> > I am using a timer to fire up calls to
> > csoundPerformBuffer, so
> > I can have transport controls. FLTK widgets and graphics
> > can be used.
> >
> > With Carbon it seems to be possible to do it  and my app
> > has worked
>
> and your app's menus etc are not hijacked by fltk?
>

No, everything seems fine, although I am only using two
menus, the Apple one and the Application menu. But
I had other menus before and they always responded.
The window interface, buttons, text fields all work fine.

> > with FLTK graphics and widget opcodes. Since I don't
> > know and am too lazy to Objective-C, I never tested
> Cocoa.
>
> i'm fairly new to it myself [ maccsound is
> carbon/powerplant ]
>
> > With regards to audio IO, I would suggest you use the
> > same system you had with MacCsound, which seems to be
> > fast enough.
> > The only thing that does not work with it is Digidesign
> > IO (at
> > least on a 10.4 with the digi 002).
>
> there is a bug in maccsound's audio io with tiger that i
> need to fix,   which might
> be the problem with the digi device.
> >
>
> -m
> ___________________
> matt ingalls
> development@gvox.com
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep
> through log files for problems?  Stop!  Download the new
> AJAX search engine that makes searching your log files as
> easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

i am getting a message everytime i request a termination  from within  
my Yield
callback:

"early return from csoundPerformBuffer " [ or something like that, i  
am presently on a different computer ]


also, when i run more than one instance, i am getting an error  
message not able to delete the tmp file.
maybe because one render deletes the other's temp file?
i thought all that was fixed in csound5??



Matt;
________________________
           matt  ingalls
http://sonomatics.com



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

>>
>> "early return from csoundPerformBuffer " [ or something like that, i
>> am presently on a different computer ]
>
> This is normal behavior. The message can be removed if people find
> it annoying, but it is not an error.

it is annoying.  if it is ok with you, please remove!

>
>> also, when i run more than one instance, i am getting an error
>> message not able to delete the tmp file.
>> maybe because one render deletes the other's temp file?
>> i thought all that was fixed in csound5??
>
> This may be a Mac specific problem related to temporary files
> (e.g. all instances trying to use the same file name - this may
> be due to the code that adds an extension to the file name,
> replacing the originl suffix which may very well have been a
> random number).

i will investigate further.  in csound4 i do something like
sprintf(orcname, "/tmp/%d%d", getuid(), getpid());
and then add on a unique number

i also seem to remember that in order for the temp files to be deleted
on reboot, they needed to be in /tmp directory and not where
tmpfil wanted to put it.

> While I do not think this is related to the temporary file
> problem, note that some specific opcodes and features may not
> work correctly in multiple instances; these include Cscore,
> some of the graph code (e.g. PostScript or FLTK), and many of
> the plugin opcodes that are not in libstdopcod.dylib.

you mean these things still use global variables? or is there some other
reason?

using multiple instances is top importance to me -- i am willing to  
help out
on this if possible.

-m@
________________________
           matt  ingalls
http://sfsound.org/matt.html



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

well the default tmpnam is doing nothing as far as i can see:

UnifiedCSD:  /Users/matt/Documents/Csound/csd/Bach.csd
STARTING FILE
Creating orchestra
Creating score
orchname:  /var/tmp/tmp.orc
scorename: /var/tmp/tmp.sco



i think best would be to add this to one_file:

#ifdef __MACH__
char *mytmpnam(CSOUND *csound, char *a)
{
	static unsigned long count = 0;
  	sprintf(a, "/tmp/%d%d%d", getuid(), getpid(), count++);

     return a;
}
#else

this way every tmpnam is prettymuch guaranteed to be unique, and it  
is put
in /tmp so that if crashes etc occur and temp files are not removed,  
they
will be deleted by the OS [ on OSX, they stay forever in /var ]







On Dec 28, 2005, at 12:48 PM, Istvan Varga wrote:

> On Wednesday 28 December 2005 21:24, matt wrote:
>
>> i will investigate further.  in csound4 i do something like
>> sprintf(orcname, "/tmp/%d%d", getuid(), getpid());
>> and then add on a unique number
>>
>> i also seem to remember that in order for the temp files to be  
>> deleted
>> on reboot, they needed to be in /tmp directory and not where
>> tmpfil wanted to put it.
>
> Would it improve things if I #ifdef out code like the one shown
> below for OS X ? I recall some issue with tmpnam generating file
> names like "tmp.000001" or similar.
>
>       mytmpnam(csound, sconame);              /* Generate score  
> name */
> /* -------------------------- */
>       if ((p=strchr(sconame, '.')) != NULL) *p='\0'; /* with  
> extention */
>       strcat(sconame, ".sco");
> /* -------------------------- */
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

-m@
________________________
           matt  ingalls
http://sfsound.org/matt.html



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

> Cscore uses many global variables and it is large, ugly, old, and
> rarely used so no one bothered cleaning it up so far. The PostScript
> graphs should be easy to fix (there are 1-2 statics there or so, it
> is just rarely used and got no attention). Of the plugins that do not
> work in multiple instances, many depend on third party libraries and
> the libraries themselves may often use global variables. It should
> be possible to fix the Maldonado soundfont opcodes, as those do not
> use any external libraries, although the code is really awful and
> global variables is just one of the issues that need to be cleaned
> up. Finally, some of the utility plugins may use globals; at least
> cvanal and pvanal should be OK, however.

OK i could probably live with all that -- i hope sndinfo is good too??



i am seeing another problem today... periodically i get strange  
segmentation faults with
strange output like:

[merror:  no legal opcode, line 671:
i10  84.06  9.8   
[m   0      4.002   
[m  1900   0.3    1 
[m.4    220
error:  no legal opcode, line 672:
i10  84.618 9.8   
[m   0      5.005   
[m  900    0.4    4 
[m.1    180
error:  no legal opcode, line 673:
e
error:  illegal character <, line 678:
</CsScore>
^
error:  illegal character >, line 678:
</CsScore>
         ^
error:  no legal opcode, line 678:
</CsScore>
error:  illegal character <, line 679:
</CsoundSynthesiz 
[mer>
^
error:  illegal character >, line 679:
</CsoundSynthesiz 
[mer>
                  
[m  ^
error:  no legal opcode, line 679:

[ btw, this is printout from the console, not my own output callback ]

My only thought is maybe there is struct alignment discrepancy  
between my host and the csoundlib??
[ and i would not know how to fix it ]

Any other ideas?

THANKS!!!
-m




-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

these messages are from a single instance
  calling csoundPerform inside application thread.

they didn't paste into the email right -- i think they are really non- 
printing characters


On Dec 29, 2005, at 2:00 AM, Istvan Varga wrote:

> On Thursday 29 December 2005 09:22, matt wrote:
>
>> My only thought is maybe there is struct alignment discrepancy
>> between my host and the csoundlib??
>> [ and i would not know how to fix it ]
>
> You can easily check that by compiling CsoundLib from sources.
> The problem may also be caused by thread safety issues if more
> than one thread attempts to use the same CSOUND* object at the
> same time (which is not safe). Last, but not least, this may very
> well be related to the tmp file issues, as the messages look
> much like parsing a binary file as an orchestra.
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

-m@
________________________
           matt  ingalls
http://sfsound.org/matt.html



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

>
> So, what does the following program print for you on the Mac ?

/var/tmp/tmp.0.mYibpZ
/var/tmp/tmp.1.nlNQPM


as i remember, somewhere in the code the first '.' is found and that's
why the unique stuff is removed
>
> I made some changes to the CVS sources so that the original suffix of

i think you should just add my code from my last email.
it has proven to work in my csound4 sources


Matt;
________________________
           matt  ingalls
http://sonomatics.com



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

On Dec 29, 2005, at 2:54 AM, Istvan Varga wrote:

> That is now already #ifdef'd out for __MACH__ so you get
> /var/tmp/tmp.0.mYibpZ.orc etc. I may consider adding a custom
> tmpnam implementation if this still does not work, though.

it still will not work because:

1) somewhere in the one_file code the first '.' is found and then  
everything is replaced with orc/sco after
	so the temp orc&sco for a .csd file becomes /var/tmp/tmp.orc and / 
var/tmp/tmp.sco accordingly.

2) temp files HAVE to be in /tmp root directory for the OS to delete  
them on OSX reboot

USE MY CODE I GAVE YOU!!!!!!!!!!!!

-m@
________________________
           matt  ingalls
http://sfsound.org/matt.html



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

On Dec 29, 2005, at 10:51 AM, Istvan Varga wrote:

> There are #ifdef's there, too. Do you have up to date sources ?

yes, but never figured out how to compile.  [ you have to use a  
commandline!?!??  :) ]
but Victor just sent me some helpful instructions, so hopefully i can  
do it..

>
>> 2) temp files HAVE to be in /tmp root directory for the OS to delete
>> them on OSX reboot
>
> The files should be deleted when csoundReset() is called.

yes, but if there is a crash or something
then they are not, and those can pile up over time, especially
with shitty host applications like mine.

Matt;
________________________
           matt  ingalls
http://sonomatics.com



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net