I'm not sure what there is to do here.  It seems like the fuzzer is generating a path that when tesselated requires allocating a ton of edges.

reed@ fixed a previous deserialization oom by sanitizing the inputs (https://skia-review.googlesource.com/c/26805) but that seems harder to do here when there's a layer of tesselation indirection.

Could you attach a deserialized (HTML) version of the test case? I'm not sure how to extract the path from it.

There is no html for this fuzzer.  It's a fuzzer that takes random memory, deserializes it into cc::PaintOps (some of which contain paths) and then rasters them into a ganesh-backed SkCanvas.

If you download this file, build paint_op_buffer_fuzzer, and then pass the file as an arg to the fuzzer executable, it should repro.  It says it's only on mac, but I don't see why this would be platform specific.

Where "this file" is the clusterfuzz test case, sorry.

ClusterFuzz has detected this issue as fixed in range 493200:493235.

Detailed report: https://clusterfuzz.com/testcase?key=6640411786608640

Fuzzer: libFuzzer_paint_op_buffer_fuzzer
Job Type: mac_libfuzzer_chrome_asan
Platform Id: mac

Crash Type: Out-of-memory (exceeds 2048 MB)
Crash Address: 
Crash State:
  paint_op_buffer_fuzzer
  
Sanitizer: address (ASAN)

Regressed: https://clusterfuzz.com/revisions?job=mac_libfuzzer_chrome_asan&range=491869:492014
Fixed: https://clusterfuzz.com/revisions?job=mac_libfuzzer_chrome_asan&range=493200:493235

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=6640411786608640

See https://chromium.googlesource.com/chromium/src/+/master/testing/libfuzzer/reproducing.md 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 6640411786608640 is verified as fixed, so closing issue as verified.

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