Closed Bug 438784 Opened 12 years ago Closed 12 years ago

URL of loaded page no longer exposed as the accValue for the Doc accessible.


(Core :: Disability Access APIs, defect, major)

1.9.0 Branch
Not set





(Reporter: MarcoZ, Assigned: MarcoZ)




(Keywords: access, fixed1.9.0.2, regression)


(2 files)

Regression range is

Regression introduced after beta 5. This causes the NVDA developers not to be able to look at a document's URL without resorting to the iSimple* interfaces to determine if a link is a same page link. Using IAccessible2 should not under any circumstances require the use of any iSimple* interfaces.

Alexander, any obvious candidates from that day sticking out at you?
Flags: blocking1.9.0.1?
It's broken in the platform independent code, not any MSAA specific stuff.
OS: Windows XP → All
Hardware: PC → All
Only MSAA needs to do expose the doc's URL in the value. In ATK/AT-SPI it's exposed as a document attribute.
Regression from bug 425693.
Blocks: 425693
This restores functionality. With nsDocAccessibleWrap::get_accValue being the only instance overriding nsAccessibleWrap::get_accValue, and from tests, I don't see any negative impact from this. Textfields, comboboxes still work as expected.
Assignee: nobody → marco.zehe
Attachment #324761 - Flags: review?(aaronleventhal)
Comment on attachment 324761 [details] [diff] [review]
Only allocate memory if the value is not empty, but return S_OK in any case.

r+ for trunk.
Attachment #324761 - Flags: review?(aaronleventhal) → review+
Try-Server builds will appear here:

Jamie, please check it out and see if this is what you want, and if you see any negative side effects with NVDA.
This also fixes the problem and reinstates part of the backout I did in bug 425693 from bug 421650. Reminder: The original reason to do this for the MSAA-related methods was bug 424971, which was fixed by backing out the "if isempty return false" part for get_accname, to fix a problem JAWS 8 was having with this in alerts.
Marco, if your alternative way doesn't bring any regression for Jaws8 then I like it more. Just please add comment for get_accName why it's a different. Also what do you think about another methods?
(In reply to comment #8)
> Also what do you think about another methods?

You mean the other methods that were changed in bug 425693? These changes didn't introduce any regressions AFAIK, the only one that regressed was this one.
Flags: wanted1.9.0.x?
Flags: blocking1.9.0.1?
Flags: blocking1.9.0.1-
Comment on attachment 324787 [details] [diff] [review]
Alternative: Revert part of fix for bug 425693.

Officially asking review for the alternative and more safe patch. I'll add a comment to get_accName to clarify why that one behaves differently.
Attachment #324787 - Flags: review?(surkov.alexander)
Comment on attachment 324787 [details] [diff] [review]
Alternative: Revert part of fix for bug 425693.

Attachment #324787 - Flags: review?(surkov.alexander) → review+
Depends on: 441737
Pushed to mozilla-central in changeset:
Closed: 12 years ago
Resolution: --- → FIXED
No longer depends on: 441737
When we get to platform-specific tests, revisit this bug to add a testcase.
Flags: in-testsuite?
Verified fixed in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1a1pre) Gecko/2008062603 Minefield/3.1a1pre
Comment on attachment 324787 [details] [diff] [review]
Alternative: Revert part of fix for bug 425693.

Has been baking on 1.9.1a1 for weeks, didn't cause any regressions, and would help the NVDA people tremendously with identifying links that point to a spot within the same page.
Attachment #324787 - Flags: approval1.9.0.2?
Comment on attachment 324787 [details] [diff] [review]
Alternative: Revert part of fix for bug 425693.

Approved for Please land in CVS. a=ss.
Attachment #324787 - Flags: approval1.9.0.2? → approval1.9.0.2+
Checking in accessible/src/msaa/nsAccessibleWrap.cpp;
/cvsroot/mozilla/accessible/src/msaa/nsAccessibleWrap.cpp,v  <--  nsAccessibleWrap.cpp
new revision: 1.119; previous revision: 1.118
Keywords: fixed1.9.0.2
Flags: wanted1.9.0.x? → wanted1.9.0.x+
Flags: in-testsuite?
You need to log in before you can comment on or make changes to this bug.