TestRunner interface broken, causing layout test failures in local build.

[Justin Novosad]

Hi All,

I have a clean build with no local code changes and most layout test are failing. Looking at the results, it looks as if the testRunner interface does not exist.  For example, tests that call testRunner.dumpAsText() are producing render tree dumps.  I restarted with a brand new checkout did a clean build -> same problem.

Has anyone else experienced this? Does anyone know what may be causing this?

Thanks, 

    -Justin

[Rick Byers]

Thanks for posting!  I saw this too starting a couple hours ago and was very confused by it - was trying to figure out why it was only me (as far as I could see the tree was still green).  I didn't get very far in debugging - just that all layouttests were failing on ToT Debug (in both gn and gyp builds).  I started looking at recent commits and didn't see an obvious culprit - presumably someone (in a timezone that isn't now end-of-day) will want to bisect?

[Justin Novosad]

I have been experiencing this since last Friday. So I am pretty sure by now that it is an issue that does not affect everyone.

[Raymond Toy]

I have this problem.  But only on linux; mac appears to be fine.

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

[Justin Novosad]

I am experiencing this on linux too. My backup plan is to abuse the try bots.

[Rick Byers]

I've realized this happens whenever giving content_shell a file to run that's inside a directory that is a symlink (as my chromium checkout is, since I've explicitly placed it on my SSD).  If I do: "cd `pwd -P`" first to get to a true physical path that doesn't include any symlinks it works fine again.

I filed https://bugs.chromium.org/p/chromium/issues/detail?id=614778 to track.  Unless someone else beats me to it, I'll bisect to find the CL that broke this.

[TAMURA, Kent]

We found paths starting with "/media/" caused the problem.  I'm making a fix.

--

TAMURA Kent
Software Engineer, Google

[Raymond Toy]

Thanks so much for fixing this. run-webkit-tests works again for me.  

(And for the record, I had the sources on my SSD at /media/...)

[Justin Novosad]

Yeah, same here.  Thanks!

[Nate Chapin]

I have a lot fewer tests failing than I used to, but I'm still seeing certain directories (e.g., http/tests/inspector) all time out with a clean checkout.  Is that likely to be fallout from the same change?

[TAMURA, Kent]

Yeah, we still have the issue for inspector tests, which access local files outside of LayoutTests/.

[Jochen Eisinger]

any update on this?

[Rick Byers]

Yep, it's all fixed.  See the bug: https://bugs.chromium.org/p/chromium/issues/detail?id=614778

[Jochen Eisinger]

hum, so at least inspector/ still times out for me (while running content_shell --run-layout-tests manually works)

[Rick Byers]

That's probably something unrelated (are you sure it's a problem only with /media?).

I've got a CL up now to completely remove the special-casing of /media from content shell (it's no longer necessary) - you could verify the issue still exists with that applied (it should since the logic is now restricted to imported/wpt).