[Freecycle-users] Re: freecycle 0.51 errors

[Viceic Predrag]

Hi Marek,


It seems that the portmidi includes are not found.

If I remember well, the portmidi includes should be copied manually 
to /usr/include/(portmidi.h). Could you please check?

Thanks,
Predrag

Le Mercredi 15 Février 2006 00:50, vous avez écrit :
> Hi Predrag,
>
> i just tried to compile latest tarball release and here's the output
>
> Project MESSAGE: Jack includes found!
> Project MESSAGE: Portmidi includes found!
> Project MESSAGE: Portaudio includes found!
> Project MESSAGE: ALSA includes found!
> Project MESSAGE: libinstpatch includes found!
> Project MESSAGE: aubio includes found!
> Project MESSAGE: SoundTouch includes found!
> Project MESSAGE: Jack will be linked dynamically
> Project MESSAGE: Building Freecycle with midi support
> WARNING: Failure to find: portmidimidiinput.h
> WARNING: Failure to find: alsaseqmidiinput.h
> WARNING: Failure to find: portmidimidiinput.cpp
> WARNING: Failure to find: alsaseqmidiinput.cpp
>
> g++ -c -pipe -Wall -W -O2 -march=athlon-xp -pipe -g -D_REENTRANT
> -DHAS_PORTMIDI -DHAS_JACK -DHAS_ALSA -DHAS_PORTAUDIO -DHAS_LIBINSTPATCH
> -DHAS_AUBIO -DHAS_SOUNDTOUCH -DHAS_MIDI -DQT_NO_DEBUG
> -DQT_THREAD_SUPPORT -DQT_SHARED -DQT_TABLET_SUPPORT
> -I/usr/qt/3/mkspecs/linux-g++ -I. -I../src -I../../../../../include
> -I../../../../../local/include -I/opt/gnome/include/glib-2.0
> -I../../../../../include/glib-2.0 -I../../../../../lib/glib-2.0/include
> -I/usr/qt/3/include -o freecycle.o freecycle.cpp
> In file included from freecycle.h:43,
>                  from freecycle.cpp:21:
> soundplayer.h:140: error: expected `,' or `...' before '*' token
> soundplayer.h:141: error: ISO C++ forbids declaration of
> `PaStreamCallbackTimeInfo' with no type
> soundplayer.h:144: error: expected `,' or `...' before '*' token
> soundplayer.h:145: error: ISO C++ forbids declaration of
> `PaStreamCallbackTimeInfo' with no type
> make[1]: *** [freecycle.o] Error 1
> make[1]: Leaving directory `/usr/src/Software/FreeCycle/freecycle/src'
> make: *** [sub-src] Error 2
>
> Qt is at is 3.3.4. Gcc at 3.4.4.
>
> thanks,
>
> Marek
>
> On Thu, 2005-09-08 at 09:56 +0200, Viceic Predrag wrote:
> > Le Mercredi, 7 Septembre 2005 23.45, vous avez écrit :
> > > On Wed, 2005-09-07 at 18:53 +0200, Viceic Predrag wrote:
> > > > Hi Marek,
> > > >
> > > > I'm glad that it works fine for you. You should really try the
> > > > 0.5alpha!
> > > >
> > > > As for the alsaseq, I want to keep Freecycle portable. Portmidi
> > > > really isn't difficult to compile. You should make the modifications
> > > > described in README though.
> > >
> > > I can understand that but i think it's more important to support core
> > > linux technologies than support OSes that already have numerous
> > > alternatives available even if commercial.
> >
> > This argument is valuable, but I want that people use my app, and the
> > percentage of people working with audio on linux is raher small..
> >
> > > Moreover, once jack gains
> > > midi_over_jack functionality, you should be able to easily update
> > > freecycle and it would become crossplatform with respect to macOSx.
> > > You can still have portaudio/mid support optional.
> >
> > Optional compilation takes time to implement (and knowledge).. :)
> >
> > > Couple of remarks:
> > > Ok, so that slowdown still is visible, though it only slows down once
> > > the beatlines appear on the screen, but it's not that slow, it's only
> > > not superfast :)
> > > Second thing - i wasn't able to move/lock/remove the beatlines, all it
> > > does is always make a selection, which seems to be a bug.
> >
> > RTFM (I'm maybe repeating myself, but the manual was rather long and
> > tedious to make.. :) )
> >
> > > If there's anything that i can do to make freecycle better, let me
> > > know, any bugreports, benchmarks etc. Since i've been waiting for an
> > > application like this a long time :) (if if it's not gtk+ :D
> >
> > Well you could hack qmake files (freecycle.pro and src.pro) so that they
> > find libinstpatch and portmidi and maybe set the define that they found
> > it (for conditional compilation)
> >
> > > One more question - the time/freq domain analysis is not a part of
> > > aubio? I'm asking because it seems to be more precise than aubio (if
> > > you zoom in and see the slice positions).
> >
> > Yes, my algorithms (the first two) are doing zero-crossing once the beat
> > position is found, which is not the case for aubio.
> >
> > > Though there are some cases it
> > > can't seem to handle well either(i could upload a file if you're
> > > interested). I wonder if it'd be possible to implement the analysis
> > > algorithms as plugins so that more can be added later on - i hjave a
> > > couple of more beatslicing papers :))
> >
> > Well, you can look at the implementation (soundmanager.h). I'm not sure
> > that the plugins is a good idea, but new algorithms are definitelly
> > welcome!
> >
> > > thanks,
> >
> > Cheers,
> > Predrag
> >
> > > Marek
> > >
> > > > P
> > > >
> > > > Le Mercredi, 7 Septembre 2005 19.12, vous avez écrit :
> > > > > Hi Predrag,
> > > > >
> > > > > sorry for previous remarks, it seems it works just fine under jack,
> > > > > the slowdown was probably caused by something else before (run
> > > > > under /dev/dsp).
> > > > >
> > > > > Marek
> > > > >
> > > > > PS: Pleeease make also direct alsaseq support :)))