Re: problem with sound driver in iBook2
> ...
> They aren't conflicting. HAS_RECORD defines whether the code for
> recording is built in at all, dmasound.mach.record says whether
> recording is supported for the machine it's running on.
Yes, but we have two different concepts for the same functionality
(en-/disabling recording support). What is the use of checking
HAS_RECORD during compile time, if you can check easily during runtime?
Two concepts are harder to maintain, you will have some overhead to
keep them consistent.
> ...
> Well, this just avoids the problem by pretending that recording is
> supported, when in fact it isn't.
Yes, I was just looking what happens. Deep inside I was hoping that
recording capability might perhaps be added with not too much effort.
How about Darwin, couldn't we borrow some code of their driver, or
is this code not freely available?
> ...
> However, as some apps open the device with O_RDWR even when they're
> never going to read from it, this might not be the worst approach. This
> patch does more or less what you describe, but reads will return
> -EINVAL.
That's how it should be, if we keep it as is.
Heinz
Reply to: