Bug 214279

Summary: [ews] watchos should build both arm64_32 and armv7k architectures
Product: WebKit Reporter: Aakash Jain <aakash_jain>
Component: Tools / TestsAssignee: Aakash Jain <aakash_jain>
Status: RESOLVED FIXED    
Severity: Normal CC: aakash_jain, ap, darin, jbedard, keith_miller, thorton, webkit-bug-importer, ysuzuki
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
See Also: https://bugs.webkit.org/show_bug.cgi?id=214147
https://bugs.webkit.org/show_bug.cgi?id=214306
https://bugs.webkit.org/show_bug.cgi?id=214310
Attachments:
Description Flags
Patch none

Description Aakash Jain 2020-07-13 15:05:22 PDT
watchos queue in EWS should build both arm64_32 and armv7k architectures. This will help in preventing build breakage on both architectures.
Comment 1 Aakash Jain 2020-07-13 15:09:34 PDT
Created attachment 404181 [details]
Patch
Comment 2 Aakash Jain 2020-07-13 15:12:14 PDT
Sample run: https://ews-build.webkit-uat.org/#/builders/42/builds/81
Comment 3 Keith Miller 2020-07-13 15:18:31 PDT
(In reply to Aakash Jain from comment #2)
> Sample run: https://ews-build.webkit-uat.org/#/builders/42/builds/81

I don't see any clang invocations using armv7k? Are you sure that's right? Does it need to be a clean build?
Comment 4 Darin Adler 2020-07-13 15:28:42 PDT
I did review+ because the change looks logical and the EWS is green for watchOS.

But maybe it’s not working. Keith’s question does seem relevant.
Comment 5 Aakash Jain 2020-07-13 18:08:57 PDT
(In reply to Keith Miller from comment #3)
> (In reply to Aakash Jain from comment #2)
> I don't see any clang invocations using armv7k? Are you sure that's right? Does it need to be a clean build?
That was due to missing parameter 'ONLY_ACTIVE_ARCH=NO'. Fixed that in https://trac.webkit.org/r264327. Current patch should be good now.

However, it seems like the build is broken for armv7k, e.g.: https://ews-build.webkit-uat.org/#/builders/42/builds/93

In file included from Source/WebCore/editing/TextIterator.h:28:
Source/WebCore/editing/CharacterRange.h:74:26: error: implicit conversion loses integer precision: 'const uint64_t' (aka 'const unsigned long long') to 'CFIndex' (aka 'long') [-Werror,-Wshorten-64-to-32]
Comment 6 Darin Adler 2020-07-13 18:11:14 PDT
We had a lot of breakage like this a while back. Someone changed the warning settings I think? Maybe Jonathan Bedard?
Comment 7 Darin Adler 2020-07-13 18:11:40 PDT
There are other kinds of fixes for those warnings, but I think changing the settings is what someone chose to do.
Comment 8 Jonathan Bedard 2020-07-14 10:11:39 PDT
(In reply to Darin Adler from comment #6)
> We had a lot of breakage like this a while back. Someone changed the warning
> settings I think? Maybe Jonathan Bedard?

That was the fix, although we need a bit more that that for armv7k, I'll have a patch ready on https://bugs.webkit.org/show_bug.cgi?id=214306 shortly.
Comment 9 EWS 2020-07-14 11:38:50 PDT
Committed r264358: <https://trac.webkit.org/changeset/264358>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 404181 [details].
Comment 10 Radar WebKit Bug Importer 2020-07-14 11:39:14 PDT
<rdar://problem/65555319>
Comment 11 Aakash Jain 2020-07-14 12:31:02 PDT
Deployed on the EWS server. Seems to be working fine, e.g.: https://ews-build.webkit.org/#/builders/41/builds/342/steps/7/logs/stdio