[webkit-help] Automatic Layout Test Execution for Cross Compiled Platforms

Fri Nov 9 01:05:35 PST 2012

On 9 November 2012 11:35, Dirk Pranke <dpranke at chromium.org> wrote:

> Hi Arun,
>
> If you have a port of Python that runs on your embedded platform (some
> kind of Linux, it sounds like) then in theory you could run the
> buildbot tester there, but it would probably be a fair amount of work.
> Also, I thought you didn't have python available?
>

Yes. As of now we don't have python run-time available.

> I would probably choose instead to run a single buildbot instance that
> did both the building and testing, and just figure out how to get the
> few files you need over to the embedded device.
>
> The new-run-webkit-tests test driver is designed (in theory) to work
> in the sort of environment we're talking about, and as we've said
> people have followed an approach much like this on at least two other
> ports. I would definitely be open to changing it to make it work more
> easily / better for you. Feel free to follow up with me off-list if
> you get started hacking on it and have questions.
>

Thanks you very much. I will investigate further about NRWT & related
stuffs & let you know for further help. I'm going to take Android Chromium
as a sample for my work.

Thanks to you both Dirk & Ryosuke for giving some hints.

> -- Dirk
>
> On Thu, Nov 8, 2012 at 9:25 PM, Arunprasad Rajkumar
> <ararunprasad at gmail.com> wrote:
> > Dirk, Thank you very much for giving more information.
> >
> > I understood like, once the completion of build, from host we can trigger
> > the target device(a remote embedded device) to pick up the compiled DRT &
> > other relevant files to start layout test. Upon completion of each
> result it
> > has to be notified to the host which hosts the buildbot slave. Another
> way
> > would be having two buildbot slaves, one for cross compiling the webkit
> > which runs on host machine(x86_64+Linux) and another would be running on
> > embedded device(ARM/MIPS/SH4 + Linux) which only runs layout tests. Final
> > result would be based on these two bots.
> >
> > I will investigate further about what are all the run-time(perl,python)
> need
> > for the things to run. If you found some time, please comment whether my
> > approach is good or too bad.
> >
> > Kind Regards,
> > Arun.
> >
> > On 8 November 2012 23:06, Dirk Pranke <dpranke at chromium.org> wrote:
> >>
> >> Hi Arun,
> >>
> >> Ryosuke's suggestion is a good one; in addition, I believe there is a
> >> group at Sony (?) that has done something similar, but their patches
> >> haven't been upstreamed yet.
> >>
> >> On Thu, Nov 8, 2012 at 3:39 AM, Arunprasad Rajkumar
> >> <ararunprasad at gmail.com> wrote:
> >> > Thanks Ryosuke.
> >> >
> >> > Are you guys running Tools/Scripts/new-run-webkit-tests script in
> >> > Android
> >> > device or it is been triggered from the host?
> >> >
> >>
> >> It runs on the host.
> >>
> >> > Usually things are cross compiled right? How the build system & target
> >> > system(where webkit runs) are interconnected?
> >> >
> >>
> >> Yes, things are compiled on the host. If you're developing
> >> interactively, you just compile as normal, and then run
> >> new-run-webkit-tests, which will either expect the binaries to have
> >> been downloaded to the device, or download them itself. The script
> >> contains all the hooks you need to implement whatever process you want
> >> (it can even do a build to ensure things are up to date).
> >>
> >> The major wrinkle you need to implement is to figure out some way for
> >> the binary on the embedded driver to be able to access the test files
> >> that live on the host. Usually either you can cross-mount a filesystem
> >> or wrap file:// access urls in something like an http server that can
> >> retrieve things.
> >>
> >> In an automated system (like buildbot), the bot scripts run on the
> >> host and just launch new-run-webkit-tests when appropriate.
> >>
> >> -- Dirk
> >>
> >> > On 8 November 2012 16:14, Ryosuke Niwa <rniwa at webkit.org> wrote:
> >> >>
> >> >> You might be interested in looking into how Chromium Android port
> solve
> >> >> this problem. See
> >> >> Tools/Scripts/webkitpy/layout_tests/port/chromium_android.py.
> >> >>
> >> >> On Thu, Nov 8, 2012 at 2:36 AM, Arunprasad Rajkumar
> >> >> <ararunprasad at gmail.com> wrote:
> >> >>>
> >> >>>
> >> >>> Hello Developers,
> >> >>>
> >> >>> I'm working on a embedded platform where I need to automate the
> Layout
> >> >>> Tests. From the Tools/Scripts/new-run-webkit-tests &
> >> >>> https://www.webkit.org/blog/1456/layout-tests-practice/ I
> understood
> >> >>> like it
> >> >>> requires python to run layout tests, since the web server & related
> >> >>> things
> >> >>> implemented in python. But the platform I'm working on is a resource
> >> >>> constraint one & doesn't have these run-times(python,perl,..etc).
> >> >>>
> >> >>> Do you guys have any suggestions for my question? I'm thinking to
> use
> >> >>> js-test-driver for some of the non webkit specific tests. Will the
> >> >>> same
> >> >>> thing can be used for WebKit Layout tests.
> >> >>>
> >> >>> Thanks in advance.
> >> >>>
> >> >>> Kind Regards,
> >> >>> Arun.
> >> >>> --
> >> >>> Arunprasad Rajkumar
> >> >>> http://in.linkedin.com/in/ararunprasad
> >> >>>
> >> >>>
> >> >>> _______________________________________________
> >> >>> webkit-help mailing list
> >> >>> webkit-help at lists.webkit.org
> >> >>> http://lists.webkit.org/mailman/listinfo/webkit-help
> >> >>>
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > Arunprasad Rajkumar
> >> > http://in.linkedin.com/in/ararunprasad
> >> >
> >> >
> >> > _______________________________________________
> >> > webkit-help mailing list
> >> > webkit-help at lists.webkit.org
> >> > http://lists.webkit.org/mailman/listinfo/webkit-help
> >> >
> >
> >
> >
> >
> > --
> > Arunprasad Rajkumar
> > http://in.linkedin.com/in/ararunprasad
> >
>

-- 
*Arunprasad Rajkumar*
http://in.linkedin.com/in/ararunprasad
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-help/attachments/20121109/64a864cb/attachment-0001.html>