Re: [wmii] feature request: ClientClose event

From: Jeremy Maitin-Shepard <jeremy_AT_jeremyms.com>
Date: Fri, 18 Aug 2006 19:17:29 -0400

"Adam Ritter" <xiph_AT_freemail.hu> writes:

> On 8/17/06, Denis Grelich <denis_AT_grelich.de> wrote:
>> On Thu, 17 Aug 2006 18:56:55 +0200
>> "Anselm R. Garbe" <arg_AT_10kloc.org> wrote:
>>
>> > On Thu, Aug 17, 2006 at 04:32:33PM +0200, Adam Ritter wrote:
>> > > It would be nice if a script could be notified if a client is
>> > > closed. It would be easy to remember kill commands, but if the user
>> > > closes the window the script won't be notified. I.e. if I create a
>> > > key for focusing the ncmpc window and optionally creating a client,
>> > > after the client is closed, it won't create an ncmpc client again
>> > > (the script could always check if it is the right client, but that
>> > > has limitations).
>> >
>> > Why not writing a wrapper script for the kill command? You could
>> > check the client's props (class:inst) and decide what todo,
>> > before writing kill. I see no need for the event you request.
>>
>> A kill command is not the only way to close a client/destroy a window,
>> isn't it.
> Yes, that's the problem. And the indexes of the clients change (when
> destroy_client calls cext_vdetach), so the script must know about
> every DestroyClient event to trace the index changes.

The solution to this though is to fix the client identification to use
an identifier that persists for the life of the client and is never
reused, since without that you still have race conditions even if you
track destroy client.

-- 
Jeremy Maitin-Shepard
Received on Sat Aug 19 2006 - 01:18:12 UTC

This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:13:01 UTC