382007 - Focus events inconsistent during tab browsing

Status: VERIFIED WORKSFORME

(Core :: Disability Access APIs, defect)

Description

[Scott Haeger]

When a control has focus in page 1 (forms.html) and you tab browse to page 2 (tables.html) which has focus on the doc frame (probably anywhere for page 2) you receive the following focus events as seen by accerciser:

1) focus(0, 0, None)
	source: [entry | ]
	application: [application | Minefield]
2) object:state-changed:focused(1, 0, None)
	source: [entry | ]
	application: [application | Minefield]
3) focus(0, 0, None)
	source: [document frame | Forms]
	application: [application | Minefield]
4) object:state-changed:focused(1, 0, None)
	source: [document frame | Forms]
	application: [application | Minefield]
5) focus(0, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]
6) object:state-changed:focused(1, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]

I would expect not to see focus on the doc frame from the page I am leaving (forms.html) when a control has focus.  Events 3 and 4 should not occur.  I don't know when this change occurred but it has deviated from the events laid out in http://live.gnome.org/LSR/FocusEvents

Comment 1

[Scott Haeger]

Further testing shows that this problem does not occur for all controls.  I have seen this problem for multiline textboxes, and single and multiselection lists.  The problem does not exist for single line textboxes, buttons, check boxes, and radio buttons.

Comment 2

[Aaron Leventhal]

This is caused by this.mCurrentBrowser.focusedElement.blur(); from bug 175893
That line also causes bug 323805 and bug 323806.

Comment 3

[Aaron Leventhal]

Joanie, is this something we need to fix for Orca support of Firefox 3?

Comment 4

[Joanmarie Diggs]

Aaron, I can't reproduce the bug which makes it hard to know what the impact would be on Orca.

Scott, can you still reproduce it?

Comment 5

[Scott Haeger]

I ran the same test that I ran back in May.  Although the results are different, they still show a problem.  The following are two sets of events looking only at focus: and object:state-changed events.  The first set of events is when the focus is on the multiline textbox in page 1.  The second set of events is when the focus in on a single line textbox in page 1.  In both cases the focus is on the doc frame of page 2.

Multiline textbox:
focus:(0, 0, None)
	source: [document frame | Forms]
	application: [application | Minefield]
object:state-changed:focused(1, 0, None)
	source: [document frame | Forms]
	application: [application | Minefield]
focus:(0, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]
object:state-changed:focused(1, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]
object:state-changed:active(0, 0, None)
	source: [frame | Tables - Minefield]
	application: [application | Minefield]

Single line textbox:
focus:(0, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]
object:state-changed:focused(1, 0, None)
	source: [document frame | Tables]
	application: [application | Minefield]

Once again, I would not expect to see the focus events on the doc frame of page 1 as seen in the multiline textbox.

I will study the impact this has on Orca and report back later.

Comment 6

[Scott Haeger]

Attachment: test page 1

forms.html test page

Comment 7

[Scott Haeger]

Attachment: second test page

Nothing special about this test page.  Attached for completeness.  I believe you can use any page here just as long at focus is on the doc frame.

Comment 8

[Scott Haeger]

My tests show that Orca is not affected by this problem.

Comment 9

[David Bolter [:davidb] (NeedInfo me for attention)]

Mass un-assigning bugs assigned to Aaron.

Comment 10

[alexander :surkov (:asurkov)]

Could you check if this valid still?

Comment 11

[Fernando Herrera]

No. Now we only get focus on the document frame of the second tab when activating tab 2, and the previously focused element in doc 1 when activating tab 1. So probably "works for me"

Comment 12

[alexander :surkov (:asurkov)]

ok, great, thank you. worksforme per comment 11.

Comment 13

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1527381