369511 - Focus jumps outside of the tabbed widget completely

Status: RESOLVED INVALID

(Core :: Disability Access APIs, defect)

Description

[Wayne DeAngelo]

Attachment: Testcase

Env: FF3 Minefield Nightly build 20070206. 

Problem: When focus is in a tabbed widget and ctrl+tab is pressed, focus can jump outside of the tabbed widget completely. 

Steps to Recreate:
1. Start the file called Testcases, Next click on JAWS failures when type='content" (Note tcase is in the JAWS failures even though it is clearly a FF3 issue).
2. When focus is on a tab1 checkbox, and ctrl+tab is pressed, focus jumps away from the tab-panel. This occurs whether or not type='content'
3. The focus can end up anywhere in the application.

Comment 3

[Wayne DeAngelo]

Attachment: complete testcase

Finally got the testcase down to a reasonable size.

Comment 4

[Aaron Leventhal]

What we need is a test case of just a simple XUL file.

Comment 5

[Aaron Leventhal]

Wayne, I see a bunch of folders and XUL files there. We should really be posting just one XUL file for each test case.

Comment 6

[Tom Brunet]

Attachment: The XUL file of interest

Comment 7

[Tom Brunet]

Attachment: Referred to by main.xul

Comment 8

[Tom Brunet]

Attachment: The two files required to see the problem.

Sorry - the two files packaged up a little cleaner.

Comment 9

[Aaron Leventhal]

This occurs only in JAWS -- please file with Freedom Scientific.

Comment 10

[Tom Brunet]

I see the issue without JAWS running when type='content' is not specified (the bottom set of tabs) on the latest minefield:
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a3pre) Gecko/20070213 Minefield/3.0a3pre

The type='content' case (top set of tabs) seems to be fixed as of Firefox 2.0.  Both cases are broken when using xulrunner 1.8.0.4.

Comment 11

[Aaron Leventhal]

Someone please open a new bug with an accurate description and easy testcase.

Comment 12

[Wayne DeAngelo]

Closing as invalid, talked to the TB ( the person that saw the problem orginally )and he stated that he will open a new defect with a new "easy tcase".

Comment 13

[Tom Brunet]

New write-up and clarified as bug 370396.