Doesn't seem to repro for me in ChromeOS 71.0.3578.94 on a Chromebook Pixel 2.  Maybe it's a recent regression?  James or Kevin, are you able to repro?

Let me build ToT CrOS for Linux and check ...

On 73.0.3674.0 CrOS-on-Linux I can't repro. Both with and w/o pinch-zoom mouse and touch seem fine.

On many video players (e.g. otakustream.tv) starting from 72.0.3626.59 (beta), clicking on controls on video players does nothing most of the time using mouse or touch pad. Reproduced on Pixelbook and Chromebook Pixel 2015. Touch screen touches are still working. Clicks sometimes still get through.

I'm afk right now, could someone please check if it has pointer-events:none set? https://chromium-review.googlesource.com/c/chromium/src/+/1415840 was just merged back to 72 so might be related.

dcheng@ and wberoux@ -

Could you please each go to chrome://version and copy all the information there into the bug? In particular, the Variations list may indicate a field-trial that is affecting the issue.

riajiang@'s CL from comment 6 is in 72.0.3626.65 and higher (which aren't available yet).  See https://chromium.googlesource.com/chromium/src/+log/6bc20f393dd046df31738365b49d52cab2781505.  It's possible that will fix this issue.

However, lfg@ was not able to repro this in 72.0.3626.49, which doesn't have that fix.  dcheng@ reported this in 72.0.3623.3.  So either a fix was merged between those versions or the bug is specific to a field trial.

dcheng@: Can you try updating to 72.0.3626.49 or newer (like lfg@) to see if it still repros?  That would help indicate whether this is likely to be a field trial or not.

CC'ing TPMs since M72 ships to stable soon.

Is this bug specific to Chrome OS or Desktop and Android too?

Ah, comment 4 points to a repro with 72.0.3626.59.  That's newer than lfg@'s version, so there's a good chance it's field trial related.  As wjmaclean@ notes, maybe it's worth pasting the list of variations here.

wberoux@: Can you confirm that you're able to repro the bug using the steps dcheng@ posted (with https://www.reddit.com/r/movies/comments/7dtn5i/pixars_incredibles_2_trailer/)?  That will help us know that it's possible to spot it in the right conditions, and that it isn't nondeterministic.

Comment 10: We haven't seen any repros of it on desktop or Android yet.

Comment 10: Just to confirm, there's no repro for me on Windows 72.0.3626.53.

Also, this shouldn't affect Android (where OOPIFs are not yet deployed on Beta).  Just for completeness, I tested the repro steps on Android 72.0.3626.64 with SitePerProcess manually enabled, and the bug doesn't seem to repro there either.  So far, this seems ChromeOS specific.

This may be the in process mash experiment, piman suggests.

Assuming the version number is correct, 72.0.3623.3, it isn't mash. SingleProcessMash was enabled for a couple of dev builds in 73, but not at all in 72.

The repro is a bit flaky, but if you fast forward to the end of the Youtube video I linked, try to click the [Replay] button.

Once you do that, the Play/Pause button mostly stops responding to mouse interactions. From that point on, even after reloading or opening the repro link in a new tab, the Play/Pause button is mostly unresponsive to pointer interactions.

This is true on version 72.0.3626.59.

Interestingly enough, while it happens consistently on my Pixelbook, it doesn't happen consistently on my Chromebook Flip, which is on version 72.0.3626.49 (dev).

I'm unable to repro on a Pixelbook with 72.0.3626.59.

VizHitTestDrawQuad was disabled on CrOS since Jan 16, is it reproable on 72.0.3626.59 with explicitly turning VizHitTestDrawQuad on?

Even with viz hit test draw quad explicitly enabled, I can't repro.