Bug 6385

Summary:

REGRESSION: Unnamed frames return a generated name as window.frameElement.name

Product:

WebKit

Reporter:

Alexey Proskuryakov <ap>

Component:

Frames

Assignee:

Alexey Proskuryakov <ap>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

alice.barraclough

Priority:

Version:

420+

Hardware:

Mac

OS:

OS X 10.4

Attachments:

Description	Flags
test case	none
patch that fixes this regression	mjs: review+
correct the test case	darin: review+

Alexey Proskuryakov

Reported 2006-01-05 10:56:08 PST

This is a rather old regression (at least from September).

Attachments
test case (2.71 KB, application/zip) 2006-01-05 10:56 PST, Alexey Proskuryakov	no flags	Details
patch that fixes this regression (3.30 KB, patch) 2006-01-22 23:17 PST, Darin Adler	mjs: review+	Details Formatted Diff Diff
correct the test case (2.37 KB, patch) 2006-01-23 21:56 PST, Alexey Proskuryakov	darin: review+	Details Formatted Diff Diff
View All Add attachment proposed patch, testcase, etc.

Alexey Proskuryakov

Comment 1 2006-01-05 10:56:58 PST

Created attachment 5501 [details] test case I wonder if it's possible to create self-contained tests for such issues...

Alexey Proskuryakov

Comment 2 2006-01-05 12:25:02 PST

Ah, the actual returned name is "-->", which gets displayed as "-->". Firefox returns an empty name, not a generated one like this... Interestingly, HTMLElement::frameGetter is not invoked to get the name at all, although it has code for FrameName case; Window::getValueProperty is invoked, and retrieves the name as m_part->name().

Darin Adler

Comment 3 2006-01-06 11:17:03 PST

Are you sure this is a regression? We've always generated names like this internally.

Alexey Proskuryakov

Comment 4 2006-01-06 12:58:27 PST

(In reply to comment #3) Yes; with stock Safari 2.0.2, window.frameElement.name returns the original empty name (I've double- checked using escape(), to make sure the returned name wasn't just invisible because of its comment-like form).

Alice Liu

Comment 5 2006-01-11 17:31:08 PST

<rdar://problem/4400230>

Joost de Valk (AlthA)

Comment 6 2006-01-22 04:36:53 PST

Adding Regression keyword.

Darin Adler

Comment 7 2006-01-22 23:17:28 PST

Created attachment 5865 [details] patch that fixes this regression

Darin Adler

Comment 8 2006-01-22 23:18:06 PST

The regression is fixed, but window.name still gives the strange frame name. We might not want that and it's worth doing some additional testing and perhaps filing additional bugs.

Maciej Stachowiak

Comment 9 2006-01-23 13:41:15 PST

Perhaps generated names should be kept as a purely internal thing. I'm not really sure what we use them for.

Alexey Proskuryakov

Comment 10 2006-01-23 21:55:48 PST

Filed bug 6751 for the window.name issue. Here, the test case lacks "resources/" for the left frame, reopening because of this. Also, perhaps it would be better to use escape() on the names, to make sure the right one is really empty.

Alexey Proskuryakov

Comment 11 2006-01-23 21:56:18 PST

Created attachment 5901 [details] correct the test case

Darin Adler

Comment 12 2006-01-23 23:18:28 PST

Comment on attachment 5901 [details] correct the test case You're welcome to update to use escape() -- I did the other half.

Darin Adler

Comment 13 2006-01-23 23:18:48 PST

Remaining bit is not P1.

Eric Seidel (no email)

Comment 14 2006-01-31 21:20:41 PST

Removing Regression keyword from bugs already fixed.

Note You need to log in before you can comment on or make changes to this bug.