Hit MOZ_CRASH(called `Option::unwrap()` on a `None` value) at libcore/option.rs:345
Categories
(Core :: Graphics: WebRender, defect, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox-esr60 | --- | unaffected |
firefox65 | --- | wontfix |
firefox66 | --- | wontfix |
firefox67 | --- | fixed |
firefox68 | --- | fixed |
People
(Reporter: tsmith, Assigned: gw)
References
(Blocks 2 open bugs)
Details
(Keywords: assertion, crash, testcase)
Attachments
(2 files)
150 bytes,
text/html
|
Details | |
47 bytes,
text/x-phabricator-request
|
pascalc
:
approval-mozilla-beta+
|
Details | Review |
Updated•7 years ago
|
Comment 2•6 years ago
|
||
This crash still happens, I just repro'd on the latest macOS nightly. It produced the report at https://crash-stats.mozilla.org/report/index/c0351dcd-a07b-4e84-a964-7bfe00190413 with the useless [@ GeckoCrash]
signature. I've filed bug 1544246 to add that frame to the skiplist.
Glenn, since we have a testcase (attached above), could you take a look? The GeckoCrash
signature was pretty high on the list of crashes for 67 beta, although there might be multiple different crashes lumped into that.
Assignee | ||
Comment 3•6 years ago
|
||
Assignee | ||
Comment 4•6 years ago
|
||
This is a fix for crash bug https://bugzilla.mozilla.org/show_bug.cgi?id=1513019.
Ideally I'd get Dzmitry to take a look over this, as I don't know this code too well, but he's on PTO right now. I think what this code tries to do is use a simpler / more accurate method of plane splitting when the transform is simple (e.g. the case of stacked planes on the z-axis). However, the simple path check was also passing for transforms that were failing to produce a reasonable inverted matrix. Instead, we now only run that simple path for matrices that are pure translations.
This fixes the crash and the try run looks good. I think this should be safe to merge, but we should probably wait until Dzmitry can take a look before uplift to beta, if time permits.
Try run looks good:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=a3465dd26f6e455de2893d77be2db13eac79a574
Assignee | ||
Updated•6 years ago
|
Comment 6•6 years ago
|
||
bugherder |
Updated•6 years ago
|
Updated•6 years ago
|
Updated•6 years ago
|
Assignee | ||
Comment 7•6 years ago
|
||
Comment on attachment 9058185 [details]
Bug 1513019 - Fix plane splitting with complex, axis-aligned transforms.
Beta/Release Uplift Approval Request
- Feature/Bug causing the regression: None
- User impact if declined: Crashes when WebRender is enabled, on pages that contain transforms with very large scale values. These are fairly rare, but are one of the top crash bugs in the WebRender release experiment.
- Is this code covered by automated tests?: Yes
- Has the fix been verified in Nightly?: Yes
- Needs manual test from QE?: No
- If yes, steps to reproduce:
- List of other uplifts needed: None
- Risk to taking this patch: Low
- Why is the change risky/not risky? (and alternatives if risky): It's a small patch that only affects users enrolled in WebRender, and only a code path that is hit on a small number of pages. The patch has been in nightly for a couple days now. The patch itself is very small and easy to back out if any issues are caused.
- String changes made/needed:
Updated•6 years ago
|
Comment 8•6 years ago
|
||
Comment on attachment 9058185 [details]
Bug 1513019 - Fix plane splitting with complex, axis-aligned transforms.
Fix for a webrender crash, uplift approved for 67 beta 12, thanks.
![]() |
||
Comment 9•6 years ago
|
||
bugherder uplift |
Updated•6 years ago
|
Updated•6 years ago
|
Description
•