Earliest build that failed in this way was actually https://luci-milo.appspot.com/buildbot/chromium.infra.cron/rebaseline-o-matic/518334 (yesterday).

Okay - what's happening here is that rebaseline-o-matic is trying to rebaseline its own commit (0837888cabb8dd018af80020c9fa359cc93f5729), and failing somehow because of that.

And the reason why its own commit involved adding a NeedsRebaseline line is because Precise is still one of the supported ports, but rebaseline-o-matic has been unable to get builds for this builder for several days (since it was removed).

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8a23d79fa048dbd0a7309ac17c91e91afc9a04a3

commit 8a23d79fa048dbd0a7309ac17c91e91afc9a04a3
Author: dpranke <dpranke@chromium.org>
Date: Sat Nov 19 18:38:20 2016

Skip tests on Precise that were marked NeedsRebaseline for Precise.

Background: In the past few days, rebaseline-o-matic has been re-adding lines
to TestExpectations with the [ Precise ] specifier, likely because the WebKit
Linux Precise builder was removed but the linux-precise port isn't yet removed,
and the builder WebKit Linux Precise was not yet removed from from
webkitpy/config/builders.py.

This is happening because we're dropping support for Linux Precise
( http://crbug.com/660580 ).

Now, webkitpy/config/builders.py has been updated, and the precise port should
be removed next week; after that happens, all of these lines can be removed
from TestExpectations.

(Patch from https://codereview.chromium.org/2518693002/#ps1, merged forward).

TBR=qyearsley@chromium.org, tkent@chromium.rg
BUG= 666901 

Review-Url: https://codereview.chromium.org/2515573004
Cr-Commit-Position: refs/heads/master@{#433423}

[modify] https://crrev.com/8a23d79fa048dbd0a7309ac17c91e91afc9a04a3/third_party/WebKit/LayoutTests/TestExpectations

Update: I believe that this *should* be resolved after Precise support is dropped (which should be done after https://codereview.chromium.org/2514253002).

Note, `webkit-patch auto-rebaseline` does still sort-of work (example run https://codereview.chromium.org/2522893002) but it adds [ Precise ] lines, and then the on the next run, rebaseline-o-matic tries to rebaseline the lines that it modified, and fails.

Hey Quinten, can we disable this until it's fixed? It's currently sending mail every few minutes to two lists.

That would be OK with me - what's the most convenient way to disable it for a few hours? Note that I'm expecting that it will probably be fixed this morning.

If it will be fixed today I think we can leave it.

Alright, looks like https://build.chromium.org/p/chromium.infra.cron/builders/rebaseline-o-matic is green again :-)