albatross-users] Albatross 1.10 pre4 released
In-Reply-To: <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
Message-ID: <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>

So much for 1.10pre3.  For my next trick I have packaged up the
1.10pre4 release of Albatross.  This resolves the problems encountered
in the 1.10pre2 release.

The changes since 1.10pre3 are:

* ResponseMixin redirect() renamed to send_redirect().


The changes since 1.10pre2 are:

* Request classes now have a get_path_info() method.

* Redirect now sends the session id cookie when it is defined.  The
  logic has been moved into ResponseMixin.redirect().

* SessionBase now has a session_vars() method that returns a list of
  the names in the session.

* Random page URI handling is back to 1.10pre1 state.  The base_path
  used to split the requested URI will always end with '/'.

* Random page URI cracking has been moved into the get_page_from_uri()
  method.  This allows application to redefine the cracking method.

* More documentation updates - especially in the prepackaged classes
  chapter.

* Most samples have been changed to allow them to be run using the
  al-httpd program.

There still is no documentation of the FastCGI and Standalone
BaseHTTPServer support.

The pre2 was supposed to be the last pre release of 1.10.  This time
for sure...  If there are no issues reported with this one we will
release the real 1.10 by the end of the week.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Sun Jul 13 07:18:55 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 13 Jul 2003 17:18:55 +1000
Subject: [albatross-users] crafting get arg urls
In-Reply-To: <3F10E4BA.90802@harvee.org>
References: <3F10E4BA.90802@harvee.org>
Message-ID: <m3llv2lwrk.fsf@ferret.object-craft.com.au>

> I'm hitting another cognitive brick wall.  It's probably because I'm
> up too late but here is what's messing with my head.
> 
> Need to create a get style URL for the spamtrap entry display.  was
> using a single argument and manually generated the anchor record
> just fine.  Tried adding an argument and things got confused with
> the & and all.
> 
> here's what I'm generating for albatross:
> <al-a
> href="'spamtrap_display.cgi?repository=dumpster&token='+line.value().token"
> target="spamtrap_display">

Try:
  <al-a
  expr="'repository=dumpster&token='+line.value().token"
  target="spamtrap_display">

The expr attribute is used to evaluate an href attribute.  The href
attribute is then mostly untouched.  It gets the current_url()
prefixed when '?' cannot be found in the value you specify.

If you specify the href directly then the prefix is still added if
necessary.

> The real problem is getting clarity on quoting mixed in with
> variable execution. I realize as I'm working on these problems that
> I don't know or can't find easily a simple statement on how to
> retrieve values and substitute them into strings simply and easily
> as the case above demonstrates.
> 
> it's a real simple thing and common to virtually every aspect of
> albatross.  The answer to this problem will probably occurred to me in
> the morning but for now, it's irritating the hell out of me.

More explanation is probably necessary in the documentation,

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Sun Jul 13 14:16:02 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Sun, 13 Jul 2003 10:16:02 -0400
Subject: [albatross-users] crafting get arg urls
In-Reply-To: <m3llv2lwrk.fsf@ferret.object-craft.com.au>
References: <3F10E4BA.90802@harvee.org> <m3llv2lwrk.fsf@ferret.object-craft.com.au>
Message-ID: <3F1169A2.6000501@harvee.org>

Dave Cole wrote:
> Try:
>   <al-a
>   expr="'repository=dumpster&token='+line.value().token"
>   target="spamtrap_display">
> 

thanks.  That put me in the right area.

> 
> More explanation is probably necessary in the documentation,

yes, I would suggest adding a complete section on this topic (data extraction 
from the context storage and merging bits of strings and data in expr statements

---eric


From sheila at thinkspot.net  Sun Jul 13 16:17:36 2003
From: sheila at thinkspot.net (Sheila King)
Date: Sun, 13 Jul 2003 09:17:36 -0700
Subject: [albatross-users] pre4 - Problems still ...
In-Reply-To: <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
Message-ID: <1900622.1058087856@SHEILA-LAPTOP>

--On Sunday, July 13, 2003 4:52 PM +1000 Dave Cole 
<djc@object-craft.com.au> wrote:

> So much for 1.10pre3.  For my next trick I have packaged up the
> 1.10pre4 release of Albatross.  This resolves the problems encountered
> in the 1.10pre2 release.

This does prevent me from getting 500 Internal Server error now on my test 
app.

Current Problem:
Session Variables are still being lost, as before. If you go to the initial 
log in page, it shows four session variables, including _last_access and 
_authstatus but when you go to the next page (after entering log in 
information and attempting to authenticate) you go to the Expired page 
(which is an error) and the two session variables mentioned above are no 
longer in the session. The code contains NO statements that remove the 
session or delete session variables.


Live demo here:
http://mathxy.com/cgi-bin/test/mgr.py

account num = 123456
email = test@example.com
passwd = test1234
(only the login page should/would work. create login and lost password 
disabled.)

Code (still) available here:
http://mathxy.com/temp/


-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From djc at object-craft.com.au  Sun Jul 13 22:29:44 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 14 Jul 2003 08:29:44 +1000
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <1900622.1058087856@SHEILA-LAPTOP>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
Message-ID: <m33cha6ox3.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

Sheila> --On Sunday, July 13, 2003 4:52 PM +1000 Dave Cole
Sheila> <djc@object-craft.com.au> wrote:

>> So much for 1.10pre3.  For my next trick I have packaged up the
>> 1.10pre4 release of Albatross.  This resolves the problems
>> encountered in the 1.10pre2 release.

Sheila> This does prevent me from getting 500 Internal Server error
Sheila> now on my test app.

Sheila> Current Problem: Session Variables are still being lost, as
Sheila> before. If you go to the initial log in page, it shows four
Sheila> session variables, including _last_access and _authstatus but
Sheila> when you go to the next page (after entering log in
Sheila> information and attempting to authenticate) you go to the
Sheila> Expired page (which is an error) and the two session variables
Sheila> mentioned above are no longer in the session. The code
Sheila> contains NO statements that remove the session or delete
Sheila> session variables.


Sheila> Live demo here: http://mathxy.com/cgi-bin/test/mgr.py

Sheila> account num = 123456 email = test@example.com passwd =
Sheila> test1234 (only the login page should/would work. create login
Sheila> and lost password disabled.)

Sheila> Code (still) available here: http://mathxy.com/temp/

The problem is a bit difficult to spot.  If you run ngrep you will see
it though.

It looks like you have some <al-img> tags somewhere (haven't looked at
the source yet) because I can see requests like this:

        GET http://mathxy.com/cgi-bin/test/mgr.py/spacer.gif

This is followed by a traceback response which you do not see because
the browser thinks it should be getting an image.

After the traceback Albatross deletes the session, so you get dropped
back to a session expired state.

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Mon Jul 14 13:35:46 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 14 Jul 2003 09:35:46 -0400
Subject: [albatross-users] unscriptable object
Message-ID: <3F12B1B2.3090805@harvee.org>

Traceback (most recent call last):
   File "/usr/lib/python2.2/site-packages/albatross/app.py", line 221, in run
     self.merge_request(ctx)
   File "/usr/lib/python2.2/site-packages/albatross/app.py", line 301, in 
merge_request
     ctx.merge_request()
   File "/usr/lib/python2.2/site-packages/albatross/context.py", line 306, in 
merge_request
     self.set_value(name, value)
   File "/usr/lib/python2.2/site-packages/albatross/context.py", line 336, in 
set_value
     if name[0] == '_':
TypeError: unsubscriptable object



the Q is where and why?  _ is a legal character and used in other places.


From sheila at thinkspot.net  Mon Jul 14 14:08:30 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 14 Jul 2003 07:08:30 -0700
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F12B1B2.3090805@harvee.org>
References: <3F12B1B2.3090805@harvee.org>
Message-ID: <305789.1058166510@SHEILA-LAPTOP>

--On Monday, July 14, 2003 9:35 AM -0400 "Eric S. Johansson" 
<esj@harvee.org> wrote:

> Traceback (most recent call last):
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
> 221, in run      self.merge_request(ctx)
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
> 301, in  merge_request
>      ctx.merge_request()
>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
> line 306, in  merge_request
>      self.set_value(name, value)
>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
> line 336, in  set_value
>      if name[0] == '_':
> TypeError: unsubscriptable object
>
>
>
> the Q is where and why?  _ is a legal character and used in other
> places.

When I have received this error in the past, it is usually because 
the object was a None type.

I bet that your name sequence is empty (i.e. None). OK, well that is 
my guess.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org





From sheila at thinkspot.net  Mon Jul 14 14:12:09 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 14 Jul 2003 07:12:09 -0700
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F12B1B2.3090805@harvee.org>
References: <3F12B1B2.3090805@harvee.org>
Message-ID: <525505.1058166729@SHEILA-LAPTOP>

--On Monday, July 14, 2003 9:35 AM -0400 "Eric S. Johansson" 
<esj@harvee.org> wrote:

> Traceback (most recent call last):
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
> 221, in run      self.merge_request(ctx)
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
> 301, in  merge_request
>      ctx.merge_request()
>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
> line 306, in  merge_request
>      self.set_value(name, value)
>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
> line 336, in  set_value
>      if name[0] == '_':
> TypeError: unsubscriptable object
>
>
>
> the Q is where and why?  _ is a legal character and used in other
> places.
>

Another thought (this has bit me before):

Do all of your form elements have a name field? I had a problem once 
where I forgot to put names in the HTML template for my form 
elements, and IIRC got this type of error, especially if you have 
more than one form element without a name, as then albatross thinks 
it is a list type.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org





From esj at harvee.org  Mon Jul 14 14:36:01 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 14 Jul 2003 10:36:01 -0400
Subject: [albatross-users] unscriptable object
In-Reply-To: <305789.1058166510@SHEILA-LAPTOP>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>
Message-ID: <3F12BFD1.3090009@harvee.org>

Sheila King wrote:

> --On Monday, July 14, 2003 9:35 AM -0400 "Eric S. Johansson" 
> <esj@harvee.org> wrote:
> 
>> Traceback (most recent call last):
>>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
>> 221, in run      self.merge_request(ctx)
>>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
>> 301, in  merge_request
>>      ctx.merge_request()
>>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
>> line 306, in  merge_request
>>      self.set_value(name, value)
>>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
>> line 336, in  set_value
>>      if name[0] == '_':
>> TypeError: unsubscriptable object
>>
>>
>>
>> the Q is where and why?  _ is a legal character and used in other
>> places.
> 
> 
> When I have received this error in the past, it is usually because the 
> object was a None type.
> 
> I bet that your name sequence is empty (i.e. None). OK, well that is my 
> guess.

OK, I missing something here.  What name sequence are you talking about?

and as for missing names on form elements, I have all input types named. 
Especially the submit buttons.

FYI, here's the essential form elements and the albatross glue code I used to 
connect to my data extraction code.


<h2>Instructions:</h2>

<p>Red messages strongly suspected to be spam.  Green messages are strongly 
suspected to be not spam.</p>

<p>Your job is: Correct message with wrong classification.</p>


<p>There are only two ways a message can have the wrong classification.  Either:

<ul>
<li>The email <b>is</b> spam, and was classified as good e-mail</li>
<li>Or the email <b>is not</b> spam, and was classified as spam.</li>
<li>If you are not sure, check the message detail by following the subject 
link.</li>
</ul>

In order to change the classification, press the button on the line
labeled "change classification".  Repeat until all messages have been
processed.

</p>

<al-form action="reclaim.cgi" method="post">

<p>
The dumpster contains <b><al-value expr="spamtrap_size"></b>  messages.
</p>

<p>
Sort display by: <al-select name="sort_selection" optionexpr="sort_list"/> 
<al-input type="submit" name="sort_action" value="Change Sort" >
</p>

<p>
go to <al-input type="submit" name="pagedirection"  value="next" > or <al-input 
type="submit" name="pagedirection"  value="previous" > page
</p>

<table width=<al-value expr="page_width">
<tr align=left>
<th width=37>Select</th>
<th width=50 align=right>Score</th>
<th width=50 align=center>To</th>
<th>From</th>
<th>Subject (for detail follow link)
</tr>
<al-for iter="line" expr="display_list">

<tr class=<al-value expr="line.value().style">>
<td  bgcolor="#ffffff" align=center >

<al-if expr="line.value().priority == 3">
<al-exec expr="categorize='was_green'">
<al-else>
<al-exec expr="categorize='was_red'">
</al-if>

<al-input type="submit" nameexp="categorize"  value="change classification" >
<td>
<al-value expr="line.value().truncated_score()">
<al-if expr="line.value().list_type != None">
<br>
<al-value expr="line.value().list_type">
</al-if>
</td>

<td><al-value expr="line.value().destination"></td>
<td><al-value expr="line.value().from_address"></td>
<td><!--- <al-value expr="line.value().token"><br> --->
<al-a 
expr="'spamtrap_display.cgi?repository=dumpster&token='+line.value().token" 
target="spamtrap_display">
<al-value expr="line.value().subject"></al-a></td>
</tr>
</al-for>
</table>
</al-form>
<p>
aaaa
</p>


class correction:
     def page_enter(self, context):

         if not configuration_data['is_global_spamtrap']:
             user =  os.environ["REMOTE_USER"]
         else:
             user = None

         syslog.syslog("page_enter")

         context.locals.screen_buffer = {}
         context.add_session_vars('screen_buffer')

         context.locals.current_line_index = 0
         context.add_session_vars('current_line_index')

         context.locals.spamtrap_size = 0
         context.add_session_vars('spamtrap_size')

         context.locals.sort_item = "Date"
         context.add_session_vars('sort_item')

         context.locals.sort_list = [ 'Date', 'From', 'To', 'Subject' ]

         context.locals.spamtrap_size, context.locals.screen_buffer = 
generate_spamtrap_matrix(user)

         #syslog.syslog("spamtrap_size, screen_buffer %s %s" % 
(context.locals.spamtrap_size, context.locals.screen_buffer))

     def page_process(self, context):

         if not configuration_data['is_global_spamtrap']:
             user =  os.environ["REMOTE_USER"]
         else:
             user = None

         #if context.locals.train:
         #    syslog.syslog("train button pressed")
         syslog.syslog("any? button pressed")
         for entry in context.locals.__dict__:
             syslog.syslog("key %s" % (entry))

	# old code remnant
         if context.req_equals("submitaction"):
             if os.environ["REMOTE_USER"] == "e.johansson" and \
                    os.environ["REMOTE_ADDR"] == "192.168.25.1" or \
                    os.environ["REMOTE_USER"] == "albatross" :
                 context.locals.action_denied= "no-go"
             else:
                 context.locals.action_denied= "go-go"
                 pass

         #if context.req_equals("sort_item"):

     def page_display (self, context):

         context.locals.page_width="100%"
         syslog.syslog("type = %s" % type(context.locals.screen_buffer))
         sort_buffer = []
         for token in context.locals.screen_buffer.keys():
             line = context.locals.screen_buffer[token]

             #syslog.syslog("page_display line == %s" %line)
             if context.locals.sort_item  == "Date":
                 sort_buffer.append(str(line.date)+":"+token)

             elif context.locals.sort_item  == "To":
                 sort_buffer.append(str(line.destination)+":"+token)

             elif context.locals.sort_item  == "From":
                 sort_buffer.append(str(line.from_address)+":"+token)

             elif context.locals.sort_item  == "Subject":
                 sort_buffer.append(str(line.destination)+":"+token)
             else:
                 syslog.syslog("no such item on select list %s" % 
context.locals.sort_item)
                 sort_buffer.append(str(line.date)+":"+token)

         sort_buffer.sort()

         display_limit = 0
         display_list = []
         for index in sort_buffer:
             display_limit = display_limit + 1
             if display_limit > configuration_data['spamtrap_display_lines']:
                 break

             # extract items from screen_buffer according to the sort_buffer
             (junk,token) = index.split(":")

             display_list.append(context.locals.screen_buffer[token])

         # transfer display list of X lines to template variable space
         context.locals.display_list = display_list

         context.run_template("reclaim.html")
         syslog.syslog("emd of display")

app = SimpleApp(base_url = 'reclaim.py',
                 template_path = '.',
                 start_page = 'reclaim',
                 secret = '-=-secret-=-')

app.register_page('reclaim', correction())

app.run(Request())




From sheila at thinkspot.net  Mon Jul 14 18:14:56 2003
From: sheila at thinkspot.net (Sheila King)
Date: Mon, 14 Jul 2003 11:14:56 -0700
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F12BFD1.3090009@harvee.org>
References: <3F12B1B2.3090805@harvee.org>
 <305789.1058166510@SHEILA-LAPTOP> <3F12BFD1.3090009@harvee.org>
Message-ID: <3220060.1058181296@SHEILA-LAPTOP>

--On Monday, July 14, 2003 10:36 AM -0400 "Eric S. Johansson" 
<esj@harvee.org> wrote:

> Sheila King wrote:
>
>> --On Monday, July 14, 2003 9:35 AM -0400 "Eric S. Johansson"
>> <esj@harvee.org> wrote:
>>
>>> Traceback (most recent call last):
>>>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
>>> 221, in run      self.merge_request(ctx)
>>>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line
>>> 301, in  merge_request
>>>      ctx.merge_request()
>>>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
>>> line 306, in  merge_request
>>>      self.set_value(name, value)
>>>    File "/usr/lib/python2.2/site-packages/albatross/context.py",
>>> line 336, in  set_value
>>>      if name[0] == '_':
>>> TypeError: unsubscriptable object
>>>
>>>
>>>
>>> the Q is where and why?  _ is a legal character and used in other
>>> places.
>>
>>
>> When I have received this error in the past, it is usually because
>> the  object was a None type.
>>
>> I bet that your name sequence is empty (i.e. None). OK, well that
>> is my  guess.
>
> OK, I missing something here.  What name sequence are you talking
> about?

OK, never mind about the "name sequence" nonsense I wrote earlier.

But the error is being caused because whatever is being passed in to 
the set_value function in context.py is an object that cannot be 
referenced by indexes.

That's why it says "unsubscriptable"

I would recommend having your program print out or display somewhere 
what the value of "name" is before it gets to the line where the 
error is occurring. Hopefully that will give you better insight into 
what is occurring. But you can be assured that "name" is not a 
string. Yes, the "_" character is allowed in names, but the object 
that is assigned to "name" at that point in your program is not a 
string type object.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org


From djc at object-craft.com.au  Mon Jul 14 21:37:45 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 07:37:45 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F12BFD1.3090009@harvee.org>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>
 <3F12BFD1.3090009@harvee.org>
Message-ID: <m3fzl86b86.fsf@ferret.object-craft.com.au>

> <al-input type="submit" nameexp="categorize"  value="change classification" >

Try using nameexpr="categorize"

It looks like it may be a good idea to check that input fields have a
non-None name in the template interpreter...

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Mon Jul 14 21:39:04 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 07:39:04 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <525505.1058166729@SHEILA-LAPTOP>
References: <3F12B1B2.3090805@harvee.org> <525505.1058166729@SHEILA-LAPTOP>
Message-ID: <m3brvw6b5z.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

Sheila> Another thought (this has bit me before):

Sheila> Do all of your form elements have a name field? I had a
Sheila> problem once where I forgot to put names in the HTML template
Sheila> for my form elements, and IIRC got this type of error,

Exactly what I was thinking.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Mon Jul 14 23:51:13 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 09:51:13 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F12B1B2.3090805@harvee.org>
References: <3F12B1B2.3090805@harvee.org>
Message-ID: <m38yr0sm4u.fsf@ferret.object-craft.com.au>

> Traceback (most recent call last):
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line 221, in run
>      self.merge_request(ctx)
>    File "/usr/lib/python2.2/site-packages/albatross/app.py", line 301,
> in merge_request
>      ctx.merge_request()
>    File "/usr/lib/python2.2/site-packages/albatross/context.py", line
> 306, in merge_request
>      self.set_value(name, value)
>    File "/usr/lib/python2.2/site-packages/albatross/context.py", line
> 336, in set_value
>      if name[0] == '_':
> TypeError: unsubscriptable object
> 
> the Q is where and why?  _ is a legal character and used in other places.

What release of Albatross are you using?

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Mon Jul 14 23:50:43 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 14 Jul 2003 19:50:43 -0400
Subject: [albatross-users] unscriptable object
In-Reply-To: <m3fzl86b86.fsf@ferret.object-craft.com.au>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>	<3F12BFD1.3090009@harvee.org> <m3fzl86b86.fsf@ferret.object-craft.com.au>
Message-ID: <3F1341D3.4000001@harvee.org>

Dave Cole wrote:

>><al-input type="submit" nameexp="categorize"  value="change classification" >
> 
> 
> Try using nameexpr="categorize"

bingo.  that was it.


> It looks like it may be a good idea to check that input fields have a
> non-None name in the template interpreter...

you have an interesting problem.  How do you syntax check when you pass through 
all of the things that the manager keywords.  Maybe we should start building a 
"safe-to-pass" dictionary?

---eric




From djc at object-craft.com.au  Mon Jul 14 23:58:09 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 09:58:09 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F1341D3.4000001@harvee.org>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>
 <3F12BFD1.3090009@harvee.org>
 <m3fzl86b86.fsf@ferret.object-craft.com.au>
 <3F1341D3.4000001@harvee.org>
Message-ID: <m3wuekr78u.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Dave Cole wrote:
>>> <al-input type="submit" nameexp="categorize" value="change
>>> classification" >
>> Try using nameexpr="categorize"

Eric> bingo.  that was it.

>> It looks like it may be a good idea to check that input fields have
>> a non-None name in the template interpreter...

Eric> you have an interesting problem.  How do you syntax check when
Eric> you pass through all of the things that the manager keywords.
Eric> Maybe we should start building a "safe-to-pass" dictionary?

Rather than try to validate HTML all we do is look for the attributes
that are important to Albatross.  Validating HTML is a pretty big job.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul 15 00:14:25 2003
From: matt at pollenation.net (Matt Goodall)
Date: 15 Jul 2003 01:14:25 +0100
Subject: [albatross-users] unscriptable object
In-Reply-To: <m3wuekr78u.fsf@ferret.object-craft.com.au>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>
 <3F12BFD1.3090009@harvee.org> <m3fzl86b86.fsf@ferret.object-craft.com.au>
 <3F1341D3.4000001@harvee.org>  <m3wuekr78u.fsf@ferret.object-craft.com.au>
Message-ID: <1058228065.10475.16.camel@localhost>

On Tue, 2003-07-15 at 00:58, Dave Cole wrote:
> >>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:
> 
> Eric> Dave Cole wrote:
> >>> <al-input type="submit" nameexp="categorize" value="change
> >>> classification" >
> >> Try using nameexpr="categorize"
> 
> Eric> bingo.  that was it.
> 
> >> It looks like it may be a good idea to check that input fields have
> >> a non-None name in the template interpreter...
> 
> Eric> you have an interesting problem.  How do you syntax check when
> Eric> you pass through all of the things that the manager keywords.
> Eric> Maybe we should start building a "safe-to-pass" dictionary?
> 
> Rather than try to validate HTML all we do is look for the attributes
> that are important to Albatross.  Validating HTML is a pretty big job.

Only "pretty big"? ;-)

The way Albatross only processes known attributes, passing through
unknown attributes untouched, is crucial to its success as a templating
engine. I have used (and discarded) a number of templating engines that
tamper with attributes or plain refuse to let them through ... that's so
annoying.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From esj at harvee.org  Tue Jul 15 00:30:00 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 14 Jul 2003 20:30:00 -0400
Subject: [albatross-users] unscriptable object
In-Reply-To: <m38yr0sm4u.fsf@ferret.object-craft.com.au>
References: <3F12B1B2.3090805@harvee.org> <m38yr0sm4u.fsf@ferret.object-craft.com.au>
Message-ID: <3F134B08.4070104@harvee.org>

Dave Cole wrote:

> 
> What release of Albatross are you using?

1.10pre1


From djc at object-craft.com.au  Tue Jul 15 00:36:31 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 10:36:31 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F134B08.4070104@harvee.org>
References: <3F12B1B2.3090805@harvee.org>
 <m38yr0sm4u.fsf@ferret.object-craft.com.au>
 <3F134B08.4070104@harvee.org>
Message-ID: <m3brvwr5gw.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Dave Cole wrote:
>> What release of Albatross are you using?

Eric> 1.10pre1

Can you please try 1.10pre4 and see if it catches the problem?

- Dave

-- 
http://www.object-craft.com.au


From esj at harvee.org  Tue Jul 15 00:43:26 2003
From: esj at harvee.org (Eric S. Johansson)
Date: Mon, 14 Jul 2003 20:43:26 -0400
Subject: [albatross-users] unscriptable object
In-Reply-To: <m3wuekr78u.fsf@ferret.object-craft.com.au>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>	<3F12BFD1.3090009@harvee.org>	<m3fzl86b86.fsf@ferret.object-craft.com.au>	<3F1341D3.4000001@harvee.org> <m3wuekr78u.fsf@ferret.object-craft.com.au>
Message-ID: <3F134E2E.2090406@harvee.org>

Dave Cole wrote:
> Eric> you have an interesting problem.  How do you syntax check when
> Eric> you pass through all of the things that the manager keywords.
> Eric> Maybe we should start building a "safe-to-pass" dictionary?
> 
> Rather than try to validate HTML all we do is look for the attributes
> that are important to Albatross.  Validating HTML is a pretty big job.

yes, validating HTML is a huge job.  However, the task gets much smaller if you 
ask the question "is this a valid HTML/albatross string".  then you would catch 
misspellings like the one I made without adding a huge amount of weight.  The 
hard part then becomes combing the HTML documentation for valid symbols or 
strings or whatever you want to call them.

additionally, this does not need to be a hard and fast error but can simply be a 
syslog directed warning turned on by explicit call.

just an idea and worth exactly what you paid for it.  ;-)

---eric



From matt at pollenation.net  Tue Jul 15 00:52:22 2003
From: matt at pollenation.net (Matt Goodall)
Date: 15 Jul 2003 01:52:22 +0100
Subject: [albatross-users] New get_page_from_uri functionality
Message-ID: <1058230342.30947.55.camel@localhost>

I've just made use of the new method, get_page_from_uri, that Dave added
a week or so ago. It works very nicely - thanks Dave.

I now have a page module, assets.py, which handles all requests for the
/assets uri tree, i.e. /assets/styles/default.css sends the content of
styles/default.css straight from the file system to the browser.

I'll try to add this as an example to the Wiki when I have a few
minutes. Feel free to hassle me to do this sooner if it would be useful
to you.

Thanks again Dave.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From andrewm at object-craft.com.au  Tue Jul 15 00:58:33 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 15 Jul 2003 10:58:33 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: Message from "Eric S. Johansson" <esj@harvee.org>
 of "Mon, 14 Jul 2003 20:43:26 -0400." <3F134E2E.2090406@harvee.org>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP> <3F12BFD1.3090009@harvee.org> <m3fzl86b86.fsf@ferret.object-craft.com.au> <3F1341D3.4000001@harvee.org> <m3wuekr78u.fsf@ferret.object-craft.com.au>  <3F134E2E.2090406@harvee.org>
Message-ID: <20030715005833.D26CA3C421@coffee.object-craft.com.au>

>yes, validating HTML is a huge job.  However, the task gets much smaller
>if you ask the question "is this a valid HTML/albatross string".  then you
>would catch misspellings like the one I made without adding a huge amount
>of weight.  The hard part then becomes combing the HTML documentation for
>valid symbols or strings or whatever you want to call them.
>
>additionally, this does not need to be a hard and fast error but can
>simply be a syslog directed warning turned on by explicit call.

Validation is something that's only of interest to the developer - it's
probably better to have a stand-alone validator that you run over the
templates while developing them, than to wear the cost of validation
every time to run a template.

I'd presume there are HTML validators out there than can be taught
about albatross's tags, although it would be difficult with things like
<al-if><al-else></al-endif>.

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From djc at object-craft.com.au  Tue Jul 15 01:06:58 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 11:06:58 +1000
Subject: [albatross-users] unscriptable object
In-Reply-To: <3F134E2E.2090406@harvee.org>
References: <3F12B1B2.3090805@harvee.org> <305789.1058166510@SHEILA-LAPTOP>
 <3F12BFD1.3090009@harvee.org>
 <m3fzl86b86.fsf@ferret.object-craft.com.au>
 <3F1341D3.4000001@harvee.org>
 <m3wuekr78u.fsf@ferret.object-craft.com.au>
 <3F134E2E.2090406@harvee.org>
Message-ID: <m3y8z0pphp.fsf@ferret.object-craft.com.au>

>>>>> "Eric" == Eric S Johansson <esj@harvee.org> writes:

Eric> Dave Cole wrote: you have an interesting problem.  How do you
Eric> syntax check when you pass through all of the things that the
Eric> manager keywords.  Maybe we should start building a
Eric> "safe-to-pass" dictionary?
>> Rather than try to validate HTML all we do is look for the
>> attributes that are important to Albatross.  Validating HTML is a
>> pretty big job.

Eric> yes, validating HTML is a huge job.  However, the task gets much
Eric> smaller if you ask the question "is this a valid HTML/albatross
Eric> string".  then you would catch misspellings like the one I made
Eric> without adding a huge amount of weight.  The hard part then
Eric> becomes combing the HTML documentation for valid symbols or
Eric> strings or whatever you want to call them.

Eric> additionally, this does not need to be a hard and fast error but
Eric> can simply be a syslog directed warning turned on by explicit
Eric> call.

Eric> just an idea and worth exactly what you paid for it.  ;-)

I just did some research...

On Debian you can install two packages; tidy-proxy and
wdg-html-validator.  This will probably show you all sorts of things
about the HTML coming out of your application.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Tue Jul 15 01:17:39 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 15 Jul 2003 11:17:39 +1000
Subject: [albatross-users] New get_page_from_uri functionality
In-Reply-To: <1058230342.30947.55.camel@localhost>
References: <1058230342.30947.55.camel@localhost>
Message-ID: <m3n0fgpozw.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> I've just made use of the new method, get_page_from_uri, that
Matt> Dave added a week or so ago. It works very nicely - thanks Dave.

It seems like such an obvious feature in retrospect.

Matt> I now have a page module, assets.py, which handles all requests
Matt> for the /assets uri tree, i.e. /assets/styles/default.css sends
Matt> the content of styles/default.css straight from the file system
Matt> to the browser.

Cool.  It does open up all sorts of weird and wonderful URL schemes.

Matt> I'll try to add this as an example to the Wiki when I have a few
Matt> minutes. Feel free to hassle me to do this sooner if it would be
Matt> useful to you.

It would be excellent to get an example up there.

Matt> Thanks again Dave.

My pleasure.  Thanks for taking it for a spin.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Tue Jul 15 01:27:32 2003
From: matt at pollenation.net (Matt Goodall)
Date: 15 Jul 2003 02:27:32 +0100
Subject: [albatross-users] Feature request
Message-ID: <1058232452.31165.90.camel@localhost>

It would be nice if the code for multiples pages in a random app could
be kept in a single Python module. Something like a mix of page object
and page module (see below) might work although I am aware that page
objects are created once and are registered with the application.

users.py
--------
# Page for uri of /users/list
class ListUsers:
    name = 'list'
    def page_display(self, ctx):
        pass

# Page for uri of /users/add
class AddUser:
    name = 'add'
    def page_process(self, ctx):
        pass
    def page_display(self, ctx):
        pass
-----

Perhaps there is a better way than this?

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From sheila at thinkspot.net  Tue Jul 15 20:28:34 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 15 Jul 2003 13:28:34 -0700
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <m33cha6ox3.fsf@ferret.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
Message-ID: <6834747.1058275714@SHEILA-LAPTOP>

--On Monday, July 14, 2003 8:29 AM +1000 Dave Cole
<djc@object-craft.com.au> wrote:

>>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:
> 
> Sheila> --On Sunday, July 13, 2003 4:52 PM +1000 Dave Cole
> Sheila> <djc@object-craft.com.au> wrote:

> Sheila> Current Problem: Session Variables are still being lost, as
> Sheila> before. If you go to the initial log in page, it shows four
> Sheila> session variables, including _last_access and _authstatus but
> Sheila> when you go to the next page (after entering log in
> Sheila> information and attempting to authenticate) you go to the
> Sheila> Expired page (which is an error) and the two session variables
> Sheila> mentioned above are no longer in the session. The code
> Sheila> contains NO statements that remove the session or delete
> Sheila> session variables.

> The problem is a bit difficult to spot.  If you run ngrep you will see
> it though.
> 
> It looks like you have some <al-img> tags somewhere (haven't looked at
> the source yet) because I can see requests like this:
> 
>         GET http://mathxy.com/cgi-bin/test/mgr.py/spacer.gif
> 
> This is followed by a traceback response which you do not see because
> the browser thinks it should be getting an image.
> 
> After the traceback Albatross deletes the session, so you get dropped
> back to a session expired state.

OK, I've been working on this, and I do agree now that there is/are
error(s) in my code causing the session to drop.

I've installed Ethereal (<http://www.ethereal.com/>) and am trying to learn
how to use it to see the http response packets so that I can see these
errors you are mentioning.

There is definitely something different about the error handling in pre4 as
opposed to previous versions of Albatross.

The error you mention above is essentially a 404 error. There was no al-img
tag, but my commonmacros.html file did have a normal HTML <img> tag and the
image refereced was not on the server, and so should generate a 404 error.

However, it has apparently been like that for a long time (this image is
also called in my full application program) and has never caused session to
drop before.

Maybe it should cause session to drop. Dunno.

But, even if I fix that problem, there are other errors I've found, such as
expressions not being defined in the local context name space. Before these
always took me to the default Albatross error page (which I've not tried to
overwrite) with a nice traceback. When this happened, I KNEW that something
in my code was wrong.

Now it is not doing that. It is taking me to the expired session page in my
application and the only way I can see the tracebacks for anything (it
seems) is by sniffing packets.

I'm not a neophyte cgi programmer, although I'm no expert either. But I
have a fair/modest amount of experience in this area, and I've never had to
sniff packets before to read the error messages sent by a Python cgi script.

:(

Is this again something peculiar to the redirect/random URL type of
applications?

It looks to me like this new send_redirect function could be the culprit?
Or is at least involved in some way?

    def send_redirect(self, loc):
        cookie = self.get_header('Set-Cookie')
        if cookie:
            self.request.write_header('Set-Cookie', cookie)
        return self.request.redirect(loc) 

Could this be the reason that after the handle_exception method prints out
the traceback, that the request is still redirected to the previously
requested redirect page? If so, IMO this behavior should be modified.
handle_exception should print the output (traceback) and then halt.

I'm just grasping about to try and think why it is so flipping hard for me
to see that there are errors in my code, when with previous versions of
Albatross it was easy.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org








From matt at pollenation.net  Tue Jul 15 21:12:12 2003
From: matt at pollenation.net (Matt Goodall)
Date: 15 Jul 2003 22:12:12 +0100
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <6834747.1058275714@SHEILA-LAPTOP>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP>
Message-ID: <1058303532.8094.4.camel@localhost>

On Tue, 2003-07-15 at 21:28, Sheila King wrote:
> I've installed Ethereal (<http://www.ethereal.com/>) and am trying to learn
> how to use it to see the http response packets so that I can see these
> errors you are mentioning.

If you're using Mozilla you might find the LiveHTTPHeaders tool a bit
more convenient than Ethereal. http://livehttpheaders.mozdev.org/

If you're not using Mozilla then you should be ;-).

- Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From sheila at thinkspot.net  Tue Jul 15 21:32:46 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 15 Jul 2003 14:32:46 -0700
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <1058303532.8094.4.camel@localhost>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP> <1058303532.8094.4.camel@localhost>
Message-ID: <10687297.1058279566@SHEILA-LAPTOP>

--On Tuesday, July 15, 2003 10:12 PM +0100 Matt Goodall
<matt@pollenation.net> wrote:

> On Tue, 2003-07-15 at 21:28, Sheila King wrote:
>> I've installed Ethereal (<http://www.ethereal.com/>) and am trying to
>> learn how to use it to see the http response packets so that I can see
>> these errors you are mentioning.
> 
> If you're using Mozilla you might find the LiveHTTPHeaders tool a bit
> more convenient than Ethereal. http://livehttpheaders.mozdev.org/
> 
> If you're not using Mozilla then you should be ;-).

Thanks for the link. That looks quite interesting, and of course I'm using
Mozilla. Although technically I will need to test my app in many different
browsers before deployment. It does look like LiveHTTPHeaders may have some
debugging/testing functionality that I cannot (easily) achieve by other
means.

It looks like that LiveHTTPHeaders doesn't provide all of the content,
though? And I need to see the content to see the tracebacks, which lets me
know exactly what is wrong with the script. So, I guess it's ethereal for
now.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org






From matt at pollenation.net  Tue Jul 15 21:48:41 2003
From: matt at pollenation.net (Matt Goodall)
Date: 15 Jul 2003 22:48:41 +0100
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <6834747.1058275714@SHEILA-LAPTOP>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP>
Message-ID: <1058305720.28841.19.camel@localhost>

On Tue, 2003-07-15 at 21:28, Sheila King wrote:
> But, even if I fix that problem, there are other errors I've found, such as
> expressions not being defined in the local context name space. Before these
> always took me to the default Albatross error page (which I've not tried to
> overwrite) with a nice traceback. When this happened, I KNEW that something
> in my code was wrong.
> 
> Now it is not doing that. It is taking me to the expired session page in my
> application and the only way I can see the tracebacks for anything (it
> seems) is by sniffing packets.

I just tried this here by deliberately changing the name in an
<al-value> expr. Is that what you mean by "expressions not being defined
in the local context"? I get a traceback for a NameError exception as
expected. 

I'm not sure what's happening for you but it might be worth
**completely** removing the albatross directory from site-packages and
reinstalling 1.10pre4. IIRC, you patched 1.10pre2 so maybe there is
something that 1.10pre4 setup.py didn't replace. This is pure
speculation but at least you will definitely know what's installed.

If it still fails after that could you isolate the error in a totally
trivial example and post it to the list. I'll _try_ to take a look if
Dave and Andrew are too busy although I can't promise too much - lots to
do myself.

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From todd at slack.net  Tue Jul 15 22:10:36 2003
From: todd at slack.net (Todd Grimason)
Date: Tue, 15 Jul 2003 18:10:36 -0400
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <10687297.1058279566@SHEILA-LAPTOP>; from sheila@thinkspot.net on Tue, Jul 15, 2003 at 02:32:46PM -0700
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au> <m3znjttehl.fsf@ferret.object-craft.com.au> <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au> <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au> <1900622.1058087856@SHEILA-LAPTOP> <m33cha6ox3.fsf@ferret.object-craft.com.au> <6834747.1058275714@SHEILA-LAPTOP> <1058303532.8094.4.camel@localhost> <10687297.1058279566@SHEILA-LAPTOP>
Message-ID: <20030715181036.B18619@crank.slack.net>

* Sheila King <sheila@thinkspot.net> [2003-07-15 17:34]:
> 
> It looks like that LiveHTTPHeaders doesn't provide all of the content,
> though? And I need to see the content to see the tracebacks, which lets me
> know exactly what is wrong with the script. So, I guess it's ethereal for
> now.

tcpflow -- A TCP Flow Recorder 

http://www.circlemud.org/~jelson/software/tcpflow/

very handy, assuming(?) you're on a linux/bsd/osx platform.


-- 

___________________________
toddgrimason*todd@slack.net



From sheila at thinkspot.net  Tue Jul 15 22:52:17 2003
From: sheila at thinkspot.net (Sheila King)
Date: Tue, 15 Jul 2003 15:52:17 -0700
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <1058305720.28841.19.camel@localhost>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP> <1058305720.28841.19.camel@localhost>
Message-ID: <15458017.1058284337@SHEILA-LAPTOP>

--On Tuesday, July 15, 2003 10:48 PM +0100 Matt Goodall
<matt@pollenation.net> wrote:

> I just tried this here by deliberately changing the name in an
> <al-value> expr. Is that what you mean by "expressions not being defined
> in the local context"? I get a traceback for a NameError exception as
> expected. 
>
> I'm not sure what's happening for you but it might be worth
> **completely** removing the albatross directory from site-packages and
> reinstalling 1.10pre4. IIRC, you patched 1.10pre2 so maybe there is
> something that 1.10pre4 setup.py didn't replace. This is pure
> speculation but at least you will definitely know what's installed.
> 
> If it still fails after that could you isolate the error in a totally
> trivial example and post it to the list. I'll _try_ to take a look if
> Dave and Andrew are too busy although I can't promise too much - lots to
> do myself.

Thank Matt for the tips and ideas. I will try to dig in a bit more tonight
(providing my other tasks permit the time) and see if I can isolate what
the problem is, exactly.

-- 
Sheila King
http://www.thinkspot.net/sheila/
http://www.k12groups.org






From andrewm at object-craft.com.au  Tue Jul 15 23:09:31 2003
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Wed, 16 Jul 2003 09:09:31 +1000
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: Message from Sheila King <sheila@thinkspot.net>
 of "Tue, 15 Jul 2003 13:28:34 MST." <6834747.1058275714@SHEILA-LAPTOP>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au> <m3znjttehl.fsf@ferret.object-craft.com.au> <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au> <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au> <1900622.1058087856@SHEILA-LAPTOP> <m33cha6ox3.fsf@ferret.object-craft.com.au>  <6834747.1058275714@SHEILA-LAPTOP>
Message-ID: <20030715230931.4FDE23C421@coffee.object-craft.com.au>

>OK, I've been working on this, and I do agree now that there is/are
>error(s) in my code causing the session to drop.

One thing I find useful is to comment out the call to remove_session in
app.py, Application.handle_exception(). I'm not sure if this will be as
helpful in a randompage app, but it's very handy for other deployment
options.

>There is definitely something different about the error handling in pre4 as
>opposed to previous versions of Albatross.

The error *handling* largely hasn't changed, but it could be that pre4
is picking up an error (and raising an exception) that previous versions
let pass silently, or that pre4 is botching headers.

>The error you mention above is essentially a 404 error. There was no al-img
>tag, but my commonmacros.html file did have a normal HTML <img> tag and the
>image refereced was not on the server, and so should generate a 404 error.

I wouldn't have thought that would cause the symptoms you are seeing.

>But, even if I fix that problem, there are other errors I've found, such as
>expressions not being defined in the local context name space. Before these
>always took me to the default Albatross error page (which I've not tried to
>overwrite) with a nice traceback. When this happened, I KNEW that something
>in my code was wrong.
>
>Now it is not doing that. It is taking me to the expired session page in my
>application and the only way I can see the tracebacks for anything (it
>seems) is by sniffing packets.

What can cause this is if the rendered html template contains references
back to other albatross rendered content (so, for example, frames or
images). What happens is that one frame or image gets an exception, and
when the next one is fetched by the browser, the session has been removed.
The fact that you are seeing a traceback going over the wire, but not
seeing anything rendered in your browser strongly suggests a non-html
object is to blame (an image?).

Another possibility is that the traceback you're seeing go past is being
interpreted by the browser as HTTP headers and hidden for that reason
(which would tie in with the theory you mention below). Maybe you could
mail a text dump of the trace to me and I see if I can tell which
option it is? If you can't get ethereal to produce a text dump, do
something like:

    # tcpdump -s 2000 -n -w /tmp/alba-dump port 80

Then gzip the /tmp/alba-dump file and mail it to me.

>I'm not a neophyte cgi programmer, although I'm no expert either. But I
>have a fair/modest amount of experience in this area, and I've never had
>to sniff packets before to read the error messages sent by a Python cgi
>script.
>
>:(

It does tend to be a technique best used when all others have failed. The
last time I used it in this context was debugging the "file" input
handling (to see the mime headers)... 8-)

>Is this again something peculiar to the redirect/random URL type of
>applications?

Well, they do have a lot of gotchyas that don't apply to other application
models - I tend to try to avoid them for that reason.

>It looks to me like this new send_redirect function could be the culprit?
>Or is at least involved in some way?
>
>    def send_redirect(self, loc):
>        cookie = self.get_header('Set-Cookie')
>        if cookie:
>            self.request.write_header('Set-Cookie', cookie)
>        return self.request.redirect(loc) 
>
>Could this be the reason that after the handle_exception method prints out
>the traceback, that the request is still redirected to the previously
>requested redirect page? If so, IMO this behavior should be modified.
>handle_exception should print the output (traceback) and then halt.

Ah - that sounds like a good theory. Dave?

>I'm just grasping about to try and think why it is so flipping hard for me
>to see that there are errors in my code, when with previous versions of
>Albatross it was easy.

It certainly should be easy...

-- 
Andrew McNamara, Senior Developer, Object Craft
http://www.object-craft.com.au/

From djc at object-craft.com.au  Tue Jul 15 23:42:13 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 16 Jul 2003 09:42:13 +1000
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <6834747.1058275714@SHEILA-LAPTOP>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP>
Message-ID: <m3ptkbfjca.fsf@ferret.object-craft.com.au>

>>>>> "Sheila" == Sheila King <sheila@thinkspot.net> writes:

>> It looks like you have some <al-img> tags somewhere (haven't looked
>> at the source yet) because I can see requests like this:
>> 
>> GET http://mathxy.com/cgi-bin/test/mgr.py/spacer.gif
>> 
>> This is followed by a traceback response which you do not see
>> because the browser thinks it should be getting an image.
>> 
>> After the traceback Albatross deletes the session, so you get
>> dropped back to a session expired state.

Sheila> There is definitely something different about the error
Sheila> handling in pre4 as opposed to previous versions of Albatross.

I am interested to know what you see as being the difference.

Sheila> The error you mention above is essentially a 404 error. There
Sheila> was no al-img tag, but my commonmacros.html file did have a
Sheila> normal HTML <img> tag and the image refereced was not on the
Sheila> server, and so should generate a 404 error.

Another way that you might see the browser requesting the image from
the application via http://mathxy.com/cgi-bin/test/mgr.py/spacer.gif
is if you put a plain IMG tag in your template like this:

        <img src="spacer.gif">

The browser will treat this as a relative URL and will expand it to a
full URL. 

Just looked at your code.  commonmacros.html has this:

        <img src="spacer.gif" width="600" height="1" vspace="0" hspace="0">

Sheila> However, it has apparently been like that for a long time
Sheila> (this image is also called in my full application program) and
Sheila> has never caused session to drop before.

Sheila> Maybe it should cause session to drop. Dunno.

Your application does not recognise the request for the image and
treat it as such - it responds with a python traceback.

Sheila> But, even if I fix that problem, there are other errors I've
Sheila> found, such as expressions not being defined in the local
Sheila> context name space. Before these always took me to the default
Sheila> Albatross error page (which I've not tried to overwrite) with
Sheila> a nice traceback. When this happened, I KNEW that something in
Sheila> my code was wrong.

Did you manage to see the content being returned for the image
requests?  They certainly looked like a normal Albatross traceback to
me.  I did not look at the content of the traceback.

Sheila> Now it is not doing that. It is taking me to the expired
Sheila> session page in my application and the only way I can see the
Sheila> tracebacks for anything (it seems) is by sniffing packets.

This is not so much a problem of Albatross.  The browser requests some
image data and receives some text in response.  It handles this by
silently discarding the data from what I can see.

Sheila> I'm not a neophyte cgi programmer, although I'm no expert
Sheila> either. But I have a fair/modest amount of experience in this
Sheila> area, and I've never had to sniff packets before to read the
Sheila> error messages sent by a Python cgi script.

You don't have to sniff packets - I just do it sometimes when I do not
want to modify the source.  Given that not all of what happens between
the browser and application is visible, packet sniffing is a useful
tool at times.

Sheila> Is this again something peculiar to the redirect/random URL
Sheila> type of applications?

Nope.  If you cause the browser to request an image of an application
that is not prepared to serve images then you will not get good
results.

Sheila> It looks to me like this new send_redirect function could be
Sheila> the culprit?  Or is at least involved in some way?

Shiela>     def send_redirect(self, loc):
Shiela>         cookie = self.get_header('Set-Cookie')
Shiela>         if cookie:
Shiela>             self.request.write_header('Set-Cookie', cookie)
Shiela>         return self.request.redirect(loc) 

I would be a little surprised if it was having any impact on the
problem.

Sheila> Could this be the reason that after the handle_exception
Sheila> method prints out the traceback, that the request is still
Sheila> redirected to the previously requested redirect page? If so,
Sheila> IMO this behavior should be modified.  handle_exception should
Sheila> print the output (traceback) and then halt.

The redirect is caused by the Location header, not the Set-Cookie.

Sheila> I'm just grasping about to try and think why it is so flipping
Sheila> hard for me to see that there are errors in my code, when with
Sheila> previous versions of Albatross it was easy.

I am a bit confused about this.  I cannot see how the 1.10pre4 release
is different to previous releases in how it handles and reports
errors.  That does not mean that there are no differences, it just
means I don't think anything that has changed to impact that
behaviour.

- Dave

-- 
http://www.object-craft.com.au


From djc at object-craft.com.au  Wed Jul 16 00:12:09 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 16 Jul 2003 10:12:09 +1000
Subject: [albatross-users] Re: pre4 - Problems still ...
In-Reply-To: <20030715230931.4FDE23C421@coffee.object-craft.com.au>
References: <m3n0gukjvb.fsf@ferret.object-craft.com.au>
 <m3znjttehl.fsf@ferret.object-craft.com.au>
 <m3vfu7xhhv.fsf_-_@ferret.object-craft.com.au>
 <m3r84ulxz6.fsf_-_@ferret.object-craft.com.au>
 <1900622.1058087856@SHEILA-LAPTOP>
 <m33cha6ox3.fsf@ferret.object-craft.com.au>
 <6834747.1058275714@SHEILA-LAPTOP>
 <20030715230931.4FDE23C421@coffee.object-craft.com.au>
Message-ID: <m3lluzfhye.fsf@ferret.object-craft.com.au>

>>>>> "Andrew" == Andrew McNamara <andrewm@object-craft.com.au> writes:

>> Could this be the reason that after the handle_exception method
>> prints out the traceback, that the request is still redirected to
>> the previously requested redirect page? If so, IMO this behavior
>> should be modified.  handle_exception should print the output
>> (traceback) and then halt.

Andrew> Ah - that sounds like a good theory. Dave?

I can't see how setting a cookie would cause a redirect to occur when
previously it didn't.  Maybe some packet dumps will shed some light.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Wed Jul 16 10:04:55 2003
From: matt at pollenation.net (Matt Goodall)
Date: Wed, 16 Jul 2003 11:04:55 +0100
Subject: [albatross-users] Levelling the "random vs non-random" field
Message-ID: <3F152347.1060002@pollenation.net>

The other day I was thinking about what _I_ think could make Albatross 
better still. At some point (i.e. if I ever have the time) there are a 
few things I might experiment with in Albatross:

    * Provide random apps with the page lifecycle (specifically
      page_enter and page_leave) that non-random apps benefit from -
      it's a wonderful way to manage session data. This is probably more
      difficult but I suspect it's not impossible. The session already
      has __page__ so I think Albatross has enough to work out what's
      going on, it's just putting the hooks in the right place?
    * Make ctx.set_page() and ctx.redirect() work for both types of
      apps. I think this is one of the few cases where Albatross forces
      you to decide on an application type right at the start and I
      think that's a shame. Perhaps it's even unnecessary? I don't know yet.
    * Make a RandomPageObject mixin. Or better still, separate the
      random vs non-random and page object vs page module behaviours to
      allow whatever combination makes sense. I don't think this would
      be too difficult and it may be quite useful.
    * More request adapters. The one I'm probably most interested in is
      Twisted. Twisted is a lot of fun and has some some really nice
      features but woven (the web part of the framework) is more complex
      than something like Albatross.

What are your thoughts on the above? Have you explored any of these 
already and decided its too difficult? Dave/Andrew, is there anything 
that you simply don't want to see happening in Albatross?

Of course, I may never actually find any time to look into any of these ;-)

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenation.net



From djc at object-craft.com.au  Wed Jul 16 10:56:43 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 16 Jul 2003 20:56:43 +1000
Subject: [albatross-users] Levelling the "random vs non-random" field
In-Reply-To: <3F152347.1060002@pollenation.net>
References: <3F152347.1060002@pollenation.net>
Message-ID: <m3znjewxhw.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> The other day I was thinking about what _I_ think could make
Matt> Albatross better still. At some point (i.e. if I ever have the
Matt> time) there are a few things I might experiment with in
Matt> Albatross:

Matt>     * Provide random apps with the page lifecycle (specifically
Matt> page_enter and page_leave) that non-random apps benefit from -
Matt> it's a wonderful way to manage session data. This is probably
Matt> more difficult but I suspect it's not impossible. The session
Matt> already has __page__ so I think Albatross has enough to work out
Matt> what's going on, it's just putting the hooks in the right place?

I am not sure that this would be general enough to rely upon in your
application - users would be jumping all over the place.  Does that
matter though?

Matt> * Make ctx.set_page() and ctx.redirect() work for both types of
Matt> apps. I think this is one of the few cases where Albatross
Matt> forces you to decide on an application type right at the start
Matt> and I think that's a shame. Perhaps it's even unnecessary? I
Matt> don't know yet.

That seems like a really good idea.  So good it makes me wonder why it
does not already do that.  Maybe I have forgotten why I thought the
two were sufficiently different enough to warrant different names.

Matt> * Make a RandomPageObject mixin. Or better still, separate the
Matt> random vs non-random and page object vs page module behaviours
Matt> to allow whatever combination makes sense. I don't think this
Matt> would be too difficult and it may be quite useful.

If we required Python 2.2 and changed RandomPageModuleMixin so that it
did not inherit from any other class, then I think we could change the
code to use the super() builtin function where it is currently calling
PageModuleMixin.page_enter() and other methods.  We would still need
to do something about the call to self.load_page_module() though.
Maybe just implement a stubbed method in the PageObjectMixin.

Matt> * More request adapters. The one I'm probably most interested in
Matt> is Twisted. Twisted is a lot of fun and has some some really
Matt> nice features but woven (the web part of the framework) is more
Matt> complex than something like Albatross.

That would be cool to have.  I spent a little time looking at Twisted
when I had to build a forking server.  I could not get Twisted to do
what I wanted and did not have the time to figure it out.

Matt> What are your thoughts on the above? Have you explored any of
Matt> these already and decided its too difficult? Dave/Andrew, is
Matt> there anything that you simply don't want to see happening in
Matt> Albatross?

My immediate response to that is no.  Each time the classes are pushed
in a new direction we are closer to my original goal with Albatross.
I wanted to build a collection of classes that could be used to
construct many different types of web applications.

In my day dreams I imagined different layers on top of Albatross being
used to narrow the focus and raise the functionality for different and
smaller problem domains.  It is still a day dream though.

Matt> Of course, I may never actually find any time to look into any
Matt> of these ;-)

If only people would pay us money to play with this stuff all day :-)

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Wed Jul 16 12:54:56 2003
From: matt at pollenation.net (Matt Goodall)
Date: Wed, 16 Jul 2003 13:54:56 +0100
Subject: [albatross-users] Levelling the "random vs non-random" field
In-Reply-To: <m3znjewxhw.fsf@ferret.object-craft.com.au>
References: <3F152347.1060002@pollenation.net> <m3znjewxhw.fsf@ferret.object-craft.com.au>
Message-ID: <3F154B20.2020203@pollenation.net>

Dave Cole wrote:

>>>>>>"Matt" == Matt Goodall <matt@pollenation.net> writes:
>>>>>>            
>>>>>>
>Matt>     * Provide random apps with the page lifecycle (specifically
>Matt> page_enter and page_leave) that non-random apps benefit from -
>Matt> it's a wonderful way to manage session data. This is probably
>Matt> more difficult but I suspect it's not impossible. The session
>Matt> already has __page__ so I think Albatross has enough to work out
>Matt> what's going on, it's just putting the hooks in the right place?
>
>I am not sure that this would be general enough to rely upon in your
>application - users would be jumping all over the place.  Does that
>matter though?
>
In my eyes, I see the real benefit of page_enter and page_leave as 
somewhere to set up and tear down any session data for the page. This 
need is equally true for random apps as non-random apps. Are there any 
other uses of page_enter/leave that you have come across? Obviously, 
there's no way of knowing to clear up session data if someone closes 
their browser but at least you get the chance during normal application 
flow.

Hmm, just a random thought, but would it be useful/possible to take the 
page_enter/leave concept and make pages with persistent state? I'm 
thinking along the lines of pickling the non-transient properties of a 
page instance to the session rather than mucking about with 
add_session_vars. Woah, best stop this train of thought now or I'll want 
to play even more ;-).

>Matt> * Make ctx.set_page() and ctx.redirect() work for both types of
>Matt> apps. I think this is one of the few cases where Albatross
>Matt> forces you to decide on an application type right at the start
>Matt> and I think that's a shame. Perhaps it's even unnecessary? I
>Matt> don't know yet.
>
>That seems like a really good idea.  So good it makes me wonder why it
>does not already do that.  Maybe I have forgotten why I thought the
>two were sufficiently different enough to warrant different names.
>
Please don't get rid of one, both are required: set_page() essentially 
just changes which page actually gets displayed, redirect() allows you 
to change the URL or jump away from a form post.

IIRC, Sheila mentioned that she didn't like _having_ to redirect in 
random apps, set_page() would help here but redirect() definitely has 
its uses as described above.

>Matt> * Make a RandomPageObject mixin. Or better still, separate the
>Matt> random vs non-random and page object vs page module behaviours
>Matt> to allow whatever combination makes sense. I don't think this
>Matt> would be too difficult and it may be quite useful.
>
>If we required Python 2.2 and changed RandomPageModuleMixin so that it
>did not inherit from any other class, then I think we could change the
>code to use the super() builtin function where it is currently calling
>PageModuleMixin.page_enter() and other methods.  We would still need
>to do something about the call to self.load_page_module() though.
>Maybe just implement a stubbed method in the PageObjectMixin.
>
Ah, super() ... so nice. I've lost count of the number of times I've 
tried to use it with Albatross before remembering that it's only 
available for new style classes ;-).

>Matt> * More request adapters. The one I'm probably most interested in
>Matt> is Twisted. Twisted is a lot of fun and has some some really
>Matt> nice features but woven (the web part of the framework) is more
>Matt> complex than something like Albatross.
>
>That would be cool to have.  I spent a little time looking at Twisted
>when I had to build a forking server.  I could not get Twisted to do
>what I wanted and did not have the time to figure it out.
>
Twisted doesn't fork, could that be why you had trouble? ;-). At least, 
I don't think it forks?!

Seriously though, once Twisted is documented and there are more, well 
written example apps it could be a very nice application platform. The 
thought of integrating smtp, nntp, http etc into one application is just 
so lovely. But right now, I'm in the same boat as you - looking 
longingly in with not enough time to play and learn. By the way, have 
you looked at Twisted's LivePage stuff? All I can say is wow!

I think scgi forks. I was wondering about that as yet another request 
adapter but its lacking that "itch to scratch" quality for me now that 
there is fastcgi support.

Perhaps a FeatureRequest page to go alongside the ProposedChanges page 
on the wiki would be a good way of gathering ideas for future versions? 
That might also help keep track of things as people wander in and out of 
the project, as invariably happens with open source.

>Matt> What are your thoughts on the above? Have you explored any of
>Matt> these already and decided its too difficult? Dave/Andrew, is
>Matt> there anything that you simply don't want to see happening in
>Matt> Albatross?
>
>My immediate response to that is no.  Each time the classes are pushed
>in a new direction we are closer to my original goal with Albatross.
>I wanted to build a collection of classes that could be used to
>construct many different types of web applications.
>
>In my day dreams I imagined different layers on top of Albatross being
>used to narrow the focus and raise the functionality for different and
>smaller problem domains.  It is still a day dream though.
>
>Matt> Of course, I may never actually find any time to look into any
>Matt> of these ;-)
>
>If only people would pay us money to play with this stuff all day :-)
>
Yeah, well that's why I've gone a bit quiter on the list this last week 
- I did too much "playing" last week. Python is often like playing 
anyway though :).

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenation.net



From djc at object-craft.com.au  Thu Jul 17 01:34:32 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 17 Jul 2003 11:34:32 +1000
Subject: [albatross-users] Levelling the "random vs non-random" field
In-Reply-To: <3F154B20.2020203@pollenation.net>
References: <3F152347.1060002@pollenation.net>
 <m3znjewxhw.fsf@ferret.object-craft.com.au>
 <3F154B20.2020203@pollenation.net>
Message-ID: <m3smp6ndg7.fsf@ferret.object-craft.com.au>

>>>>> "Matt" == Matt Goodall <matt@pollenation.net> writes:

Matt> Hmm, just a random thought, but would it be useful/possible to
Matt> take the page_enter/leave concept and make pages with persistent
Matt> state? I'm thinking along the lines of pickling the
Matt> non-transient properties of a page instance to the session
Matt> rather than mucking about with add_session_vars. Woah, best stop
Matt> this train of thought now or I'll want to play even more ;-).

You could get a poor mans version of the same thing by just placing
all persistent information about each page in a single object.  That
single object could then be placed into the session.

Some funny untested code:

    class MyCtx(albatross_ctx):
        def set_page_state(self, name):
            # Create a page state dictionary in the session if it does
            # not already exist
            if not hasattr(self.locals, '_page_state'):
                self.locals._page_state = {}
                self.add_session_vars('_page_state')

            # Find the state for the current page
            state = self.locals._page_state.get(name)
            if not state:
                state = Vars()
                self.locals._page_state[name] = state

            # Place current page state information in _state for
            # template access
            self.locals._state = state
            return state

Then in your page module or whatever...

    def page_enter(ctx):
        ctx.set_page_state('blah')

    def page_process(ctx):
        ctx.locals._state.something = 'or other'

    def page_display(ctx):
        ctx.run_template('blah.html')

The in blah.html...

    <al-value expr="_state.something">

>>  That would be cool to have.  I spent a little time looking at
>> Twisted when I had to build a forking server.  I could not get
>> Twisted to do what I wanted and did not have the time to figure it
>> out.
>> 
Matt> Twisted doesn't fork, could that be why you had trouble? ;-). At
Matt> least, I don't think it forks?!

I used the wrong term.  What I meant to say was that the server had to
talk a TCP protocol then fork and monitor child processes.

Matt> I think scgi forks. I was wondering about that as yet another
Matt> request adapter but its lacking that "itch to scratch" quality
Matt> for me now that there is fastcgi support.

Matt> Perhaps a FeatureRequest page to go alongside the
Matt> ProposedChanges page on the wiki would be a good way of
Matt> gathering ideas for future versions? That might also help keep
Matt> track of things as people wander in and out of the project, as
Matt> invariably happens with open source.

Good idea.

- Dave

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Thu Jul 17 09:55:52 2003
From: matt at pollenation.net (Matt Goodall)
Date: Thu, 17 Jul 2003 10:55:52 +0100
Subject: [albatross-users] Levelling the "random vs non-random" field
In-Reply-To: <m3smp6ndg7.fsf@ferret.object-craft.com.au>
References: <3F152347.1060002@pollenation.net>	<m3znjewxhw.fsf@ferret.object-craft.com.au>	<3F154B20.2020203@pollenation.net> <m3smp6ndg7.fsf@ferret.object-craft.com.au>
Message-ID: <3F1672A8.7050103@pollenation.net>

Dave Cole wrote:

>>>>>>"Matt" == Matt Goodall <matt@pollenation.net> writes:
>>>>>>            
>>>>>>
>Perhaps a FeatureRequest page to go alongside the
>Matt> ProposedChanges page on the wiki would be a good way of
>Matt> gathering ideas for future versions? That might also help keep
>Matt> track of things as people wander in and out of the project, as
>Matt> invariably happens with open source.
>
>Good idea.
>
Done. It only contains the ideas from this thread for now. I'm sure 
there are other ideas floating around out there though.

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: matt@pollenation.net



From djc at object-craft.com.au  Sun Jul 20 10:58:24 2003
From: djc at object-craft.com.au (Dave Cole)
Date: 20 Jul 2003 20:58:24 +1000
Subject: [albatross-users] Albatross 1.10 released
Message-ID: <m3r84ltqgf.fsf@ferret.object-craft.com.au>

OVERVIEW

Albatross is a small toolkit for developing highly stateful web
applications.

The toolkit has been designed to take a lot of the pain out of
constructing intranet applications although you can also use Albatross
for deploying publicly accessed web applications.

In slightly less than 3000 lines of Python (according to pycount) you
get the following:

  * An extensible HTML templating system similar to DTML including tags for:
    - Conditional processing.
    - Macro definition and expansion.
    - Sequence iteration and pagination.
    - Tree browsing.
    - Lookup tables to translate Python values to arbitrary template
      text.
  * Application classes which offer the following features:
    - Optional server side or browser side sessions.
    - The ability to place Python code for each page in a dynamically
      loaded module, or to place all page processing code in a single
      mainline.
  * The ability to deploy applications either as CGI or via mod_python
    by changing less than 10 lines of code.

The toolkit application functionality is defined by a collection of
fine grained mixin classes.  Nine different application types and five
different execution contexts are prepackaged, you are able to define
your own drop in replacements for any of the mixins to alter any
aspect of the toolkit semantics.

Application deployment is controlled by your choice of either cgi,
FastCGI, mod_python, or BaseHTTPServer Request class.  It should be
possible to develop a Request class for Medusa or Twisted to allow
applications to be deployed on those platforms with minimal changes.

Albatross comes with over 160 pages of documentation.  HTML, PDF and
PostScript formatted documentation is available from the toolkit
homepage.

The toolkit homepage:

        http://www.object-craft.com.au/projects/albatross/

The Albatross mailing list subscription and archives:

	http://object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users


CHANGES SINCE 1.01

There have been many improvements and bug fixes since release 1.01.
The following page describes the changes in detail.

    http://www.object-craft.com.au/projects/albatross/albatross/rel-1.10.html

(There were no changes to 1.10pre4)

-- 
http://www.object-craft.com.au


From matt at pollenation.net  Sun Jul 20 15:44:04 2003
From: matt at pollenation.net (Matt Goodall)
Date: 20 Jul 2003 16:44:04 +0100
Subject: [albatross-users] Albatross 1.10 released
In-Reply-To: <m3r84ltqgf.fsf@ferret.object-craft.com.au>
References: <m3r84ltqgf.fsf@ferret.object-craft.com.au>
Message-ID: <1058715843.14480.74.camel@localhost>

Dave, thanks for all the time you put into the release at weekends ...
but don't assume that this means I won't be sending you anymore patches!
;-)

Cheers, Matt

-- 
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: matt@pollenation.net

From gnb at itga.com.au  Sun Jul 20 23:33:24 2003
From: gnb at itga.com.au (Gregory Bond)
Date: Mon, 21 Jul 2003 09:33:24 +1000
Subject: [albatross-users] Albatross 1.10 released
In-Reply-To: Your message of 20 Jul 2003 20:58:24 +1000.
Message-ID: <200307202333.JAA17216@lightning.itga.com.au>

This is a multipart MIME message.

--==_Exmh_14668846380
Content-Type: text/plain; charset=us-ascii

I've sent a PR to update the FreeBSD port (pr ports/54680).  Or you can use 
the attached ports framework.

To use it:  
 - cd /usr/port (or somewhere handy)
 - run the attached file through "sh"
 - cd www/py-albatross && make 
 - "make install" as root




--==_Exmh_14668846380
Content-Type: application/x-shar ; name="albatross.shar"
Content-Description: albatross.shar
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="albatross.shar"
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