RE: [fluid-dev] fluidsynth compiles for use with coreaudio!

[Joe Corneli]

Are there any special techniques needed to debug a program of this
size and scope with gdb?

Joe

On Wed, 4 Jun 2003, Joe Corneli wrote:

> Hi Markus & al.,
> 
> This is what I got:
> 
> %fluidsynth --verbose --audio-driver=coreaudio Unison.SF2 prog 0 0 noteon 0 
> 64 127
> fluidsynth: warning: Ignoring sample *KPianoB5: can't use ROM samples
> fluidsynth: prog        0       0       0
> fluidsynth: prog        1       0       1
> fluidsynth: prog        2       0       2
> fluidsynth: prog        3       0       3
> fluidsynth: prog        4       0       4
> fluidsynth: prog        5       0       5
> fluidsynth: prog        6       0       6
> fluidsynth: prog        7       0       7
> fluidsynth: prog        8       0       8
> fluidsynth: prog        9       128     0
> fluidsynth: prog        10      0       10
> fluidsynth: prog        11      0       11
> fluidsynth: prog        12      0       12
> fluidsynth: prog        13      0       13
> fluidsynth: prog        14      0       14
> fluidsynth: prog        15      0       15
> Bus error
> 
> *no sound*
> 
> Note that when fluidsynth isn't waiting for input, as in, no "-" at the
> end, it crashes almost immediately:
> 
> %fluidsynth --audio-driver=coreaudio Unison.SF2
> fluidsynth: warning: Ignoring sample *KPianoB5: can't use ROM samples
> Bus error
> 
> Joe
> 
> On Wed, 4 Jun 2003 address@hidden wrote:
> 
> > Hi,
> > 
> > try to start fluidsynth, then on the fluidsynth command line type type
> > prog 0 0 (selects program 0 on channel 0, assuming the soundfont has a 
> > program 0)
> > noteon 0 64 127 (hits a key with full speed).
> > If that produces sound, then you can verify whether input works with the 
> > --verbose command line switch. In this case, each noteon event produces one 
> > line of output.
> > 
> > Regards
> > 
> > Markus
> > 
> > > -----Original Message-----
> > > From: ext Joe Corneli [mailto:address@hidden
> > > Sent: 04 June 2003 10:49
> > > To: FluidSynth Devel
> > > Subject: [fluid-dev] fluidsynth compiles for use with coreaudio!
> > > 
> > > 
> > > Dear FluidSynth Developers -
> > > 
> > > Here is the configure command I used:
> > > 
> > > ./configure -enable-coreaudio 
> > > CPPFLAGS="-I/sw/include
> > > -I/System/Library/Frameworks/CoreAudio.framework/Versions/A/Headers" 
> > > LDFLAGS="-L/sw/lib -framework CoreAudio"
> > > 
> > > I guessed that if I could get fluidsynth to compile for jack 
> > > on mac using
> > > -framework instead of -l, the samething should work for coreaudio.
> > > 
> > > Aside from this, I did a little bit of find and grep in the directory
> > > /System/Library/Frameworks/ -- this made my guess a more educated one.
> > > 
> > > For the record, recall that I needed several changes to 
> > > various files -- I
> > > didn't make any more changes than previously suggested by Josh.
> > > 
> > > Ok, this is all fine and good... but the problem I have now is
> > > that both
> > > 
> > >    clavier -o - | fluidsynth --audio-driver=coreaudio -
> > > 
> > >    clavier -o - | fluidsynth --audio-driver=coreaudio Unison.SF2 -
> > > 
> > > (where Unison.SF2 is a free soundfont file I found on the web at
> > > www.personalcopy.com) *do not produce* sound.  Can someone please
> > > help me figure out what to do next?  Do I need to pass a 
> > > --midi-driver 
> > > command to fluidsynth, and if so, what command would you suggest?
> > > 
> > > Lastly, I noticed the following when I ran make -- maybe 
> > > these warnings
> > > end up having something to do with the muteness I'm now 
> > > experiencing? I
> > > totally don't know.
> > > 
> > > 
> > > gcc -DHAVE_CONFIG_H -I. -I. -I. -I./../include -I/sw/include
> > > -I/System/Library/Frameworks/CoreAudio.framework/Versions/A/He
> > > aders -O2
> > > -fomit-frame\
> > > -pointer -funroll-all-loops -finline-functions -Wall -W 
> > > -Wpointer-arith
> > > -Wbad-function-cast -Wcast-qual -Wcast-align 
> > > -Wstrict-prototypes -Wno-unu\
> > > sed -Winline -c fluid_coreaudio.c -Wp,-MD,.deps/fluid_coreaudio.TPlo
> > > -fno-common -DPIC -o .libs/fluid_coreaudio.lo                 
> > >               
> > > fluid_coreaudio.c: In function 
> > > `new_fluid_core_audio_driver2':                               
> > >                                                      
> > > fluid_coreaudio.c:159: warning: passing arg 2 of 
> > > `AudioDeviceAddIOProc'
> > > from incompatible pointer type                                
> > >             
> > > fluid_coreaudio.c:165: warning: passing arg 2 of 
> > > `AudioDeviceStart' from
> > > incompatible pointer type                                     
> > >            
> > > fluid_coreaudio.c: In function 
> > > `delete_fluid_core_audio_driver':                             
> > >                                                      
> > > fluid_coreaudio.c:191: warning: passing arg 2 of 
> > > `AudioDeviceStop' from
> > > incompatible pointer type                                     
> > >             
> > > fluid_coreaudio.c:192: warning: passing arg 2 of 
> > > `AudioDeviceRemoveIOProc'
> > > from incompatible pointer type 
> > >  
> > > 
> > > Joe
> > > 
> > > PS.  Here are the details of my elementary find and grep 
> > > algorithm in case
> > > they happen to be helpful later.  Like I said, all this really did was
> > > point me in the right direction, it didn't find anything specifically
> > > helpful in and of itself.  I like it when I learn something useful
> > > you can do with simple command line utilities, so that's 
> > > another reason
> > > I want to share it.
> > > 
> > > In the above-mentioned directory,
> > > 
> > >   find . | grep "Audio.*\.h"  > list
> > > 
> > > then
> > >   grep -v "html" list > audioh
> > > 
> > > (I know you should be able to do this all in one go but for 
> > > some reason I
> > > got an error message when I tried that).  Then 
> > > 
> > >   grep -n "AudioDeviceAddIOProc" <audioh>,
> > > 
> > > for example, where <audioh> denotes the expansion of the file 
> > > audioh into
> > > its entries.  I found several occurances of "AudioDeviceAddIOProc" and
> > > similarly also of (most of) the other terms that came up as 
> > > undefined in
> > > previous attempts to make the file.  This, combined with the 
> > > advice from
> > > Stephane about how to compile for jack support ("ha ha, he said 'jack
> > > support'") wos enough for me to guess about -framework CoreAudio.
> > > 
> > > 
> > > 
> > > 
> > > 
> > > 
> > > _______________________________________________
> > > fluid-dev mailing list
> > > address@hidden
> > > http://mail.nongnu.org/mailman/listinfo/fluid-dev
> > > 
> > 
> 
>