From: Björn Stenberg <bjorn_at_haxx.se>
Date: Mon, 12 Jan 2004 09:30:09 +0100

[IDC]Dragon wrote:
> The playtime is updated in the playback_tick() I'd like to get rid of, and
> is reset on a track change. But we don't need to update the playtime with
> each DMA transfer, this is overhead. Instead it can be queried with
> mp3_get_playtime() where needed.

I don't think this overhead is anything to worry about, though. It's only updated a few times per second.

> But you indirectly point me to the fact that this function may rather belong
> to the mpeg.c code. We can do the pause and reset-on-trackchange business
> there.

Yes, I think that is better. The application does not directly control the transition of tracks and should therefore not be responsible for keeping the accurate time either.

> External players can keep their own time if they have to, it's not
> really a function of the (low level) sink.

On the other hand it's not unreasonable to expect the sink to be able to tell the source for how long the data has been playing. The sink knows this better than any other code.

> My intention was to give direct access to the unmodified function, but well.

The transparent wrapper function can have the same signature, in which case it doesn't matter if it's a direct call or not.

> If we restore the volume for plugin playback, there is still the question of
> how the plugin can increse/decrease that if it doesn't know the original
> value. What's your position to making the config available? Or is there
> another way to read the volume?

I can't think of any bad side effects from allowing plugins access to the settings, so I guess I'm in favor of it. :-)

-- 
Björn

Received on 2004-01-12