[webkit-dev] Looking to remove cygwin dependency for javascript tests for Windows ports

[Myles C. Maxfield]

Alex’s claim should be broadened: The right direction is to eventually not use Cygwin for anything at all.

> On Aug 2, 2017, at 8:35 PM, Alex Christensen <achristensen at apple.com> wrote:
> 
> I think eventually running all tests on Windows without Cygwin would be a step in the right direction.
> 
>> On Aug 2, 2017, at 9:53 AM, Szabo, Stephan (San Francisco) <Stephan.Szabo at sony.com> wrote:
>> 
>> Hi,
>> 
>> As part of Sony's work on getting the jsconly build for Windows, we're also looking at the possibility of trying to remove/reduce the dependency on Cygwin for the javascript tests from run-javascript-tests since the build itself runs from a normal windows shell. Internally we did a very preliminary POC of a version of run-jsc-stress-tests for windows which built perl scripts for the test scripts rather than shell scripts and were able to get a bunch of the tests running. We attached that to https://bugs.webkit.org/show_bug.cgi?id=174985 . Obviously, that's not a reasonable version for inclusion, but we wanted to discuss whether there was support for removing the dependency before doing too much more down this path.
>> 
>> Our current thinking is that if we go forward with this, we'd probably step it as:
>> 1. Move the test script and test runner code from run-jsc-stress-tests into a ruby file that is included from the main script
>> 2. Make an option to allow using an alternate version of the above
>> 3. Make an alternate version that didn't rely on shell
>> 
>> Thanks,
>> Stephan
>> _______________________________________________
>> webkit-dev mailing list
>> webkit-dev at lists.webkit.org
>> https://lists.webkit.org/mailman/listinfo/webkit-dev
> 
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-dev