[Webware-discuss] cvs update: pre 0.6b1 -- please test

[Chuck E. <Chu...@ya...>]

Hi,

I've made several updates to CVS tonight. I have no other updates planned 
before cutting 0.6b1. If you have the opportunity, please consider syncing 
to CVS and doing some basic testing. If I don't hear anything bad in the 
next day or two, then I will cut the release.

Regarding 0.6 final, the only item on my TO DO list is comprehensive 
release notes.

CVS info is at https://sourceforge.net/cvs/?group_id=4866


Here are the release notes I have collected for changes since 0.6a1:

Webware 0.6b1 Release Notes
---------------------------
Since 0.6a1:

WebKit

     - The app server will no longer start up if you haven't run
       Webware/install.py. This will help ensure that people use WebKit
       properly, and also eliminate reports of the Docs context in WebKit
       not working (it requires install.py).

     - Fixed bugs in cookie handling for Python 2.1 and earlier.

     - Fixed wkcgi.exe so that it works with IIS as well as Apache.
         - IIS runs cgi's in a different directory than the directory
           containing the cgi executable itself.  So it couldn't find the
           webkit.cfg file since it was looking in the wrong directory.
         - IIS doesn't close stdin after writing CONTENT_LENGTH bytes to
           it. So we have to make sure we read in CONTENT_LENGTH bytes,
           and no more, otherwise it hangs forever waiting for more input.

     - Removed the unstable Async*.py classes which already had their
       instructions removed from the Install Guide.

     - Extracted Cans into a preliminary CanKit, delivered to Jay Love
       <js...@js...>.

     - Some general code cleanup.


MiddleKit

     - MKBrowser now loads again. This is a WebKit context for browsing
       any MiddleKit store. However, much work remains to be done on this.
       I would be happy if someone volunteered to work on this as it would
       be valuable and easily approached as a separate project.

     - MiddleObject.setValueForKey() now raises LookupError instead of
       NameError if there is a problem. This is consistent with
       valueForKey() and fits the doc strings of the exceptions better.