Bug 139811 - REGRESSION(r177535): It broke the GTK performance bot
Summary: REGRESSION(r177535): It broke the GTK performance bot
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P1 Blocker
Assignee: Csaba Osztrogonác
URL:
Keywords:
Depends on:
Blocks: 139786
  Show dependency tree / graph
 
Reported: 2014-12-19 00:10 PST by Csaba Osztrogonác
Modified: 2014-12-20 19:40 PST (History)
6 users (show)

See Also:

Attachments
Patch (2.34 KB, patch)
2014-12-19 00:15 PST, Csaba Osztrogonác 		no flags Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Csaba Osztrogonác 2014-12-19 00:10:44 PST 
After http://trac.webkit.org/r177535 performance bots remove 
WebKitBuild directory unconditionally. It was a workaround to
fix a Mac only issue, but it broke the GTK bot:
https://build.webkit.org/builders/GTK%20Linux%2064-bit%20Release%20%28Perf%29/builds/1522
5.) jhbuild updated gtk dependencies 
6.) delete WebKitBuild directory_1 deleted WebKitBuild directory 
--> It removes the necessary dependencies built by step 5.)

This clean build is incorrect and unnecessary on the GTK bot.
Comment 1 Csaba Osztrogonác 2014-12-19 00:15:51 PST 
Created attachment 243547 [details]
Patch
Comment 2 Csaba Osztrogonác 2014-12-19 00:24:21 PST 
Comment on attachment 243547 [details]
Patch

Clearing flags on attachment: 243547

Committed r177569: <http://trac.webkit.org/changeset/177569>
Comment 3 Csaba Osztrogonác 2014-12-19 00:24:28 PST 
All reviewed patches have been landed.  Closing bug.
Comment 4 Alexey Proskuryakov 2014-12-19 00:34:28 PST 
I now think that the cleanest fix would be to set is_clean in these two before calling Factory.__init__. With this, CleanBuildIfScheduled would clean the directory at the right time.
Comment 5 Ryosuke Niwa 2014-12-19 00:37:43 PST 
Or pass in --no-build to run-perf-tests...
Comment 6 Csaba Osztrogonác 2014-12-19 00:44:45 PST 
(In reply to comment #2)
> Comment on attachment 243547 [details]
> Patch
> 
> Clearing flags on attachment: 243547
> 
> Committed r177569: <http://trac.webkit.org/changeset/177569>

Lucas or Ryosuke, could you possibly push this change to buildmaster
and then restart to fix the GTK performance bot? Thanks.
Comment 7 Lucas Forschler 2014-12-19 08:02:14 PST 
I've restarted the master, let me know if things don't look correct.
Comment 8 Csaba Osztrogonác 2014-12-19 08:25:03 PST 
(In reply to comment #7)
> I've restarted the master, let me know if things don't look correct.

Thanks, the GTK performance bot is happy now. :)
Comment 9 Darin Adler 2014-12-20 19:39:00 PST 
Comment on attachment 243547 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=243547&action=review

> Tools/ChangeLog:8
> +        Trigger clean build only on mac platform.

Why? If the problem is on GTK, then why isn’t the condition here about GTK, rather than Mac?
Comment 10 Darin Adler 2014-12-20 19:40:08 PST 
Comment on attachment 243547 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=243547&action=review

>> Tools/ChangeLog:8
>> +        Trigger clean build only on mac platform.
> 
> Why? If the problem is on GTK, then why isn’t the condition here about GTK, rather than Mac?

I guess it’s not an important question; looks like Alexey removed this code a few commits later.