[buildbot] [GTK][WPE] Move JavaScript related tests to own bot

[Diego Pino]

[buildbot] [GTK][WPE] Move JavaScript related tests to own bot

[Diego Pino]

Created attachment 400489 [details]
Patch

[Diego Pino]

Created attachment 400490 [details]
Patch

[Diego Pino]

Created attachment 400577 [details]
Patch

[Diego Pino]

Created attachment 400587 [details]
Patch

[Diego Pino]

Created attachment 400805 [details]
Patch

[Carlos Alberto Lopez Perez]

Comment on attachment 400805 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=400805&action=review

> Tools/BuildSlaveSupport/build.webkit.org-config/config.json:307
> +                    {
> +                      "name": "GTK Linux 64-bit Release Wayland (Tests JS)", "type": "TestJS", "builddir": "gtk-linux-64-release-wayland-tests-js",
> +                      "platform": "gtk", "configuration": "release", "architectures": ["x86_64"],
> +                      "additionalArguments": ["--display-server=wayland"],
> +                      "slavenames": ["gtk-linux-bot-14"]
> +                    },

I don't think we need a wayland-js bot.
AFAIK JS tests don't give any different output or have any different behavior in relation to the display server used.

> Tools/BuildSlaveSupport/build.webkit.org-config/steps_unittest.py:567
> +    'GTK Linux 64-bit Debug (Tests)': ['configure build', 'svn', 'kill old processes', 'delete WebKitBuild directory', 'delete stale build files', 'jhbuild', 'download-built-product', 'extract-built-product', 'jscore-test', 'layout-test', 'webkitpy-test', 'webkitperl-test', 'bindings-generation-tests', 'builtins-generator-tests', 'dashboard-tests', 'archive-test-results', 'upload', 'MasterShellCommand', 'API tests', 'webdriver-test'],

Something seems wrong here. The "GTK Linux 64-bit Debug (Tests)" shouldn't run the step 'jscore-test' now... should it?

[Diego Pino]

Created attachment 400810 [details]
Patch

[Diego Pino]

(In reply to Carlos Alberto Lopez Perez from comment #6)
> Comment on attachment 400805 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=400805&action=review
> 
> > Tools/BuildSlaveSupport/build.webkit.org-config/config.json:307
> > +                    {
> > +                      "name": "GTK Linux 64-bit Release Wayland (Tests JS)", "type": "TestJS", "builddir": "gtk-linux-64-release-wayland-tests-js",
> > +                      "platform": "gtk", "configuration": "release", "architectures": ["x86_64"],
> > +                      "additionalArguments": ["--display-server=wayland"],
> > +                      "slavenames": ["gtk-linux-bot-14"]
> > +                    },
> 
> I don't think we need a wayland-js bot.
> AFAIK JS tests don't give any different output or have any different
> behavior in relation to the display server used.
> 

It makes sense. In the patch update I removed jscore-test in the gtk-wayland-release bot, since it has been running until now.

> > Tools/BuildSlaveSupport/build.webkit.org-config/steps_unittest.py:567
> > +    'GTK Linux 64-bit Debug (Tests)': ['configure build', 'svn', 'kill old processes', 'delete WebKitBuild directory', 'delete stale build files', 'jhbuild', 'download-built-product', 'extract-built-product', 'jscore-test', 'layout-test', 'webkitpy-test', 'webkitperl-test', 'bindings-generation-tests', 'builtins-generator-tests', 'dashboard-tests', 'archive-test-results', 'upload', 'MasterShellCommand', 'API tests', 'webdriver-test'],
> 
> Something seems wrong here. The "GTK Linux 64-bit Debug (Tests)" shouldn't
> run the step 'jscore-test' now... should it?

Right, it seems there's a factory called TestAllButJSC. I needed to update the current bots to run with this factory instead.

[Carlos Alberto Lopez Perez]

Comment on attachment 400810 [details]
Patch

Looks good to me. Thanks!

[EWS]

Committed r262417: <https://trac.webkit.org/changeset/262417>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 400810 [details].

[Radar WebKit Bug Importer]

<rdar://problem/63872948>