Dispatch of DOMActivate from trigger elements is not reliable

[GoogleCodeExporter]

The attached form contains a trigger and an output. The output displays a 
count of the number of times that the trigger is activated, which is 
maintained by a setvalue action on DOMActivate of the trigger. If the trigger 
is activated repeatedly without a sufficient pause between activations, it 
can easily be seen that the counter is not updated sanely. Perhaps every 
second or third time (ish), the counter can be seen to increment by two 
rather than one.

Original issue reported on code.google.com by phil.boo...@gtempaccount.com on 21 Sep 2009 at 11:20

• Blocking: #531, #561, #564, #565, #566, #567

Attachments:

• trigger-domactivate.html

[GoogleCodeExporter]

Some more observations on this issue.

In Internet Explorer, I can't reproduce the problem. No matter how quickly I 
repeatedly 
activate the trigger, the counter only ever increments in single units.

In Chrome, the situation is worse. Every time the trigger is activated the 
counter 
increments by at least two, even when there is a lengthy pause between 
activations. 
Sometimes the counter is incremented by three for a single activation.

Original comment by phil.boo...@gtempaccount.com on 21 Sep 2009 at 11:32

[GoogleCodeExporter]

Updating the labels.

Original comment by phil.boo...@gtempaccount.com on 23 Sep 2009 at 10:52

• Added labels: Browser-FF, Browser-Safari

[GoogleCodeExporter]

Original comment by phil.boo...@gtempaccount.com on 28 Sep 2009 at 3:09

• Changed state: Started
• Added labels: Iteration-20090925

[GoogleCodeExporter]

Original comment by phil.boo...@gtempaccount.com on 28 Sep 2009 at 3:11

[GoogleCodeExporter]

There are actually two problems at work here.

Firstly, the click and dblclick events do not appear to be mutually exclusive 
in 
Firefox and Safari, as they are in Internet Explorer. Thus, by mapping both 
events to 
DOMActivate as we currently do, we cause three DOMActivate events to be 
dispatched 
when a user double clicks on a trigger.

Secondly, Safari actually dispatches the DOMActivate correctly itself, unlike 
the 
other two browsers. Therefore, when we call mapclick2domactivate() in Safari, 
we are 
ensuring that every click results in two DOMActivate events being dispatched.

Original comment by phil.boo...@gtempaccount.com on 30 Sep 2009 at 12:10

[GoogleCodeExporter]

A proposed fix is awaiting review at r3022.

Original comment by phil.boo...@gtempaccount.com on 30 Sep 2009 at 12:33

• Changed state: InReview

[GoogleCodeExporter]

The fix proposed at r3022 actually causes a weird refresh problem, described in 
issue 
574. So I'm reverting the status of this issue to Started and working on a new 
proposal 
for this issue now.

Original comment by phil.boo...@gtempaccount.com on 1 Oct 2009 at 2:23

• Changed state: Started

[GoogleCodeExporter]

A revised proposed fix is awaiting review at r3028.

Original comment by phil.boo...@gtempaccount.com on 1 Oct 2009 at 3:00

• Changed state: InReview

[GoogleCodeExporter]

In trunk at r3030.

Original comment by phil.boo...@gtempaccount.com on 2 Oct 2009 at 1:58

• Changed state: Fixed