Re: Please approve python-xmpp
Satturday, 14 May 2005 11:31 Steve Langasek wrote:
Hello.
I have corrected list of bugfixes: some were gone as insignigicant, other were
upgraded :)
Really, most of these bug are really serious enough. They were found while
developing GUI client based on this library and this client were just not
going to release without these fixes.
So here is new list (bug 307988):
important: SASL fix. SASL authentication is a standart for xmpp so unability
to properly authenticate will have major effect on usability of the library
important: bugfix in browser module that renders this module dangerous to use
and hence greatly affects usability of library for transports
normal: allows user to detect servers with broken tls implementation and
disable tls for them. Very cheap solution that have no impact on the code
that is not aware about this feature. I think it is worths to go in sarge
documentation fixes: docstrings fixes in different modules, examples
improvements.
important: roster fix. Renders roster unsuitable for general use. Greatly
affect client scripts that use python-xmpp.
important: memory leak fix for transports.
important: hang on big incoming data chunks while working with tls enabled. In
fact, this is gate for DoS condition.
If you feel that this list is [not] ok please let me know - I'll ask my
sponsor to upload these/corrected changes to the archive.
> Alexey,
>
> On Sat, May 14, 2005 at 08:44:19AM +0400, Alexey Nezhdanov wrote:
> > This release (0.2-rc3-1) closes the following:
> >
> > Bug http://bugs.debian.org/307988 :
> > minor: Compartibility fix in xsend example.
> > important: Fix in browser to prevent error looping
> > minor: Fix in browser to raise NodeProcessed on handled iq
> > minor: Changes in auth code to complain RFC (not use NonSASL after
> > failed SASL) important: TLS fix: there was a hangs on big chunks of data
> > important: getRegInfo traceback of certain servers fix
> > normal: More gentle T/NT classes workaround. Not solved yet though :(
> > normal: several disconnect traceback fixes
>
> This does not meet the freeze guidelines as listed at
> <http://lists.debian.org/debian-devel-announce/2005/05/msg00001.html>.
--
Respectfully
Alexey Nezhdanov
Reply to: