You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Garrett Rooney <ro...@electricjellyfish.net> on 2005/11/14 19:06:03 UTC
Time for 1.3.0-rc3?
So there appear to have been a fair number of fixes merged into the
1.3.x branch since rc2 was rolled. Since we haven't even officially
released rc2 to the world yet, would it make sense to roll an rc3 to
pick them up.
Off the top of my head, at least one of the fixes was for a crash bug,
one was to maintain a feature we had in 1.2.x, and there were several
fixes for annoying bugs found by the GCC developers. I don't see any
reason to not have these fixes in 1.3.0.
There's one major fix remaining in status that hasn't been approved
yet (the JavaHL reentrant calls thing), and if some of the JavaHL
people could decide if that should or shouldn't be moved, that would
be great, but other than that we seem to be pretty much caught up on
the critical fixes waiting to be merged front.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by Branko Čibej <br...@xbc.nu>.
Barry Scott wrote:
>
> On Nov 21, 2005, at 06:07, Branko Čibej wrote:
>
>> Barry Scott wrote:
>>>
>>> On Nov 15, 2005, at 15:28, kfogel@collab.net wrote:
>>>
>>>> I already posted a complete list of changes in another thread. I'll
>>>> repost it in this thread, since I think some people might not see the
>>>> other one. Here is the list again:
>>>>
>>>
>>> Did the fixes to allow compiling with Windows .net compiler get in?
>> AFAIK a fix is only necessary if you use neon-0.25.3 or later. I'm
>> not aware of any such fixes being committed, so most probably 1.3.0
>> on Windows will still use neon-0.24. However, we don't have to wait
>> for 1.4 to upgrade Neon on Windows; so, theoretically, this could be
>> fixed in 1.3.1.
>
> Why is rc3 missing the fixed ezt files that I provided against rc2?
Because I didn't get around to committing them? I wish the subject of
that mail had said [PATCH}...
> Did you fixed up the generator as you said you hoped to do?
No. I started on a fix for APR, and have a prototype working, but I had
to drop it due to lack of time.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by Barry Scott <ba...@barrys-emacs.org>.
On Nov 21, 2005, at 06:07, Branko Čibej wrote:
> Barry Scott wrote:
>>
>> On Nov 15, 2005, at 15:28, kfogel@collab.net wrote:
>>
>>> I already posted a complete list of changes in another thread. I'll
>>> repost it in this thread, since I think some people might not see
>>> the
>>> other one. Here is the list again:
>>>
>>
>> Did the fixes to allow compiling with Windows .net compiler get in?
> AFAIK a fix is only necessary if you use neon-0.25.3 or later. I'm
> not aware of any such fixes being committed, so most probably 1.3.0
> on Windows will still use neon-0.24. However, we don't have to wait
> for 1.4 to upgrade Neon on Windows; so, theoretically, this could
> be fixed in 1.3.1.
Why is rc3 missing the fixed ezt files that I provided against rc2?
Did you fixed up the generator as you said you hoped to do?
Barry
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by Barry Scott <ba...@barrys-emacs.org>.
On Nov 21, 2005, at 06:07, Branko Čibej wrote:
> Barry Scott wrote:
>>
>> On Nov 15, 2005, at 15:28, kfogel@collab.net wrote:
>>
>>> I already posted a complete list of changes in another thread. I'll
>>> repost it in this thread, since I think some people might not see
>>> the
>>> other one. Here is the list again:
>>>
>>
>> Did the fixes to allow compiling with Windows .net compiler get in?
> AFAIK a fix is only necessary if you use neon-0.25.3 or later. I'm
> not aware of any such fixes being committed, so most probably 1.3.0
> on Windows will still use neon-0.24. However, we don't have to wait
> for 1.4 to upgrade Neon on Windows; so, theoretically, this could
> be fixed in 1.3.1.
With out the bug fixes in 0.25.3 svn lock run against an old server
prompts for a password rather then failing. Use the buggy neon seems
like
a bad idea.
Barry
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by Branko Čibej <br...@xbc.nu>.
Barry Scott wrote:
>
> On Nov 15, 2005, at 15:28, kfogel@collab.net wrote:
>
>> I already posted a complete list of changes in another thread. I'll
>> repost it in this thread, since I think some people might not see the
>> other one. Here is the list again:
>>
>
> Did the fixes to allow compiling with Windows .net compiler get in?
AFAIK a fix is only necessary if you use neon-0.25.3 or later. I'm not
aware of any such fixes being committed, so most probably 1.3.0 on
Windows will still use neon-0.24. However, we don't have to wait for 1.4
to upgrade Neon on Windows; so, theoretically, this could be fixed in 1.3.1.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by Barry Scott <ba...@barrys-emacs.org>.
On Nov 15, 2005, at 15:28, kfogel@collab.net wrote:
> I already posted a complete list of changes in another thread. I'll
> repost it in this thread, since I think some people might not see the
> other one. Here is the list again:
>
Did the fixes to allow compiling with Windows .net compiler get in?
Barry
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by kf...@collab.net.
Garrett Rooney <ro...@electricjellyfish.net> writes:
> So there appear to have been a fair number of fixes merged into the
> 1.3.x branch since rc2 was rolled. Since we haven't even officially
> released rc2 to the world yet, would it make sense to roll an rc3 to
> pick them up.
>
> Off the top of my head, at least one of the fixes was for a crash bug,
> one was to maintain a feature we had in 1.2.x, and there were several
> fixes for annoying bugs found by the GCC developers. I don't see any
> reason to not have these fixes in 1.3.0.
>
> There's one major fix remaining in status that hasn't been approved
> yet (the JavaHL reentrant calls thing), and if some of the JavaHL
> people could decide if that should or shouldn't be moved, that would
> be great, but other than that we seem to be pretty much caught up on
> the critical fixes waiting to be merged front.
I already posted a complete list of changes in another thread. I'll
repost it in this thread, since I think some people might not see the
other one. Here is the list again:
--------------------8-<-------cut-here---------8-<-----------------------
Here are the significant changes merged into 1.3.x since RC2 was
rolled from r17100. Do any seem worthy of an RC3 to us?
* r17116 [merge of r17115]
Tweak some text regarding the WC format change from r16855.
* r17120 [merge of r16360 - r17117]
Norwegian messages from trunk.
* r17135 [merge of r17131]
Fix some useless stat'ing and reading of property files on commits
* r17152 [merge of r17127]
Fix the JavaHL status API for retrieving last committed revision.
* r17153 [merge of r17129]
Fix broken tests in svnmerge_test.py by disabling keyword expansion
* r17155 [merge of r17039, r17051]
Fix issue #767, where commit used a temporary file in the current
working directory (thus interfering with commits in read-only
directories).
* r17157 [merge of r17150]
Fix a bug that caused anonymous write access over ra_svn to fail
with "Authorization failed" errors.
* r17161 [merge of r17149]
Remove the possibility of a buffer overflow in keyword expansion.
* r17229 [merge of r16568, r17123]
Make svn info not error out on paths that don't exist in future revs.
* r17297 [merge of r17257]
Emit a workaround for bugs in SWIG 1.3.25 and earlier which cause
SWIG_GetModule to be called with the wrong number of arguments.
* r17309 [merge of r17190, r17271]
Fix a bug in diffing schedule delete files.
* r17350 [merge of r17214]
Fix crashes caused by reentrant calls into the Java bindings, and
closing an initialisation-time race condition.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by "Peter N. Lundblad" <pe...@famlundblad.se>.
On Mon, 14 Nov 2005, Garrett Rooney wrote:
> So there appear to have been a fair number of fixes merged into the
> 1.3.x branch since rc2 was rolled. Since we haven't even officially
> released rc2 to the world yet, would it make sense to roll an rc3 to
> pick them up.
>
+1. I'm ready to sign the non-windows tarballs...
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Time for 1.3.0-rc3?
Posted by David James <ja...@gmail.com>.
On 11/14/05, Garrett Rooney <ro...@electricjellyfish.net> wrote:
> So there appear to have been a fair number of fixes merged into the
> 1.3.x branch since rc2 was rolled. Since we haven't even officially
> released rc2 to the world yet, would it make sense to roll an rc3 to
> pick them up.
>
> Off the top of my head, at least one of the fixes was for a crash bug,
> one was to maintain a feature we had in 1.2.x, and there were several
> fixes for annoying bugs found by the GCC developers. I don't see any
> reason to not have these fixes in 1.3.0.
r17257 is an important fix, which is necessary in order for our SWIG
bindings to compile with GCC 2.95.x.
> There's one major fix remaining in status that hasn't been approved
> yet (the JavaHL reentrant calls thing), and if some of the JavaHL
> people could decide if that should or shouldn't be moved, that would
> be great, but other than that we seem to be pretty much caught up on
> the critical fixes waiting to be merged front.
r17279 is also quite important and is still waiting in STATUS. This
fix removes recursive symbolic links from our build tree, thus solving
two issues:
1. Python recursively imports modules, such that "libsvn.core"
imports "libsvn.libsvn.core", and "libsvn.libsvn.core" imports
"libsvn.libsvn.libsvn.core", and so on.
2. The Subversion build tree can't be copied from one machine to
another using "scp", because "scp" enters into an infinite copying
loop when it sees recursive symbolic links.
I think it would be a very good idea to include this fix in 1.3.0.
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james
Re: Time for 1.3.0-rc3?
Posted by David James <ja...@gmail.com>.
On 11/14/05, Garrett Rooney <ro...@electricjellyfish.net> wrote:
> So there appear to have been a fair number of fixes merged into the
> 1.3.x branch since rc2 was rolled. Since we haven't even officially
> released rc2 to the world yet, would it make sense to roll an rc3 to
> pick them up.
>
> Off the top of my head, at least one of the fixes was for a crash bug,
> one was to maintain a feature we had in 1.2.x, and there were several
> fixes for annoying bugs found by the GCC developers. I don't see any
> reason to not have these fixes in 1.3.0.
r17257 is an important fix, which is necessary in order for our SWIG
bindings to compile with GCC 2.95.x.
> There's one major fix remaining in status that hasn't been approved
> yet (the JavaHL reentrant calls thing), and if some of the JavaHL
> people could decide if that should or shouldn't be moved, that would
> be great, but other than that we seem to be pretty much caught up on
> the critical fixes waiting to be merged front.
r17279 is also quite important and is still waiting in STATUS. This
fix removes recursive symbolic links from our build tree, thus solving
two issues:
1. Python recursively imports modules, such that "libsvn.core"
imports "libsvn.libsvn.core", and "libsvn.libsvn.core" imports
"libsvn.libsvn.libsvn.core", and so on.
2. The Subversion build tree can't be copied from one machine to
another using "scp", because "scp" enters into an infinite copying
loop when it sees recursive symbolic links.
I think it would be a very good idea to include this fix in 1.3.0.
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james