You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by kf...@collab.net on 2005/10/06 20:43:09 UTC
Setting a date for 1.3 RC tarball.
Let's set a date for the first 1.3.0 release candidate tarball. How
about Wednesday, October 12th? Ben Reser, does that work for you?
Details, and possible items for discussion:
The 1.3.x/STATUS file looks pretty good. Most of what's in there is
bugfixes, and I think we'll be able to ensure that the important ones
get enough review & votes by Wedesday to go in. No problems there.
On http://subversion.tigris.org/svn_1.3_releasenotes.html there are
still a few pending items. Some of them might be done now, others I'm
pretty sure are not. Here are the relevant items:
- Is the Python bindings Trac test crash on Windows, solved? I
think not; at least, the last message in the thread is
http://svn.haxx.se/dev/archive-2005-09/1047.shtml, and it's a
disgruntled statement from Brane: "Ah. Well. Yes, the trac tests
crash again.. :("
- We haven't upgraded server-side Python scripts to use the new
automatic memory management API. On the one hand, this doesn't
need to block 1.3 -- it's a nice-to-have, not a must-have. On the
other hand, David James *already* wrote the patch to do it:
http://svn.haxx.se/dev/archive-2005-09/0821.shtml
I think we should apply it to trunk, and port the change to 1.3.x,
if for no other reason than that this is exactly what release
candidates are for. It's the best way to get the new bindings the
testing they need. Thoughts?
Aside from the above, is there anything else we need on the 1.3.x
line? Are there any trunk changes that are crucial but which haven't
been nominated in 1.3.x/STATUS yet? (I'm behind on commit review, so
please speak up if you know of something.)
I've got a few other questions about 1.3, but they aren't things that
would block a release candidate, so I'll ask them in separate mails.
-Karl
--
www.collab.net <> CollabNet | Distributed Development On Demand
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Setting a date for 1.3 RC tarball.
Posted by "C. Michael Pilato" <cm...@collab.net>.
kfogel@collab.net writes:
> > No, this isn't solved. To fix the crash issues in the Windows
> > bindings, we need to compile Subversion on Windows with debug symbols,
> > and use a debugger to find out where the bindings are crashing during
> > the test suite. C. Mike, would you be able to take a look at this?
>
> Or if not CMike, is there someone else with a Windows box and a little
> bindings experience around?
>
> (I've pinged CMike privately in case he didn't see this thread,
> though.)
CMike currently has no Windows build environment. That setup is on a
computer in some foam packing in a box in a ministorage unit in
TheMiddleOfNowhere, North Carolina.
>
> > > - We haven't upgraded server-side Python scripts to use the new
> > > automatic memory management API. On the one hand, this doesn't
> > > need to block 1.3 -- it's a nice-to-have, not a must-have. On the
> > > other hand, David James *already* wrote the patch to do it:
[...]
> An open question: can we release the upgraded versions in 1.3.1 or
> later? Or should it wait until 1.4. I guess 1.4, much as I hate to
> say it.
Nah, 1.3.1 would be fine. The behavior of the scripts isn't changing,
the scripts don't themselves expose APIs, etc. This tweak would be
along the lines of a memory usage bugfix, and is a perfectly valid
candidate for a patch release.
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Setting a date for 1.3 RC tarball.
Posted by kf...@collab.net.
David James <ja...@gmail.com> writes:
> We need to fix the crashes on Windows before the Python bindings will
> be ready for a release candidate. Do you think the release could wait
> a week?
A week? I think we shouldn't need that long.
> > - Is the Python bindings Trac test crash on Windows, solved? I
> > think not; at least, the last message in the thread is
> > http://svn.haxx.se/dev/archive-2005-09/1047.shtml, and it's a
> > disgruntled statement from Brane: "Ah. Well. Yes, the trac tests
> > crash again.. :("
>
> No, this isn't solved. To fix the crash issues in the Windows
> bindings, we need to compile Subversion on Windows with debug symbols,
> and use a debugger to find out where the bindings are crashing during
> the test suite. C. Mike, would you be able to take a look at this?
Or if not CMike, is there someone else with a Windows box and a little
bindings experience around?
(I've pinged CMike privately in case he didn't see this thread, though.)
> > - We haven't upgraded server-side Python scripts to use the new
> > automatic memory management API. On the one hand, this doesn't
> > need to block 1.3 -- it's a nice-to-have, not a must-have. On the
> > other hand, David James *already* wrote the patch to do it:
> >
> > http://svn.haxx.se/dev/archive-2005-09/0821.shtml
> >
> > I think we should apply it to trunk, and port the change to 1.3.x,
> > if for no other reason than that this is exactly what release
> > candidates are for. It's the best way to get the new bindings the
> > testing they need. Thoughts?
>
> I'd like to commit the patch, but I don't know enough about the
> affected scripts to test them properly. Anyone able to help?
I'd say let's not block 1.3 with this one, then. The main script to
worry about is probably mailer.py, which isn't too hard to test (well,
individually, none of them are hard to test, but in aggregate it's a
bunch of drudge work). We should probably merge the patch to trunk.
Getting code into the bleeding edge is a good way to get it tested by
random people (I can do one or two, but couldn't promise to test all
of them).
An open question: can we release the upgraded versions in 1.3.1 or
later? Or should it wait until 1.4. I guess 1.4, much as I hate to
say it.
-Karl
--
www.collab.net <> CollabNet | Distributed Development On Demand
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Setting a date for 1.3 RC tarball.
Posted by David James <ja...@gmail.com>.
On 10/6/05, kfogel@collab.net <kf...@collab.net> wrote:
> Let's set a date for the first 1.3.0 release candidate tarball. How
> about Wednesday, October 12th? Ben Reser, does that work for you?
We need to fix the crashes on Windows before the Python bindings will
be ready for a release candidate. Do you think the release could wait
a week?
> Details, and possible items for discussion:
>
> The 1.3.x/STATUS file looks pretty good. Most of what's in there is
> bugfixes, and I think we'll be able to ensure that the important ones
> get enough review & votes by Wedesday to go in. No problems there.
>
> On http://subversion.tigris.org/svn_1.3_releasenotes.html there are
> still a few pending items. Some of them might be done now, others I'm
> pretty sure are not. Here are the relevant items:
>
> - Is the Python bindings Trac test crash on Windows, solved? I
> think not; at least, the last message in the thread is
> http://svn.haxx.se/dev/archive-2005-09/1047.shtml, and it's a
> disgruntled statement from Brane: "Ah. Well. Yes, the trac tests
> crash again.. :("
No, this isn't solved. To fix the crash issues in the Windows
bindings, we need to compile Subversion on Windows with debug symbols,
and use a debugger to find out where the bindings are crashing during
the test suite. C. Mike, would you be able to take a look at this?
> - We haven't upgraded server-side Python scripts to use the new
> automatic memory management API. On the one hand, this doesn't
> need to block 1.3 -- it's a nice-to-have, not a must-have. On the
> other hand, David James *already* wrote the patch to do it:
>
> http://svn.haxx.se/dev/archive-2005-09/0821.shtml
>
> I think we should apply it to trunk, and port the change to 1.3.x,
> if for no other reason than that this is exactly what release
> candidates are for. It's the best way to get the new bindings the
> testing they need. Thoughts?
I'd like to commit the patch, but I don't know enough about the
affected scripts to test them properly. Anyone able to help?
> Aside from the above, is there anything else we need on the 1.3.x
> line?
If the Windows crashes are fixed, then our release candidate is ready to go.
Cheers,
David
P.S. The install-swig-py target can fail in a VPATH setup, if you are
building from a release build, and make the "install-swig-py" target
before making the "swig-py" target. Max Bowsher is currently working
on a fix. I'm hoping this fix will be included, but it is a minor
issue.
--
David James -- http://www.cs.toronto.edu/~james
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Setting a date for 1.3 RC tarball.
Posted by Max Bowsher <ma...@ukf.net>.
kfogel@collab.net wrote:
> Let's set a date for the first 1.3.0 release candidate tarball. How
> about Wednesday, October 12th? Ben Reser, does that work for you?
>
> Details, and possible items for discussion:
>
> The 1.3.x/STATUS file looks pretty good. Most of what's in there is
> bugfixes, and I think we'll be able to ensure that the important ones
> get enough review & votes by Wedesday to go in. No problems there.
We are going to need a high throughput of bindings and buildsystem reviews.
In addition to just nominating several more changes, I also have a
non-trivial patch in progress concerning the SWIG bindings buildsystem,
especially where VPATH builds are concerned.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org