RFC: Reworking sound sub-system

[calris]

Alright... adding a "data->loaded = false;" after playing a sound did the trick for me. It's very ugly, but it works

Oh Yuck! - You are essentially reloading the sound into memory from disk every time it is played (although file system caching should help with any delays).

I worry about what reloading a sound mid-play might do internally - if the SDL sound system is in the middle of playing a sample, you are effectively ripping it out from underneath it.

I've never really looked at what SDL under Linux does when you play the same sample memory while it is already being played...

[PowerWyrm]

Hmm - have you had a look at how I've implemented sound for Windows?

Take a look at:

https://github.com/calris/angband/commits/sound-devel-clean

It's a much cleaner version (but in the end identical end result) to what Nick pulled. It's much easier to see what is going on

I bet the Windows sound module will work nicely. Now that the SDL module works fine in PWMAngband, I'll port that too.

[calris]

I bet the Windows sound module will work nicely. Now that the SDL module works fine in PWMAngband, I'll port that too.

Don't bet on it - I never tested it

[PowerWyrm]

Don't bet on it - I never tested it

Well I've now ported everything and everything works fine again.

The only problems I've noticed which were already present in the old system were:
- Mix_PlayMusic() not playing a sound from the beginning under Windows (fixed by reloading the sound)
- PlaySound(SND_PURGE) not stopping the current sound (didn't change that, so the Windows module actually plays sounds over the top of each other)

[PowerWyrm]

By the way, sound.cfg can be removed from \lib\sounds since it's not used anymore...

[calris]

By the way, sound.cfg can be removed from \lib\sounds since it's not used anymore...

It's still used by at least the OS X port. There may be others, not 100% sure