From andrewm at object-craft.com.au  Mon May  2 10:37:21 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Mon, 02 May 2005 10:37:21 +1000
Subject: [albatross-users] pickle problem 
In-Reply-To: <20050429151532.GD3466@odin.wolfheart.ro> 
References: <20050429151532.GD3466@odin.wolfheart.ro>
Message-ID: <20050502003721.962777940D9@longblack.object-craft.com.au>

>ApplicationError: locals "user": Can't pickle <class 'Base.User'>: it's not the same object as Base.User
>
>
>Any idea what the problem is? user of type Base.User is kept in the
>session. This happens sometimes and sometimes it works ok.

You might do better asking on python-list at python.org - this error is
coming out of cPickle, rather than Albatross (albatross has simply caught
the error and re-raised it to add more diagnostics).

If I had to guess, I would say that the pickler is unable to identify
the code object associated with that class. When you pickle an instance
of a class, only the instance attributes and a reference (the module
and class name) to the class are saved.

You could try renaming your "User" class or the "Base" module temporarily - 
this will tell you if the pickler is stumbling across the wrong module.

If you're playing tricks with the class definition at run-time, this will
also confound the pickler, I think (eg, metaclasses), or maybe changing the
instance's class at run-time.

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


From andrewm at object-craft.com.au  Mon May  2 13:10:10 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Mon, 02 May 2005 13:10:10 +1000
Subject: [albatross-users] pickle problem 
In-Reply-To: <20050502003721.962777940D9@longblack.object-craft.com.au> 
References: <20050429151532.GD3466@odin.wolfheart.ro> <20050502003721.962777940D9@longblack.object-craft.com.au>
Message-ID: <20050502031010.519ED7940D9@longblack.object-craft.com.au>

>You might do better asking on python-list at python.org - this error is
>coming out of cPickle, rather than Albatross (albatross has simply caught
>the error and re-raised it to add more diagnostics).
>
>If I had to guess, I would say that the pickler is unable to identify
>the code object associated with that class. When you pickle an instance
>of a class, only the instance attributes and a reference (the module
>and class name) to the class are saved.

Looking at the pickle source code, this error is raised when the object
can't be found by name in sys.modules. In other words (and simplified
slightly):

    obj != getattr(sys.modules[obj.__module__], obj.__name__)

So, I guess that means either:

 * the object lies about it's name or module,
 * the object has been obscured by a later binding with the same name, or,
 * the module has been obscured by something playing tricks with sys.modules.

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


From djc at object-craft.com.au  Mon May  2 14:38:40 2005
From: djc at object-craft.com.au (Dave Cole)
Date: Mon, 02 May 2005 14:38:40 +1000
Subject: [albatross-users] pickle problem
In-Reply-To: <20050502003721.962777940D9@longblack.object-craft.com.au>
References: <20050429151532.GD3466@odin.wolfheart.ro> <20050502003721.962777940D9@longblack.object-craft.com.au>
Message-ID: <4275AED0.9080309@object-craft.com.au>

Andrew McNamara wrote:
>>ApplicationError: locals "user": Can't pickle <class 'Base.User'>: it's not the same object as Base.User
>>
>>
>>Any idea what the problem is? user of type Base.User is kept in the
>>session. This happens sometimes and sometimes it works ok.
> 
> 
> You might do better asking on python-list at python.org - this error is
> coming out of cPickle, rather than Albatross (albatross has simply caught
> the error and re-raised it to add more diagnostics).
> 
> If I had to guess, I would say that the pickler is unable to identify
> the code object associated with that class. When you pickle an instance
> of a class, only the instance attributes and a reference (the module
> and class name) to the class are saved.
> 
> You could try renaming your "User" class or the "Base" module temporarily - 
> this will tell you if the pickler is stumbling across the wrong module.
> 
> If you're playing tricks with the class definition at run-time, this will
> also confound the pickler, I think (eg, metaclasses), or maybe changing the
> instance's class at run-time.

Another way to make it happen bit me a long time ago.

If you reload a module then try to pickle an object created by the old 
module before it was reloaded, then the pickle will fail.

I reported this as a Python bug, but Guido said that he does not 
consider it to be a bug.

http://sourceforge.net/tracker/index.php?func=detail&aid=451547&group_id=5470&atid=105470

- Dave

djc at echidna:~$ python
Python 2.3.4 (#2, Jan  5 2005, 08:24:51)
[GCC 3.3.5 (Debian 1:3.3.5-5)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
 >>> import pickle, copy
 >>> o = copy._EmptyClass()
 >>> pickle.dumps(o, 1)
'(ccopy\n_EmptyClass\nq\x00oq\x01}q\x02b.'
 >>> reload(copy)
<module 'copy' from '/usr/lib/python2.3/copy.pyc'>
 >>> pickle.dumps(o, 1)
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/lib/python2.3/pickle.py", line 1386, in dumps
     Pickler(file, protocol, bin).dump(obj)
   File "/usr/lib/python2.3/pickle.py", line 231, in dump
     self.save(obj)
   File "/usr/lib/python2.3/pickle.py", line 293, in save
     f(self, obj) # Call unbound method with explicit self
   File "/usr/lib/python2.3/pickle.py", line 721, in save_inst
     save(cls)
   File "/usr/lib/python2.3/pickle.py", line 293, in save
     f(self, obj) # Call unbound method with explicit self
   File "/usr/lib/python2.3/pickle.py", line 765, in save_global
     raise PicklingError(
pickle.PicklingError: Can't pickle <class copy._EmptyClass at 
0x401fb4ac>: it's not the same object as copy._EmptyClass
 >>>

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


From andrewm at object-craft.com.au  Mon May  2 15:25:08 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Mon, 02 May 2005 15:25:08 +1000
Subject: [albatross-users] pickle problem 
In-Reply-To: <4275AED0.9080309@object-craft.com.au> 
References: <20050429151532.GD3466@odin.wolfheart.ro> <20050502003721.962777940D9@longblack.object-craft.com.au> <4275AED0.9080309@object-craft.com.au>
Message-ID: <20050502052508.676FF7940D9@longblack.object-craft.com.au>

>> If you're playing tricks with the class definition at run-time, this will
>> also confound the pickler, I think (eg, metaclasses), or maybe changing the
>> instance's class at run-time.
>
>Another way to make it happen bit me a long time ago.
>
>If you reload a module then try to pickle an object created by the old 
>module before it was reloaded, then the pickle will fail.

Yes, this is one facet of what I had in mind when I said "changing the
instance's class at run-time". Sorry for not expressing it more clearly.

>I reported this as a Python bug, but Guido said that he does not 
>consider it to be a bug.
>
>http://sourceforge.net/tracker/index.php?func=detail&aid=451547&group_id=5470&atid=105470

The reload() documentation says:

    If a module instantiates instances of a class, reloading the module
    that defines the class does not affect the method definitions of the
    instances -- they continue to use the old class definition. The same
    is true for derived classes.

Unfortunately, you only look at this after wasting a day finding your
pickle bug! I'm sure if someone submitted a patch to the pickle doco...

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


From andrewm at object-craft.com.au  Thu May  5 12:51:23 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Thu, 05 May 2005 12:51:23 +1000
Subject: [albatross-users] XML and Albatross templates?
Message-ID: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>

As people are probably aware, some aspects of Albatross templating prevent
XML validation of the source - specifically the <al-else> tag:

    <al-if expr="..."> ...  <al-else> ...  </al-if>

There's no way to balance that <al-else> (normally empty tags would be
written <al-else/>, but that doesn't reflect what's happening here).

One option is to deprecate <al-if> (or <al-else>, at least), favouring an
inline version of <al-lookup> - for example:

    <al-lookup expr="...">
        <al-item expr="...."> abc </al-item>
        <al-item expr="...."> def </al-item>
        ghi
    </al-lookup>

This is quite a bit more verbose, but many of the cases where <al-if>
would formerly have been used are now better written using the "anytag"
structure. "Anytag" (which appears in the development prerelease [1])
allows any tag to be prefixed with "al-". When that is done, attributes
can selectively be the product of a python expression. For example:

    <table>
        <al-for iter="i" expr="stuff">
            <al-tr al-class="i.value().get_style()">
                    ....
            </al-tr>
        </al-for>
    </table>

Note that this also means that albatross tags within html tags are no
longer necessary. So this horrible contruct should be a thing of the past:

    <td colspan="<al-value expr='n.span()'>">

This is now written:

    <al-td al-colspan="n.span()">

Thoughts on the al-else thing? Why is XML validation important? At
some point we may wish to switch to an XML-based parser (rather than
the current regexp-based parser) - correct XML structure would make
this easier.

[1] http://www.object-craft.com.au/projects/albatross/download/albatross-1.30-dev-20050301.tar.gz

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


From gnb at itga.com.au  Thu May  5 13:04:16 2005
From: gnb at itga.com.au (Gregory Bond)
Date: Thu, 05 May 2005 13:04:16 +1000
Subject: [albatross-users] XML and Albatross templates?
In-Reply-To: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
Message-ID: <42798D30.1070304@itga.com.au>

Andrew McNamara wrote:

>Thoughts on the al-else thing? Why is XML validation important? 
>

Hmm... like the Anytag stuff a lot, would save a lot of ugly hackery.

The al-else... will take a little bit bit of conversion, but _WELL_ 
worth the effort, beacuse it should mean the Emacs html mode with do the 
&@%$*# indentation right!



From fred at polgardy.com  Thu May  5 13:28:59 2005
From: fred at polgardy.com (Frederick Polgardy Jr)
Date: Wed, 4 May 2005 22:28:59 -0500
Subject: [albatross-users] XML and Albatross templates?
In-Reply-To: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
Message-ID: <200505042229.00254.fred@polgardy.com>

On Wednesday 04 May 2005 21:51, Andrew McNamara wrote:
> As people are probably aware, some aspects of Albatross templating
> prevent XML validation of the source - specifically the <al-else> tag:
>
>     <al-if expr="..."> ...  <al-else> ...  </al-if>
>
> There's no way to balance that <al-else> (normally empty tags would be
> written <al-else/>, but that doesn't reflect what's happening here).

I forget how it goes off the top of my head, but investigate the Jakarta 
Struts logic tag library.  I know there is an if/then/else type tag 
structure, and I'm relatively certain it's valid XML.

The "anytag" scheme looks sweet, count me in.


From andrewm at object-craft.com.au  Thu May  5 14:03:31 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Thu, 05 May 2005 14:03:31 +1000
Subject: [albatross-users] XML and Albatross templates? 
In-Reply-To: <200505042229.00254.fred@polgardy.com> 
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au> <200505042229.00254.fred@polgardy.com>
Message-ID: <20050505040332.0EA367940F0@longblack.object-craft.com.au>

>I forget how it goes off the top of my head, but investigate the Jakarta 
>Struts logic tag library.  I know there is an if/then/else type tag 
>structure, and I'm relatively certain it's valid XML.

This sort of thing?

  <logic:if name="calendarBean" property="month" op="equal" value="January" >
    <logic:or name="calendarBean" property="year" op="LessEqual" value="1999"/>
    <logic:or name="calendarBean" property="day" op="equal" value="Monday"/>

    <logic:then>
      At least one of the above conditions is true.
    </logic:then>

    <logic:else>
      All of Above conditions are false
    </logic:else>
  </logic:if>

I suspect something like the following is closer to the spirit of the
current Albatross code:

  <al-if expr="...">
    abc
    <al-elif expr="...">
      def
    </al-elif>
    <al-elif expr="...">
      ghi
    </al-elif>
    <al-else>
      jkl
    </al-else>
  </al-if>

But I don't see that as any sort of improvement over the <al-lookup>
suggestion. And the parser currently only has the concept of empty or
containing elements - this would require (for backward compatability)
support for containing elements with optional close tag (as HTML has).

>The "anytag" scheme looks sweet, count me in.

It's there now in the dev snapshot - feedback, gotchyas, etc gratefully
received.

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


From andrewm at object-craft.com.au  Thu May  5 14:07:30 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Thu, 05 May 2005 14:07:30 +1000
Subject: [albatross-users] XML and Albatross templates? 
In-Reply-To: <42798D30.1070304@itga.com.au> 
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au> <42798D30.1070304@itga.com.au>
Message-ID: <20050505040730.171B17940F0@longblack.object-craft.com.au>

>The al-else... will take a little bit bit of conversion, 

At least for the medium-term, al-else would be retained - it would only
need to be dropped if we moved to an xml parser.

>but _WELL_ worth the effort, beacuse it should mean the Emacs html mode
>with do the &@%$*# indentation right!

Very true. 

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


From gabriel.cooper at mediapulse.com  Fri May  6 00:27:38 2005
From: gabriel.cooper at mediapulse.com (Gabriel Cooper)
Date: Thu, 05 May 2005 10:27:38 -0400
Subject: [albatross-users] XML and Albatross templates?
In-Reply-To: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
Message-ID: <427A2D5A.2040906@mediapulse.com>


Andrew McNamara wrote:

>Note that this also means that albatross tags within html tags are no
>longer necessary. So this horrible contruct should be a thing of the past:
>
>    <td colspan="<al-value expr='n.span()'>">
>
>This is now written:
>
>    <al-td al-colspan="n.span()">
>  
>
We implemented the al-xxx attribute change in Snakeskin a several months 
ago and really love it. Since then we've discussed making the "anytag" 
tag several times, our only roadblocks having been setting aside the 
time to do it! Once that's implemented it will completely change what 
template code looks like. /Much/ cleaner code, etc. Definitely implement 
this change! :)

Gabriel Cooper.


From miguel at yorku.ca  Fri May  6 00:28:43 2005
From: miguel at yorku.ca (Miguel Marques)
Date: Thu, 05 May 2005 10:28:43 -0400 (EDT)
Subject: [albatross-users] XML and Albatross templates? 
In-Reply-To: <20050505040332.0EA367940F0@longblack.object-craft.com.au>
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
	<200505042229.00254.fred@polgardy.com>
	<20050505040332.0EA367940F0@longblack.object-craft.com.au>
Message-ID: <20050505.102843.112610619.miguel@yorku.ca>

Greetings...
On Thu, 05 May 2005 14:03:31 +1000, Andrew McNamara <andrewm at object-craft.com.au> wrote:
> I suspect something like the following is closer to the spirit of the
> current Albatross code:
> 
>   <al-if expr="...">
>     abc
>     <al-elif expr="...">
>       def
>     </al-elif>
>     <al-elif expr="...">
>       ghi
>     </al-elif>
>     <al-else>
>       jkl
>     </al-else>
>   </al-if>
> 
> But I don't see that as any sort of improvement over the <al-lookup>
I think the improvement is that it  makes it much clearer that it's a
if/elif/else construct than the al-lookup one does.  
I often have web designers without a lot
of programming experience build/edit pages and I think the above would
be preferable.
IMHO.

							Miguel

C. Miguel Marques, Assistant Manager, Software Systems Developer
Development Services, Computing & Network Services 
York University, Toronto, Ontario, Canada
e-mail: miguel at yorku.ca, voice: (416)736-2100x22684


From fred at polgardy.com  Fri May  6 00:38:31 2005
From: fred at polgardy.com (Frederick Polgardy)
Date: Thu, 05 May 2005 09:38:31 -0500
Subject: [albatross-users] XML and Albatross templates?
In-Reply-To: <20050505.102843.112610619.miguel@yorku.ca>
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>	<200505042229.00254.fred@polgardy.com>	<20050505040332.0EA367940F0@longblack.object-craft.com.au> <20050505.102843.112610619.miguel@yorku.ca>
Message-ID: <427A2FE7.10905@polgardy.com>

Agreed.  In fact, I've always avoided the lookup scheme, because I 
haven't had the time to grok exactly how it works and what it buys you 
(maybe I've just not seen any really good examples).

My only objection to the suggestion quoted below is that the al-if is at 
a different level of nesting than the other elements.  The only thing I 
could come up with off the top of my head would be something 
structurally more like:

<cond>
    <if test="">
    </if>

    <if test="">
    </if>

    <else>
    </else>
</cond>

But I'm not sold on it, and any change like that is bound to break 
existing stuff.

Miguel Marques wrote:

>Greetings...
>On Thu, 05 May 2005 14:03:31 +1000, Andrew McNamara <andrewm at object-craft.com.au> wrote:
>  
>
>>I suspect something like the following is closer to the spirit of the
>>current Albatross code:
>>
>>  <al-if expr="...">
>>    abc
>>    <al-elif expr="...">
>>      def
>>    </al-elif>
>>    <al-elif expr="...">
>>      ghi
>>    </al-elif>
>>    <al-else>
>>      jkl
>>    </al-else>
>>  </al-if>
>>
>>But I don't see that as any sort of improvement over the <al-lookup>
>>    
>>
>I think the improvement is that it  makes it much clearer that it's a
>if/elif/else construct than the al-lookup one does.  
>I often have web designers without a lot
>of programming experience build/edit pages and I think the above would
>be preferable.
>IMHO.
>
>							Miguel
>



From andrewm at object-craft.com.au  Mon May  9 16:42:25 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Mon, 09 May 2005 16:42:25 +1000
Subject: [albatross-users] XML and Albatross templates? 
In-Reply-To: <20050505025123.F2CA27940F0@longblack.object-craft.com.au> 
References: <20050505025123.F2CA27940F0@longblack.object-craft.com.au>
Message-ID: <20050509064225.674757940F0@longblack.object-craft.com.au>

>Note that this also means that albatross tags within html tags are no
>longer necessary. So this horrible contruct should be a thing of the past:
>
>    <td colspan="<al-value expr='n.span()'>">
>
>This is now written:
>
>    <al-td al-colspan="n.span()">

I've just checked the implementation, and my example isn't correct. It
should read (note the changed attribute name):

    <al-td colspanexpr="n.span()">

In this case, the attribute is evaluated, and the result substitued, eg:

    <td colspan="3">

There is a second notation for boolean attributes:

    <al-input name="abc.value" disabledbool="abc.isdisabled()" />

which will evaluate the expression and emit the attribute if result
is true:

    <input name="abc.value" disabled />

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


From korg at darkqueen.org  Fri May 13 12:54:12 2005
From: korg at darkqueen.org (Cameron Blackwood)
Date: Fri, 13 May 2005 12:54:12 +1000
Subject: [albatross-users] Back into albatross... again... :)
Message-ID: <20050513025412.9B50A5269A@firewall.darkqueen.org>


Hi,


I tend to drift into and out of albatross coding, and Im currently
on the way 'back in' :) and I have a few statements and some
questions... :)

Hate though I do to sound like Troy McClure, but you might remember me
from such classic albatros discussions such as 'needing expr="foo(x,y)"
in <al-input type="image">'. :) :)

After a big pause in alb development, Im back :) and Ive
forgotten everything that I previously learnt. :(

So, excuse the length and amount of rambling in this post... and feel
free to comment or ignore as you see fit...


'header' applications.
----------------------
I have this weird idea that I was thinking of, but Ive hit a problem.

Lets say Im dreaming of a number of 'game' applications with a single
user/login/tracking stage. So the flow of use would be:

  1 login page
  2 user's page 
    - choice to join or start a game from a list of different ones
      (eg, tic tak toe, chess, go etc)
    - list of currently 'running' games
      (eg, turn 3 of chess game 'fred vrs jane', 
           ttt turn 2 of 'jane vrs -computer-', 
           turn 8 of chess 'jane vrs netfoo' )
  3 user selects a game and goes and plays for a while
    - input is passed to the correct 'game instance' and the
      view pages marked up as need using the templates for that
      specific game type.
  4 user drops back to 2 and logs out or plays another

Assuming the game state is stored in some other server/db and albatross
is just a front end that requests data and marks up the 'view', then
I can see two ways to do this:

  1) make an uber app with all the possible app pages loaded in with
     a truck load of 
       s.register_page( 'GameN_pageM', 'GameN_pageM.html')

     + all inside albatross (all games are just albatross pages)
     - all pages must be loaded at the start :-/ (hard to modify on
       the fly)
     - (Im assuming) performace would suck if it was an apache cgi
       (given that it would have to create page instances for every
       game for every request, which is wastefull if you have 20 games
       with 10 pages each) Hmm. looks like I need to look at al-http

  2) have some kind of 'farmer' which takes the request and delivers
     it to the specific game on the fly.

     - specific games wouldnt be albatross apps (atleast I cant think
       of any easy way to do it?)
     + lean 'albatross' application on front
     + more 'dynamic' (in that a specific game can have trouble or be
       added or removed without changing the pages that the albatross
       application knows about)

I like the idea of having the flexability to add/remove new apps as I
go without needing to take the front end down (#2), but I cant see any
nice way to pass the user info on from the front end 'game header' to
an specific albatross 'game' sub-application.

(Except maybe to have the front end store the auth info in a cookie
and then url redirect the user to the game when they select it, which
seems kinda wrong to me for some reason.) (Although I guess if you
stored them in a cookie and then url-redirected to a specific game URL
then the game could unpack the cookie and store the values in its
session context, thus only needing to check it once.)

Having written this out, Im thinking that the #1 idea seems to be the
better way to go as it keeps it keeps all the 'games' as simple
collections of albatross pages (although I guess Id have to keep all
the templates in the same directory, as you have to pass a template
directory to the albatross application constructor (I havnt tried
setting the s.register_page('gameM.pageN','gameM/pageN.html')).

Hmmm. Opinions welcome.... as is mocking laughter. :)
(Ive mostly just hacked up smaller cgi-apps, never thought of
something this 'big'.)


Dynamic images for al-input?
----------------------------
The reason that I asked my (Feb last year) question about 
<al-input type="image"> with variable image generation is that I
wanted to do something like this:

<al-input type="image" name="click" srcexpr="create_map_image(ox,oy)">

where create_map_image(ox,oy) talks to the game and generates the
image of the map at the offset ox,oy dynamically. You obviously have
scroll buttons to change ox and oy ;). Previously I used url hackery
to do this, ie, src="create_map_image/54/34" would draw the map at
ox=54 oy=53, but that just seems 'wrong'. :)

Any ideas how to do this? or should I look at hacking src/expr/srcexpr
on al-input for images? (Which I think is probably the Right Thing To
Do(tm) :)

I was thinking about using a:
  http://www.object-craft.com.au/projects/albatross/wiki/Macro_20Expand
  1 Expanding Albatross Macros in custom tags
and I think that page could do with an example of the _dtml_ that uses
the tags that it just defined. :)


Albatross learning curve.... & wrappers
---------------------------------------
Actually its not that bad, :) I picked it up again pretty quickly.

Ive ended up making a wrapper for albatross because there are some
things that I keep doing all the time and albatross doesnt help me
with (or maybe I dont know how to do them neatly).

Im wondering if anyone else has 'wrapped' the more unfriendly parts
of albatross or does everyone just use it raw? If anyone else has any
'wrapping' suggestions then please post 'em :).


Ive only just started back into albatross, so there is probably some
stuff that Im missing or that will need to be hacking in later, but number
one is that this hides 'context' from the user, so the user just ends
up with code like this:

  class Login( Albatross_wrapper.base_Page ):

    def work_on_enter(s):
      s.create_global('username')
      s.create_local('password')
  
    def work_process(s):
      login,username,passwd=s['login','username','password']
      s.DEBUG('login [%s][%s]',username,passwd)
      if login and username=='admin' and passwd=='foo':
        s.goto('Game')
      elif s['quick:korg']: # quick loging for testing.. remove later
        s.DEBUG('quick login')
        s['username']='korg'
        s['password']='a.secret'
        s.goto('Game')
      elif s['nuke']: # nuke button?
        s.nuke()    # remove session store

  class GameApp( Albatross_wrapper.base_App ): pass

  def main():
    app=GameApp( 'toplevel.cgi',
                 (Login,Game,Map)
               )
    app.DEBUG('Logging locals %s',__name__)
    app.run_cgi( )


create_local and create_global create variables (duh!). (Locals
are destroyed when you leave the page).

You can __setitem__ (will barf if the variable doesnt exist (not sure if
thats what I want just yet)).

You can __getitem__ a variable or a list of variables. Yay! Easy unpacking
without checking for ctx.local.myvariable then accessing it.
Will return 'None' if the variable doesnt exist (again, unsure if thats
what I want, but *shrugs*).


The base_App works out template directories (by looking for
Albatross_templates in your sys.path, which Im not happy with really
as Id rather just pass full path names to the register_page()
function, which I admit I havnt tried yet... and if that works, then I
wont need to pass a template directory as I can just search for each
page template as I register the page :).

It also names the pages from their class names (ie above, Login is
named as Login and uses a template that matches [Ll]ogin{.html|.dtml}.

Id like the wrapper to be able to run as a number of different
albatross servers types, but Im currently only using:
   from albatross  import SimpleSessionFileApp  as Albatross_Application
   from albatross  import SessionFileAppContext as Albatross_Context

I also want to make it better at being switched between apache cgi and
al-httpd, but I havnt played with al-httpd so Ill do that when I come
to it. (Eg, if you call it it runs al-http, if its called, it works as
an apache cgi.)

I also want to code up an ImagePage that generates page images (with
the dynamic generation above) but I havnt got _that_ far back into it
yet. ;)




Oh and I just like to end by repeating what everyone already knows...
that albatross rocks. :)


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 ----------
True happiness will be found only in true love.


From miguel at yorku.ca  Fri May 27 04:05:04 2005
From: miguel at yorku.ca (Miguel Marques)
Date: Thu, 26 May 2005 14:05:04 -0400 (EDT)
Subject: [albatross-users] Nested tags...
Message-ID: <20050526.140504.71105915.miguel@yorku.ca>

Greetings...

I have the need to do something like the following:

<al-select nameexpr="val1" optionexpr="val2" onChange="updateText('<al-value expr="val1">')" whitespace/>

Unfortunately only the enclosed al-value gets expanded.
Is there a trick I'm not aware of to make this work or is it time for
an alx-select tag?
Thanks in advance for any help...

						Miguel
							
C. Miguel Marques, Assistant Manager, Software Systems Developer
Development Services, Computing & Network Services 
York University, Toronto, Ontario, Canada
e-mail: miguel at yorku.ca, voice: (416)736-2100x22684


From gabriel.cooper at mediapulse.com  Fri May 27 05:38:39 2005
From: gabriel.cooper at mediapulse.com (Gabriel Cooper)
Date: Thu, 26 May 2005 15:38:39 -0400
Subject: [albatross-users] Nested tags...
In-Reply-To: <20050526.140504.71105915.miguel@yorku.ca>
References: <20050526.140504.71105915.miguel@yorku.ca>
Message-ID: <429625BF.7020906@mediapulse.com>


If they've released the changes that were talked about a few weeks back 
you should be able to download that and then do the following:

<al-select nameexpr="val1" optionexpr="val2" onChangeexpr="'updateText(\'' + val1 +'\')'" whitespace/>


If not you could use snakeskin (snakeskin-tools.sf.net) instead to do it 
like so:

<ss-select nameexpr="val1" optionexpr="val2" ss-onChange="'updateText(\'' + val1 +'\')'" whitespace/>


(note that I didn't check syntax on these....)

but. eh... promoting snakeskin here seems horribly subversive so just go 
look for the development version of Albatross. ;)

Gabriel.


Miguel Marques wrote:

>Greetings...
>
>I have the need to do something like the following:
>
><al-select nameexpr="val1" optionexpr="val2" onChange="updateText('<al-value expr="val1">')" whitespace/>
>
>Unfortunately only the enclosed al-value gets expanded.
>Is there a trick I'm not aware of to make this work or is it time for
>an alx-select tag?
>Thanks in advance for any help...
>
>						Miguel
>							
>C. Miguel Marques, Assistant Manager, Software Systems Developer
>Development Services, Computing & Network Services 
>York University, Toronto, Ontario, Canada
>e-mail: miguel at yorku.ca, voice: (416)736-2100x22684
>_______________________________________________
>Albatross-users mailing list
>Albatross-users at object-craft.com.au
>https://www.object-craft.com.au/cgi-bin/mailman/listinfo/albatross-users
>
>  
>



From andrewm at object-craft.com.au  Fri May 27 09:27:50 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Fri, 27 May 2005 09:27:50 +1000
Subject: [albatross-users] Nested tags... 
In-Reply-To: <429625BF.7020906@mediapulse.com> 
References: <20050526.140504.71105915.miguel@yorku.ca>
	<429625BF.7020906@mediapulse.com>
Message-ID: <20050526232750.581BE6F415E@longblack.object-craft.com.au>

>If they've released the changes that were talked about a few weeks back 
>you should be able to download that and then do the following:
>
><al-select nameexpr="val1" optionexpr="val2" onChangeexpr="'updateText(\'' + val1 +'\')'" whitespace/>

Yes, the development snapshot has this functionality, and this is the
way I would recommend getting the behaviour you want.

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


From miguel at yorku.ca  Fri May 27 22:37:31 2005
From: miguel at yorku.ca (Miguel Marques)
Date: Fri, 27 May 2005 08:37:31 -0400 (EDT)
Subject: [albatross-users] Nested tags... 
In-Reply-To: <20050526232750.581BE6F415E@longblack.object-craft.com.au>
References: <20050526.140504.71105915.miguel@yorku.ca>
	<429625BF.7020906@mediapulse.com>
	<20050526232750.581BE6F415E@longblack.object-craft.com.au>
Message-ID: <20050527.083731.74730425.miguel@yorku.ca>

Greetings...
On Fri, 27 May 2005 09:27:50 +1000, Andrew McNamara <andrewm at object-craft.com.au> wrote:
> >If they've released the changes that were talked about a few weeks back 
> >you should be able to download that and then do the following:
> >
> ><al-select nameexpr="val1" optionexpr="val2" onChangeexpr="'updateText(\'' + val1 +'\')'" whitespace/>
> 
> Yes, the development snapshot has this functionality, and this is the
> way I would recommend getting the behaviour you want.
> 
I have been hoping for a new release to come out because there's
several changes in it that would greatly simplify code I'm writing.
Including the one above.
I am, however, a bit weary about using a development snapshot to deploy
applications.  Is it mostly stable at this point?

							Miguel

C. Miguel Marques, Assistant Manager, Software Systems Developer
Development Services, Computing & Network Services 
York University, Toronto, Ontario, Canada
e-mail: miguel at yorku.ca, voice: (416)736-2100x22684



From tchur at optushome.com.au  Sat May 28 00:11:45 2005
From: tchur at optushome.com.au (Tim Churches)
Date: Sat, 28 May 2005 00:11:45 +1000
Subject: [albatross-users] Nested tags...
In-Reply-To: <20050527.083731.74730425.miguel@yorku.ca>
References: <20050526.140504.71105915.miguel@yorku.ca>	<429625BF.7020906@mediapulse.com>	<20050526232750.581BE6F415E@longblack.object-craft.com.au>
	<20050527.083731.74730425.miguel@yorku.ca>
Message-ID: <42972AA1.7030307@optushome.com.au>

Miguel Marques wrote:
> Greetings...
> On Fri, 27 May 2005 09:27:50 +1000, Andrew McNamara <andrewm at object-craft.com.au> wrote:
> 
>>>If they've released the changes that were talked about a few weeks back 
>>>you should be able to download that and then do the following:
>>>
>>><al-select nameexpr="val1" optionexpr="val2" onChangeexpr="'updateText(\'' + val1 +'\')'" whitespace/>
>>
>>Yes, the development snapshot has this functionality, and this is the
>>way I would recommend getting the behaviour you want.
>>
> 
> I have been hoping for a new release to come out because there's
> several changes in it that would greatly simplify code I'm writing.
> Including the one above.
> I am, however, a bit weary about using a development snapshot to deploy
> applications.  Is it mostly stable at this point?

Yes, it is fine for general use, but I do wish the OC guys would get
around to renaming it, because you are not the only user who is wary of it.

Tim C


From andrewm at object-craft.com.au  Tue May 31 17:40:50 2005
From: andrewm at object-craft.com.au (Andrew McNamara)
Date: Tue, 31 May 2005 17:40:50 +1000
Subject: [albatross-users] Albatross 1.30 released
Message-ID: <20050531074050.B44F57940DC@longblack.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 more than 4500 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 as CGI, FastCGI, mod_python or
    a pure python HTTP server 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 six
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 170 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 1.20

  * Any HTML tag can now be prefixed with "al-" allowing any attribute to be
    the result of python evaluation. For example:

      <al-td colspanexpr="i.span()"> could produce <td colspan="3">

    and:

      <al-input name="abc.value" disabledbool="abc.isdisabled()">
      could produce <input name="abc.value" disabled>

  * Since macros and lookups are an application global resource, they can
    only be defined once per application, however this was not previously
    checked. Redefinition of macros or lookups will now result in an
    ApplicationError exception.

  * an in-line version of the <al-lookup> tag has been introduced, which is
    expanded in place if the tag has an expr= attribute.

  * a new <al-require> tag has been added to allow templates to assert that
    specific Albatross features are available, or templating scheme version
    number is high enough.

  * "Cache-Control: no-cache" is now set in addition to "Pragma: no-cache"
    - the former is defined for HTTP/1.1, the later for HTTP/1.0.

  * Simplified session cookie handling.

BUGFIXES SINCE 1.20:

  * FastCGI apps were not being explicitly finalised, relying
    instead on their object destructor, with the result that writing
    application output (or errors) would be indefinitely deferred if
    object cycles existed. We now call "fcgi.Finish()" from the fcgiapp
    Request.return_code() method.

  * When handling exceptions, the traceback is now explicitly deleted
    from the local namespace to prevent cycles (otherwise the garbage
    collection of other objects in the local namespace will be delayed).

  * Two fixes to the <al-select> tag: the albatross-specific "list"
    attribute was leaking into resulting markup, and the use of the "expr"
    attribute would result in invalid markup being emitted.

  * Thanks to Robert Fendt for picking this up: the Albatross-generated
    hidden field input element must not appear naked inside a form
    element for strict HTML conformance.  The solution is to wrap the
    input elements in div.

  * BranchingSession sessions could not be "deleted" - the solution is
    to add a dummy "session" shared by all branches, which is deleted
    when one branch "logs out".

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