Patches pending commit for 2.3

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Patches pending commit for 2.3

Felix Schwarz
Hi,

today I committed a few trivial patches to the 2.3.x branch. Now I glanced
over the bugs for newer versions to find out if there is more stuff that needs
backporting.

This is my preliminary queue of pending patches, feel free to comment/discuss:
- #995 - CherryPy 2.3.0 regression with Python 2.5
   http://www.bitbucket.org/felixschwarz/cp2-candidates/src/tip/invalid_cookies
   Even though Python changed it's behavior for Python 2.5+, I don't see any
   actual change in the tests so I have to commit that without any test case...

   Interestingly this is fixed differently in 3.1 and trunk. Also 3.0 misses
   the fix completely. To me the fix in 3.1 (r1765) looks a bit better so I'm
   tempted to backport this instead of the patch in #995.

   Also #735 which is the original issue is still open. I think it should be
   closed again as it changed scope (it was fixed but no in the version the
   last commenter wanted)

- #868 - traceback returned when CherryPy receives invalid cookie
   http://www.bitbucket.org/felixschwarz/cp2-candidates/src/tip/invalid_cookies
   backport of r2175

- #951 - Content-Length breaks curl on 100-Continue
   no patch yet but looks interesting enough to warant a backport.

Debian also ships a patch for the autoreloader:
http://patch-tracker.debian.org/package/python-cherrypy/2.3.0-3
Basically it's a backport of later changes to _cpengine so I guess that can go
in as well...

fs


smime.p7s (7K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Patches pending commit for 2.3

Christoph Zwerschke
Am 02.04.2010 15:21 schrieb Felix Schwarz:
> Interestingly this is fixed differently in 3.1 and trunk. Also 3.0 misses
> the fix completely. To me the fix in 3.1 (r1765) looks a bit better so I'm
> tempted to backport this instead of the patch in #995.

Do you mean http://www.cherrypy.org/changeset/1765? To me that's the
uglies fix of all. Either output cookie with sep="\n" and then split
with "\n", or output cookie with no (i.e. default) sep, and then split
with splitlines() - the splitlines() handles all kinds of line seps. The
latter method looks like the cleanest solution to me, plus it also
obviates the need for the "if cookie" statement since the list will be
empty in this case.

> Also #735 which is the original issue is still open. I think it
> should be closed again as it changed scope (it was fixed but no in
> the version the last commenter wanted)

Right, that's the same issue and can be closed then.

 From my side, I don't have any more patches for 2.3.1; I noticed the
deprecation of sha and md5 in Py 2.6 was already handled in r2662.

-- Christoph

--
You received this message because you are subscribed to the Google Groups "cherrypy-devel" group.
To post to this group, send email to [hidden email].
To unsubscribe from this group, send email to [hidden email].
For more options, visit this group at http://groups.google.com/group/cherrypy-devel?hl=en.

Reply | Threaded
Open this post in threaded view
|

RE: Patches pending commit for 2.3

Robert Brewer-4
In reply to this post by Felix Schwarz
Felix Schwarz wrote:
> today I committed a few trivial patches to the 2.3.x branch.

Thanks!

> Now I glanced over the bugs for newer versions to find out
> if there is more stuff that needs backporting.
>
> This is my preliminary queue of pending patches, feel free to
> comment/discuss:
> - #995 - CherryPy 2.3.0 regression with Python 2.5
>    http://www.bitbucket.org/felixschwarz/cp2-
> candidates/src/tip/invalid_cookies
>    Even though Python changed it's behavior for Python 2.5+, I don't
> see any
>    actual change in the tests so I have to commit that without any test
> case...
>
> Interestingly this is fixed differently in 3.1 and trunk. Also 3.0
> misses the fix completely.

Right; 2.3.0 was released after 3.0.0.

> To me the fix in 3.1 (r1765) looks a bit better
> so I'm
>    tempted to backport this instead of the patch in #995.
>
>    Also #735 which is the original issue is still open. I think it
> should be
>    closed again as it changed scope (it was fixed but no in the version
> the
>    last commenter wanted)

Fine by me.

> - #868 - traceback returned when CherryPy receives invalid cookie
>    http://www.bitbucket.org/felixschwarz/cp2-
> candidates/src/tip/invalid_cookies
>    backport of r2175
>
> - #951 - Content-Length breaks curl on 100-Continue
>    no patch yet but looks interesting enough to warant a backport.
>
> Debian also ships a patch for the autoreloader:
> http://patch-tracker.debian.org/package/python-cherrypy/2.3.0-3
> Basically it's a backport of later changes to _cpengine so I guess that
> can go
> in as well...

Yes, as long as that doesn't require upgrading to CP 3's process package. The inheritance/interaction between cherrypy.server and .engine changed significantly from CP2 to CP3.


Robert Brewer
[hidden email]

--
You received this message because you are subscribed to the Google Groups "cherrypy-devel" group.
To post to this group, send email to [hidden email].
To unsubscribe from this group, send email to [hidden email].
For more options, visit this group at http://groups.google.com/group/cherrypy-devel?hl=en.