Hi!
On Sat, May 12, 2001 at 12:13:35AM -0400, Jeff Tranter wrote:
> I have doxygen comments ready to commit. I have just a few questions:
>
> - are the samples written using csl_sample_write in a fixed format or are
> they inferred from the name?
They are inferred from the name, whereas it's probably the best if you use
simple wav samples if you want it to work with every backend. Maybe we
should make some
> - can you call csl_sample_write more than once, and if so, are the
> samples concantenated together?
Yes, it's the data that gets concatenated. I.e. you should only write one
sample, with a header and some data, but you can write it in multiple write
calls.
> - ssl_sample_cache_add was not implemented. I added an empty
> function definition in cslsample.c so I could document it.
Have a look at the changes I did to your implementation (would have caused
problems with threading), in any case it should be fine now.
> - will the behaviour (e.g. caching) be the same when using the OSS
> backend as compared to the aRts backend?
If somebody writes an OSS backend for samples, then probably it will not be
exactly the same (as you for instance can't play two samples at a time), but
only "as good as possible". I think the caching implementation will just do
nothing at all there, or maybe flag the samples in memory that they don't
get unloaded after the last csl_sample_free (only after the last
csl_sample_cache_remove).
Cu... Stefan
--
-* Stefan Westerfeld, [EMAIL PROTECTED] (PGP!), Hamburg/Germany
KDE Developer, project infos at http://space.twc.de/~stefan/kde *-
