Bug 91667 – Use native (non-cygwin) binaries for perl, gperf, and bison in Chromium

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Scott Graham]

Use native (non-cygwin) binaries for perl, gperf, and bison in Chromium

[Scott Graham]

Created attachment 153084 [details]
Patch

[Kentaro Hara]

Comment on attachment 153084 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=153084&action=review

Thanks for the great speed-up.

> Source/WebCore/make-hash-tools.pl:35
>      my $colorDataGperf             = $ARGV[0];
> +    my $customGperf                = $ARGV[1];
> +    shift;
>      shift;

Nit: You can write like this:

    my $colorDataGperf = shift;
    my $customGperf = shift;

[Scott Graham]

Created attachment 153273 [details]
Patch

[Scott Graham]

Comment on attachment 153273 [details]
Patch

Done, thanks.

[Kentaro Hara]

Comment on attachment 153273 [details]
Patch

OK

[WebKit Review Bot]

Comment on attachment 153273 [details]
Patch

Clearing flags on attachment: 153273

Committed r123117: <http://trac.webkit.org/changeset/123117>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Tony Chang]

The Windows build appears to be hanging on the canary bots.  I suspect this change, but I'm not sure.  I'll try to repro on a local win machine.

[Tony Chang]

http://build.chromium.org/p/chromium.webkit/waterfall?builder=Webkit+Win+Builder&builder=Webkit+Win+Builder+%28dbg%29&builder=Win+Builder&builder=Win7+Perf&reload=none

Something between r123112 and r123118.

[Scott Graham]

(In reply to comment #9)
> http://build.chromium.org/p/chromium.webkit/waterfall?builder=Webkit+Win+Builder&builder=Webkit+Win+Builder+%28dbg%29&builder=Win+Builder&builder=Win7+Perf&reload=none
> 
> Something between r123112 and r123118.

Must be this, perhaps a missing dependency causing a missing DLL dialog. Could you revert? (I'm not a committer)

[Tony Chang]

Actually, I'm going to roll out because this breaks build-webkit --chromium.  We haven't added perl, bison, or gperf to the Source/WebKit/chromium/DEPS.

[Tony Chang]

Reverted r123117 for reason:

Breaks chromium-win compile.

Committed r123124: <http://trac.webkit.org/changeset/123124>

[Scott Graham]

Created attachment 153607 [details]
Patch

[Scott Graham]

Added entries to chromium/DEPS. I think they're right, but I haven't been able to get update-webkit/build-webkit to work on my Windows machine (unrelated this patch) so I'm not 100% sure.

[Tony Chang]

(In reply to comment #14)
> Added entries to chromium/DEPS. I think they're right, but I haven't been able to get update-webkit/build-webkit to work on my Windows machine (unrelated this patch) so I'm not 100% sure.

Did you make any other changes since the last change?  The previous failures I linked too weren't related to the DEPS problem-- they were on the canary bots which use a chromium checkout.

[Scott Graham]

(In reply to comment #15)
> (In reply to comment #14)
> > Added entries to chromium/DEPS. I think they're right, but I haven't been able to get update-webkit/build-webkit to work on my Windows machine (unrelated this patch) so I'm not 100% sure.
> 
> Did you make any other changes since the last change?  The previous failures I linked too weren't related to the DEPS problem-- they were on the canary bots which use a chromium checkout.

Oh, shoot, I see. No, I didn't change anything else. I'll try to reproduce that failure.

[Tony Chang]

I'll go ahead and land the DEPS change for you.  It should be safe and maybe make it easier to isolate the failure.

[Tony Chang]

DEPS change:
http://trac.webkit.org/changeset/123274

[Scott Graham]

(In reply to comment #17)
> I'll go ahead and land the DEPS change for you.  It should be safe and maybe make it easier to isolate the failure.

thanks!

[Scott Graham]

Created attachment 153864 [details]
Patch

[Tony Chang]

Comment on attachment 153864 [details]
Patch

Clearing flags on attachment: 153864

Committed r123387: <http://trac.webkit.org/changeset/123387>

[Tony Chang]

All reviewed patches have been landed.  Closing bug.

[Tony Chang]

I landed the patch manually so we can watch it on the bots.

[Scott Graham]

(In reply to comment #23)
> I landed the patch manually so we can watch it on the bots.

Sigh, apparently something else broke. Could I get you to revert yet again? My apologies for all the noise.

[WebKit Review Bot]

Re-opened since this is blocked by 92036

[Tony Chang]

(In reply to comment #24)
> (In reply to comment #23)
> > I landed the patch manually so we can watch it on the bots.
> 
> Sigh, apparently something else broke. Could I get you to revert yet again? My apologies for all the noise.

Reverting in https://bugs.webkit.org/show_bug.cgi?id=92036

FWIW, you should be able to run these on the chromium try bots if you're having problems running it locally:
http://www.chromium.org/developers/testing/try-server-usage#TOC-My-patch-is-in-WebKit-only

[Tony Chang]

mid-air collision

[Scott Graham]

Created attachment 154108 [details]
Patch

[Tony Chang]

Comment on attachment 154108 [details]
Patch

Actually, I want to land by hand.

[Tony Chang]

Comment on attachment 154108 [details]
Patch

Clearing flags on attachment: 154108

Committed r123496: <http://trac.webkit.org/changeset/123496>

[Tony Chang]

All reviewed patches have been landed.  Closing bug.

[Scott Graham]

Previous attempts were failing due to IDE exiting silently when building a file with \'s in #line directives. Totally bizarre, see http://build.chromium.org/p/tryserver.chromium/builders/win_layout/builds/630
"SIGKILL failed to kill process" in compile step. Reproduced/debugged locally.

Current try here: http://build.chromium.org/p/tryserver.chromium/builders/win_layout/builds/636 which succeeded compile (but will probably fail later on unrelated stuff).