a quick note on 0.9a8 and some other things

[Austin Hicks]

Hi,

There is a 0.9a8. I had a calculation backward and audio output
was consequently taking a core even when you weren't actually doing
anything. Nothing else changed.


This probably breaks audio output for someone, somewhere. I don't
know who. If you have a problem with it, please tell me. People aren't
reporting them until a month after the fact. No one on this list, mind
you, but it's worth mentioning because it keeps happening and I can't
test on every computer in the world.


This bug is not entirely fixed. It still takes more than it
should. Anyone who has been on this list for a while knows that I was
hoping to skate through 0.9 without doing the audio_io rewrite, but
fixing this permanently can't be done without it. The rewrite will
therefore be happening in what I am tentatively thinking of as the
December sprint. My work on the Rust compiler is nearing completion but
turned out to be much, much larger than expected, to the point where
people on official teams are having trouble helping me finish it. So it
might prove to be early January.

Finally, Libaudioverse doesn't currently work on XP because of two
function calls. I would like to go ahead and adopt the official stance
that it will never work on XP: I intend to remove the audio backend that
XP needs in the audio_io rewrite. There is also a chance that I will
drop Vista support. Microsoft (great, wise Microsoft) made one of the
audio APIs needed for microphone nonfunctional in Vista before SP1. We
will see what happens as I'm not sure if there are any other critical
differences. I do not believe that either of these platforms are still
significantly popular, and they will be much less so by the time we get
to 1.0. Since I can't test on them anyway, I'm not able to justify
spending significant effort to keep them.