>>>> "William" == William M Perry
<wmperry(a)aventail.com> writes:
William> karlheg(a)bittersweet.inetarena.com (Karl M. Hegbloom)
William> writes:
> From the `esound.sgml':
>
> <sect1 id="environment-variables"> <title>Environment
> Variables</title>
William> I think this should be visible from lisp with something
William> like `esd-speaker', because I might want to switch back
William> and forth during the course of running XEmacs. And
William> should the connection to esd be persistent? Right now it
William> looks like it is opening/closing the connection for every
William> sound.
Hmm. I wonder if useing `setenv' before playing a sound works. I've
not tried it.
I think that it's not persistent so that the `esd' can go idle and
release the sound device. There's an `esd' switch to set an idle
timeout. If it were not for that, `realplay' would not work. (I
tried to wrap realplay in a script that does a
`LD_PRELOAD=libesddsp.so:libesd.so', but that didn't work right. It
works for a few seconds, and then hangs.
William> This isn't a big deal for me anymore, since I have a
William> SoundBlaster Live! in the new machine, and it mixes
William> everything together for me. No more need for esd. :)
How does that work? I should look into that card... I want to buy a
new sound card soon; I've got an aging Pro Audio Studio Spectrum
(EISA). When the drives `fsck' at boot, the sound card plays a
reeling in a fish noise. The microphone port is blown too.
William> Now, if XEmacs were only multithreaded to take advantage
William> of all the CPUs. :)
Someday maybe it will be? Or a thing much like XEmacs re-written
almost from scratch... A Guile-XEmacs-GTK maybe. How many CPU's do
you have?