r325140 doesn't affect web contents.
Please investigate Blink changes.

with respect to comment 1:

Rebisected for the above issue,getting the same Narrow bisect range.

Suspecting:r325137?

This looks like a rendering issue with the menu being placed a few pixels above where it's expected. My change is not related.

Blink changes between good and bad revisions are https://chromium.googlesource.com/chromium/blink/+log/21fff74806410395494bac6f7ed128ab07a5840b..817d1655758a97ccf3e0dd07a2bfd57c2eef7771?pretty=fuller&n=10000

Suspecting e0810e020e978d8e518a3a9ffd280e69e5b97c7c, but I am not familiar with this code.

hajimehoshi, please re-assign if not related to your change.

I'm afraid I don't think my change caused this side effect.

Suspecting a9ce0ce28c5d2f9b029fcb209358260141e2f41a since the "Actual" and "Expected" videos show rounded corners and the result is off by 1px.

Please help assign to correct person if your change is not the cause.

I don't think the repro/bisect are reliable:

1) the range in c#4 doesn't match the one in c#0

2) I can repro even in M43 (that's as far back as I checked): sporadically on first page load, most of the time when reloading the page.  See attachment.


adharap - can you confirm that the issue is present on older releases also, specifically on page reload?

Deleted: chrome_43.0.2347.0.mp4 974 KB

chrome_43.0.2347.0.mp4 974 KB Download

Not a regression, not limited to Chrome. I can reliably reproduce this in multiple versions of Chrome and Firefox. Seems like a race condition between multiple js event listeners.