We are trying to bring the X86 port in JSC up to speed. We are now at a point where it would make sense to have a EWS bot to track regressions, since some of them will apply to all 32bit ports. If this is set up in a Linux x86-64 machine all you need to do is build JSC with build-jsc --32-bit (recently made to work).
Created attachment 366935 [details] patch
We are already aware about the new EWS CI (https://bugs.webkit.org/show_bug.cgi?id=196641) but , in meantime, we wish go ahead with this new EWS. The idea is to move all the EWS workers at the same time in the future.
Comment on attachment 366935 [details] patch What architecture is this? x86? Could we name it jsc-i386-ews, jsc-i686-ews, or something to make it more clear?
(In reply to Michael Catanzaro from comment #3) > Comment on attachment 366935 [details] > patch > > What architecture is this? x86? > > Could we name it jsc-i386-ews, jsc-i686-ews, or something to make it more > clear? Yes. it is i386. Since it other EWS for x86_64 has not a reference to the PC architecture in the name I assumed as a right notation just add the 32 bit and make no references to the architecture. In others words no architecture reference = PC. I have no objections to rename the worker with other nomenclature whatever you'd prefer better.
Having i386 in the name would be consistent with jsc-armv7-ews.
(In reply to Alexey Proskuryakov from comment #5) > Having i386 in the name would be consistent with jsc-armv7-ews. Yes, this makes sense.
Created attachment 367032 [details] patch
(In reply to Pablo Saavedra from comment #6) > (In reply to Alexey Proskuryakov from comment #5) > > Having i386 in the name would be consistent with jsc-armv7-ews. > > Yes, this makes sense. New patch uploaded with the required changes
Comment on attachment 367032 [details] patch Clearing flags on attachment: 367032 Committed r244076: <https://trac.webkit.org/changeset/244076>
All reviewed patches have been landed. Closing bug.
<rdar://problem/49736225>