a7 Status
Mike Schroepfer
That was a fun night! I haven't seen bonsai and tbox run out of memory
before...
Below is my understanding of our current status. I'd like to try and
wrap this up today if we can. Can I get everyones help with the following?
a) Correct anything I've got wrong here
b) If you are pre-approved for landing below can you make sure to
coordinate with others to land in free window today. BKap I think you
are up first...
c) Any opinions on the items asking for approval, the xml:id perf
regression, or cross-site XMLHttpRequest. Options for each of these:
i) hold a7 and get them in
ii) move them to 1.9.1
iii) let them slip into a8
Best,
Schrep
Landed:
1) URLClassifier
2) doc.all changes
3) oncopy/onpaste
4) SVG lighting filters
5) Full zoom (4821)
6) CSS 'text-rendering'
Ready to land and pre-approved:
1) cross-origin wrappers https://bugzilla.mozilla.org/show_bug.cgi?id=367911
2) showModalDialog() https://bugzilla.mozilla.org/show_bug.cgi?id=194404
3) XPATH generator: https://bugzilla.mozilla.org/show_bug.cgi?id=319768
Asking for approval:
1) SVG tile filter: https://bugzilla.mozilla.org/show_bug.cgi?id=373572
** My $.02 is ok
2) https://bugzilla.mozilla.org/show_bug.cgi?id=319455
** My $.02 is ok
3) history menu is slowhttps://bugzilla.mozilla.org/show_bug.cgi?id=385397
** My $.02 is we really want this
Patch not reviewed:
1) Cross-site XMLHttpRequest
https://bugzilla.mozilla.org/show_bug.cgi?id=389508
Backed out because of perf regression:
1) xml:id https://bugzilla.mozilla.org/show_bug.cgi?id=275196
Mike Schroepfer
smaug
> Hey Folks,
>
> That was a fun night! I haven't seen bonsai and tbox run out of memory
> before...
>
> Below is my understanding of our current status. I'd like to try and
> wrap this up today if we can. Can I get everyones help with the following?
>
> a) Correct anything I've got wrong here
> b) If you are pre-approved for landing below can you make sure to
> coordinate with others to land in free window today. BKap I think you
> are up first...
> c) Any opinions on the items asking for approval, the xml:id perf
> regression, or cross-site XMLHttpRequest. Options for each of these:
> i) hold a7 and get them in
> ii) move them to 1.9.1
> iii) let them slip into a8
>
> Best,
>
> Schrep
I have a bit cleaned-up patch for xml:id waiting for review. It may
speed up some things which are called a lot. The tp regression wasn't
very bad, so might be worth to try. (Or was it actually only tp2
regression. Tp is up anyway on bl-bldlnx03)
-Olli
Mike Schroepfer
Still left:
> Ready to land and pre-approved:
> 3) XPATH generator: https://bugzilla.mozilla.org/show_bug.cgi?id=319768
>
> Asking for approval:
>
> 1) SVG tile filter: https://bugzilla.mozilla.org/show_bug.cgi?id=373572
> ** My $.02 is ok
> 2) https://bugzilla.mozilla.org/show_bug.cgi?id=319455
> ** My $.02 is oks
>
> Patch not reviewed:
>
> 1) Cross-site XMLHttpRequest
> https://bugzilla.mozilla.org/show_bug.cgi?id=389508
>
> Backed out because of perf regression:
>
> 1) xml:id https://bugzilla.mozilla.org/show_bug.cgi?id=275196
And:
regression: https://bugzilla.mozilla.org/show_bug.cgi?id=389398
Cheers,
Schrep
smaug
land it before a8.
Jonas, if you have any ideas how to still improve the performance of the
patch, feel free to suggest.
-Olli
Mike Schroepfer
> If xml:id can't be get into a7, would it still be possible to try to
> land it before a8.
>
> Jonas, if you have any ideas how to still improve the performance of the
> patch, feel free to suggest.
>
Default answer is no. Why do we want this and what is the regression risk?
Best,
Schrep
Jesse Ruderman
https://bugzilla.mozilla.org/show_bug.cgi?id=389878
Gmail, Contacts and Settings links are broken
https://bugzilla.mozilla.org/show_bug.cgi?id=389428
Crash [@ gfxTextRun::CompressedGlyph::IsClusterStart] loading some
Bugzilla/Wikipedia pages
https://bugzilla.mozilla.org/show_bug.cgi?id=389969
There is no protocol handling dialog
https://bugzilla.mozilla.org/show_bug.cgi?id=389943
Loading microsummary generators throws an exception (breaks
microsummaries)
Jesse Ruderman
https://bugzilla.mozilla.org/show_bug.cgi?id=390083
Frequent crashes [@ XPCContext::GetRuntime] during GC after closing a
tab
With bug 389428, I can "dogfood" trunk the 40% of the time I'm using a
site other than Bugzilla and Wikipedia, but with bug 390083 I can't
dogfood it at all. I don't see how we can be a day away from
releasing a Gecko beta.
Justin Dolske
> These regressions should be considered for blocking A7:
I might also suggest:
https://bugzilla.mozilla.org/show_bug.cgi?id=389985
Bug 389985 - crash [@ XPCWrappedNativeScope::FindInJSObjectScope]
It's crashed ~20 times for me in the last two days, which is about as
many total crashes as I've had since early June. :(
Justin
smaug
We want it because it is a useful feature, requested by many users and
also 3rd party gecko embedders
(https://bugzilla.mozilla.org/show_bug.cgi?id=275196#c78), and because
of parity-opera.
Hard to say anything exact about regression risk. I'd say that the risk
is in performance, but the latest patch should be better.
-Olli
> Best,
>
> Schrep
Mike Schroepfer
We'll continue to hold the tree closed as we knockout the crashes.
Please flag bugs as blocking 1.9+ with a Target Milestone of mozilla1.9
M7. We will use that to track the bugs we went close before release.
Best,
Schrep
Mike Connor
https://bugzilla.mozilla.org/buglist.cgi?
query_format=advanced&target_milestone=Firefox+3
+M7&target_milestone=mozilla1.9+M7&resolution=---
&chfieldto=Now&known_name=Fx3%3F&query_based_on=Fx3%
3F&field0-0-0=flagtypes.name&type0-0-0=equals&value0-0-0=blocking-
firefox3%
2B&field0-0-1=flagtypes.name&type0-0-1=equals&value0-0-1=blocking1.9%2B
-- Mike
> _______________________________________________
> dev-planning mailing list
> dev-pl...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-planning
rocal...@gmail.com
> https://bugzilla.mozilla.org/show_bug.cgi?id=389428
> Crash [@ gfxTextRun::CompressedGlyph::IsClusterStart] loading some
> Bugzilla/Wikipedia pages
I think we should back out the patch in bug 385270; that should fix
these issues. I need to test that though.
Rob
rocal...@gmail.com
wrote:
I did the backout. Bugs 389428, 390050 and 390032 are fixed (well, at
least the critical crashing issues are fixed).
Rob
steve....@gmail.com
The backout of bug https://bugzilla.mozilla.org/show_bug.cgi?id=385270
has made https://bugzilla.mozilla.org/show_bug.cgi?id=386584 return.
(Another crasher when changing fontsize on www.shacknews.com)
rocal...@gmail.com
<steve.engl...@gmail.com> wrote:
> The backout of bughttps://bugzilla.mozilla.org/show_bug.cgi?id=385270
> has madehttps://bugzilla.mozilla.org/show_bug.cgi?id=386584return.
> (Another crasher when changing fontsize onwww.shacknews.com)
That's expected. I don't think it should block M7 release though.
Rob