[Linuxsampler-devel] Default Audio/MIDI devices

[Rui N. C. <rn...@rn...>]

Hi,

Now that the CREATE AUDIO_OUTPUT_DEVICE is supposedly working with default
parameters, as of Vladimir's latest CVS commit, isn't it worthy to put
back in operation the so-called deprecated SET CHANNEL AUDIO_OUTPUT_TYPE
command?

I mean, for instance, that these old SET CHANNEL AUDIO_OUTPUT_TYPE command
would map to the corresponding CREATE AUDIO_OUTPUT_DEVICE command, just
taking the defaults.

Most probably, a special server state check should be taken on whether the
device isn't already created by a previous request, if duplicate default
device instances are to be avoided, either of ALSA or JACK types. Note
that JACK and ALSA are all-upper-case tokens as was before.

The same would be applicable to SET CHANNEL MIDI_INPUT_TYPE and CREATE
MIDI_INPUT_DEVICE commands.

Most important, this suggestion would make older LSCP scripts, and
qsampler for that matter, forward compatible with the new server engine
protocol, so prevent the already released code from choking with the
latest server development path.

As usual, I can set my efforts on patching linuxsampler for you to try
out. However, I'm affraid this has implications on changing the parser
grammar/lexical source files (src/network/{lscp.y, lscp.l}) for which I've
been a complete failure on how to put them to regenerate a proper parser.

I though it would be sufficient to issue `make parser` under src/network,
but that's been failing on my face. Either my flex/bison installation is
bad or it is *me* that is doing it miserably wrong.

e.g. my flex doesn't recognize those --fast and --8bit command line
options, only -f and -8 respectively; then it also fails with:
"lscp.l", 33: unrecognized %option: reentrant

Currently using flex-2.5.4a and bison-1.875 (Mdk 10.0 Official).

Maybe someone could help me out, before I do some terrible mistake ;)

Any flex/bison gurus out there? Simon? ;)

Cheers.
-- 
rncbc aka Rui Nuno Capela
rn...@rn...