W3C home > Mailing lists > Public > www-dom@w3.org > January to March 2005

Re: [DOM3 Events] key repeating

From: Arnaud Le Hors <lehors@us.ibm.com>
Date: Tue, 15 Mar 2005 15:22:19 -0800
Message-ID: <42376E2B.6090506@us.ibm.com>
To: www-dom@w3.org
CC: "L. David Baron" <dbaron@dbaron.org>

Good luck with that. I looked at this in 97 when I was working on HTML 
4.0, I wanted to specify the order these events occur but found out that 
each platform (Unix/X, Windows, & Mac) had its own very specific way of 
doing so. For instance, if I remember correctly, while X sends a 
keypress only if both a keydown and keyup occured, Windows sends the 
keypress before sending keyup...

This means that the browser can no longer simply pass the events along, 
but must implement its own event mechanism on top of the system's one. I 
don't deny the gain for the application developers but unless things 
have changed you're going to get a lot of resistance from the browser 
"vendors" to do that...

Arnaud  Le Hors - Program Director, Corporate Standards, IBM

Philippe Le Hegaret wrote:

>On Thu, 2005-03-03 at 14:14 -0500, L. David Baron wrote:
>>On Monday 2005-01-24 14:04 -0500, Philippe Le Hegaret wrote:
>>>In the case of a key repeat, you should generate multiple
>>>keydown/textInput/keyup, and not just the textInput. That's exactly what
>>>the repeat functionality is about. It does have the drawback of not
>>>having the ability to differentiate between fast key input from a user
>>>on a keyboard and the key repeat functionality.
>>That drawback seems significant (especially for things like online
>>games, where holding down something that's normally a repeating key
>>could cause a state transition), and the fix for it seems reasonably
>>simple.  I propose modifying the model by:
>> 1) adding back keypress events (They would carry the same data about
>>    the key that keyup and keydown events carry now.), and 
>> 2) making keyup and keydown events fire only when a key goes up or
>>    down.
>I agree that this is a drawback, but we need to guarantee that the
>keyboard events will act the same way in multiple platforms. I would be
>interested in the feasibility of the implementation of your approach.
>From rom what I remember when doing some tests a few years ago, keypress
>events varied a lot depending on the platform. I can certainly try those
>tests again if you believe this is no longer the case.
>btw, if reintroduced, the keypress would need to be generated after a
>keyup, but before a textInput.
Received on Tuesday, 15 March 2005 23:23:08 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 20 October 2015 10:46:12 UTC