| Summary: | Crash under RemoteDisplayListRecorder::restore() | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Simon Fraser (smfr) <simon.fraser> | ||||||||||
| Component: | WebKit Process Model | Assignee: | Kimmo Kinnunen <kkinnunen> | ||||||||||
| Status: | RESOLVED FIXED | ||||||||||||
| Severity: | Normal | CC: | dino, kkinnunen, sabouhallawa, simon.fraser, webkit-bug-importer | ||||||||||
| Priority: | P2 | Keywords: | InRadar | ||||||||||
| Version: | Safari Technology Preview | ||||||||||||
| Hardware: | Unspecified | ||||||||||||
| OS: | Unspecified | ||||||||||||
| URL: | https://ews-build.s3-us-west-2.amazonaws.com/macOS-BigSur-Release-WK2-Tests-EWS/459507-7519/results.html | ||||||||||||
| Attachments: |
|
||||||||||||
|
Description
Simon Fraser (smfr)
2022-05-17 15:57:08 PDT
Created attachment 459540 [details]
Patch
I could not repro the issue, but I didn't have the exact same configuration. The strange thing is that dereferencing a disengaged std::optional should assert if the patch is fixing what it thinks it is fixing. However, I could not make std::optional assert in our builds. However, I seem to remember seeing such an assertion, so I don't know which is wrong -- my try or my recollection. I was in fact thinking of std::optional::value() which throws bad_optional_access. It appears we don't compile with libc++ debug assertions even on debug. From this perspective the patch is still consistent (potentially fixing the issue) Created attachment 459713 [details]
For landing.
Ignore that commit. It was landed incorrectly. Why isn't the commit-queue picking this up? Created attachment 459873 [details]
[fast-cq] Patch for landing
Trying to poke the cq.
I guess all cq patches have to go via github now. Somehow commit queue already applied this but never heralded. |