Tracing, it seems that there is a call to sf_close from deact/urnoff2
and after that the file descriptor has vanished.

#0  sf_close (sndfile=0x81846b8) at sndfile.c:416
#1  0x08052bcc in csoundFileClose (csound=0x80e1008, fd=0x80eb1b8)
    at Engine/envvar.c:1258
#2  0x08051cd7 in fdchclose (csound=0x80e1008, ip=0x817fb10)
    at Engine/auxfd.c:131
#3  0x0805bd01 in deact (csound=0x80e1008, ip=0x817fb10) at Engine/insert.c:507
#4  0x0808e02e in turnoff2 (csound=0x80e1008, p=0x816e538)
    at OOps/goto_ops.c:202
#5  0x0804ce4d in csoundPerform (csound=0x80e1008) at Top/csound.c:1097
#6  0x0804b0e8 in main (argc=2, argv=0xbffff084)
    at frontends/csound/csound_main.c:109

Then on the next cycle there is no file for diskin2 to read.
Anyone admit to changing turnoff2 or deact?

fdchclose certainly writes the NULL into the structure.  It looks to
me as if the turnoff2 is not removing the diskin2 call.

Anyone else?

==John ffitch

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

The problem I had was with fout and ficlose. Unless there are somehow 
related to turnoff2 or deact I don't see how those changes would have 
something to do with the current problems?

Rory.

Steven Yi wrote:
> HI John and all,
> 
> I don't know about this code, but I do remember Rory having a problem
> with some turnoff2 (I think) in an instrument with one of the
> filewriting opcodes; the fix he had certainly works but at the time I
> wasn't sure if it fixed the source of the problem or the symptoms.
> Perhaps this is also part of that and is the source of the problem he
> had?
> 
> I checked the CVS history of each of the files mentioned in the stack
> trace below (from envvvar.c to goto_ops.c) and didn't see any
> significant changes that seem to be related this issue between 5.03
> and 5.04 though only with a quick glance between versions.
> 
> steven
> 
> On 12/10/06, jpff  wrote:
>> Tracing, it seems that there is a call to sf_close from deact/urnoff2
>> and after that the file descriptor has vanished.
>>
>> #0  sf_close (sndfile=0x81846b8) at sndfile.c:416
>> #1  0x08052bcc in csoundFileClose (csound=0x80e1008, fd=0x80eb1b8)
>>     at Engine/envvar.c:1258
>> #2  0x08051cd7 in fdchclose (csound=0x80e1008, ip=0x817fb10)
>>     at Engine/auxfd.c:131
>> #3  0x0805bd01 in deact (csound=0x80e1008, ip=0x817fb10) at Engine/insert.c:507
>> #4  0x0808e02e in turnoff2 (csound=0x80e1008, p=0x816e538)
>>     at OOps/goto_ops.c:202
>> #5  0x0804ce4d in csoundPerform (csound=0x80e1008) at Top/csound.c:1097
>> #6  0x0804b0e8 in main (argc=2, argv=0xbffff084)
>>     at frontends/csound/csound_main.c:109
>>
>> Then on the next cycle there is no file for diskin2 to read.
>> Anyone admit to changing turnoff2 or deact?
>>
>> fdchclose certainly writes the NULL into the structure.  It looks to
>> me as if the turnoff2 is not removing the diskin2 call.
>>
>> Anyone else?
>>
>> ==John ffitch
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

You're right Steven, I was using turnoff2. The changes I suggested, 
which Victor added, were made to

static int outfile(CSOUND *csound, OUTFILE *p)
{
......
}

in fout.c.

All the change does is make sure there is a valid file pointer. I can't 
see how this might cause any problems, even though the the audio opcodes 
share a lot of common functions. Of course the current problem could 
have something to do with the problems I was getting with fout in the 
first place. In an earlier email from Victor he had this to say 
withregards teh problems:

"It's a bug that the opcode tries to write to a closed file. In fact it 
is probably a bug that the opcode is being called after a deinit, but 
that's another story..."

so the fix definitely made the symptoms go away, but I'm afraid the 
source of the problem may still be lurking around the place...

Rory.


Steven Yi wrote:
> I had thought that you had turnoff2 also involved in your examples,
> no?  If not, all of the audio file opcodes all use the same base
> functions so one symptom here or there might all be related I think,
> and this diskin2 was mentioned as part of David's CSD.
> 
> steven
> 
> 
> On 12/10/06, Rory Walsh  wrote:
>> The problem I had was with fout and ficlose. Unless there are somehow
>> related to turnoff2 or deact I don't see how those changes would have
>> something to do with the current problems?
>>
>> Rory.
>>
>> Steven Yi wrote:
>>> HI John and all,
>>>
>>> I don't know about this code, but I do remember Rory having a problem
>>> with some turnoff2 (I think) in an instrument with one of the
>>> filewriting opcodes; the fix he had certainly works but at the time I
>>> wasn't sure if it fixed the source of the problem or the symptoms.
>>> Perhaps this is also part of that and is the source of the problem he
>>> had?
>>>
>>> I checked the CVS history of each of the files mentioned in the stack
>>> trace below (from envvvar.c to goto_ops.c) and didn't see any
>>> significant changes that seem to be related this issue between 5.03
>>> and 5.04 though only with a quick glance between versions.
>>>
>>> steven
>>>
>>> On 12/10/06, jpff  wrote:
>>>> Tracing, it seems that there is a call to sf_close from deact/urnoff2
>>>> and after that the file descriptor has vanished.
>>>>
>>>> #0  sf_close (sndfile=0x81846b8) at sndfile.c:416
>>>> #1  0x08052bcc in csoundFileClose (csound=0x80e1008, fd=0x80eb1b8)
>>>>     at Engine/envvar.c:1258
>>>> #2  0x08051cd7 in fdchclose (csound=0x80e1008, ip=0x817fb10)
>>>>     at Engine/auxfd.c:131
>>>> #3  0x0805bd01 in deact (csound=0x80e1008, ip=0x817fb10) at Engine/insert.c:507
>>>> #4  0x0808e02e in turnoff2 (csound=0x80e1008, p=0x816e538)
>>>>     at OOps/goto_ops.c:202
>>>> #5  0x0804ce4d in csoundPerform (csound=0x80e1008) at Top/csound.c:1097
>>>> #6  0x0804b0e8 in main (argc=2, argv=0xbffff084)
>>>>     at frontends/csound/csound_main.c:109
>>>>
>>>> Then on the next cycle there is no file for diskin2 to read.
>>>> Anyone admit to changing turnoff2 or deact?
>>>>
>>>> fdchclose certainly writes the NULL into the structure.  It looks to
>>>> me as if the turnoff2 is not removing the diskin2 call.
>>>>
>>>> Anyone else?
>>>>
>>>> ==John ffitch
>>>>
>>>> -------------------------------------------------------------------------
>>>> Take Surveys. Earn Cash. Influence the Future of IT
>>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>>> opinions on IT & business topics through brief surveys - and earn cash
>>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>> -------------------------------------------------------------------------
>>> Take Surveys. Earn Cash. Influence the Future of IT
>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>> opinions on IT & business topics through brief surveys - and earn cash
>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hopefully it might explain all this, perhaps Victor has already sussed 
out what the problem is as it was he who diagnosed the opcode being 
called after a deinit.

Rory.

P.S. it was fairly obvious you were not implying that I caused the bug, 
but thanks for the confirmation!


Steven Yi wrote:
> Just to clarify, I wasn't implying that you had introduced a bug, but
> that the bug you observed initially may be related.  I think it's
> important that you brought up just now that the opcode being called
> after deinit; if that's the case, that seems likely to be a problem
> that could explain all this.
> 
> steven
> 
> 
> On 12/10/06, Rory Walsh  wrote:
>> You're right Steven, I was using turnoff2. The changes I suggested,
>> which Victor added, were made to
>>
>> static int outfile(CSOUND *csound, OUTFILE *p)
>> {
>> ......
>> }
>>
>> in fout.c.
>>
>> All the change does is make sure there is a valid file pointer. I can't
>> see how this might cause any problems, even though the the audio opcodes
>> share a lot of common functions. Of course the current problem could
>> have something to do with the problems I was getting with fout in the
>> first place. In an earlier email from Victor he had this to say
>> withregards teh problems:
>>
>> "It's a bug that the opcode tries to write to a closed file. In fact it
>> is probably a bug that the opcode is being called after a deinit, but
>> that's another story..."
>>
>> so the fix definitely made the symptoms go away, but I'm afraid the
>> source of the problem may still be lurking around the place...
>>
>> Rory.
>>
>>
>> Steven Yi wrote:
>>> I had thought that you had turnoff2 also involved in your examples,
>>> no?  If not, all of the audio file opcodes all use the same base
>>> functions so one symptom here or there might all be related I think,
>>> and this diskin2 was mentioned as part of David's CSD.
>>>
>>> steven
>>>
>>>
>>> On 12/10/06, Rory Walsh  wrote:
>>>> The problem I had was with fout and ficlose. Unless there are somehow
>>>> related to turnoff2 or deact I don't see how those changes would have
>>>> something to do with the current problems?
>>>>
>>>> Rory.
>>>>
>>>> Steven Yi wrote:
>>>>> HI John and all,
>>>>>
>>>>> I don't know about this code, but I do remember Rory having a problem
>>>>> with some turnoff2 (I think) in an instrument with one of the
>>>>> filewriting opcodes; the fix he had certainly works but at the time I
>>>>> wasn't sure if it fixed the source of the problem or the symptoms.
>>>>> Perhaps this is also part of that and is the source of the problem he
>>>>> had?
>>>>>
>>>>> I checked the CVS history of each of the files mentioned in the stack
>>>>> trace below (from envvvar.c to goto_ops.c) and didn't see any
>>>>> significant changes that seem to be related this issue between 5.03
>>>>> and 5.04 though only with a quick glance between versions.
>>>>>
>>>>> steven
>>>>>
>>>>> On 12/10/06, jpff  wrote:
>>>>>> Tracing, it seems that there is a call to sf_close from deact/urnoff2
>>>>>> and after that the file descriptor has vanished.
>>>>>>
>>>>>> #0  sf_close (sndfile=0x81846b8) at sndfile.c:416
>>>>>> #1  0x08052bcc in csoundFileClose (csound=0x80e1008, fd=0x80eb1b8)
>>>>>>     at Engine/envvar.c:1258
>>>>>> #2  0x08051cd7 in fdchclose (csound=0x80e1008, ip=0x817fb10)
>>>>>>     at Engine/auxfd.c:131
>>>>>> #3  0x0805bd01 in deact (csound=0x80e1008, ip=0x817fb10) at Engine/insert.c:507
>>>>>> #4  0x0808e02e in turnoff2 (csound=0x80e1008, p=0x816e538)
>>>>>>     at OOps/goto_ops.c:202
>>>>>> #5  0x0804ce4d in csoundPerform (csound=0x80e1008) at Top/csound.c:1097
>>>>>> #6  0x0804b0e8 in main (argc=2, argv=0xbffff084)
>>>>>>     at frontends/csound/csound_main.c:109
>>>>>>
>>>>>> Then on the next cycle there is no file for diskin2 to read.
>>>>>> Anyone admit to changing turnoff2 or deact?
>>>>>>
>>>>>> fdchclose certainly writes the NULL into the structure.  It looks to
>>>>>> me as if the turnoff2 is not removing the diskin2 call.
>>>>>>
>>>>>> Anyone else?
>>>>>>
>>>>>> ==John ffitch
>>>>>>
>>>>>> -------------------------------------------------------------------------
>>>>>> Take Surveys. Earn Cash. Influence the Future of IT
>>>>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>>>>> opinions on IT & business topics through brief surveys - and earn cash
>>>>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>> -------------------------------------------------------------------------
>>>>> Take Surveys. Earn Cash. Influence the Future of IT
>>>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>>>> opinions on IT & business topics through brief surveys - and earn cash
>>>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>> -------------------------------------------------------------------------
>>>> Take Surveys. Earn Cash. Influence the Future of IT
>>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>>> opinions on IT & business topics through brief surveys - and earn cash
>>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>> -------------------------------------------------------------------------
>>> Take Surveys. Earn Cash. Influence the Future of IT
>>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>>> opinions on IT & business topics through brief surveys - and earn cash
>>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys - and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net