Preparing next stable release - 0.11.5
Christian Boos
As some of you may have noticed, the Milestone page for 0.11.5 says
"06/23/09", i.e. tomorrow.
That timing was based on a 3 month interval between point releases, when
we decided to go for a time-based release schedule.
The actual interval between 0.11.3 and 0.11.4 was a bit shorter, so this
gave us more time for 0.11.5. But now, here we are.
This release is quite interesting, there are quite a number of new
features along the way:
- better svn support: support for Subversion 1.6, a memory leak fix,
and more
- usability improvements: icons for attachment links for direct
download, provide feedback notices after POSTs operations
- new --http11 option for HTTP/1.1 support in tracd; also a few missing
Content-Length response headers were added
- new "[mimeview] treat_as_binary" setting
- support for automatic backups of Trac database during upgrade, beyond
sqlite
- test infrastructure improvements (rework-testing, with support for
alternate database backends)
See
http://trac.edgewall.org/query?status=closed&group=component&order=priority&milestone=0.11.5
for more.
Part of the "better svn support", we will also shortly add support for
improved display of the 1.5 svn:mergeinfo property (#7715) and support
for the 1.5 style of svn:externals (#7687). There are also a number of
tickets with patches that should still go in, for fixing small issues
(e.g. #8358, #8283, #8378, this kind).
This is all well and good, with a significant number of people having
contributed various patches. Thanks!
However ...
- We're still pretty clueless about the often reported major performance
regression with Trac and Apache (#7490). The last hint to date leads to
mod_deflate... Could that be related to the Content-Length response
header not being set in some cases, and mod_deflate having issues with
that? If this is the case, then 0.11.5 should fix it (the #8020 related
fixes on the Content-Length header).
- There are still a number of svn authz issues remaining opened (#4503,
#5097, #5640, #8289)
- Certainly a good fraction of the remaining 32 opened tickets will need
to be moved to a later milestone
I don't think the above points should prevent us from releasing 0.11.5,
but is there anything like a show-stopper that I missed?
Now for the next steps, I think the tasks already mentioned above
(#7715, #7687, the other smaller tickets with a patch) can be finalized
this week. Then an upgrade of t.e.o could be done this week-end and a
"freeze" and testing period of one week could follow, as we did with
quite some success with 0.11.4(rc1,rc2), to avoid the usual Python 2.3
breakage ;-)
I won't be available during the first 3 weeks of July, but I'll try to
make 0.11.5 ready for the freeze period, until then.
Besides, there are quite a lot of tickets waiting on 0.11.6. Should I do
the usual, first move the 0.11.6 tickets to 0.11.7 and move the
unfinished 0.11.5 tickets to 0.11.6? Or is it time to consider 0.11.5 as
one of the last 0.11 releases and start focus on 0.12? (there will be
probably a 0.11.6, but do we really want a 0.11.7?). We could therefore
move the current 0.11.6 tickets to 0.12.1, as the plate is already full
for 0.12 proper.
-- Christian
PS: mod_wsgi 3.0 reached rc3 today, so it would be nice if somebody
could check if it is working fine with the latest 0.11-stable; same goes
for Subversion 1.6.3, which is now in pre-release
Remy Blank
> Now for the next steps, I think the tasks already mentioned above
> (#7715, #7687, the other smaller tickets with a patch) can be finalized
> this week.
Ok from my side for #7715, I'm almost done with the 0.11-stable backport.
> Then an upgrade of t.e.o could be done this week-end and a
> "freeze" and testing period of one week could follow, as we did with
> quite some success with 0.11.4(rc1,rc2), to avoid the usual Python 2.3
> breakage ;-)
Sounds good. How about tracking the release testing in a ticket, as I
suggested in ReleaseTesting[1]? As soon as rc1 is ready, we can put all
the people on that page in CC, and they can report back when they are
done. I'll clean up the page tomorrow.
[1] http://trac.edgewall.org/wiki/TracDev/ReleaseTesting
> We could therefore
> move the current 0.11.6 tickets to 0.12.1, as the plate is already full
> for 0.12 proper.
+1 from my side, I think it's high time that we start thinking about
planning the 0.12 release. Also, this will finally allow us to drop 2.3
for good.
-- Remy
Christian Boos
> Christian Boos wrote:
>
>> Now for the next steps, I think the tasks already mentioned above
>> (#7715, #7687, the other smaller tickets with a patch) can be finalized
>> this week.
>>
>
> Ok from my side for #7715, I'm almost done with the 0.11-stable backport.
>
Great, testing right now in our production setup, seems to be working
fine (although we don't have yet lots of svn:mergeinfo, this is going to
change soon ;-) ).
>> Then an upgrade of t.e.o could be done this week-end and a
>> "freeze" and testing period of one week could follow, as we did with
>> quite some success with 0.11.4(rc1,rc2), to avoid the usual Python 2.3
>> breakage ;-)
>>
>
> Sounds good. How about tracking the release testing in a ticket, as I
> suggested in ReleaseTesting[1]? As soon as rc1 is ready, we can put all
> the people on that page in CC, and they can report back when they are
> done. I'll clean up the page tomorrow.
>
> [1] http://trac.edgewall.org/wiki/TracDev/ReleaseTesting
>
Done: http://trac.edgewall.org/ticket/8412
Please don't hesitate to modify the description as you see fit.
>> We could therefore
>> move the current 0.11.6 tickets to 0.12.1, as the plate is already full
>> for 0.12 proper.
>>
>
> +1 from my side, I think it's high time that we start thinking about
> planning the 0.12 release. Also, this will finally allow us to drop 2.3
> for good.
>
Fine!
-- Christian
Felix Schwarz
May I suggest also including #8240? The one-line fix I gave in the
ticket should be enough - at least for issue reported in the ticket.
fs
Christian Boos
> May I suggest also including #8240? The one-line fix I gave in the
> ticket should be enough - at least for issue reported in the ticket.
>
Sure, that one was part of the "(e.g. #8358, #8283, #8378, this kind)"
list ;-)
I'll wait a bit more for feedback about the proposed move of the 0.11.6
tickets to 0.12.1, in case someone has a better idea. After that I'll
move the 0.11.5 tickets to 0.11.6 and put back those I think we can
still complete for 0.11.5. At that point, things will be clearer.
-- Christian
Christian Boos
> Felix Schwarz wrote:
>
>> May I suggest also including #8240? The one-line fix I gave in the
>> ticket should be enough - at least for issue reported in the ticket.
>>
>
> Sure, that one was part of the "(e.g. #8358, #8283, #8378, this kind)"
> list ;-)
>
Oh sorry, when writing the above, I had #8340 in mind, not #8240 (soo
many tickets, so few digits...).
#8240 looked a bit more complex last time I looked at it, but if you're
confident the fix is enough... In all cases, a few unit tests here would
definitely help, both for making the issue easier to understand and for
giving us more confidence on maintaining your patch (er, code change
suggestion here, as there's no patch ;-) ).
> I'll wait a bit more for feedback about the proposed move of the 0.11.6
> tickets to 0.12.1, in case someone has a better idea. After that I'll
> move the 0.11.5 tickets to 0.11.6 and put back those I think we can
> still complete for 0.11.5. At that point, things will be clearer.
>
It's done now.
All: feel free to go through the 0.11.6 list and move back the tickets
you think you can finish up this week.
If you think a given ticket is really important to get fixed for 0.11.5
but you're not sure you can do it yourself this week, please speak up on
the list as well.
-- Christian
Shane Caraveo
>
> It's done now.
>
> All: feel free to go through the 0.11.6 list and move back the tickets
> you think you can finish up this week.
> If you think a given ticket is really important to get fixed for 0.11.5
> but you're not sure you can do it yourself this week, please speak up on
> the list as well.
btw, there is one small bug that is a bit of a problem. ticket 8416
Shane
Remy Blank
> btw, there is one small bug that is a bit of a problem. ticket 8416
Fixed three months ago:
http://trac.edgewall.org/changeset/8061
-- Remy
Christian Boos
Some updates about the upcoming 0.11.5 release...
Last week and this week-end, things have progressed well. Let's look at
the remaining items.
There's still a minor glitch in #7715 that I hope I can fix this
evening, as well as do the usual round of update for the default wiki pages.
While testing, I also found another minor problem with the MySQL backend
(http://trac.edgewall.org/ticket/8424).
The dreaded #7490 ticket will either be reported as working with
0.11.5rc1 and then closed, or we'll postpone the ticket to a later release.
So I think tomorrow evening is a realistic target for the rc1. It'd be
nice to hear from Jonas if he could actually assist in making the rc1
packages available, otherwise we could perhaps simply attach them on the
TracDownload page.
-- Christian
Jonas Borgström
Hi All,
I've actually not fallen off the face of the earth yet so I think I'll
be able to package the release this evening :)
But I've not been able to follow the lists that well for a while so I
have some catching up to do. Please don't hesitate to resend messages to
me if you haven't received any response...
Regards,
Jonas
(I'll be available on #trac today)
Christian Boos
> On 6/29/09 12:14 PM, Christian Boos wrote:
>
>> Hello,
>>
>> Some updates about the upcoming 0.11.5 release...
>> Last week and this week-end, things have progressed well. Let's look at
>> the remaining items.
>>
>> There's still a minor glitch in #7715 that I hope I can fix this
>> evening, as well as do the usual round of update for the default wiki pages.
>> While testing, I also found another minor problem with the MySQL backend
>> (http://trac.edgewall.org/ticket/8424).
>> The dreaded #7490 ticket will either be reported as working with
>> 0.11.5rc1 and then closed, or we'll postpone the ticket to a later release.
>>
>> So I think tomorrow evening is a realistic target for the rc1. It'd be
>> nice to hear from Jonas if he could actually assist in making the rc1
>> packages available, otherwise we could perhaps simply attach them on the
>> TracDownload page.
>>
>
> Hi All,
>
> I've actually not fallen off the face of the earth yet so I think I'll
> be able to package the release this evening :)
>
Great!
Well, I won't myself be available this evening before late (I'll try to
pop up on #trac circa 10 PM for a short while), but here's my take on
the latest status:
- #8424: I think we can commit the simple fix for the first problem,
the second problem is more involved and is for 0.11.6 at the very least
(I have a fix in preparation)
- #7715: fix svn:mergeinfo for scoped repositories: suggesting making
it a known issue for rc1. II will try to fix it in the coming days but
can't be ready for this evening
- #8412 and subprocess issue with Python 2.3: pick the proposed solution 3
Otherwise, all the other testing I made didn't show any nasty issues
(well, except #7823, but that's also a kind of known issue...) and I'm
sure 0.11.5 will rock ;-)
-- Christian