Closed Bug 571443 Opened 10 years ago Closed 8 years ago

"get_fix_stack" step should not use obsolete fix-macosx-stack.pl (got "Bad output from otool" on Mac64)

Categories

(Release Engineering :: General, defect, P2)

x86
macOS
defect

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: jruderman, Assigned: coop)

References

()

Details

(Whiteboard: [disabled on buildbot-0.8.0 branch] [automation])

Attachments

(2 files)

http://mxr.mozilla.org/build/source/buildbotcustom/process/factory.py#842

When fix_stack_using_bpsyms.py is ready (bug 570287), get_fix_stack should be changed to use it instead of fix-%s-stack.pl.

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1276230745.1276232213.506.gz&fulltext=1

If this red is frequent, maybe we could temporarily skip the stack fixer, and use "cp" instead.  (Don't use fix_macosx_stack.py, it's busted on 64-bit too.)
I wonder why this suddenly became a frequent red.
Priority: -- → P3
Whiteboard: [automation]
Nothing for RelEng to do here until bug 570287 is fixed, correct?
You could skip stack-fixing entirely for now (see comment 0).
(In reply to comment #14)
> You could skip stack-fixing entirely for now (see comment 0).

Just on OSX 10.6 64-bit?
Yeah, that sounds reasonable.
I can't really test this because the alive test keeps failing prior to this....but it looks like it'll work!
Attachment #452828 - Flags: review?(nrthomas)
Comment on attachment 452828 [details] [diff] [review]
copy sdleak.tree.raw to sdleak.tree
[Checkin: Comment 21]

r+. Bustage at compare_sdleak_tree may be entertaining, assuming we ever get that far.
Attachment #452828 - Flags: review?(nrthomas) → review+
dbaron, any thoughts on using sdleak.tree.raw in place of sdleak.tree here?
Is the difference between the two the lack of fix-macosx-stack being run on them?  If fix-macosx-stack worked, that would make the output significantly less useful, but not entirely useless.  But given that it doesn't, sounds fine.
Attachment #452828 - Flags: checked-in+
Comment on attachment 452828 [details] [diff] [review]
copy sdleak.tree.raw to sdleak.tree
[Checkin: Comment 21]

buildbot-0.8.0 branch:
http://hg.mozilla.org/build/buildbotcustom/rev/3dec35ce1e44
Attachment #452828 - Attachment description: copy sdleak.tree.raw to sdleak.tree → copy sdleak.tree.raw to sdleak.tree [Checkin: Comment 21]
Whiteboard: [automation] → [disabled on buildbot-0.8.0 branch] [automation]
(In reply to comment #0)
> When fix_stack_using_bpsyms.py is ready (bug 570287), get_fix_stack should be
> changed to use it instead of fix-%s-stack.pl.
> 
> (Don't use fix_macosx_stack.py, it's busted on 64-bit too.)

Bug 570287 has been fixed 5 months ago.
Can fix_stack_using_bpsyms.py be used now?
Assignee: nobody → coop
(In reply to Serge Gautherie (:sgautherie) from comment #22) 
> Bug 570287 has been fixed 5 months ago.
> Can fix_stack_using_bpsyms.py be used now?

Given that this has sat for a further 18 months now (sorry about that), is this still needed/wanted?
Probably.  But I'm not seeing any output from diffbloatdump.pl in the logs I look at, so it's hard to be sure.

https://tbpl.mozilla.org/php/getParsedLog.php?id=12484772&tree=Firefox
Status: NEW → ASSIGNED
Priority: P3 → P2
Attachment #632292 - Flags: review?(catlee) → review+
Comment on attachment 632292 [details] [diff] [review]
Use fix_stack_using_bpsyms.py script instead of fix-%s-stack.pl

https://hg.mozilla.org/build/buildbotcustom/rev/b20f1a16297b
Attachment #632292 - Flags: checked-in?
deployed to production
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Comment on attachment 632292 [details] [diff] [review]
Use fix_stack_using_bpsyms.py script instead of fix-%s-stack.pl

https://hg.mozilla.org/build/buildbotcustom/rev/b20f1a16297b
Attachment #632292 - Flags: checked-in? → checked-in+
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.