I have confirmed that the m and n score statements (mark and repeat marked
section) are not working in the latest canonical (4.23f11).

As far as I can tell, typical usage of these statements will cause Csound to
go into a loop of reopening the score file at the marked position.  It no
longer returns to after the n-statement when the marked section is finished.
(see example from Csound book below which causes an infinite loop on my
machine).

These changes from the Change Log seems to be responsible:

2003-02-16
    * version.h (VERSION): 4.23f01

    * sread.c (sread_init): new function to separate include files
    from sections
    (sread): Removed resetting of str and the unwinding after section
    ends
    
    * scsort.c (scsort): Call sread-init

Of course, reversing these changes solves the problem, but there must be
some reason that these changes were made.  I don't yet understand how we can
fix m and n while preserving the intention of this change.  One possibility
might be to read in the marked section to a string and to revisit it that
way instead of reopening the file (like I think macros are done).

Does anyone else understand how to fix this?

Anthony Kozar
anthony.kozar@utoledo.edu
http://akozar.spymac.net/


; 511.sco from "The Csound Book"
; example by John ffitch

f1 0 4096 10 1
f2 0 4096 10 10 0 5 0 1

;1st    Verse      
i511    0   .3  15000   8.00
i511    +   .   9000    8.02
i511    +   .   12000   8.04
i511    +   .   9000    8.00
s               
m   Chorus         
i512    0   .3  20000   8.07
i512    +   .   12000   8.09
i512    +   .   18000   8.07
i512    +   .   11000   8.05
s               
;2nd    Verse      
i511    0   .3  15000   8.04
i511    +   .   9000    8.00
i511    +   .   12000   8.02
i511    +   .   9000    8.00
s               
n   Chorus         
s               
;3rd    Verse      
i511    0   .3  15000   8.04
i511    +   .   9000    8.00
i511    +   .   12000   8.04
i511    +   .   9000    8.05
s               
n   Chorus         
s
;4th    Verse      
i511    0   .3  15000   8.04
i511    +   .   9000    8.04
i511    +   .   12000   8.02
i511    +   .   9000    8.02
s               
i511    0   4   15000   8.00
i512    .05 4   10000   7.001

The bug that was fixed was that repeating section involving #include'd
code was completely broken, looping in very odd ways.  The code was
changed in order to overcome this.  This rather fills me with terror
at the prospect of fixing n/m without reintroducing that bug.  
==John ff

OK.  I will continue to ponder how it might be solved without recreating the
include problem.

Anthony

On 7/1/04 2:02 AM, jpff@codemist.co.uk etched in stone:

> The bug that was fixed was that repeating section involving #include'd
> code was completely broken, looping in very odd ways.  The code was
> changed in order to overcome this.  This rather fills me with terror
> at the prospect of fixing n/m without reintroducing that bug.
> ==John ff
> 

Earlier tonight, I committed code that fixes m and n (in ordinary usage --
there are still problems with catching bad input that does bad things).
However, looking back through emails from nearly four years ago and
re-examining the code from 4.23 and 4.23f01, my "fix" looks very similar to
some of the code that was removed to fix the problem cited by John below. I
do not understand what the original problem was and I am now afraid of
re-introducing it.

John, I know it was years ago, but do you happen to still have an example of
the input that showed the bug that you fixed as described below.  Or can you
describe the problem in more detail, please?

Thanks.

Anthony

jpff@codemist.co.uk wrote on 7/1/04 2:02 AM:

> The bug that was fixed was that repeating section involving #include'd
> code was completely broken, looping in very odd ways.  The code was
> changed in order to overcome this.  This rather fills me with terror
> at the prospect of fixing n/m without reintroducing that bug.
> ==John ff
> 

Anthony Kozar wrote on 6/30/04 8:11 PM:

>> I have confirmed that the m and n score statements (mark and repeat marked
>> section) are not working in the latest canonical (4.23f11).
 
>> 2003-02-16
>> * version.h (VERSION): 4.23f01
>> 
>> * sread.c (sread_init): new function to separate include files
>> from sections
>> (sread): Removed resetting of str and the unwinding after section
>> ends
>> 
>> * scsort.c (scsort): Call sread-init


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net