Bug 112680

Summary:

REGRESSION(r146174): It made jquery/manipulation.html fail on 32 bit platforms

Product:

WebKit

Reporter:

Csaba Osztrogonác <ossy>

Component:

JavaScriptCore

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Critical

CC:

ap, ggaren, kadam, msaboff, ossy, zarvai

Priority:

Version:

528+ (Nightly build)

Hardware:

All

OS:

All

Bug Depends on:

Bug Blocks:

79666, 112609

Attachments:

Description	Flags
diff from bot archive	none

Description Csaba Osztrogonác 2013-03-19 04:13:38 PDT

http://trac.webkit.org/changeset/146174 and http://trac.webkit.org/changeset/146178 made
jquery/manipulation.html fail on 32 bit platforms. At least on Qt, because only Qt
has 32 bit tester bot. See this diff for details:

http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20NRWT/r146178%20%2830468%29/jquery/manipulation-pretty-diff.html

Comment 1 Ádám Kallai 2013-03-19 06:27:28 PDT

Skipped in: http://trac.webkit.org/changeset/146201

Comment 2 Michael Saboff 2013-03-19 08:35:18 PDT

(In reply to comment #0)
> http://trac.webkit.org/changeset/146174 and http://trac.webkit.org/changeset/146178 made
> jquery/manipulation.html fail on 32 bit platforms. At least on Qt, because only Qt
> has 32 bit tester bot. See this diff for details:
> 
> http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20NRWT/r146178%20%2830468%29/jquery/manipulation-pretty-diff.html

Was there a crash with backtrace as well?

Comment 3 Csaba Osztrogonác 2013-03-19 08:50:13 PDT

No, It is only fail, not crash.

Comment 4 Michael Saboff 2013-03-19 13:28:58 PDT

Built r146178 32bit Mac WebKit and jquery/manipulation.html works fine for me.

Comment 5 Alexey Proskuryakov 2014-07-31 10:29:21 PDT

This test is very slow, particularly in debug builds, but there doesn't appear to be anything 32-bit specific there any more. OK to close?

Comment 6 Csaba Osztrogonác 2014-07-31 10:47:26 PDT

Created attachment 235831 [details]
diff from bot archive

Comment 7 Csaba Osztrogonác 2014-07-31 10:53:05 PDT

(In reply to comment #5)
> This test is very slow, particularly in debug builds, but there doesn't appear to be anything 32-bit specific there any more. OK to close?

It is ~1.5 years old bug, who knows what happened during this long period.
There isn't QtWebKit anymore, there isn't any EFL or GTK 32 bit tester bot.

We only know that it was a valid 32 bit regression (diff attached) caused by
r146174 and r146178. And the test runtime wasn't slow at all, it ran on the
bot in 2 seconds, so it wasn't related to slowness at all.

Comment 8 Csaba Osztrogonác 2014-07-31 10:59:27 PDT

I checked it on QtWebKit release branch (branched before removing from trunk)
and it passes now, so I assume it is already fixed on trunk too.