Summary: | Nullopt in RenderFlexibleBox::layoutFlexItems in RenderFlexibleBox::layoutBlock via RenderMultiColumnFlow::layout | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Ryosuke Niwa <rniwa> | ||||||
Component: | Layout and Rendering | Assignee: | Sergio Villar Senin <svillar> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | achristensen, bfulgham, cgarcia, ews-feeder, fred.wang, gpoo, jfernandez, koivisto, product-security, rbuis, simon.fraser, svillar, webkit-bug-importer, zalan | ||||||
Priority: | P2 | Keywords: | InRadar | ||||||
Version: | WebKit Nightly Build | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
Attachments: |
|
Description
Ryosuke Niwa
2021-03-17 01:41:30 PDT
I can reproduce nullopt with DumpRenderTree at r274459 but encountering abort in CoreAnimation if I ran the test with WebKitTestRunner so you might need to either workaround that crash or use DumpRenderTree to debug this. I'll check it out, maybe it's just a dup of bug 222584 (In reply to Sergio Villar Senin from comment #2) > I'll check it out, maybe it's just a dup of bug 222584 Errr I mean bug 222854 :) Ryosuke, which revision are you using? I'm hitting an ASSERT but a totally different one. (In reply to Sergio Villar Senin from comment #4) > Ryosuke, which revision are you using? I'm hitting an ASSERT but a totally > different one. OK I got the same trace in macOS. In Linux it hits an ASSERT in RenderLayer first. I'll upload a patch for the original issue tomorrow. Created attachment 423577 [details]
Patch
I believe this is not a security issue. We were just hitting an ASSERT that checks that the content size suggestion of a flex item is not negative, basically because it does not make sense, but I doubt this could be exploitable in any way. That's why I'm including a potential layout test that we could even upload to WPT as it's still useful for other engines even though they don't hit the assertion. Ping reviewers BTW there must be something wrong with Release EWS as I don't get any failure locally when testing this on MacOS Thanks for the review. Waiting for upstream WPT to accept the test and then I'll land this one. Committed r276835 (237186@main): <https://commits.webkit.org/237186@main> |