FWIW, in Chromium, if we had an off-by-default feature that we wanted to
test using wpt, we would probably use the
https://cs.chromium.org/chromium/src/third_party/WebKit/LayoutTests/VirtualTestSuites
mechanism.
On Sat, Jun 3, 2017 at 1:03 AM, youenn fablet wrote:
> Hi Ali,
>
> Yes this is something we should have a solution for. Right now, we have
> two ways:
> - Enabling the runtime-enabled flag only in DumpRenderTree/
> WebKitTestRunner
> - Updating WebKit testharnessreport.js to call internal APIs for some/all
> WPT tests
>
> Ideally, we should be able to set additional configuration information to
> WPT tests without modifying them.
> A configuration file alongside the WPT test might be feasible and could
> supersede the
> technique.
> Folder-based hierarchical configuration might be nice but I am unsure
> whether that would meet the complexity tradeoff bar.
> Nothing yet started there so any help would be greatly appreciated!
> y
>
> Le ven. 2 juin 2017 à 13:37, Ali Juma a écrit :
>
>> Hi all,
>>
>> While a feature is in development behind a runtime-enabled flag, is there
>> a way to write tests in web-platform-tests that enable the feature when
>> they're run as part of LayoutTests? Adding "" tags would technically work but seems a bit
>> questionable.
>>
>> The point of this would be to be able to use web platform tests as
>> regression tests while developing a feature, allowing the tests to be
>> upstreamed sooner and saving having to write them outside of
>> web-platform-tests first and then move them over later.
>>
>> Thanks,
>> Ali
>>
>
> ___
> webkit-dev mailing list
> webkit-dev@lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-dev
>
>
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
https://lists.webkit.org/mailman/listinfo/webkit-dev
