Bug 150755 - Make CMake builder build with CMake
Summary: Make CMake builder build with CMake
Status: RESOLVED WORKSFORME
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Alex Christensen
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-10-31 01:15 PDT by Alex Christensen
Modified: 2015-10-31 17:39 PDT (History)
4 users (show)

See Also:


Attachments
Patch (3.31 KB, patch)
2015-10-31 01:16 PDT, Alex Christensen
no flags Details | Formatted Diff | Diff
Patch (4.78 KB, patch)
2015-10-31 01:26 PDT, Alex Christensen
ossy: review+
ossy: commit-queue-
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Alex Christensen 2015-10-31 01:15:29 PDT
Make CMake builder build with CMake
Comment 1 Alex Christensen 2015-10-31 01:16:04 PDT
Created attachment 264471 [details]
Patch
Comment 2 WebKit Commit Bot 2015-10-31 01:18:03 PDT
Attachment 264471 [details] did not pass style-queue:


ERROR: Tools/BuildSlaveSupport/build.webkit.org-config/config.json:0:  No JSON object could be decoded  [json/syntax] [5]
Total errors found: 1 in 3 files


If any of these errors are false positives, please file a bug against check-webkit-style.
Comment 3 Alex Christensen 2015-10-31 01:18:32 PDT
This makes https://bugs.webkit.org/show_bug.cgi?id=150738 do the right thing.  The build master will need to be restarted after landing this change.  I'm not sure how to make this show up on the buildbot dashboard
Comment 4 Alex Christensen 2015-10-31 01:26:55 PDT
Created attachment 264473 [details]
Patch
Comment 5 Csaba Osztrogonác 2015-10-31 04:41:23 PDT
Comment on attachment 264473 [details]
Patch

LGTM, but please add the missing unittest too before landing.
Comment 6 Alexey Proskuryakov 2015-10-31 08:14:14 PDT
Comment on attachment 264473 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=264473&action=review

> Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg:843
> +class BuildWithCMakeFactory(Factory):

I'm not a big fan of adding factories. This creates a combinatorial explosion, as you'll need a separate factory to build with ASan, and for every other existing variant.

> Tools/BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/WebKitBuildbot.js:40
> +        "Apple El Capitan CMake Debug (Build)": {platform: Dashboard.Platform.MacOSXElCapitan, debug: true, builder: true, architecture: Buildbot.BuildArchitecture.SixtyFourBit, heading: "CMake"},

This presumably made it appear. 

The question I have, did we agree to keep it green? I have not seen any discussions to that effect. I'd hate to roll out patches for breaking Mac CMake build if we don't get any benefits from it.
Comment 7 Alex Christensen 2015-10-31 11:31:23 PDT
Alexey, could you commit this and restart the master?
(In reply to comment #6)
> Comment on attachment 264473 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=264473&action=review
> 
> > Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg:843
> > +class BuildWithCMakeFactory(Factory):
> 
> I'm not a big fan of adding factories. This creates a combinatorial
> explosion, as you'll need a separate factory to build with ASan, and for
> every other existing variant.
I see your point, but right now there is just one builder.  I think we can reorganize things when this becomes a problem.
> 
> > Tools/BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/WebKitBuildbot.js:40
> > +        "Apple El Capitan CMake Debug (Build)": {platform: Dashboard.Platform.MacOSXElCapitan, debug: true, builder: true, architecture: Buildbot.BuildArchitecture.SixtyFourBit, heading: "CMake"},
> 
> This presumably made it appear. 
I didn't test this.  I assume this is all that's needed.
> 
> The question I have, did we agree to keep it green? I have not seen any
> discussions to that effect. I'd hate to roll out patches for breaking Mac
> CMake build if we don't get any benefits from it.
It's probably not critical enough to roll out patches that break it right now, and we don't get benefits from it right now, but we will soon.  Right now I'm the only one to see build failures after building on my own computer.  I want it to be easier to not lose progress on this work.
Comment 8 Lucas Forschler 2015-10-31 15:02:29 PDT
Comment on attachment 264473 [details]
Patch

I have this working already, by using the 'additionalArguments'  parameter in config.json.
I'm not sure why we need a second factory?
Comment 9 Alexey Proskuryakov 2015-10-31 15:39:59 PDT
> It's probably not critical enough to roll out patches that break it right now

In this case, it's probably better to not add it to the dashboard for now.
Comment 10 Alex Christensen 2015-10-31 17:39:36 PDT
(In reply to comment #8)
> Comment on attachment 264473 [details]
> Patch
> 
> I have this working already, by using the 'additionalArguments'  parameter
> in config.json.
> I'm not sure why we need a second factory?
This is just the way I thought of.  http://trac.webkit.org/changeset/191843 solves this problem.