Safari 1.3, 2.0 and the 2005-05-09 webkit CVS all send pairs of the required key events for the arrow keys. The attached test page records the Safari event sequence for a single tap as "keydown, keypress, keydown, keypress, keyup, keyup". I happen to have a 10.2.8 machine handy with Safari 1.0.3, which does not exhibit this behavior. Nor do Firefox 1.0.4, IE/Mac 5.2.2, or IE/Win 6. Safari sends single "keydown, keypress, keyup" events for keys other than arrows. Safari 2.0 and the CVS version were tested on 10.4.1.
*** Bug 7876 has been marked as a duplicate of this bug. ***
I wonder if bug 8136 (a regression) is caused by redundant keyboard events.
Reassigning to webkit-unassigned, to make sure more people see this.
Hyatt: I had a crazy idea...could this be the cause of some of the performance problems with arrowing down to scroll a page?
*** Bug 10450 has been marked as a duplicate of this bug. ***
(In reply to comment #4) > Hyatt: I had a crazy idea...could this be the cause of some of the performance > problems with arrowing down to scroll a page? > I have seen the same problem with the function keys, escape key, home key and the page up / page down keys. not just the arrow keys.
I wasn't aware there were perf problems when arrowing... do you have an example?
(In reply to comment #7) > I wasn't aware there were perf problems when arrowing... do you have an > example? > see http://trac.mochikit.com/ticket/182 I'm using OS X 10.4.7 and Safari 2.0.4 on PowerPC (also behaves the same on my Intel with, I think, the same versions) I have also tried webkit nightly build from 21/09/06
To work around this problem, one can cancel default processing of the event by returning false from its handler.
Created attachment 11487 [details] proposed fix
<rdar://problem/4125399>
Committed revision 17760.