| Summary: | REGRESSION(r275267): [GTK][WPE] 12 new crashes on service-workers layout tests | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Carlos Alberto Lopez Perez <clopez> | ||||||||
| Component: | Service Workers | Assignee: | Chris Dumez <cdumez> | ||||||||
| Status: | RESOLVED FIXED | ||||||||||
| Severity: | Normal | CC: | benjamin, bugs-noreply, cdumez, cgarcia, cmarcelo, ews-watchlist, ggaren, mcatanzaro, webkit-bug-importer, zdobersek | ||||||||
| Priority: | P2 | Keywords: | InRadar | ||||||||
| Version: | WebKit Nightly Build | ||||||||||
| Hardware: | Unspecified | ||||||||||
| OS: | Unspecified | ||||||||||
| See Also: | https://bugs.webkit.org/show_bug.cgi?id=223808 | ||||||||||
| Attachments: |
|
||||||||||
|
Description
Carlos Alberto Lopez Perez
2021-04-01 11:22:39 PDT
And it have made also crash the test http/wpt/service-workers/clone-opaque-being-loaded-response.html (same backtrace) Sadly, the issue is not obvious to me. I don't see what's different about the glib port here, expect that it is using a different version of FileSystem::writeToFile(). I will likely need help from someone able to run the glib port to debug this. I think this means the buffer that is being passed to g_output_stream_write_all() is nullptr, which I guess is possible if the bufferSize is 0. Created attachment 424919 [details]
Patch
I have not been able to validate my fix but I *think* this is what the glib port implementation of fileSystem::writeToFile() did not like. Comment on attachment 424919 [details]
Patch
Have another idea.
Created attachment 424921 [details]
Patch
Comment on attachment 424921 [details]
Patch
Tested! It fixes the issue! thanks :)
Comment on attachment 424921 [details]
Patch
Great, thanks for validating the fix.
Committed r275379: <https://commits.webkit.org/r275379> All reviewed patches have been landed. Closing bug and clearing flags on attachment 424921 [details]. |