[webkit-dev] NRWT Migration Update

Eric Seidel eric at webkit.org
Fri Oct 21 10:29:37 PDT 2011


I didn't even know that page existed. :)

NRWT is Python.  It's split into many files.  It is part of webkitpy
(which is our python library for dealing with webkit).

NRWT was written independently (over several years) at Google.  It
predates webkitpy, but it now lives as part of webkitpy:
http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests
but it doesn't fit 100% into the webkitpy model (it's not a Command
subclass, it doesn't have a Host object) so hacking on it can be
slightly confusing.

As a port author the files of most interest to you are likely the Port
subclasses:
http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests/port
qt.py, gtk.py, etc.

I'm happy to answer any and all questions as they come up.


http://trac.webkit.org/browser/trunk/Tools/Scripts/run-webkit-tests
is still the preferred way to run the tests.  Eventually that wrapper
will be deleted and replaced with NRWT directly.  Until then, it is
smart enough to run ORWT vs NRWT depending on the platform and
configuration.

Also, although NRWT supports running tests in parallel, we have that
feature disabled by default when run from run-webkit-tests.  Running
tests in parallel can cause some tests to flake.  If you're developing
on your local machine and don't care about possible flakes, just want
the fastest test run possible, it's best to run new-run-webkit-tests
directly for now.

-eric


On Fri, Oct 21, 2011 at 6:41 AM, SravanKumar Sandela
<sravan.ken at gmail.com> wrote:
> Hi Eric,
>
> Can you please inform us, what are the major difference between NRWT and
> ORWT. I think it would of lot help for people like me who worked on ORWT
> long back and now to catch up with NRWT.
>
> Thanks a lot in anticipation.
>
> -Sravan.
>
> On Fri, Oct 21, 2011 at 5:03 PM, Eric Seidel <eric at webkit.org> wrote:
>>
>> The WK2 bot looks stable (possibly more stable than before) after the
>> NRWT switchover:
>>
>> http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(WebKit2%20Tests)/r98091%20(15558)/results.html
>>
>> I'll work on switching the --leaks bot back to NRWT next week, and
>> then the only remaining builder to switch is Apple's Windows.  (I
>> likely won't be able to takle that one myself.)
>>
>> Thanks again for your patience in this process.
>>
>> -eric
>>
>> On Thu, Oct 20, 2011 at 9:54 PM, Eric Seidel <eric at webkit.org> wrote:
>> > After further testing tonight, I was able to reproduce the exact set
>> > of 81 WebKit 2 failures with both NRWT and ORWT this evening.  I'm
>> > going to move the bots to NRWT shortly.
>> >
>> > -eric
>> >
>> > On Wed, Oct 19, 2011 at 9:59 PM, Eric Seidel <eric at webkit.org> wrote:
>> >> After a long hiatus, I'm back working on NRWT.
>> >>
>> >> As of this evening all of the blocking issues to switching the WK2 bot
>> >> are resolved:
>> >> https://bugs.webkit.org/show_bug.cgi?id=56729
>> >>
>> >> Just waiting for the WK2 bot to have some smaller number of failures,
>> >> then I'll pull the trigger.
>> >>
>> >>
>> >> After WK2 is switched to NRWT, there is only --leaks and Windows to
>> >> switch, before we can consider deleting ORWT.
>> >>
>> >> Interested parties can continue to track the NRWT migration here:
>> >> https://bugs.webkit.org/show_bug.cgi?id=34984
>> >>
>> >> -eric
>> >>
>> >
>> _______________________________________________
>> webkit-dev mailing list
>> webkit-dev at lists.webkit.org
>> http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
>
>
>
> --
> Living for the unseen and undone....
>


More information about the webkit-dev mailing list