Hi bsheedy@,

 can you please give me the instructions on how to run these tests.

thank you

See the Building and Running sections in https://chromium.googlesource.com/chromium/src/+/b4064e96f0a3e6df677d8e35b5958f693e34b7a4/chrome/android/javatests/src/org/chromium/chrome/browser/vr/README.md#Building

You will need a rooted Pixel device of some sort in order to run those particular tests. If you don't have access to one, you can run the tests on swarming.

Hi sadrul@ and penghuang@,

There are some VR tests failing due to my CL - https://chromium-review.googlesource.com/c/chromium/src/+/1106600/14/ui/gl/gl_surface_egl.cc#1447

This is because the presentation time is reported as null in my CL when  eglGetFrameTimestampsANDROID() call fails.(line 1467 in gl_surface_egl.cc)

There is a DCHECK in blink/renderer code which is expecting presentation time to be not null.

Can you please let me know what should be expected behavior here. Should presentation time be reported as TimeTicks::Now() in those cases or it can be reported as null ?

Also i am not sure why eglGetFrameTimestampsANDROID() fails. There is no error code apart from true/false. I will double check that this api call is passing for telemetry tests or not.

thanks



 

For cases where we couldn't get the correct presentation-time, we generally just use the current time (TimeTicks::Now()) as the presentation-time (e.g. [1]). So perhaps that's what we could do too before returning in gl_surface_egl.cc:1451 (and probably before the early return in line 1471 too)?

[1] https://cs.chromium.org/chromium/src/ui/gl/gl_surface_presentation_helper.cc?sq=package:chromium&dr=CSs&g=0&l=63

ok.. sounds good.

It sounds like this was fixed based on the comments. Can someone close it if so?

The crash was fixed with a revert, and it's been long enough that I think it's safe to assume that either it got relanded without issue or has been abandoned.