New issue
Advanced search Search tips

Issue 906325 link

Starred by 2 users

Issue metadata

Status: Assigned
Owner:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 1
Type: Bug



Sign in to add a comment

CHECK failure: layout_object_mapped_result.EqualWithinEpsilon(result.BoundingBox(), 1.1f) || la

Project Member Reported by ClusterFuzz, Nov 17

Issue description

Detailed report: https://clusterfuzz.com/testcase?key=4896103620935680

Fuzzer: inferno_twister
Job Type: linux_debug_chrome
Platform Id: linux

Crash Type: CHECK failure
Crash Address: 
Crash State:
  layout_object_mapped_result.EqualWithinEpsilon(result.BoundingBox(), 1.1f) || la
  blink::LayoutGeometryMap::MapToAncestor
  blink::LayoutGeometryMap::AbsoluteRect
  
Sanitizer: address (ASAN)

Regressed: https://clusterfuzz.com/revisions?job=linux_debug_chrome&range=538990:538995

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=4896103620935680

Issue filed automatically.

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

Comment 1 by ClusterFuzz, Nov 17

Components: Blink>Compositing Blink>Layout
Labels: Test-Predator-Auto-Components
Automatically applying components based on crash stacktrace and information from OWNERS files.

If this is incorrect, please apply the Test-Predator-Wrong-Components label.
Owner: chrishtr@chromium.org
Status: Assigned (was: Untriaged)
Assigning to chrishtr@ based on this CL:
https://chromium-review.googlesource.com/910116

When working on https://crbug.com/905914 I found that I could hit this reliably. Here's a minimized testcase:
<!doctype html>
<style>
@keyframes bounceInLeft {
    0% {
      transform: none;
    }
}
</style>
<div id="container" class="bounceInLeft" style="animation-name: bounceInLeft; animation-fill-mode: forwards;">
  <div id="node" style="opacity: 0.9; width: 100px; height: 100px; background: lightblue;">
    <div id="descendant" style="position: absolute; width: 10px; height: 10px; background: lightgreen;"></div>
  </div>
</div>
If the clusterfuzz regression range is correct, this is a regression from turning on CSSTypedOM.
I'm not sure my testcase in comment #3 is relevant to this bug after all. It's another bug that is a way to trigger this crash and nothing more.
Owner: xidac...@chromium.org
I think this was fixed by https://chromium-review.googlesource.com/c/chromium/src/+/1344813.

Assigning to xidachen to verify and add this testcase to any re-land.
Mergedinto: 905914
Status: Duplicate (was: Assigned)
Cannot repro this locally with ToT code, closing this.
Owner: pdr@chromium.org
Status: Assigned (was: Duplicate)
This does not seem like a dupe of 905914 after all. Investigating the real cause now.
Labels: OS-Mac OS-Windows
I can repro this locally and I think it is a real bug, unlike some other DCHECKS at this location, because the values are not extreme: "1072,144 0x0" vs "1166,144 0x0"

This was probably a regression from turning on CSSOM but that's not likely to be the root cause. I've clicked minimize on the testcase to try to get a more minimized testcase to see if we can do something about this.
the problem with minimization is this testcase is flaky as per CF.

[2018-11-17 03:30:38 UTC] clusterfuzz-linux-high-end-jq06: Minimize task started.
[2018-11-17 03:46:17 UTC] clusterfuzz-linux-high-end-jq06: Minimize task errored out: Unable to reproduce crash reliably, skipping minimization (crashed 3/10)

i can reupload testcase with higher timeout (Re-upload testcase with like 30 sec timeout).
Project Member

Comment 11 by ClusterFuzz, Dec 13

ClusterFuzz is analyzing your testcase. Developers can follow the progress at https://clusterfuzz.com/testcase?key=4812456813395968.

Comment 12 Deleted

Project Member

Comment 13 by ClusterFuzz, Dec 14

Labels: -Reproducible Unreproducible
ClusterFuzz testcase 4812456813395968 appears to be flaky, updating reproducibility label.

Sign in to add a comment