|Summary:||The response of 'GET CHANNEL INFO' command does not conform to LSCP specification|
|Product:||LinuxSampler||Reporter:||Grigor Iliev <gr.iliev>|
|Component:||LSCP||Assignee:||Christian Schoenebeck <cuse>|
|Attachments:||patch for lscpserver.cpp|
Description Grigor Iliev 2005-05-24 21:30:27 CEST
The following categories returns -1 instead of "NONE" when they are not set. - AUDIO_OUTPUT_DEVICE - MIDI_INPUT_DEVICE The following categories are set to -1 but they are not specified in LSCP draft, so not sure about them. - INSTRUMENT_NR - MIDI_INPUT_PORT btw INSTRUMENT_NAME also is not specified in LSCP and return "NONE" Sending patch for lscpserver.cpp
Comment 1 Grigor Iliev 2005-05-24 21:32:06 CEST
Created attachment 6 [details] patch for lscpserver.cpp
Comment 2 Christian Schoenebeck 2005-05-24 22:14:14 CEST
I will postpone this after I got the release packaged and out. Because it might break compatibility with liblscp and qsampler.
Comment 3 Rui Nuno Capela 2005-05-25 23:54:14 CEST
Altough it might not conform to current LSCP spec, the AUDIO_OUTPUT_DEVICE and MIDI_INPUT_DEVICE fields of the GET CHANNEL INFO command result set are assumed to be integers on liblscp. The proposed change will surely break liblscp (0.2.9) and qsampler (0.1.0) behavior, as reading atoi("NONE") leads to a device number zero (0), which is a valid device-id and not minus-one (-1), which is not. However, the point is taken, and should be addressed in near term.
Comment 4 Christian Schoenebeck 2008-12-02 22:19:55 CET
What should we do with this issue? I guess there's no point in modifying the implementation anymore, since it would introduce the mentioned backward compatibility problems. Should we just update the LSCP specs regarding the mentioned lack of clearification and mark this report as resolved?
Comment 5 Grigor Iliev 2008-12-03 12:16:18 CET
ok, I'm fine with that.
Comment 6 Christian Schoenebeck 2008-12-10 23:16:58 CET
Ok, fixed the LSCP sepcs as suggested.