Port of commit 667785 from kjs

• Component: JavaScriptCore
• Status: RESOLVED FIXED

[Luciano Montanaro]

The log for that change says:
"fixed special no-args case I discovered when writing tests
more or less accidentally".

[Luciano Montanaro]

Created attachment 14700 [details]
The two-line patch

The patch for the backported fix

[David Kilzer (:ddkilzer)]

Comment on attachment 14700 [details]
The two-line patch

Please set the "review?" flag on patches you'd like reviewed.  Thanks!

This patch also needs a ChangeLog and a layout test.  See: http://webkit.org/coding/contributing.html

[Darin Adler]

Comment on attachment 14700 [details]
The two-line patch

Needs a test.

[Darin Adler]

Comment on attachment 14700 [details]
The two-line patch

Patch is great.

We need both a change log and a layout test for this, as Dave says.

[Luciano Montanaro]

Created attachment 14712 [details]
Result of the localeCompare testcases with IE6

testcase log showing IE6 has a different behavior vs. Firefox.

[Harri Porten]

Given the IE6 test results that Luciano kindly attached to this report I suddenly developed some doubts about this patch. I only discovered this special case by accident when writing a test. But if Firefox is the only browser showing this behavior I am tempted to remove it again from Konqueror. Might base this decision on what you guys do, though.

[Sam Weinig]

Created attachment 14728 [details]
patch with updated test case

This patch includes the changes and updates fast/js/kde/StringObject.html to match kde which include new localeCompare tests.  It should be noted that both IE and Opera do not match this behavior, but Firefox does.

[Sam Weinig]

My concerns mirror Harri's.

[Darin Adler]

Comment on attachment 14728 [details]
patch with updated test case

r=me

[Sam Weinig]

Landed in r21806.