Closed Bug 703017 Opened 13 years ago Closed 13 years ago

Incompatibility with Firefox 3.6.24pre and Mozmill 1.5.6

Categories

(Core :: XPConnect, defect)

Product:
Core
Component:
XPConnect
Version:
1.9.2 Branch
Type:
defect
Priority:
Not set
Severity:
major

Tracking

()

Status:
RESOLVED WORKSFORME

People

(Reporter: u279076, Unassigned)

References

(
URL
)

Details

(Keywords: regression, Whiteboard: [WFM?]) 

Starting on 2011-10-29 all tests with mozilla-1.9.2 have started failing. Vlad has done some investigation locally and found that it is reproducible using Firefox 3.6.25pre with Mozmill 1.5.6 but not with Mozmill 1.5.5.

There is a parallel issue going on with Firefox 3.6.*pre not being served updates (bug 702996), but I'm not convinced this is related. 

I have attached an example report of the failures.
Seems like the FTP directory contains both 3.6.24pre and 3.6.25pre:
ftp://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-1.9.2/

Not sure if this is related.
This is not a Mozmill regression. Not sure why Vlad is not able to reproduce it with Mozmill 1.5.5 but for me it's clearly also broken. Same applies to Mozmill 1.5.4. So I don't blame Mozmill here.

Further testing has shown that the issue does not appear with a 2011-10-28 build but started with 2011-10-29.

Pass: http://hg.mozilla.org/releases/mozilla-1.9.2/rev/60d2c636b778
Fail: http://hg.mozilla.org/releases/mozilla-1.9.2/rev/2d3d72d5868d

Pushlog: http://hg.mozilla.org/releases/mozilla-1.9.2/pushloghtml?fromchange=60d2c636b778&tochange=2d3d72d5868d

There are only two single check-ins and it really smells like the patch from Blake:

author	Blake Kaplan <mrbkap@gmail.com>
	Fri Oct 28 13:47:46 2011 -0700 (at Fri Oct 28 13:47:46 2011 -0700)
changeset 35208	2d3d72d5868d
parent 35207	9b11e087b691
child 35209	b5d60f395ce5
pushlog:	2d3d72d5868d
Fix bug 653926. r=jst a=LegNeato 


This is a serious regression on mozilla-1.9.2 which blocks all of our tests for Nightly builds with Mozmill. I'm not sure why this issue has not been seen when running our tests for the 3.6.24 release candidate builds. Could this be a nightly build issue and does not affect release builds?
Blocks: CVE-2011-3004
Severity: normal → blocker
blocking1.9.2: --- → ?
Component: Mozmill → XPConnect
Keywords: regression
Product: Testing → Core
QA Contact: mozmill → xpconnect
Version: unspecified → 1.9.2 Branch
Our tests are working again at least with the following nightly build:

Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.25pre) Gecko/20111110 Namoroka/3.6.25pre.

There are no newer nightly builds available yet, so I can't the most recently build. Blake, do you have any idea which check-in could have fixed it?
Severity: blocker → major
Whiteboard: [WFM?]
I have updated all 1.9.2 builds on qa-horus for our daily testrun. Please check todays results and mark as WFM if none of those failures appear anymore.
I'm not seeing any results in brasstacks passed the 16th. Is this a separate issue?
(In reply to Anthony Hughes, Mozilla QA (irc: ashughes) from comment #5)
> I'm not seeing any results in brasstacks passed the 16th. Is this a separate
> issue?

The Brasstacks couchdb ran itself out of disk space on the 16 or 17th, that is now fixed.  Long term, (next quarter) we are going to move this database away from couchdb and onto elastic search which we've been using for a while and found far more stable.
Todays results are clearly visible here:
http://mozmill-release.brasstacks.mozilla.com/#/functional/reports?branch=3.6&platform=All&from=2011-11-18&to=2011-11-18

No failures anymore. That means that another checkin has really fixed the problem which we have seen with the Oct 29th build.

Blake, if you can add a comment we would appreciate. For now I will mark this bug as WFM.
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → WORKSFORME
blocking1.9.2: ? → ---
You need to log in before you can comment on or make changes to this bug.