lemon-webmodules Mailing List for Lemon - Python Web Modules

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Hi Duarte,

X_jnduarte schrieb:
> Reading the WebModules docs aparently all to make are include a line with
> AddHandler /usr/bin/python .py in "my /etc/apache/httpd.conf".
>  I'm made this and restarted apache. But it don't work!
> When I acess my http://localhost/hello.py the system don't execute the 
> example and prompts for the program to use for open the file "hello.py".

First of all I think, your problem could be better described as
"Python CGI scripts do not work with Apache".

I did not use the AddHandler directive but:
ScriptAlias /cgi-bin/ "/path/to/cgi-bin/"

Please have a look at the apache log files (error.log and suexec.log)
as they will give you valuable hints.

-- 
Felix




Hi, I'm a experienced programmer in a Desktop environment (C++, Python, 
Java), but I a newbie in WEB  programming.
So I decided make my first incursion in WEB  programming with Web 
Modules and Python.
I've a Apache 1.3 installed and running in a Debian Box (I use 
PHPMyAdmin and others PHP based utils).
My Python is 2.35 version installed from deb packages, using Synaptic 
and runs fine (I develop with WxDesigner,

Reading the WebModules docs aparently all to make are include a line with
AddHandler /usr/bin/python .py in "my /etc/apache/httpd.conf".
 I'm made this and restarted apache. But it don't work!
When I acess my http://localhost/hello.py the system don't execute the 
example and prompts for the program to use for open the file "hello.py".

I am made some variations how:
AddHandler .py /usr/bin/python
AddHandler py /usr/bin/python
AddHandler /usr/bin/python py
No success...

Can someone help me?

Greetings
Duarte
dua...@mi...




Hi Felix,

Usually if you needed to do this you would set your session expire 
parameter fairly large (eg a day) and then if the application itself 
needed to keep accurate track of when it needed to expire, it would 
store an expire variable in the session store and behave accordingly. 
You wouldn't use the session expire time so directly to affect 
application behavior.

The way you would normally write a pythonweb.org application is to use 
the web.auth module which has an expire parameter which is completely 
independent of the web.session expire parameter as long as the session 
doesn't expire. This itself is implemented using the suggestion above of 
storing an expire variable in a session store as a session variable. Are 
you not using web.auth?

Perhaps we should think about how to implement this as a feature by 
setting multiple cookies or multiple session ids in the same cookie for 
different applications but at the moment the only other way I can think 
you could do it if you don't want to use the method suggested above is 
to use two different sets of database tables for the two applications 
and implement two session managers in the code. There is a parameter 
wher you can name a database table prepend. I believe the two session 
managers would behave completely independently unless the extremely 
unlikely event of the two managers randomly creating the same session ID 
occurs.

If none of the ideas meets your requirements I'd be happy to discuss an 
extension to the API?

James



Felix Schwarz wrote:

>Hi all,
>
>I'm trying to have several applications which share the same session
>id. While looking a it I found some problems in the 0.5.3 API of
>web.session.
>
>For example I have an application A which is an admin interface with
>quite high security considerations and an application B which is only
>used by normal users.
>
>I want to have a short session expiry time for application A (10
>minutes of inactivity) but a longer one for application B (more than
>30 minutes). Both applications are using separate stores as suggested
>by the docs. I don't see how this can be implemented in the current
>API.
>
>The solution for this would be that every store has an expiry time and
>one checks if a store is still valid. How do you solve that problem?
>
>  
>




Hi all,

I'm trying to have several applications which share the same session
id. While looking a it I found some problems in the 0.5.3 API of
web.session.

For example I have an application A which is an admin interface with
quite high security considerations and an application B which is only
used by normal users.

I want to have a short session expiry time for application A (10
minutes of inactivity) but a longer one for application B (more than
30 minutes). Both applications are using separate stores as suggested
by the docs. I don't see how this can be implemented in the current
API.

The solution for this would be that every store has an expiry time and
one checks if a store is still valid. How do you solve that problem?

-- 
Felix




It works by importing it I'm afraid (using teud which uses pydoc). I was
really suprised how quick it was to write.. took me about 5 hours so I
can't really understand why docutils doesn't have something similar.

My plan is to just use it until docutils has something similar and then
because the API componets are quite separate, I can then adapt it to use
the finalised docutils code. docutils doesn't seem to generate XML either,
just psuedo-xml. If it generated XML then I could drop the pydoc and teud
parts of pigeon and just use the docutils XML format which would be a bit
cleaner. You can also write a new XML parser class very easily to make the
output look completely different, or indeed generate a different output
format altogether or use an XSLT transform if you prefer and have an XSLT
processer installed.

I've still got to document it properly, add an index in the bottom right
frame (when you load the index.html file generated) and document code that
is inherited or overrided as specifically being that but then it is
finished so I guess if the docutils people are interested in it then it
would be great to include it as part of docutils.

The thing that excites me about it is that I can integrate it into the
eggs format so that when a module is installed it can also be
automatically documented. I can also use it to document eggs, bricks
applications and other things rather than just modules alone.. all quite
exciting!

James


> James Gardner wrote:
> It sounds *very* interesting. There isn't a working part of docutils to
> extract docstrings from code and generate API documentation. Maybe this
> is it ;-)
>
> By the way - does it work by importing the code, or by compiling it ?
>
> Regards,
>
> Fuzzy
> http://www.voidspace.org.uk/python
>
>
>> James
>>
>>
>>
>> Michael Foord wrote:
>>
>>> Sounds good James.
>>>
>>> To save me having to look *too far* ;-) - I assume pigeon uses
>>> docstrings ? Can you use reStructured Text format, or have you created
>>> your own markup convention ?
>>>
>>> Regards,
>>>
>>> Fuzzy
>>> http://www.voidspace.org.uk/python
>>> -----Original Message-----
>>>   >From: "James Gardner"<ja...@py...>
>>>   >Sent: 07/08/05 17:46:55
>>>   >To:
>>> "lem...@li..."<lem...@li...>
>>>
>>>   >Subject: [PythonWeb.org] New Releases and Roadmap
>>>   >
>>>   >Hi All,
>>>   >
>>>   >I'm pleased to announce a new release of the web modules, well sort
>>> of!   >I've written a document about where we were and where I'd like
>>> to go:
>>>   >
>>>   >http://www.pythonweb.org/doc/pythonweb/
>>>   >
>>>   >So far I've split all the code up into different modules and
>>> released   >them in the eggs format (described in the link above) and
>>> also made a   >SVN snapshot of all the code:
>>>   >
>>>   >http://www.pythonweb.org/doc/pythonweb/release/
>>>   >http://www.pythonweb.org/snapshot/
>>>   >
>>>   >I'm hoping to document them all over the next few weeks using a new
>>>   >documentation format named pigeon where the documentation is
>>> written as   >part of the code itself so hopefully it will be easier
>>> to keep the   >documentation up to date. Examples here:
>>>   >
>>>   >http://www.pythonweb.org/doc/ (looks like I can't spell pigeon on
>>> the   >front screen.. oops)
>>>   >http://www.pythonweb.org/doc/web.doc.html (web module specific code)
>>>   >
>>>   >There is a lot of work to do documenting everything before it can
>>> really   >be used so please consider everything Alpha until the
>>> documentation is   >
>>>   >
>>>
>>> [Message truncated. Tap Edit->Mark for Download to get remaining
>>> portion.]
>>>
>>>
>>>
>>> -------------------------------------------------------
>>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>>> Practices
>>> Agile & Plan-Driven Development * Managing Projects & Teams * Testing
>>> & QA
>>> Security * Process Improvement & Measurement *
>>> http://www.sqe.com/bsce5sf
>>> _______________________________________________
>>> Lemon-webmodules mailing list
>>> Lem...@li...
>>> https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>>>
>>>
>>
>>
>>
>> -------------------------------------------------------
>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>> Practices
>> Agile & Plan-Driven Development * Managing Projects & Teams * Testing &
>> QA
>> Security * Process Improvement & Measurement *
>> http://www.sqe.com/bsce5sf
>> _______________________________________________
>> Lemon-webmodules mailing list
>> Lem...@li...
>> https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>>
>>
>>
>
>
>
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
> Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Lemon-webmodules mailing list
> Lem...@li...
> https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>




James Gardner wrote:
> Thanks Michael,
> 
> That's right, it uses reStructuredText in docstrings, uses Teud to turn
> them to XML and then parses the XML into the API documentation, very
> simple really. The idea is that the first docstring contains user
> documentation (in grey at the top) and the rest provides API
> documentation. I modeled it on the way the Ruby on Rails documentation
> looks.  I may make it so that the API docs and user docs are in separate
> files. It isn't too clever but the docutils project didn't seem to have
> a working version so I tweaked the best tools I could find to create
> pigeon!
> 

It sounds *very* interesting. There isn't a working part of docutils to
extract docstrings from code and generate API documentation. Maybe this
is it ;-)

By the way - does it work by importing the code, or by compiling it ?

Regards,

Fuzzy
http://www.voidspace.org.uk/python


> James
> 
> 
> 
> Michael Foord wrote:
> 
>> Sounds good James.
>>
>> To save me having to look *too far* ;-) - I assume pigeon uses
>> docstrings ? Can you use reStructured Text format, or have you created
>> your own markup convention ?
>>
>> Regards,
>>
>> Fuzzy
>> http://www.voidspace.org.uk/python
>> -----Original Message-----
>>   >From: "James Gardner"<ja...@py...>
>>   >Sent: 07/08/05 17:46:55
>>   >To:
>> "lem...@li..."<lem...@li...>
>>
>>   >Subject: [PythonWeb.org] New Releases and Roadmap
>>   >
>>   >Hi All,
>>   >
>>   >I'm pleased to announce a new release of the web modules, well sort
>> of!   >I've written a document about where we were and where I'd like
>> to go:
>>   >
>>   >http://www.pythonweb.org/doc/pythonweb/
>>   >
>>   >So far I've split all the code up into different modules and
>> released   >them in the eggs format (described in the link above) and
>> also made a   >SVN snapshot of all the code:
>>   >
>>   >http://www.pythonweb.org/doc/pythonweb/release/
>>   >http://www.pythonweb.org/snapshot/
>>   >
>>   >I'm hoping to document them all over the next few weeks using a new
>>   >documentation format named pigeon where the documentation is
>> written as   >part of the code itself so hopefully it will be easier
>> to keep the   >documentation up to date. Examples here:
>>   >
>>   >http://www.pythonweb.org/doc/ (looks like I can't spell pigeon on
>> the   >front screen.. oops)
>>   >http://www.pythonweb.org/doc/web.doc.html (web module specific code)
>>   >
>>   >There is a lot of work to do documenting everything before it can
>> really   >be used so please consider everything Alpha until the
>> documentation is   >
>>   >
>>
>> [Message truncated. Tap Edit->Mark for Download to get remaining
>> portion.]
>>
>>
>>
>> -------------------------------------------------------
>> SF.Net email is Sponsored by the Better Software Conference & EXPO
>> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
>> Practices
>> Agile & Plan-Driven Development * Managing Projects & Teams * Testing
>> & QA
>> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>> _______________________________________________
>> Lemon-webmodules mailing list
>> Lem...@li...
>> https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>>  
>>
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Lemon-webmodules mailing list
> Lem...@li...
> https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
> 
> 
> 




Thanks Michael,

That's right, it uses reStructuredText in docstrings, uses Teud to turn 
them to XML and then parses the XML into the API documentation, very 
simple really. The idea is that the first docstring contains user 
documentation (in grey at the top) and the rest provides API 
documentation. I modeled it on the way the Ruby on Rails documentation 
looks.  I may make it so that the API docs and user docs are in separate 
files. It isn't too clever but the docutils project didn't seem to have 
a working version so I tweaked the best tools I could find to create pigeon!

James



Michael Foord wrote:

>Sounds good James.
>
>To save me having to look *too far* ;-) - I assume pigeon uses docstrings ? Can you use reStructured Text format, or have you created your own markup convention ?
>
>Regards,
>
>Fuzzy
>http://www.voidspace.org.uk/python 
>
>-----Original Message-----
>   >From: "James Gardner"<ja...@py...>
>   >Sent: 07/08/05 17:46:55
>   >To: "lem...@li..."<lem...@li...>
>   >Subject: [PythonWeb.org] New Releases and Roadmap
>   >
>   >Hi All,
>   >
>   >I'm pleased to announce a new release of the web modules, well sort of! 
>   >I've written a document about where we were and where I'd like to go:
>   >
>   >http://www.pythonweb.org/doc/pythonweb/
>   >
>   >So far I've split all the code up into different modules and released 
>   >them in the eggs format (described in the link above) and also made a 
>   >SVN snapshot of all the code:
>   >
>   >http://www.pythonweb.org/doc/pythonweb/release/
>   >http://www.pythonweb.org/snapshot/
>   >
>   >I'm hoping to document them all over the next few weeks using a new 
>   >documentation format named pigeon where the documentation is written as 
>   >part of the code itself so hopefully it will be easier to keep the 
>   >documentation up to date. Examples here:
>   >
>   >http://www.pythonweb.org/doc/ (looks like I can't spell pigeon on the 
>   >front screen.. oops)
>   >http://www.pythonweb.org/doc/web.doc.html (web module specific code)
>   >
>   >There is a lot of work to do documenting everything before it can really 
>   >be used so please consider everything Alpha until the documentation is 
>   >
>   >
>
>[Message truncated. Tap Edit->Mark for Download to get remaining portion.]
>
>
>
>-------------------------------------------------------
>SF.Net email is Sponsored by the Better Software Conference & EXPO
>September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
>Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
>Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>_______________________________________________
>Lemon-webmodules mailing list
>Lem...@li...
>https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>  
>




Sounds good James.

To save me having to look *too far* ;-) - I assume pigeon uses docstrings ? Can you use reStructured Text format, or have you created your own markup convention ?

Regards,

Fuzzy
http://www.voidspace.org.uk/python 

-----Original Message-----
   >From: "James Gardner"<ja...@py...>
   >Sent: 07/08/05 17:46:55
   >To: "lem...@li..."<lem...@li...>
   >Subject: [PythonWeb.org] New Releases and Roadmap
   >
   >Hi All,
   >
   >I'm pleased to announce a new release of the web modules, well sort of! 
   >I've written a document about where we were and where I'd like to go:
   >
   >http://www.pythonweb.org/doc/pythonweb/
   >
   >So far I've split all the code up into different modules and released 
   >them in the eggs format (described in the link above) and also made a 
   >SVN snapshot of all the code:
   >
   >http://www.pythonweb.org/doc/pythonweb/release/
   >http://www.pythonweb.org/snapshot/
   >
   >I'm hoping to document them all over the next few weeks using a new 
   >documentation format named pigeon where the documentation is written as 
   >part of the code itself so hopefully it will be easier to keep the 
   >documentation up to date. Examples here:
   >
   >http://www.pythonweb.org/doc/ (looks like I can't spell pigeon on the 
   >front screen.. oops)
   >http://www.pythonweb.org/doc/web.doc.html (web module specific code)
   >
   >There is a lot of work to do documenting everything before it can really 
   >be used so please consider everything Alpha until the documentation is 
   >
   >

[Message truncated. Tap Edit->Mark for Download to get remaining portion.]




Hi All,

I'm pleased to announce a new release of the web modules, well sort of! 
I've written a document about where we were and where I'd like to go:

http://www.pythonweb.org/doc/pythonweb/

So far I've split all the code up into different modules and released 
them in the eggs format (described in the link above) and also made a 
SVN snapshot of all the code:

http://www.pythonweb.org/doc/pythonweb/release/
http://www.pythonweb.org/snapshot/

I'm hoping to document them all over the next few weeks using a new 
documentation format named pigeon where the documentation is written as 
part of the code itself so hopefully it will be easier to keep the 
documentation up to date. Examples here:

http://www.pythonweb.org/doc/ (looks like I can't spell pigeon on the 
front screen.. oops)
http://www.pythonweb.org/doc/web.doc.html (web module specific code)

There is a lot of work to do documenting everything before it can really 
be used so please consider everything Alpha until the documentation is 
finished. If you would like to use the new code one important change is 
that import web does not automatically load the web.cgi variable because 
doing so reads POST information. POST information can only be read once 
and you might choose to import web for a reason other than reading POST 
data and then not be able to use access it later (it was a problem I had 
when writing Bricks). To load CGI information you must now do:

import web
web.loadCGI()

then everything works as before.

Finally I've made the first release of Bricks. There is some 
documentation and a brief tutorial here:

http://www.pythonweb.org/bricks/

Bricks is actually being used in a production site with mod_fastcgi 
here: http://www.whatshouldireadnext.com The site fetches XML feeds from 
Amazon which significantly slows performance but it gives you an idea.

All the best,

James




Hi All,

> I was reading through the mailing list archives and saw the post about
> Bricks, is that available to test out yet?

Bricks isn't available yet but hopefully will be within the next month.

I thought I might as well give you all an update on the other projects
too:

I've now split the current pythonweb project into the following
sub-projects and re-factored all the code accordingly. 

The new modules will work with the easy install system from
http://peak.telecommunity.com/DevCenter/EasyInstall so that they can be
installed as follows with the dependencies automatically installed too:

> easy_install.py web

Pretty simple eh?

The modules are:

bricks - full WSGI framework taking a similar approach to ruby on rails
         for rapidly developing web applications

database - formally web.database for portable/simplified database access

engine - formally SnakeSQL - now a toolkit to write your own SQL engine 
         too.

model - formally database.object, an object-relational mapper

serve - classes for writing servers with

web - same as before without all the external packages included

All the modules are currently in an SVN repository and the documentation
and distribution files get correctly built using makefiles. All the new
modules work correctly together and still support Python 2.2.

Still to do:

* Write sensible documentation for the packages now they are standalone
* Restructure the pythonweb.org website for the new modules

I hope to have a pre-release of all the modules except Bricks at the end
of next week. The code of the pre-releases is unlikely to change much
before the final 0.6 releases but the documentation, tests and examples
will.

I hope to have a release of Bricks in early August. Bricks will be
Python 2.4 and above only because it will use decorators. I am currently
improving Ajax support (which already works) and will be using it for a
commercial project over the next few weeks which will mean it gets
developed quickly. 

If anyone is really keen I can provide an SVN snapshot so you can take a
peak?

Once I've got all the 0.6 modules released I'd really like to open up
the source a bit more and give more people access so improvements/bug
reports etc don't all have to come through me. I'll pretty much have
implemented everything I would like to see in the modules by then and am
more than happy for the project to take a new/better course with more
people involved.

Hope to have something to show you all soon anyway.

Cheers,

James









Hi Wayne,

Sorry for the delay in my reply.. I'm afraid I don't know what is wrong
with the old webserver.py script and I'd started writing a replacement
anyway so I've packaged it so that you can use this instead. It is
fairly sophisticated and can run WSGI appliactions as well as CGI
scripts.

See file attached.

I will be releasing a final version as part of the 0.6 releases,
hopefully in a few days.

Cheers,

James


Here is some brief documentation:


                                   1.1 Serve

   Warning: This is simply very breif documentation to get you started.
   More to come later

   XXX More here

   The serve module contains a number of classes which can be used to
   quickly build powerful servers. Included in the distribution is a
   simple implementation of such a server named server.py.

   server.py can be used as an effective server to test the examples
   distributed with the Python Web Modules. The program takes the
   following options:

usage: server.py [options] database

options:
  --version             show program's version number and exit
  -h, --help            show this help message and exit
  -p PORT, --port=PORT  port on which to start the server
  -a APP, --app-name=APP
                        name of WSGI application to execute
  -c CGI, --cgi=CGI     the cgi-bin directory
  -i CGIURL, --cgi-url=CGIURL
                        the cgi-bin directory url
  -d, --debug           allow the server to fall into debug mode when
                        environ['bricks.error.debug'] is raised
  -r, --reload          attempt to reload the application on each
request
  -l LOGFILE, --log=LOGFILE
                        log file to use
  -u PUBLIC, --public=PUBLIC
                        public directory to serve files from
  -v PRIVATE, --private=PRIVATE
                        private directory where bricks app, data, lib,
log and
                        script directories can be found
  --no-files            prevent public files being served
  --no-log              disable file logging
  -b, --browser         launch a web browser when started

   The options should be fairly self explanatory.

1.1.1 Using server.py to serve Web Module examples

     * Unzip and install is as follows:
        > tar zxfv serve-0.6.0pre0.tgz
        > cd serve-0.6.0pre0
        > python setup.py install
     * Run the server with the following commands:
        > python C:\Python24\Scripts\server.py -u "C:\htdocs"
         -i "C:\web\doc\src\lib" -c cgi-bin -p 8080

       (all on one line)

       Where C:\Python24\Scripts\server.py is the path to the installed
       script, C:\htdocs is the base directory of the webserver,
       C:\web\doc\src\lib
       is where the examples are, cgi-bin is the URL you want to access
       the examples at and 8080 is the port.
     * visit http://localhost:8080/cgi-bin/test.py replacing 
       test.py with the name of the example you wish to run.

Hi Cecil,

I think I ran into something similar on a FreeBSD machine with version 
0.3 of the web modules. Are you sure you are using the same version of 
MySQL and the same version of MySQLdb on both machines? I can't remember 
exactly what caused the problem but I remember I fixed it by manually 
copying the old MySQLdb drivers to my web modules installation so they 
used the old drivers not the new ones. In your case it sounds like the 
production machine has the newer version of MySQLdb so you could copy 
across the old MySQLdb version? Which version of web.database are you 
using, I'm sure the most recent version works with the recent versions 
of MySQLdb, but they might have introduced more incompatibilities.

I think pickle is a red herring in this case.

Hope that gives you something to start looking at.

James


Cecil Westerhof wrote:

>Op wo, 06-07-2005 te 08:47 +0200, schreef Cecil Westerhof:
>  
>
>>Op wo, 06-07-2005 te 08:03 +0200, schreef Cecil Westerhof:
>>    
>>
>>>When I do a print of read before 'return pickle.load(read)'  I get:
>>>	array('c', "S'help'\np0\n.")
>>>      
>>>
>>(It was a 'print read.getvalue()' offcourse.)
>>I also put this statement on the development machine. (I got immediatly
>>angry messages, because it destroyed the headers for the webserver.) And
>>I found the problem. On the development machine I do not get an array,
>>but just the sting, it gives:
>>        S'help'
>>        p0
>>        .
>>So on the develpment machine I get a string and on the production
>>machine I get an array. How can I solve this?
>>    
>>
>
>I have a 'solution'. It works for the moment. It is better to find out
>where the difference in behaviour comes from. But I am in the dark at
>the moment and needed a fix.
>
>The changed code:
>           def _get(self, key, app=None, pickleValue=True):
>                "Gets the variable 'key' from the session store."
>                if not app:
>                    app = self.params['app']
>                rows=self.params['cursor'].select(('ItemValue'),self.params['table']+'Store',where="SessionID='"+self.params['sessionID']+"' and ItemKey='"+key+"' and App='"+app+"'")
>                if not rows:
>                    raise SessionError("The session store doesn't have a key named '%s' for the app '%s'."%(key,app))
>                else:
>                    if type(rows[0]) == type({}):
>                        read = StringIO.StringIO(rows[0]['ItemValue'])
>                    else: # Assume it can be accessed as a tuple
>                        read = StringIO.StringIO(rows[0][0])
>                    if( read.getvalue().startswith('array') ):
>                            read = StringIO.StringIO(read.getvalue()[12:-2].replace('\\n', '\n'))
>                    if pickleValue:
>                        return pickle.load(read)
>                    else:
>                        return read
>
>I have added the lines:
>        if( read.getvalue().startswith('array') ):
>                read = StringIO.StringIO(read.getvalue()[12:-2].replace('\\n', '\n'))
>
>It seems to work. ;-}
>
>  
>




Hi Frithiof

Sorry for the long delay in replying. I hope you've managed to get the 
web modules table across to your production database.

>If I specify the foreign key constraint "DELETE CASCADE", the correct thing happens; when I delete an Account f.ex. I do not have to search through all related tables and delete manually. It is done automatically.  
>  
>
That's a bit more involved than foreign keys then!

>What I understand is this: 
>
>I create a dummy database in web modules with the same structure as the "real" one, this creates a special table structure that describes the database to web modules.
>
>I create the "real" database with SQL, using DB-API.
>
>I then create a description table for the real database copying the table created together with the dummy database.
>
>The web modules now understand the real database and uses it normally and because the constraints are hidden inside the database, the net effect is the occasional exeption.
>
>If I change the structure of the database later, I have to do the entire song and dance again.
>  
>
That's all correct and would be the easiest way of doing it at the moment

>>What I would ask though is, if you need all the database specific 
>>functionality why are you trying to use the web modules to access it, 
>>why not just write your code with the DB-API and not use the 
>>web.database module at all? 
>>    
>>
>
>I have already tried that.
>
>My experience is that the all the string pasting and parameter conversions back and forth - and checks on said conversions - *do* get very, very, repetitive in the not so very long run. ... Not to mention the many versions of DateTime to choose from and results formats, depending on who did the DB-API. 
>
>web modules is a great improvement in code briefness and clarity - even when using the web SQL drivers (I don't quite like the dict format yet - maybe because I am used to SQL ;-). 
>
>The ability to specify the format of an result as a dict or list is really useful for those functions that are not in SQL, like running balances, the automatic handling of special characters and the "stubbing" of functions that might not be in the DB-API but should be(commit/rollback), the code looks the same with different databases.
>
>... And I forgot the web forms part. Which I need too. 
>  
>
Well that's great.. exactly what I created the database modules for so 
I'm pleased they are also useful when used in conjunction with more 
advanced SQL code. I think it would be useful to be able to have a code 
alternative to the structure table to make this sort of thing much 
easier so I'll try to get that coded into the release after the one I'm 
working on at the moment!

All the best,

James





Op wo, 06-07-2005 te 08:47 +0200, schreef Cecil Westerhof:
> Op wo, 06-07-2005 te 08:03 +0200, schreef Cecil Westerhof:
> > When I do a print of read before 'return pickle.load(read)'  I get:
> > 	array('c', "S'help'\np0\n.")
> 
> (It was a 'print read.getvalue()' offcourse.)
> I also put this statement on the development machine. (I got immediatly
> angry messages, because it destroyed the headers for the webserver.) And
> I found the problem. On the development machine I do not get an array,
> but just the sting, it gives:
>         S'help'
>         p0
>         .
> So on the develpment machine I get a string and on the production
> machine I get an array. How can I solve this?

I have a 'solution'. It works for the moment. It is better to find out
where the difference in behaviour comes from. But I am in the dark at
the moment and needed a fix.

The changed code:
           def _get(self, key, app=None, pickleValue=True):
                "Gets the variable 'key' from the session store."
                if not app:
                    app = self.params['app']
                rows=self.params['cursor'].select(('ItemValue'),self.params['table']+'Store',where="SessionID='"+self.params['sessionID']+"' and ItemKey='"+key+"' and App='"+app+"'")
                if not rows:
                    raise SessionError("The session store doesn't have a key named '%s' for the app '%s'."%(key,app))
                else:
                    if type(rows[0]) == type({}):
                        read = StringIO.StringIO(rows[0]['ItemValue'])
                    else: # Assume it can be accessed as a tuple
                        read = StringIO.StringIO(rows[0][0])
                    if( read.getvalue().startswith('array') ):
                            read = StringIO.StringIO(read.getvalue()[12:-2].replace('\\n', '\n'))
                    if pickleValue:
                        return pickle.load(read)
                    else:
                        return read

I have added the lines:
        if( read.getvalue().startswith('array') ):
                read = StringIO.StringIO(read.getvalue()[12:-2].replace('\\n', '\n'))

It seems to work. ;-}

-- 
Cecil Westerhof <Cec...@xs...>



Op wo, 06-07-2005 te 08:03 +0200, schreef Cecil Westerhof:
> When I do a print of read before 'return pickle.load(read)'  I get:
> 	array('c', "S'help'\np0\n.")

(It was a 'print read.getvalue()' offcourse.)
I also put this statement on the development machine. (I got immediatly
angry messages, because it destroyed the headers for the webserver.) And
I found the problem. On the development machine I do not get an array,
but just the sting, it gives:
        S'help'
        p0
        .
So on the develpment machine I get a string and on the production
machine I get an array. How can I solve this?

-- 
Cecil Westerhof <Cec...@xs...>



I have a strange problem. I am trying to bring my application from a
development enviroment to a production enviroment. (Should have done it
sooner. Then I would have had more time to solve this problem.)
I get a very strange problem. I have pasted what comes in the logfile
(only on the production enviroment, development enviroment does not have
a problem):
        Traceback (most recent call last):
          File "SMS.cgi", line 10, in ?
            SMS.start(BASE_DIR)
          File "/home/users/decebalc/python//SMS/SMS.py", line 1549, in start
            Init()
          File "/home/users/decebalc/python//SMS/SMS.py", line 1372, in Init
            MenuEntry = CheckTotal('MenuEntry')
          File "/home/users/decebalc/python//SMS/SMS.py", line 200, in CheckTotal
            return CheckSession(ThisParam)
          File "/home/users/decebalc/python//SMS/SMS.py", line 169, in CheckSession
            if( Session.has_key(ThisParam) ):
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionBase.py", line 104, in has_key
            self.get(key, block)
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionBase.py", line 85, in get
            return self._get(key)
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionDatabase.py", line 148, in _get
            return pickle.load(read)
          File "/usr/local/lib/python2.4/pickle.py", line 1390, in load
            return Unpickler(file).load()
          File "/usr/local/lib/python2.4/pickle.py", line 872, in load
            dispatch[key](self)
          File "/usr/local/lib/python2.4/pickle.py", line 1200, in load_append
            value = stack.pop()
        IndexError: pop from empty list

I do not know if it is important, but on the development machine I am
using
python 2.2.2 and the production machine has python 2.4.1 (and it is a
BSD machine).

Attached a simple program that illustrates the problem. (username and
password have to be changed offcourse)

On my development system this gives:
        $ python  testing.py
        Set-Cookie: SMS=ZpsMOrBZGDPvg893bs5lPw; Comment=Built in Python using web.session from pythonweb.org; Max-Age=21600; Path=/; Version=1;
        help

On the production system it gives:
        $ python testing.py
        Set-Cookie: SMS=Gg2dhrRcrf5tEGTLUoSJVw; Comment=Built in Python using web.session from pythonweb.org; Max-Age=21600; Path=/; Version=1;
        Traceback (most recent call last):
          File "testing.py", line 20, in ?
            print Session['dummy']
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionBase.py", line 222, in __getitem__
            return self.get(key)
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionBase.py", line 85, in get
            return self._get(key)
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionDatabase.py", line 148, in _get
            return pickle.load(read)
          File "/usr/local/lib/python2.4/pickle.py", line 1390, in load
            return Unpickler(file).load()
          File "/usr/local/lib/python2.4/pickle.py", line 872, in load
            dispatch[key](self)
          File "/usr/local/lib/python2.4/pickle.py", line 1200, in load_append
            value = stack.pop()
        IndexError: pop from empty list

When I do a print of read before 'return pickle.load(read)'  I get:
	array('c', "S'help'\np0\n.")

When I use 'python -i testing.py' and then type Session, I get:
on develpment
        <web.session.sessionDatabase.SessionDatabase instance at 0x825e044>
and
on production
        Traceback (most recent call last):
          File "<stdin>", line 1, in ?
          File "/home/users/decebalc/python//PythonWeb/web/session/sessionBase.py", line 240, in __getattr__
            raise SessionError("The session does not have a parameter named '%s'."%name)
        web.errors.SessionError: The session does not have a parameter named '__repr__'.

Does anyone have any idea what the problem (and the solution) could be?
        
-- 

Sanatate si numai bine,



Cecil Westerhof
Decebal Computing & Services, the Netherlands
Prins Hendrikweg 7
3941 GC Doorn
the Netherlands
Tel: +31 - 6 - 27 03 28 22
Fax. +31 - 84 - 740 58 53
Email: Ce...@De...
URL: http://www.DecebalComputing.nl
-- 
Cecil Westerhof <Cec...@xs...>

Hi all,

one thing I was missing was a more comfortable possibility to query
the authAdmin object if a user has a specific role. Patch below.

Index: C:/Programme/eclipse3.1M7/workspace/Junkkiller/cgi-bin/pythonweb/web/auth/__init__.py
===================================================================

--- C:/Programme/eclipse3.1M7/workspace/Junkkiller/cgi-bin/pythonweb/web/auth/__init__.py       (revision 81)
+++ C:/Programme/eclipse3.1M7/workspace/Junkkiller/cgi-bin/pythonweb/web/auth/__init__.py       (working copy)
@@ -216,6 +216,14 @@
             username = username.lower()
         return self._driver.roles(username, app)
 
+    def hasRole(self, username, role, app=None):
+        if username == None or role == None: 
+            return False
+        else:
+            username = username.lower()
+            roles = self._driver.roles(username, app)
+            return (role in roles)
+
     def setRole(self, username, app, role):
         username = username.lower()
         return self._driver.setRole(username, app, role)

-- 
Felix




I was reading through the mailing list archives and saw the post about
Bricks, is that available to test out yet?

Thanks,

Wayne



Hi James,

Thanks for the help.=20

> The web modules do support simple foreign key constraints but nothing=20
> particularly complicated,=20

It will *work* - What I am trying to do is to build a framework to test =
different betting strategies. I create many agents. Each agent owning an =
account and following a strategy. Then the agents run and I see who =
blows up and who survives. The "survivers" might be lucky so they get to =
run with different portfolios. =20

What I have is tables of "Accounts" containing a Cash Account and a =
Portfolio, all containing positions/entries (tables of credits and =
debits, ) -  i.e. LOTS of linked tables that needs to be created and =
destroyed often.=20

If I specify the foreign key constraint "DELETE CASCADE", the correct =
thing happens; when I delete an Account f.ex. I do not have to search =
through all related tables and delete manually. It is done =
automatically. =20

Some related table entries I need to keep - f.ex. stock quotes related =
to trades that should stick around for accounting.=20

This is also easily done by constraints.

> but it would be easier to copy it from another database as=20
> described above.

What I understand is this:=20

I create a dummy database in web modules with the same structure as the =
"real" one, this creates a special table structure that describes the =
database to web modules.

I create the "real" database with SQL, using DB-API.

I then create a description table for the real database copying the =
table created together with the dummy database.

The web modules now understand the real database and uses it normally =
and because the constraints are hidden inside the database, the net =
effect is the occasional exeption.

If I change the structure of the database later, I have to do the entire =
song and dance again.

That is sort-of what I hoped to do.

> What I would ask though is, if you need all the database specific=20
> functionality why are you trying to use the web modules to access it,=20
> why not just write your code with the DB-API and not use the=20
> web.database module at all?=20

I have already tried that.

My experience is that the all the string pasting and parameter =
conversions back and forth - and checks on said conversions - *do* get =
very, very, repetitive in the not so very long run. ... Not to mention =
the many versions of DateTime to choose from and results formats, =
depending on who did the DB-API.=20

web modules is a great improvement in code briefness and clarity - even =
when using the web SQL drivers (I don't quite like the dict format yet - =
maybe because I am used to SQL ;-).=20

The ability to specify the format of an result as a dict or list is =
really useful for those functions that are not in SQL, like running =
balances, the automatic handling of special characters and the =
"stubbing" of functions that might not be in the DB-API but should =
be(commit/rollback), the code looks the same with different databases.

... And I forgot the web forms part. Which I need too.=20

Web interfaces solve a lot of annoying firewall issues one might =
encounter in the corporate environment and on lame ISP accounts. Like =
the system management station I have on our IPv6 test network. It's an =
Apache2 box and it is reached via HTTP on the global Internet to get =
information across about 15 meters physical distance blocked by several =
meters of "IT-policy" documents ;-) =20

> Would you find this useful?

Very Useful.
=20
> Hope that is of some help,

Certainly, Thanks.=20



I was looking at using the webserver.py to test locally, rather than
work directly on my hosting provider's servers.  However when I run
'python webserver.py -b public_html/' and then load a file named
'home.py' I get the source of home.py.

I also created a cgi-bin directory under public_html as well but that
didn't work either.  home.py is executable, running 'python home.py'
at a terminal works fine and gives the correct output.

Any thoughts on what could be wrong?

Thanks,

Wayne



James,

Thanks for the response.  I did figure out the problem...sorta.  The
code I used was from
http://www.pythonweb.org/projects/webmodules/doc/0.5.3/html_multipage/lib/e=
xample-file-web-template-xyaptu.html
which dind't have the headers or a 'print web.header()'.  Which is
what was causing the Apache error.

Oddly enough, when I put a blank line between the closing BODY and
closing HTML tags everything works fine.  I then removed the space and
got the same error, but it went away when I put the blank line back
in.  I'm going to test the same code on a different system tomorrow to
see if I can narrow it down.

Wayne

On 6/23/05, James Gardner <ja...@py...> wrote:
> Hi Wayne,
>=20
> I've tested your code on both Python 2.2 and Python 2.4 and it works as
> it should for me, I couldn't reproduce your error I'm afraid. Cheetah is
> the most powerful and best supported templating engine that the web
> modules support so I'd recommend using that instead. The example is at
> http://www.pythonweb.org/projects/webmodules/doc/0.5.3/html_multipage/lib=
/example-file-web-template-cheetah.html
>=20
> As for the Internal Server Error on Apache, that is usually caused
> because you have forgotten to print the HTTP header before the main
> content. Add the line:
>=20
> print "Content-type: text/html\n\n"
>=20
> after the import web.template line and you should avoid the error.
>=20
> Hope that is of some use!
>=20
> James
>=20
> Wayne Pierce wrote:
>=20
> >I just found PythonWeb and have started working through the examples.
> >Unfortunately I can't get a simple template working properly, parts of
> >the HTML disappear or are printed out of order in the final output.
> >
> >I have the following files:
> >
> >(webtest.py)
> >
> >#!/usr/bin/python
> >
> >import web.template
> >
> >dict =3D {
> >       'welcomeMessage': 'Welcome to the test page.',
> >       'testVar': True,
> >       'title': 'Example'
> >}
> >
> >print web.template.parse(
> >       type=3D'xyaptu',
> >       file=3D'web-template.tmpl',
> >       dict=3Ddict
> >)
> >
> >(web-template.tmpl)
> >
> ><html>
> >       <head>
> >               <title>$title</title>
> >       </head>
> >
> >       <body>
> >               <p>$welcomeMessage</p>
> >                       <py-open code=3D"if testVar:" />
> >                               The variable is: True
> >                       <py-clause code=3D"else:" />
> >                               The variable is: False
> >                       <py-close/>
> >       </body>
> ></html>
> >
> >When I run 'python webtest.py' from the command line (Fedora 3) I get
> >the following:
> >
> ><html>    <head>    <title>Example</title>
> >                <body>>    <p>Welcome to the test page.</p>
> >
> >The variable is: True
> ></html>   </body>
> >
> >
> >Two things happen, </head> disappears and the last two items are
> >flipped.  Running webtest.py through Apache gives me an Internal
> >Server Error due to malformed headers.
> >
> >Does anyone have suggestions on what could be causing this?  I am
> >using Python 2.3.4 with the latest version of PythonWeb.
> >
> >Thanks for any help,
> >
> >Wayne
> >
> >
> >-------------------------------------------------------
> >SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> >from IBM. Find simple to follow Roadmaps, straightforward articles,
> >informative Webcasts and more! Get everything you need to get up to
> >speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id=16492&op=3Dclick
> >_______________________________________________
> >Lemon-webmodules mailing list
> >Lem...@li...
> >https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
> >
> >
>=20
>



Is anyone planning on going to EuroPython in Sweden this week? I'm going 
so if any of you are, do let me know as it would be great to meet up!

James



Hi Frithiof,

The web modules do support simple foreign key constraints but nothing 
particularly complicated, I'm not sure if the support is sufficient for 
your purposes though. The documentation is at 
http://www.pythonweb.org/projects/webmodules/doc/0.5.3/html_multipage/lib/ForeignKeys.html

You certainly can create the database with all the advanced 
functionality and then make a table to fool the web modules into 
thinking the database was created by them. The web modules store all the 
information about the structure of the database in table named, by 
default, "StructureTable". My suggestion would be to create two 
databases, one using the web.database module and one with all the extra 
database-specific code. You can then copy the information from the 
StructureTable created by the web.database module to the other database 
and reconnect to the real one with the web.database modules. The 
structure is fairly simple. It is created in the code like this:

self.cursor.create(
                    table='StructureTable',
                    columns = [
                        self.cursor.column(name='TableName', 
type='String', required=True),
                        self.cursor.column(name='ColumnName', 
type='String', required=True),
                        self.cursor.column(name='ColumnType', 
type='String', required=True),
                        self.cursor.column(name='IsRequired', type='Bool'),
                        self.cursor.column(name='IsUnique', type='Bool'),
                        self.cursor.column(name='IsPrimaryKey', 
type='Bool'),
                        self.cursor.column(name='ForeignKeyTable', 
type='Text'),
                        self.cursor.column(name='DefaultValue', 
type='Text'),
                        self.cursor.column(name='Position', 
type='Integer', required=True),
                    ],
                    execute=True,
                     mode='direct'
                )

but it would be easier to copy it from another database as described above.

What I would ask though is, if you need all the database specific 
functionality why are you trying to use the web modules to access it, 
why not just write your code with the DB-API and not use the 
web.database module at all? You can still use the web.auth and 
web.session modules in alongside other tables, the web modules will be 
unaware of the more complicated tables in your database but the DB-API 
will be able to access all the information. Could you give a bit more 
information about what you are trying to achieve?

For the next release of the modules, the web.database module will be 
separated out into a completely separate package. I'm planning on 
including functionality to allow you to tell the web modules your 
database structure by passing parameters to the connect() method as well 
as obtaining it from the database. Would you find this useful?

Hope that is of some help,

James

Frithiof Jensen (AH/LMD) wrote:

>Hi List,
>
>I would like to manually create a database with CHECK and DELETE constraints e.t.c. so that I do not need to perpetually manage this all over my application and also to ensure that other tools respect the rules of the database.
>
>The limited SQL subset in webmodules does not allow this for compatibility between different databases, so I cannot use webmodules for that.
>
>"the path of least resistance", I Think, is to create the database with raw SQL using the Python DBI and all the advanced constructs and then somehow import/open it in webmodules. 
>
>That way webmodules do not need to "know/learn" any advanced database functionality but it will need to deal with exceptions generated by constraints - or rather pass them up to the application (which I think it does anyway). 
>
>My understanding is that webmodules maintains it's own tables about the structure of the databases it works with. Those are created automatically when creating the database in webmodules. 
>
>My problem is: How do I make webmodules aware of a database structure created outside webmodules?
>
>mvh,
>Frithiof Jensen. 
>
>
>-------------------------------------------------------
>SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
>from IBM. Find simple to follow Roadmaps, straightforward articles,
>informative Webcasts and more! Get everything you need to get up to
>speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id492&op=click
>_______________________________________________
>Lemon-webmodules mailing list
>Lem...@li...
>https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>  
>




Hi Wayne,

I've tested your code on both Python 2.2 and Python 2.4 and it works as 
it should for me, I couldn't reproduce your error I'm afraid. Cheetah is 
the most powerful and best supported templating engine that the web 
modules support so I'd recommend using that instead. The example is at 
http://www.pythonweb.org/projects/webmodules/doc/0.5.3/html_multipage/lib/example-file-web-template-cheetah.html

As for the Internal Server Error on Apache, that is usually caused 
because you have forgotten to print the HTTP header before the main 
content. Add the line:

print "Content-type: text/html\n\n"

after the import web.template line and you should avoid the error.

Hope that is of some use!

James

Wayne Pierce wrote:

>I just found PythonWeb and have started working through the examples. 
>Unfortunately I can't get a simple template working properly, parts of
>the HTML disappear or are printed out of order in the final output.
>
>I have the following files:
>
>(webtest.py)
>
>#!/usr/bin/python
>
>import web.template
>
>dict = {
>	'welcomeMessage': 'Welcome to the test page.',
>	'testVar': True,
>	'title': 'Example'
>}
>
>print web.template.parse(
>	type='xyaptu',
>	file='web-template.tmpl',
>	dict=dict
>)
>
>(web-template.tmpl)
>
><html>
>	<head>
>		<title>$title</title>
>	</head>
>	
>	<body>
>		<p>$welcomeMessage</p>
>			<py-open code="if testVar:" />
>				The variable is: True
>			<py-clause code="else:" />
>				The variable is: False
>			<py-close/>
>	</body>
></html>
>
>When I run 'python webtest.py' from the command line (Fedora 3) I get
>the following:
>
><html>    <head>    <title>Example</title>
>                <body>>    <p>Welcome to the test page.</p>
>                                                                    
>The variable is: True
></html>   </body>
>
>
>Two things happen, </head> disappears and the last two items are
>flipped.  Running webtest.py through Apache gives me an Internal
>Server Error due to malformed headers.
>
>Does anyone have suggestions on what could be causing this?  I am
>using Python 2.3.4 with the latest version of PythonWeb.
>
>Thanks for any help,
>
>Wayne
>
>
>-------------------------------------------------------
>SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
>from IBM. Find simple to follow Roadmaps, straightforward articles,
>informative Webcasts and more! Get everything you need to get up to
>speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id492&op=click
>_______________________________________________
>Lemon-webmodules mailing list
>Lem...@li...
>https://lists.sourceforge.net/lists/listinfo/lemon-webmodules
>  
>





2005	Jan (21)	Feb (40)	Mar (5)	Apr (18)	May (3)	Jun (12)	Jul (8)	Aug (7)	Sep	Oct	Nov (2)	Dec
2010	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug (1)	Sep	Oct	Nov	Dec

lemon-webmodules Mailing List for Lemon - Python Web Modules

lemon-webmodules — Discussions on the Python Web Modules at PythonWeb.org