Bug 50169 - REGRESSION (r53857): Removing focus from area element causes unwanted scrolling
Summary: REGRESSION (r53857): Removing focus from area element causes unwanted scrolling
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: 528+ (Nightly build)
Hardware: Mac OS X 10.5
: P1 Normal
Assignee: Darin Adler
URL: http://strickr.de/tests/imagemap.html
Keywords: InRadar, Regression
Depends on:
Reported: 2010-11-29 12:03 PST by Martin Stricker
Modified: 2011-04-07 11:04 PDT (History)
6 users (show)

See Also:

Patch (14.94 KB, patch)
2011-02-01 13:55 PST, Darin Adler
cfleizach: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Stricker 2010-11-29 12:03:39 PST
See Testpage:

Der dots on the map are connected via anchor links to entries in a list below:

- clicking on a dot scrolls down to entry in list (as expected)
- on next click in window (no matter if on link or not) page jumps up again to image (this is not expected)

This behaviour is present in current WebKit Nightly (5533.19.4, r72783) and Chrome Beta (8.0.552.210 beta), but not in current Safari (5533.19.4).
Comment 1 Darin Adler 2010-11-30 12:20:48 PST
Seems likely I broke this with recent image map changes.
Comment 2 komoroske 2010-12-09 16:09:15 PST
A number of users have reported this in Chrome's user forums, too, using Chrome 8.0.552.215.

See http://www.menchies.com/locations.php and http://www.longtermcarelink.net/a7homecare.htm for other examples.
Comment 3 Darin Adler 2011-01-19 11:08:40 PST
If someone else wants to fix this, they are welcome to. I won’t have time to work on it for a while. I’m keeping it assigned to me to remind myself about it.
Comment 4 Darin Adler 2011-01-19 11:08:56 PST
Comment 5 Darin Adler 2011-01-31 17:34:04 PST
I am not so sure this has anything to do with my change. Seems to have something to do with focusing area elements. I did not realize we did that!
Comment 6 Darin Adler 2011-01-31 17:48:08 PST
I suspect this problem was introduced with r53857.
Comment 7 Darin Adler 2011-01-31 17:49:31 PST
The bad code seems to be HTMLAreaElement::dispatchBlurEvent.
Comment 8 chris fleizach 2011-01-31 18:00:04 PST
(In reply to comment #7)
> The bad code seems to be HTMLAreaElement::dispatchBlurEvent.

I only see this happen, if you first tab to the image map item, then click on it. the following click will then jump back to the image
Comment 9 chris fleizach 2011-01-31 18:02:19 PST
(In reply to comment #8)
> (In reply to comment #7)
> > The bad code seems to be HTMLAreaElement::dispatchBlurEvent.

It looks like


that is called from dispatchBlurEvent() -> updateFocusAppearance

is causing the scroll back to occur. Since it's a blur event, it seems like it should not be scrolling on the update focus
Comment 10 Darin Adler 2011-02-01 09:27:20 PST
Yes, that’s right. The problem is that updateFocusAppearance is implemented and used incorrectly in HTMLMapElement. The job of that function, despite its misleading name, is to scroll to reveal the focused element, or something along those lines. HTMLMapElement tries to use it to trigger appropriate repainting of the image element.

I’m working on a patch that redoes that map/image focus code to work with a correct model.
Comment 11 Darin Adler 2011-02-01 13:55:44 PST
Created attachment 80820 [details]
Comment 12 chris fleizach 2011-02-01 14:17:25 PST
Comment on attachment 80820 [details]

Comment 13 Darin Adler 2011-02-01 14:49:24 PST
Committed r77313: <http://trac.webkit.org/changeset/77313>
Comment 14 gmmail2000-bugzilla 2011-03-14 11:29:01 PDT
(In reply to comment #13)
> Committed r77313: <http://trac.webkit.org/changeset/77313>

Will this patch be integrated in Chrome's new version, or what do I have to do to get my imagemap working in Chrome 10.0.648.133? Thanks for the answer.