I tried reverting the CL, but that broke the build.

Nico, can you help me figuring out what causes the deterministic test to fail? I have little knowledge about that part (and how it could be connected to the  change).

That cl is Android only. This is a Linux bot. It's not due to that cl.

looked a bit deeper. the diffs come from webkit.
from those CLs https://chromium-review.googlesource.com/c/chromium/src/+/737913 might be related. Trigger a build at that revision. let's see.

The buildbot already fails at that CL (a883c34fd8fead12740ba246c530eb854c1f3694).
Triggered another run at the CL before and will reach out to the authors.

+other sheriffs

the builds take forever :-/ Leaving it up to you to wait for the next one (should be 8204 on https://uberchromegw.corp.google.com/i/chromium.linux/builders/Deterministic%20Linux to verify the culprit or just revert the CL in question)

Sorry for my newbie question but you mentioned "buildbot already fails at that CL (a883c34fd8fead12740ba246c530eb854c1f3694)" which is "refs/heads/master@{#512039}".

Shouldn't we be building #512038 next, instead of 8204 which is #512040?

Already left the office, but I can totally imagine I'm the noob in here and
probably mixed up the direction of the history...

No worries I will try to lock it down. Triggered a build on #512038.

This is also reproducible locally w/ something like:

$ ./tools/mb/mb.py -m chromium.linux -b "Deterministic Linux" out-mb/determinism1
$ ./tools/mb/mb.py -m chromium.linux -b "Deterministic Linux" out-mb/determinism2
$ # sync to revision of your choice

$ ninja -C out-mb/determinism1 [goma args] blink_heap_unittests  # or one of the many other targets failing this
$ ninja -C out-mb/determinism2 [goma args] blink_heap_unittests  # same target
$ tools/determinism/compare_build_artifacts.py --first-build-dir out-mb/determinism1 --second-build-dir out-mb/determinism2 --target-platform linux


Limiting the targets makes it quite a bit faster than the full builds that Deterministic Linux does.

Re #9: Thanks for the help! I can confirm locally that the CL in #4 is causing the problem.

Testing locally to see if reverting could fix the problem.

Actually it's not trivial to revert due to conflicts.

meade@ Can you take a look if this is related to your change? Thanks!
(a883c34fd8fead12740ba246c530eb854c1f3694)

Additional Info: refs/heads/master@{#512038} passed the build.

https://uberchromegw.corp.google.com/i/chromium.linux/builders/Deterministic%20Linux/builds/8207

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2d5100869a8ae5a0ec3a355485e502a60bc12d6c

commit 2d5100869a8ae5a0ec3a355485e502a60bc12d6c
Author: Eddy Mead <meade@chromium.org>
Date: Mon Oct 30 05:24:03 2017

Remove errant brackets and unnecessary calls to "str" in make_computed_style_base.py

I think this should fix the non-deterministic build issues

Bug:  779001 , 779033 
Change-Id: I0779b54d38a3fc8db2c71206fdafbcd2dc428dac
Reviewed-on: https://chromium-review.googlesource.com/743163
Commit-Queue: meade_UTC10 <meade@chromium.org>
Reviewed-by: Darren Shen <shend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#512443}
[modify] https://crrev.com/2d5100869a8ae5a0ec3a355485e502a60bc12d6c/third_party/WebKit/Source/build/scripts/make_computed_style_base.py

Bots are green, again. https://chromium-review.googlesource.com/743163 indeed seems to have fixed it. Thanks everyone!