Steven> I started getting back into csound5 and thought I'd ask:

 Steven> -Any objections to removing 'makefile' from the Opcodes directory?  This
 Steven> interferes with the 'Makefile' that is generated with configure

makefile should not be there

 Steven> -a while ago i asked about statements that were generated with csound5
 Steven> that looked like

 Steven> 42005342 <- 42530042

 Steven> I remember John was in the middle of using these for testing some byte
 Steven> reversal code. I've commented them from my InOut/aiff.c file and was
 Steven> wondering if anyone mind if I check in my aiff.c

OK by me -- although better would be to get this code working...

 Steven> -a couple things about the zak stuff that was moved to cglob.  first
 Steven> thing is that the cvs has:

 Steven> #define zkstart cglob.zastart
 Steven> #define zastart cglob.zastart
 Steven> #define zklast  cglob.zalast
 Steven> #define zalast  cglob.zalast

 Steven> which I think should be:

 Steven> #define zkstart cglob.zkstart
 Steven> #define zastart cglob.zastart
 Steven> #define zklast  cglob.zklast
 Steven> #define zalast  cglob.zalast

i thought I had fixed that.  Do ahead

 Steven> if so, I'll check that in.  Also, there's something funny going on with
 Steven> trying to compile widgets.cpp in regards to the this stuff, as somehow
 Steven> zkstart is ending up being defined as a long somewhere before it gets
 Steven> defined as a MYFLT.  for the time being I'm commenting that section of
 Steven> code out in widgets.cpp (in the fl_slider_bank function, around line
 Steven> 2773).  I'll look into that later, but thought I'd mention it.

I had that OK here, but as I still think that widgets.cpp does not
belong in the csound base I probably lost whatever it was.

 Steven> -I noticed that csoundAppendOpcode in jpff_glue.c is currently:

 Steven> printf("DO NOT USE THIS INTERFACE\n");
 Steven>     return -1;

When the interface was first added all the functions were marked like
that.  As far as I am aware adding opcodes is supported -- that is how
the modules work -- but I never looked at this code

 Steven> so I was wondering if dynamically loaded opcodes is working, and if so,
 Steven> where the function is for that. (I was looking for what directory
 Steven> csound5 is looking in for it's dynamically loaded opcodes, as I wanted
 Steven> to see if the shared objects from the Opcodes dir would work).  

Not checked on OSX although I incorporated code which according to
Apple implements the DL library there.  Just have not had the Mac on
this summer (too busy teaching and writing)

==John ffitch

Okay, I cleared the makefile and updated cs.h for the cglob.zkstart and
cglob.zklast stuff.  I left aiff.c as-is as it is not done yet (which
was what I wasn't sure about).  

As for the shared libraries, I'll look through the code again to see
where the loading is going on.  I still don't understand what directory
csound is looking in for it's libraries, or how to set that directory. 
Any help there appreciated!


Also, this error has been a thorn in my side for a while:

Csound Command ERROR:   output soundfile cannot be both AIFF and WAV

as it comes up as the error for cases which seem to not be related. 
Will be looking at this today.

Thanks!
steven


On Wed, 2003-08-27 at 05:12, jpff@codemist.co.uk wrote:
>  Steven> I started getting back into csound5 and thought I'd ask:
> 
>  Steven> -Any objections to removing 'makefile' from the Opcodes directory?  This
>  Steven> interferes with the 'Makefile' that is generated with configure
> 
> makefile should not be there
> 
>  Steven> -a while ago i asked about statements that were generated with csound5
>  Steven> that looked like
> 
>  Steven> 42005342 <- 42530042
> 
>  Steven> I remember John was in the middle of using these for testing some byte
>  Steven> reversal code. I've commented them from my InOut/aiff.c file and was
>  Steven> wondering if anyone mind if I check in my aiff.c
> 
> OK by me -- although better would be to get this code working...
> 
>  Steven> -a couple things about the zak stuff that was moved to cglob.  first
>  Steven> thing is that the cvs has:
> 
>  Steven> #define zkstart cglob.zastart
>  Steven> #define zastart cglob.zastart
>  Steven> #define zklast  cglob.zalast
>  Steven> #define zalast  cglob.zalast
> 
>  Steven> which I think should be:
> 
>  Steven> #define zkstart cglob.zkstart
>  Steven> #define zastart cglob.zastart
>  Steven> #define zklast  cglob.zklast
>  Steven> #define zalast  cglob.zalast
> 
> i thought I had fixed that.  Do ahead
> 
>  Steven> if so, I'll check that in.  Also, there's something funny going on with
>  Steven> trying to compile widgets.cpp in regards to the this stuff, as somehow
>  Steven> zkstart is ending up being defined as a long somewhere before it gets
>  Steven> defined as a MYFLT.  for the time being I'm commenting that section of
>  Steven> code out in widgets.cpp (in the fl_slider_bank function, around line
>  Steven> 2773).  I'll look into that later, but thought I'd mention it.
> 
> I had that OK here, but as I still think that widgets.cpp does not
> belong in the csound base I probably lost whatever it was.
> 
>  Steven> -I noticed that csoundAppendOpcode in jpff_glue.c is currently:
> 
>  Steven> printf("DO NOT USE THIS INTERFACE\n");
>  Steven>     return -1;
> 
> When the interface was first added all the functions were marked like
> that.  As far as I am aware adding opcodes is supported -- that is how
> the modules work -- but I never looked at this code
> 
>  Steven> so I was wondering if dynamically loaded opcodes is working, and if so,
>  Steven> where the function is for that. (I was looking for what directory
>  Steven> csound5 is looking in for it's dynamically loaded opcodes, as I wanted
>  Steven> to see if the shared objects from the Opcodes dir would work).  
> 
> Not checked on OSX although I incorporated code which according to
> Apple implements the DL library there.  Just have not had the Mac on
> this summer (too busy teaching and writing)
> 
> ==John ffitch
> 
> 

I was getting this error for awhile whenever I was trying to send the text
listing output to a file instead of the screen.  I eventually discovered
that the command line option for this had changed from -- to -O but the Mac
front-end had not been updated.

I suppose you may be getting the error for some other reason though.

Anthony Kozar
anthony.kozar@utoledo.edu


On 8/27/03 9:04 AM, stevenyi etched in stone:

> Also, this error has been a thorn in my side for a while:
> 
> Csound Command ERROR:   output soundfile cannot be both AIFF and WAV
> 
> as it comes up as the error for cases which seem to not be related.
> Will be looking at this today.

On Wed, 27 Aug 2003, Anthony Kozar wrote:

> I was getting this error for awhile whenever I was trying to send the text
> listing output to a file instead of the screen.  I eventually discovered
> that the command line option for this had changed from -- to -O but the Mac
> front-end had not been updated.
> 

When was this changed?  I was not aware of it.
==John ff

On 8/27/03 2:00 PM, John ffitch etched in stone:

> On Wed, 27 Aug 2003, Anthony Kozar wrote:
> 
>> I was getting this error for awhile whenever I was trying to send the text
>> listing output to a file instead of the screen.  I eventually discovered
>> that the command line option for this had changed from -- to -O but the Mac
>> front-end had not been updated.
>> 
> 
> When was this changed?  I was not aware of it.
> ==John ff
> 

Dunno.

But here is part of a comment from argdecode.c (Mac src 4.23):

/* alphabetically, so i dont have to hunt for a letter next time...
**********************************************************************
[...]
-o fnam sound output filename
-O fnam log output to file
[...]
--sched real-time scheduling and memory lock (LINUX only)
[...]
*/

The "help output" when you make a mistake typing options still claims "--"
is for log output, though.

Anthony Kozar
anthony.kozar@utoledo.edu

Hi all,

In looking for the reason behind the "output soundfile cannot be both 
AIFF and WAV", I realized that the argdecode was manually done.  I had 
been looking into using a port of GNU Getopts for Java in blue, and 
thought maybe that it might be useful to use getopts in Csound as well, 
but I wasn't sure if getopts was just a shell tool or if it was a 
programming function to use in C as well.  If it is, would there be any 
objections to having argdecode redone with it?

steven