I never tried instrument 0, so I don't know about that.

The function is called during cleanup if the function pointer in opds is
not null.

Original Message:
-----------------
From: stevenyi stevenyi@csounds.com
Date: Sun, 30 Nov 2003 00:11:14 -0800
To: csound-dev@eartha.mills.edu
Subject: [CSOUND-DEV:3589] Opcode deinitialization


Hi all,

Could someone explain when opcode deinitialization functions are
called?  I tried building an opcode lib with an opcode that had thread
of 1 (i-rate) and the deinitialization function was not called.  This
opcode was also in instrument 0, so I'm having a hard time figuring out
if deinitialization functions only work for k/a-rate opcodes or only for
opcodes not in instrument 0 or what.

Any help on this greatly appreciated!!!

Thanks,
steven


--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .

On Sun, 2003-11-30 at 08:37, gogins@pipeline.com wrote:
> I never tried instrument 0, so I don't know about that.
> 
> The function is called during cleanup if the function pointer in opds is
> not null.

It seems that insert.c::orcompact only walks the nxtp chain of
performance opcodes and not the nxti chain of init-time opcodes.  I
added a walk of the nxti chain and it correctly called the
deinitialization functions for my i-time opcodes.  The code is as
follows:

            off = ip->nxti;
            while (off != NULL) {
            	if(off->dopadr) (*off->dopadr)(off);
            	off = off->nxti;	
            }

and is added after the nxtp walk in orcompact.  I'll be checking this
code into csound5 cvs in a moment as I think it's a bug-fix.  If it
seems good to everyone else, I'll add it to csound4 cvs as well.  

Thanks,
steven

(who's happy to have his opcode plugin happily cleaning up after
itself!  btw: I've learned from this opcode I've built how to redo the
deinitialization for sfont correctly for csound5 (which was in a
function added to mainRESET but is no longer accessible as sfont stuff
is in an opcode lib), so will be reworking that code too).

Hi All,

I mixed up when committing and ended up also committing my change in
insert.c::orcompact in csound4 CVS.  I will leave it in as I think it's
a bug fix as mentioned in my previous email; if anyone wants it out I'll
take it out promptly.

Thanks,
steven

Hmm... I'm not sure my first email got through, which would make my
second one not make any sense.  I've attached the first email I sent
below.

steven




On Sun, 2003-11-30 at 08:37, gogins@pipeline.com wrote:
> I never tried instrument 0, so I don't know about that.
> 
> The function is called during cleanup if the function pointer in opds
is
> not null.

It seems that insert.c::orcompact only walks the nxtp chain of
performance opcodes and not the nxti chain of init-time opcodes.  I
added a walk of the nxti chain and it correctly called the
deinitialization functions for my i-time opcodes.  The code is as
follows:

            off = ip->nxti;
            while (off != NULL) {
                if(off->dopadr) (*off->dopadr)(off);
                off = off->nxti;        
            }

and is added after the nxtp walk in orcompact.  I'll be checking this
code into csound5 cvs in a moment as I think it's a bug-fix.  If it
seems good to everyone else, I'll add it to csound4 cvs as well.  

Thanks,
steven

(who's happy to have his opcode plugin happily cleaning up after
itself!  btw: I've learned from this opcode I've built how to redo the
deinitialization for sfont correctly for csound5 (which was in a
function added to mainRESET but is no longer accessible as sfont stuff
is in an opcode lib), so will be reworking that code too).