Bug 54644 – [Chromium] set_version_blocked.html and set_version_queue.html started crashing on Chromium dbg bots after WK roll 78748:78777

• Component: New Bugs
• Status: RESOLVED FIXED

[Mikhail Naganov]

Both tests started crashing consistently on all 3 platforms (dbg):

http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Mac10.5%20(deps)(dbg)(2)/builds/607
http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Win%20(deps)(dbg)(2)/builds/1010
http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Linux%20(deps)(dbg)(2)/builds/1148

I tried to repro on Linux dbg locally, with no success.

[Mikhail Naganov]

...and the tests were OK on WebKit canary bots.

[Jeremy Orlow]

They all look green after http://trac.webkit.org/changeset/78781

[WebKit Review Bot]

http://trac.webkit.org/changeset/78843 might have broken Qt Linux Release
The following tests are not passing:
media/controls-without-preload.html

[Jeremy Orlow]

In the future, please don't roll crashes into Chromium unless you really know what your'e doing (i.e. you're positive it won't affect anyone...usually you'd only do this after talking to people who are experts on the code).

In this case, there was a fix just a few revisions after what you rolled to.  But even if there wan't, you should take a quick look at what's wrong and, if you can't figure it out, you should probably just roll out the patch.

[Mikhail Naganov]

I didn't intentionally roll crashing webkit tests into chromium. I have now reconstructed what had happened:
  - I came to work, found latest WK revision that looked green on canaries (78777), and rolled to it;
  - then I cleaned up canaries, and did another roll up to 78814;
  - but chromium bots at that moment had only started chewing my previous roll, and crashes appeared (I'm still unsure, why they weren't failing on canaries);
  - at that moment I filed this issue -- as it appears, this wasn't really needed.

I should pay more attentions to revisions in bot logs next time.
Sorry for bothering you with a false alarm.

[Jeremy Orlow]

(In reply to comment #5)
> I didn't intentionally roll crashing webkit tests into chromium. I have now reconstructed what had happened:
>   - I came to work, found latest WK revision that looked green on canaries (78777), and rolled to it;
>   - then I cleaned up canaries, and did another roll up to 78814;
>   - but chromium bots at that moment had only started chewing my previous roll, and crashes appeared (I'm still unsure, why they weren't failing on canaries);
>   - at that moment I filed this issue -- as it appears, this wasn't really needed.
> 
> I should pay more attentions to revisions in bot logs next time.
> Sorry for bothering you with a false alarm.

No worries on the false alarm.  In the future, it's probably worth following up on the bug after things settle down to make sure everyone understands the situation though.

Btw, I hope you're not reading my comments as passive agressive.  Gardening is very difficult and there's a lot going on, so it's hard to do 100% right.  I just point stuff out because I assume it'll be useful in the future.