On Sat, Jun 24, 2006 at 10:22:47AM +0200, Anselm R. Garbe wrote:
>On Sat, Jun 24, 2006 at 04:07:41AM -0400, Kris Maglione wrote:
>> On Sat, Jun 24, 2006 at 09:57:53AM +0200, Anselm R. Garbe wrote:
>> >We already had this. Please don't clutter the wm with such kind
>> >of stuff. The current wmiimenu approach is much more simplier
>> >and can be used without 9P as well. I don't want that change.
>> >There is no single sane reason why this would be more
>> >advantageous. You can access the 9P service to build up menu
>> >lists, that's all we need.
>> Like I said, I'll write it as a patch. The idea as it's been presented
>> iterests me because the latency of starting a menu is unbearable at times.
>> I've often enough wound up entering an entire tag name in vim before the
>> menu has opened and I'm opened to exploring any available way to avoid it.
>If there are any performance issues, we should investigate to
>eliminate them. But for me the menu has been quite responsible
>so far... It might be related that my /tmp is a tmpfs...
I think that the problem is the time it takes to launch wmiimenu once it's
left the cache, especially when there is significant disk io at the time.
There's also the fact that wmiimenu can't grab keyboard focus until it's
finished reading input. Recall that the performance is a problem when
switching tags, so the proglist file isn't an issue. The only viable solution
I see is to keep the wmiimenu program resident, and, ideally, the lists it's
to present also. The easiest way I can think of to do this is to integrate it
into the wm. It will be simple enough to maintain this as a patch given the
new fs.c, so I don't see any harm in writing a test implementation. Anyway,
it's of no immediate importance. Much work still has to be done before I'll
even be ready to start work on it.
-- Kris Maglione Nobody goes to the theatre unless suffering from acute bronchitis.Received on Sat Jun 24 2006 - 10:34:09 UTC
This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:10:06 UTC