1.9.2 Beta Freeze Status
Damon Sicore
Just a quick update on freeze schedule. At this morning's dev
meeting, we decided that it was reasonable to freeze for beta tonight
(11/6) as soon as the last P1 blocker was landed. On deck was a
series of JS bugs that were seemingly in reach for landing tonight and
the JEP landing. At least two issues have popped up that may/probably
will cause a delay:
* Smichaud encountered an issue with the JEP patch bitrotting (bug
517355), and now he's seeing JS issues that are yet to be revealed.
He's indicated he will work on it more, and we're hoping to have more
details soon so others (in JS) can take a look.
* There's a perf regression in tracing getters that under attack by
jorendorff. See bug 517909. At this point, we may not have that fix
by end of day.
There's still a bit of churn in other JS bugs, but I haven't
encountered any show stoppers. Still, we'll need the above. If
others know of more, please bring them up!
Sincerely,
Damon
Mike Beltzner
We were hoping for a code freeze on mozilla-1.9.2 tonight in order to
start builds tomorrow, but that's no longer looking likely. We have
five blockers remaining:
https://bugzilla.mozilla.org/show_bug.cgi?id=517355 - restoring OJI to
OSX
- smichaud is still bisecting to figure out the bitrot
https://bugzilla.mozilla.org/show_bug.cgi?id=517909 - SunSpider
regression
- seems to be related to tracing getters
- jorendorff is continuing his analysis
https://bugzilla.mozilla.org/show_bug.cgi?id=520444 - update plugin
check URL
- Blair needs to write a simple patch, have it reviewed and landed
https://bugzilla.mozilla.org/show_bug.cgi?id=520801 - taskbar previews
don't work
- Jim Mathies is looking into a problem where some Windows 7 users
aren't seeing the Aero Peek taskbar previews
https://bugzilla.mozilla.org/show_bug.cgi?id=520837 - Aero Peek perf
regression
- Rob thinks he knows what this is, Jimm's helping figure it out
We also have two bugs aside from the TraceMonkey merge which need to
land on mozilla-1.9.2:
https://bugzilla.mozilla.org/show_bug.cgi?id=514435
https://bugzilla.mozilla.org/show_bug.cgi?id=513186
cheers,
mike
Blake Kaplan
> https://bugzilla.mozilla.org/show_bug.cgi?id=514435
For the record, I landed this today and am helping out with the
LiveConnect/OJI bitrot.
--
Blake Kaplan
Damon Sicore
For the record, you rock! :)
> --
> Blake Kaplan
> _______________________________________________
> dev-planning mailing list
> dev-pl...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-planning
John J. Barton
> There's still a bit of churn in other JS bugs, but I haven't encountered
> any show stoppers. Still, we'll need the above. If others know of
> more, please bring them up!
We are trying to get Firebug 1.5b1 for FF 3.6b1, but
https://bugzilla.mozilla.org/show_bug.cgi?id=520421
will make the JS debugging UI partly broken.
It may be worth looking into the cause: the problem we see is in normal
HTML/CSS in Firebug's panel (browser) element, so it may well impact web
pages.
jjb
Johnathan Nightingale
Hmm - the recent comments in that bug seem to say that it's specific
to the new HTML5 parser, which is off by default. Does that match your
experience for the FB bustage?
Cheers,
J
---
Johnathan Nightingale
Human Shield
joh...@mozilla.com
John J. Barton
No, we don't change any Firefox settings in our tests.
jjb
John J. Barton
I tested 3.6 nightly and this bug does not occur. html5.enable is false.
jjb
Henrik Skupin
>> No, we don't change any Firefox settings in our tests.
>
> I tested 3.6 nightly and this bug does not occur. html5.enable is false.
I'm on it. You will have the regression range in a couple of hours
(depending on the build time with hg bisect)..
Henrik
Mike Beltzner
I think you misundersood John; he's saying that the bug does *not* occur when the HTML5 parser is disabled, which is the default. He also un-nominated the bug in question for blocking, so I think this turned out to be an issue specific to the HTML5 parser after all.
cheers,
mike
Henrik Skupin
>>> I tested 3.6 nightly and this bug does not occur. html5.enable
>>> is
>> false.
>>
>> I'm on it. You will have the regression range in a couple of hours
>> (depending on the build time with hg bisect)..
>
> I think you misundersood John; he's saying that the bug does *not*
> occur when the HTML5 parser is disabled, which is the default. He
> also un-nominated the bug in question for blocking, so I think this
> turned out to be an issue specific to the HTML5 parser after all.
See bug 520421. Given by John it is a trunk only issue and happens with
html.enabled = false. I made a hg bisect after I was able to reproduce
its original issue. The result is listed on bug 520421. Nothing which
blocks us on 1.9.2.
Henrik
Tim Riley
landed yet:
https://bugzilla.mozilla.org/show_bug.cgi?id=517355 - restoring OJI to OSX
- smichaud is still bisecting to figure out the bitrot
https://bugzilla.mozilla.org/show_bug.cgi?id=517909 - SunSpider regression
- seems to be related to tracing getters
- jorendorff is continuing his analysis
https://bugzilla.mozilla.org/show_bug.cgi?id=520837 - Aero Peek perf
regression
- Rob thinks he knows what this is, Jimm's helping figure it out
Did this tracemonkey one get landed on 1.9.2?
https://bugzilla.mozilla.org/show_bug.cgi?id=513186
Neal said he checked it in, but I don't see the change on
http://hg.mozilla.org/releases/mozilla-1.9.2/
--Tim
Tim Riley
landed yet:
https://bugzilla.mozilla.org/show_bug.cgi?id=517355 - restoring OJI to OSX
- smichaud is still bisecting to figure out the bitrot
https://bugzilla.mozilla.org/show_bug.cgi?id=517909 - SunSpider regression
- seems to be related to tracing getters
- jorendorff is continuing his analysis
https://bugzilla.mozilla.org/show_bug.cgi?id=520837 - Aero Peek perf
regression
- Rob thinks he knows what this is, Jimm's helping figure it out
Did this tracemonkey one get landed on 1.9.2?
Reed Loden
Tim Riley <ti...@mozilla.com> wrote:
> Did this tracemonkey one get landed on 1.9.2?
> https://bugzilla.mozilla.org/show_bug.cgi?id=513186
> Neal said he checked it in, but I don't see the change on
> http://hg.mozilla.org/releases/mozilla-1.9.2/
That bug has nothing to do with tracemonkey, but it was indeed
committed to the 1.9.2 branch:
http://hg.mozilla.org/releases/mozilla-1.9.2/rev/812c49feab66
~reed
--
Reed Loden - <re...@reedloden.com>
Rob Arnold
> Looking for a quick update on the 1.9.2 freeze. These don't seem to be
> landed yet:
> https://bugzilla.mozilla.org/show_bug.cgi?id=520837 - Aero Peek perf
> regression
> - Rob thinks he knows what this is, Jimm's helping figure it out
>
Jim has a patch that he thinks might fix the regression. We think that since
the talos machines reboot each cycle, their registry cache is cold so the
CoCreateInstance call takes a while because it has to hit the disk.
Workaround is to insert a (cached) version check before the call. Fix was
reviewed and should be landing on m-c to see if it fixes the regression.
-Rob
Damon Sicore
On Oct 9, 2009, at 10:52 AM, Tim Riley wrote:
> Looking for a quick update on the 1.9.2 freeze. These don't seem to
> be landed yet:
>
> https://bugzilla.mozilla.org/show_bug.cgi?id=517355 - restoring OJI
> to OSX
> - smichaud is still bisecting to figure out the bitrot
More patches for the above have been produced and are being attached
to the bug right now. They need to be reviewed by gal/bkap.
According to smichaud, they fix all remaining issues.
>
> https://bugzilla.mozilla.org/show_bug.cgi?id=517909 - SunSpider
> regression
> - seems to be related to tracing getters
> - jorendorff is continuing his analysis
Boris has been helping with this for the past 1.5 days or so. They
haven't been able to reproduce the regression since Tuesday (sorta).
Still digging/profiling, etc. Please stand by.
>
> https://bugzilla.mozilla.org/show_bug.cgi?id=520837 - Aero Peek perf
> regression
> - Rob thinks he knows what this is, Jimm's helping figure it out
>
> Did this tracemonkey one get landed on 1.9.2?
> https://bugzilla.mozilla.org/show_bug.cgi?id=513186
> Neal said he checked it in, but I don't see the change on http://hg.mozilla.org/releases/mozilla-1.9.2/
>
> --Tim
> _______________________________________________
> Release-drivers mailing list
> Release...@mozilla.org
> https://mail.mozilla.org/listinfo/release-drivers
Blake Kaplan
> More patches for the above have been produced and are being attached
> to the bug right now. They need to be reviewed by gal/bkap.
> According to smichaud, they fix all remaining issues.
smichaud landed this on 1.9.2 at around 1 today so liveconnect is officially
back in action. For some reason, I feel a strong need to take a shot of
tequila (but at least this is no longer blocking the beta!).
--
Blake Kaplan
Damon Sicore
On Oct 9, 2009, at 12:21 PM, Damon Sicore wrote:
>>
>>
>> https://bugzilla.mozilla.org/show_bug.cgi?id=517909 - SunSpider
>> regression
>> - seems to be related to tracing getters
>> - jorendorff is continuing his analysis
>
> Boris has been helping with this for the past 1.5 days or so. They
> haven't been able to reproduce the regression since Tuesday
> (sorta). Still digging/profiling, etc. Please stand by.
Andreas was called in here, did some analysis w/ Boris and others.
They have an idea and a patch for the idea. Keep standing by. Still
hacking.
Damon Sicore
And! The patch leaks. Andreas is looking into it.