Unable to find actual suspect through code search and also observing no suspected CL's under regression range, hence adding appropriate label and requesting someone from blink->paint team to look in to this issue.

Thanks!

This should be harmless for M-69 because the scroll nodes are not used, but this needs to be fixed for BlinkGenPropertyTrees at a minimum.

I am having trouble reproducing.

Tried:
Running the testcase under asan
Running the clusterfuzz repro script

I just clicked "redo task" in clusterfuzz and will report back with what I find.

ClusterFuzz has detected this issue as fixed in range 571830:571831.

Detailed report: https://clusterfuzz.com/testcase?key=6676244020854784

Fuzzer: marty_html_twiddler
Job Type: linux_debug_chrome
Platform Id: linux

Crash Type: CHECK failure
Crash Address: 
Crash State:
  *original_properties_->Scroll() == *object_properties->Scroll(). Property was up
  blink::FindObjectPropertiesNeedingUpdateScope::~FindObjectPropertiesNeedingUpdat
  blink::FragmentPaintPropertyTreeBuilder::UpdateForChildren
  
Sanitizer: address (ASAN)

Fixed: https://clusterfuzz.com/revisions?job=linux_debug_chrome&range=571830:571831

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=6676244020854784

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.

ClusterFuzz testcase 6676244020854784 is verified as fixed, so closing issue as verified.

If this is incorrect, please add ClusterFuzz-Wrong label and re-open the issue.

I think clusterfuzz is wrong and the test is flaky and/or requires a slow computer. That said, I can't make progress on this without being able to minimize the testcase (see: comment #3 where I sank half a day into trying to repro). I think the best course of action is to wait for a more reliable repro to surface.