Automatically applying components based on crash stacktrace and information from OWNERS files.

If this is incorrect, please apply the Test-Predator-Wrong-Components label.

Automatically adding ccs based on suspected regression changelists:

Enable root layer scrolling by bokan@chromium.org - https://chromium.googlesource.com/chromium/src/+/39cb84e649e045d2233a395009d5ccf1d08854a0

Lazily initialize TTS voices on low-end Android devices. by dskiba@chromium.org - https://chromium.googlesource.com/chromium/src/+/d10a5836167d1fa92b23ee0f2b74ddce95b0b872

If this is incorrect, please let us know why and apply the Test-Predator-Wrong-CLs label.

Very probably related to RLS. Taking a look now.

Confirmed: repro'd locally - turning off RLS solves the issue.

I'm not sure how big of an issue this is. This is a DCHECK that's firing due to a loss of precision in converting to absolute coordinates. The test case has a very high scrollTop offset (~1.8 million) which only has to be accounted for when transforming to absolute when RLS is turned on - hence the blame on the RLS patch. On its own, that shouldn't lose precision but I'm guessing there's some multiplies along the way so that we might have intermediate computations that lose precision.

Steve/Stefan, as layout experts, could you decide what to do with this?

I think we generally WontFix bugs like this (overflow / precision loss with extreme inputs).

ClusterFuzz has detected this issue as fixed in range 545318:545320.

Detailed report: https://clusterfuzz.com/testcase?key=5168381628776448

Fuzzer: marty_html_twiddler
Job Type: linux_debug_chrome
Platform Id: linux

Crash Type: CHECK failure
Crash Address: 
Crash State:
  layout_object_mapped_result.EqualWithinEpsilon(result.BoundingBox(), 0.1f) || la
  blink::LayoutGeometryMap::MapToAncestor
  blink::LayoutGeometryMap::AbsoluteRect
  
Sanitizer: address (ASAN)

Regressed: https://clusterfuzz.com/revisions?job=linux_debug_chrome&range=536888:536894
Fixed: https://clusterfuzz.com/revisions?job=linux_debug_chrome&range=545318:545320

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=5168381628776448

See https://github.com/google/clusterfuzz-tools for more information.

If you suspect that the result above is incorrect, try re-doing that job on the test case report page.