W3C home > Mailing lists > Public > public-webapi@w3.org > May 2006

Re: ACTION-119: dblclick analysis

From: Maciej Stachowiak <mjs@apple.com>
Date: Sat, 6 May 2006 22:35:55 -0700
Message-Id: <C2E24BE5-6D90-4B23-8FB7-C4D4FA06F208@apple.com>
Cc: "Web APIs WG (public)" <public-webapi@w3.org>
To: Bjoern Hoehrmann <derhoermi@gmx.net>


On May 6, 2006, at 7:16 PM, Bjoern Hoehrmann wrote:

> * Maciej Stachowiak wrote:
>> It looks like Mac browsers are inconsistent on this. I used this test
>> case:
>
> For click.detail this is fine as I explained in the previous message.
> For dblclick.detail it seems only Safari produces unexpected results,
> i.e. dblclick.detail can be != 2. If you think we should not require
> .detail to equal 2, a new issue would be due. [1]

Presumably Firefox would do the same if it actually fired more than  
one dblclick event per click sequence on the Mac, as per the platform  
convention.

I think dblclick.detail should be the same as the detail for the  
corresponding click event; this matches every browser that supports  
DOM events on every platform. It's not clear why making it always be  
2 would be better. I think if you want to change from the existing  
behavior it is up to you to justify it.

>> All browsers seem to agree that dblclick.detail should match
>> click.detail for the corresponsing click event, although Opera does
>> not appear to be following Mac platform conventions for click count,
>> and Firefox does not appear to be following the proposed detail % 2
>> == 0 rule on the Mac.
>
> I assumed this wasn't a proposed rule but some illustration.

I won't try to speak for Anne but I assumed he intended it as a  
normative requirement.

> It is not clear to me that we want to have this relationship. For  
> example,
> the environment might be configured to report clicks for both left
> and right mouse button clicks, but double clicks can only occur if
> performed with the left mouse button. Or the environment might be
> such that in a sequence of clicks the double click can only occur
> once (as Firefox apparently does on your platform).

Notwithstanding all of these hypothetical examples, I don't think a  
dblclick ever can or should occur when the click count is odd. And I  
think the Firefox behavior is a bug, probably due to Firefox getting  
more attention on Windows and Linux where click counts >= 4 do not  
normally occur. Certainly in other cases Firefox attempts to follow  
platform conventions.

> And in fact, the platform convention on Windows is that you get the  
> double click
> before you release the button the second time [2], while we require
> the button to be released for a "click" to occur.

On Windows, both Internet Explorer and Firefox do not fire the  
dblclick event until the mouse button is released, so this convention  
does not apply to web content. (I tested this with my earlier posted  
test case.) IE fires dblclick instead of click, rather than in  
addition to it, which is contrary to the DOM Events requirement for  
click, but all other browsers I could find to test always fire  
dblclick after the corresponding click.

Given all this, I think the mod 2 rule would be good as a normative  
requirement, as I thought we'd already decided.

> So I was going to say that the notion of a double click depends on  
> the environment configuration, just like we do for 'click'.

Can you give specific examples where Anne's proposed "even click  
count" rule would not lead to the right result?

Regards,
Maciej
Received on Sunday, 7 May 2006 05:36:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:55 GMT