perma comm/mail/components/enterprisepolicies/tests/browser/browser_policy_disable_safemode.js | application crashed [@ MOZ_Crash(char const*, int, char const*)]
Categories
(Thunderbird :: General, defect)
Tracking
(thunderbird_esr102 unaffected)
Tracking | Status | |
---|---|---|
thunderbird_esr102 | --- | unaffected |
People
(Reporter: intermittent-bug-filer, Assigned: mconley)
References
Details
(Keywords: crash, intermittent-failure)
Crash Data
Attachments
(1 file)
Bug 1805619 - Check for MigrationUtils before calling into it in ResetProfile.sys.mjs. r?NeilDeakin!
48 bytes,
text/x-phabricator-request
|
Details | Review |
Filed by: mkmelin [at] thunderbird.net
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=399595247&repo=comm-central
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/YjvD4LspRzSn6vjVsKS0MQ/runs/0/artifacts/public/logs/live_backing.log
Comment 1•2 years ago
|
||
Perma fail. Must be something from https://hg.mozilla.org/mozilla-central/pushloghtml?changeset=7c1415829dcee8814c161503d0f32a438fed118a
Comment 2•2 years ago
•
|
||
Mozilla crash reason: Missing chrome or resource URLs: resource:///modules/MigrationUtils.sys.mjs
Which makes me think it's https://hg.mozilla.org/mozilla-central/rev/912f29a6243b2bf5199462a6ecaf499aeacb9789#l12.13 (bug 718280)
Sadly MigrationUtils.sys.mjs lives in browser and is referenced in toolkit. Maybe the fix here is to update our migration implementation to also provide a local MigrationUtils.sys.mjs
?
Also I think bug 1805620 is the same crash.
Assignee | ||
Comment 3•2 years ago
|
||
I suspect it's this part: https://hg.mozilla.org/mozilla-central/diff/912f29a6243b2bf5199462a6ecaf499aeacb9789/toolkit/modules/ResetProfile.sys.mjs
The test that's failing is loading about:support, which loads ResetProfile.sys.mjs to see if a FirefoxProfileMigrator exists, which is a pre-requisite to doing a profile reset.
Before, it used to check for the existence of the profile by asking XPCOM to see if the migrator was registered as a component. Now that MigrationUtils is doing all of that management, we now ask it instead. And this module is only registered for browser, so crash.
So a few options to fix this:
- comm stubs out a fake MigrationUtils and returns false for
migratorExists
. This seems like a kludge and probably an option of last resort. - ResetProfile.sys.mjs is modified so that it doesn't assume
MigrationUtils
exists. It'll check for existence as a first step, and then check that an appropriate migrator exists.
I suggest (2) is probably the best way forward.
Comment 4•2 years ago
|
||
Thank you for jumping in here!
I'm not familiar with our migration code nor the toolkit migration paths at all, but I wanted to note that we appear to have our own implementations of the migrator that was previously called, or at least I think we did? https://searchfox.org/comm-central/source/mail/components/migration/src/components.conf
Though I'm not sure we'd ever see any of those "types" from this specific code path.
Assignee | ||
Comment 5•2 years ago
|
||
From a quick glance, that should still probably work - the mail migrator code still tries to pull its migrators from XPCOM, and that's what it's using to register them, so you should be good there.
Assignee | ||
Comment 6•2 years ago
|
||
Assignee | ||
Comment 7•2 years ago
|
||
Can you check to see if the patch here fixes the crash for TB?
Comment 8•2 years ago
|
||
since this is a testing specific crash code path a comm-try build: https://treeherder.mozilla.org/jobs?repo=try-comm-central&revision=5d893ff3fb8a5d3216c0a635a30c6ed3cf3bb3d9
Comment 9•2 years ago
|
||
Still crashes: https://treeherder.mozilla.org/logviewer?job_id=399653134&repo=try-comm-central&lineNumber=2736
I think the issue is that even trying to resolve the URI triggers the crash: https://searchfox.org/mozilla-central/rev/17aeb39742eba71e0936ae44a51a54197100166d/netwerk/base/nsNetUtil.cpp#3898
So I guess the solution is to check against the app constant?
Assignee | ||
Comment 10•2 years ago
|
||
Alright, added a condition for MOZ_BUILD_APP
to the lazy getter. Does that fix it?
Comment 11•2 years ago
|
||
Comment 12•2 years ago
|
||
Comment 13•2 years ago
|
||
Comment 14•2 years ago
|
||
bugherder |
Comment hidden (Intermittent Failures Robot) |
Updated•2 years ago
|
Description
•