Bug 34240 – LayoutTests/fast/forms/input-valueasnumber-*.html failures on some platforms

• Status: RESOLVED
• Resolution: FIXED
• Component: Forms

[Kent Tamura]

* input-valueasnumber-datetime.html fails on Windows
> FAIL setValueAsNumberAndGetValue(275760, 8, 13, 0, 0, 0, 1) should be 275760-09-13T00:00:00.001Z. Was 275760-09-13T00:00:00.002Z.

* input-valueasnumber-datetimelocal.html fails on Windows
> FAIL setValueAsNumberAndGetValue(275760, 8, 13, 0, 0, 0, 1) should be 275760-09-13T00:00:00.001. Was 275760-09-13T00:00:00.002.

* input-valueasnumber-{datetime,date,month,datetimelocal}.html fail on all Chromium platforms.
 Issues on dates later than 275760-09-13T00:00Z

[Kent Tamura]

Created attachment 47591 [details]
Patch

[Kent Tamura]

Failures on Windows were caused by a bug of round() in MathExtra.h. I'll fix it later.

[WebKit Commit Bot]

Comment on attachment 47591 [details]
Patch

Rejecting patch 47591 from commit-queue.

Failed to run "['git', 'svn', 'dcommit']" exit_code: 1
Committing to http://svn.webkit.org/repository/webkit/trunk ...
	D	LayoutTests/platform/win/fast/forms/input-valueasnumber-datetime-expected.txt
RA layer request failed: CHECKOUT of '/repository/webkit/!svn/ver/53959/trunk/LayoutTests/platform/win/fast/forms': 500 Internal Server Error (http://svn.webkit.org) at /usr/local/git/libexec/git-core/git-svn line 558


Full output: http://webkit-commit-queue.appspot.com/results/216564

[Kent Tamura]

Landed manually as r53991.

[Darin Adler]

Comment on attachment 47591 [details]
Patch

Why are we disabling these tests? Are we OK with WebKit working differently on these edge cases with and without V8?

[Kent Tamura]

(In reply to comment #5)
> (From update of attachment 47591 [details])
> Why are we disabling these tests? Are we OK with WebKit working differently on
> these edge cases with and without V8?

This difference is not acceptable.
Both of JSC behavior and V8 behavior don't conform to ECMA-262.  I'll fix them and revive these tests.