Lars Luthman wrote:
> OK, maybe. It's trivial to add a "safe" RDF indicator in a separate
> extension and just assume that an event type is not safe if it doesn't
> have it.
Fair enough.
> For that case I think it would be cleaner if the host just splits the
> processing period in smaller parts if it needs to fit in more events per
> time unit. The plugin doesn't need to worry about it.
Yes. But that makes it impossible to use a fixed size buffer. If we
already have a fixed size buffer extension, why break it this way?
(note: I'm not saying that fixed size buffers are always a nice thing,
but let's think about it before it's too late)
> And mine is wrong too. Embarassing. This works:
> index += (events[index].size + 23) / 16;
Yes, that's the correct one. Mine would only be correct if 'size' meant
size of the whole event including 8-byte header, not the size of the
payload.
> all. So I guess the cleanest way would be to not list the uri-map thing
> as a separate lv2:Feature in the RDF data but require that a host that
> handles events passes that LV2_Feature to the plugin's instantiate
> callback if it is going to connect a non-NULL buffer to any event ports.
Why not use lv2:optionalFeature?
Krzysztof
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev@email-addr-hidden
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-dev
Received on Thu Nov 29 16:15:06 2007
This archive was generated by hypermail 2.1.8 : Thu Nov 29 2007 - 16:15:06 EET