A few days ago, Jonas proposed that we make the 0.11 release this
Sunday. I'm OK with that as I think we don't have any critical issue
left. I'll be able to assist him a bit during the afternoon, but if
anyone could try to test the current 0.11-stable status (r7219) a bit
more before that day, all the better.
Then, a few notes on 0.11.1 maintenance:
* Too bad I didn't have time to test the mysql_unicode.patch on #4378,
as it looks very promising. That should go in 0.11.1 just after the
release, I think.
* upgrade to jQuery 1.2.6 - seems to work fine, but didn't test it
enough for 0.11, so also something to do just after the release.
* the new pool code (already in trunk) could also go in 0.11.1 just
after the release
-- Christian
If I'm online at the time you're doing the release I'm happy to help.
2008/6/20 Christian Boos <cb...@neuf.fr>:
--
Evolution: Taking care of those too stupid to take care of themselves.
I ran for a couple of days Trac 0.11rc1 and SVN 1.5rc (7 or 9, can't
remember) and it seemed to work properly.
However, I would not recommend to upgrade any production system to SVN
1.5 right now, you may want to wait for a couple of weeks to verify
that it is mature enough.
FWIW, I've crashed the 1.5 final client using ra_serf this morning,
checking out a large changeset (thousands of files). Ok, ra_serf is
not used in Trac (this is a library for remote client access)...
Cheers,
Manu
More seriously, take GCC: you're more likely to find bugs with a 4.Y.0
release than with 4.X.n (X < Y): the code is more mature.
> I'd expect the (several) release candidates to have exposed any show-
> stopper issues. Then again, i'm not that familiar with subversion's
> development process. Really think I should wait? I'd rather upgrade it
> all now, and not have to revisit this issue again in some week.
It's not about SVN: it is a general remark. This applies to GCC, it
applies to M$ ServicePacks or Apple Software Update and virtually to
every piece of software.
> I agree that crash you've found in 1.5 final is not very
> inspiring... :(
I think it is due to the new "serf" library, but I have no time for
now trying to reproduce this issue.
I moved back to SVN 1.5 with the old library (Neon) and it worked as a
charm. Guess what: libserf has been introduced w/ SVN 1.5, so it's
support is less mature. If you remember the "old" days of SVN, when
FSFS has been introduced in an official release (maybe around SVN
1.1): the authors were confident in FSFS, but they wrote that the BDB
backend was still more mature.
Even if SVN code was 100% bug free, there may have compatibility
issues coming from dependencies. There are tons of dependencies, which
vary from one OS version to another, from one distribution to another,
from one platform to another.
The question is: do you really want to experience a compatibility
issue with a brand new official version on your production system, or
can you wait for a couple of weeks till someone else bumps (or not)
into the issue and reports it ?
You can be both guys, but I'm not sure you want to take the chance
with a production system; however you can try it for yourself on a
test system...
Cheers,
Manu
I am in a similar situation to you, I am planning on upgrading our
install and am planning to roll out svn 1.5 and trac 0.11 at the same
time. I know that the Tigris servers had been running pre-release code
for quiet some time (months I think) prior to the 1.5 release cycle.
2008/6/20 Filipe Correia <fcor...@gmail.com>:
> I agree that crash you've found in 1.5 final is not very
> inspiring... :(
I wouldn't worry about that too much, ra_serf has experimental support
only (currently), same as in 1.4, as such it shouldn't be used in a
production environment.
http://subversion.tigris.org/svn_1.5_releasenotes.html#dav-modules
Cheers,
Roo.