Re: [Denemo-devel] jackmidi compile warnings

[Richard Shann]

On Fri, 2009-10-02 at 07:53 -0500, Jeremiah Benham wrote:
> On Fri, 02 Oct 2009 10:12:56 +0100
> Richard Shann <address@hidden> wrote:
> 
> > J. I am getting some warnings,
> > 
> > ../../denemo/src/jackmidi.c: In function ‘jack_midi_playback_start’:
> > ../../denemo/src/jackmidi.c:684: warning: ignoring return value of
> > ‘smf_seek_to_seconds’, declared with attribute warn_unused_result
> > ../../denemo/src/jackmidi.c: At top level:
> > ../../denemo/src/jackmidi.c:701: warning: conflicting types for
> > ‘jack_midi_playback_stop’
> > ../../denemo/src/jackmidi.c:327: warning: previous implicit
> > declaration of ‘jack_midi_playback_stop’ was here
> > 
> > I haven't had time to try jack out recently - are things working
> > generally? 
> 
> Mostly. Transport is a bit flaky. I am considering removing all
> transport code and re-adding it at a later date. It starts the
> transport and stops it. That is it. If it is stopped the play button
> must but hit twice to get it going again. I have not figured out why
> yet. I have started to study an example of jack transport here:
> 
> http://jackit.sourceforge.net/cgi-bin/lxr/http/source/example-clients/transport.c
> 
> Apparently it allows for tempo change. I am going to study it to help
> me to understand how the jack transport works better. If we are
> releasing soon I will wait until the next release to change anything. 
> 
> There is a bug with (PlayNote) caused by eliminating substitute_values
> from view.c. I had modified:
> *c = '0'+channel-1 
> 
> to be this in view.c:
> *c = '0'+channel;
> 
> There reason is if the user has set channel 0 it does not substitute
> correctly giving a value of -1 instead of 0;
well, is this a bug in both uses of substitute_midi_values? That is, was
it a bug already? Should have been. This is only to do with the dratted
distinction between user channel numbering 1-16 and midi encoding. IIRC
denemo uses MIDI encoding internally. Perhaps exportmidi.c uses user
numbering?

> 
> I am also confused on how you propose a user is to play notes in
> succession via script.
> 
> (PlayNote "0x3c" 1000)
> (wait 1000)
> (PlayNote "0x3e" 1000)
> 
> Where (wait) is the code not yet implemented that is like sleep but
> keeps the main loop active. 
> Would it look like the above. If not, what
> would the user do to have one note followed by the next? 
I think this is easy: you just do

(d-OneShotTimer 1000 "(PlayNote \"0x3c\" duration)")

when it executes it will schedule its own off event.

right?
Of course, if you are going to keep running your script in some loop
(i.e. not doing d-Get* or something) then you will need to check for
events, which is where that other stuff comes in - it all depends what
you want to be doing while the notes are sounding.

Richard



> 
> Jeremiah
> 
> 
> 
> >(I am thinking forward to a release, and so to keeping
> > things tidy etc).
> > 
> > Richard
> > 
> > 
> > 
> > 
> > _______________________________________________
> > Denemo-devel mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/denemo-devel
>