Able to reproduce the issue on Windows 7,Mac 10.13.3 & Debian using chrome latest stable-65.0.3325.181 & Canary-67.0.3376.1 as per C#0.

Observed all 3 errors same as in step 2 when we navigate to below URL.

https://ci.chromium.org/buildbot/chromium.infra.cron/publish_tarball/

Same issue observed from M60 .As it is non regression issue , marking it as Untriaged to get it addressed from dev.

Please find the attached screencast for reference.

Thanks..!

Deleted: 823808-Linux.mp4 2.6 MB

	823808-Linux.mp4 2.6 MB View Download

There's no crash going on, this is a bug about a failure in the publish_tarball bot.

gclient sync is failing like this:

src/third_party/harfbuzz-ng/src (ERROR)
----------------------------------------
[0:00:03] Started.
[0:00:03] running "git cat-file -e 957e7756634a4fdf1654041e20e883cf964ecac9^{commit}" in "/b/c/git/chromium.googlesource.com-external-github.com-harfbuzz-harfbuzz"
----------------------------------------
Error: 32> 
32> ____ src/third_party/harfbuzz-ng/src at 957e7756634a4fdf1654041e20e883cf964ecac9
32> 	You have unstaged changes.
32> 	Please commit, stash, or reset.
step returned non-zero exit code: 1

That's happening because 67.0.3376.* is broken due to https://chromium-review.googlesource.com/c/chromium/src/+/937502 -- this CL later got reverted and re-landed correctly. You can see from https://ci.chromium.org/buildbot/chromium.infra.cron/publish_tarball/10189 that 67.0.3377.* (which contains the right changes) has been produced correctly.

I think this will stop failing as soon as omahaproxy.appspot.com is updated and the bot no longer attempts to build 67.0.3376.{0,1}.

Reopening, http://omahaproxy.appspot.com/history has entries dating as far back as December. I think the bot will continue trying to create those buggy tarballs (and spamming me multiple times every day with the failures) for months if left alone.

https://chromium-review.googlesource.com/c/chromium/src/+/928761 is the CL which broke things
https://chromium-review.googlesource.com/c/chromium/src/+/969185 is the revert
https://chromium-review.googlesource.com/c/chromium/src/+/970242 is the reland

Why isn't the correct fix here to cherry-pick the revert+reland to the broken release branch?

c#7: It's not possible to cherry-pick because the tarballs are built at a specific version.  The ones it keeps trying to build are 67.0.3376.{0,1}.  The only thing we can really do is blacklist those versions in the tarball creator.

Re c#3: I'm getting spammed multiple times a day too, it's time to put a stop to this :)

The following revision refers to this bug:
  https://chromium.googlesource.com/infra/infra/+/457d4a13756261daea08dddc050459bd6fb67fbc

commit 457d4a13756261daea08dddc050459bd6fb67fbc
Author: Tom Anderson <thomasanderson@chromium.org>
Date: Thu Mar 29 00:46:51 2018

Add a version blacklist to publish_tarball

R=mmoss
BUG= 823808 

Change-Id: I23d0003bd95c6a00025373fc9b5dc38c0bd95d5a
Reviewed-on: https://chromium-review.googlesource.com/985498
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Commit-Queue: Michael Moss <mmoss@chromium.org>
Reviewed-by: Michael Moss <mmoss@chromium.org>

[modify] https://crrev.com/457d4a13756261daea08dddc050459bd6fb67fbc/recipes/README.recipes.md
[modify] https://crrev.com/457d4a13756261daea08dddc050459bd6fb67fbc/recipes/recipes/publish_tarball.py