Hi All,

I was debugging an issue found with Blue and the image opcodes.  The
bug turned out not to be related to the image opcodes, but rather had
to do with string channels.  I'm finding this behavior:

1. If I compile a CSD, performKsmps once (so that instr 0 is
executed), then call SetChannel, all is well
2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.

I've attached a python test file that illustrates this.  The crash
comes from strlen:

SECTION 1:
new alloc for instr 1:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: 13 at address: 0x0000000000000000
0x00007fff86acb812 in strlen ()
(gdb) bt
#0  0x00007fff86acb812 in strlen ()
#1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
str=0x732f73726573552f 
) at
csound_orc_semantics.c:65
#2  0x000000010008a864 in _PyEval_SliceIndex ()
#3  0x00000001000874d4 in PyEval_EvalFrameEx ()
#4  0x0000000100084093 in PyEval_EvalCodeEx ()
#5  0x0000000100083a24 in PyEval_EvalCode ()
#6  0x00000001000a2c2c in PyParser_ASTFromFile ()
#7  0x00000001000a2cd3 in PyRun_FileExFlags ()
#8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
#9  0x00000001000b3363 in Py_Main ()
#10 0x00007fff938bb5fd in start ()

I'm not sure yet how the problem should be solved.  To note, the API
related Java code in Blue had not changed from what I was using in
CS5, so this behavior is new in CS6.

Thanks!
steven

Doesn't sound good! Maybe we can ask Santa Claus ;)

On 24 December 2013 23:07, Steven Yi  wrote:
> Hi All,
>
> I was debugging an issue found with Blue and the image opcodes.  The
> bug turned out not to be related to the image opcodes, but rather had
> to do with string channels.  I'm finding this behavior:
>
> 1. If I compile a CSD, performKsmps once (so that instr 0 is
> executed), then call SetChannel, all is well
> 2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.
>
> I've attached a python test file that illustrates this.  The crash
> comes from strlen:
>
> SECTION 1:
> new alloc for instr 1:
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: 13 at address: 0x0000000000000000
> 0x00007fff86acb812 in strlen ()
> (gdb) bt
> #0  0x00007fff86acb812 in strlen ()
> #1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
> str=0x732f73726573552f 
) at
> csound_orc_semantics.c:65
> #2  0x000000010008a864 in _PyEval_SliceIndex ()
> #3  0x00000001000874d4 in PyEval_EvalFrameEx ()
> #4  0x0000000100084093 in PyEval_EvalCodeEx ()
> #5  0x0000000100083a24 in PyEval_EvalCode ()
> #6  0x00000001000a2c2c in PyParser_ASTFromFile ()
> #7  0x00000001000a2cd3 in PyRun_FileExFlags ()
> #8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
> #9  0x00000001000b3363 in Py_Main ()
> #10 0x00007fff938bb5fd in start ()
>
> I'm not sure yet how the problem should be solved.  To note, the API
> related Java code in Blue had not changed from what I was using in
> CS5, so this behavior is new in CS6.
>
> Thanks!
> steven
>
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT
> organizations don't have a clear picture of how application performance
> affects their revenue. With AppDynamics, you get 100% visibility into your
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Looks like strlen being passed a null pointer?



------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

The crash is in printksset_S (prints opcode), but the bug is in chn_export. If you use chnget instead, it does not crash.
On 24 Dec 2013, at 23:07, Steven Yi  wrote:

> Hi All,
> 
> I was debugging an issue found with Blue and the image opcodes.  The
> bug turned out not to be related to the image opcodes, but rather had
> to do with string channels.  I'm finding this behavior:
> 
> 1. If I compile a CSD, performKsmps once (so that instr 0 is
> executed), then call SetChannel, all is well
> 2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.
> 
> I've attached a python test file that illustrates this.  The crash
> comes from strlen:
> 
> SECTION 1:
> new alloc for instr 1:
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: 13 at address: 0x0000000000000000
> 0x00007fff86acb812 in strlen ()
> (gdb) bt
> #0  0x00007fff86acb812 in strlen ()
> #1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
> str=0x732f73726573552f 
) at
> csound_orc_semantics.c:65
> #2  0x000000010008a864 in _PyEval_SliceIndex ()
> #3  0x00000001000874d4 in PyEval_EvalFrameEx ()
> #4  0x0000000100084093 in PyEval_EvalCodeEx ()
> #5  0x0000000100083a24 in PyEval_EvalCode ()
> #6  0x00000001000a2c2c in PyParser_ASTFromFile ()
> #7  0x00000001000a2cd3 in PyRun_FileExFlags ()
> #8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
> #9  0x00000001000b3363 in Py_Main ()
> #10 0x00007fff938bb5fd in start ()
> 
> I'm not sure yet how the problem should be solved.  To note, the API
> related Java code in Blue had not changed from what I was using in
> CS5, so this behavior is new in CS6.
> 
> Thanks!
> steven
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT 
> organizations don't have a clear picture of how application performance 
> affects their revenue. With AppDynamics, you get 100% visibility into your 
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

My impression is that this line is dodgy  (bus.c:1171)

    if(type == CSOUND_STRING_CHANNEL)
      chn->data = (MYFLT *) ((STRINGDAT *)p->arg)->data;

because when a new string is set, chn->data is liable to point elsewhere and
((STRINGDAT *)p->arg)->data  does not point to the contents of the channel anymore.

I think the solution might be to turn “chn->data” into STRINGDAT pointers (and update 
all of the code relating to string channels). At the moment, chn->data is carrying a pointer
to char* for string channels and that is not going to work for chnexport.

Victor

On 25 Dec 2013, at 16:16, Victor Lazzarini  wrote:

> The crash is in printksset_S (prints opcode), but the bug is in chn_export. If you use chnget instead, it does not crash.
> On 24 Dec 2013, at 23:07, Steven Yi  wrote:
> 
>> Hi All,
>> 
>> I was debugging an issue found with Blue and the image opcodes.  The
>> bug turned out not to be related to the image opcodes, but rather had
>> to do with string channels.  I'm finding this behavior:
>> 
>> 1. If I compile a CSD, performKsmps once (so that instr 0 is
>> executed), then call SetChannel, all is well
>> 2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.
>> 
>> I've attached a python test file that illustrates this.  The crash
>> comes from strlen:
>> 
>> SECTION 1:
>> new alloc for instr 1:
>> Program received signal EXC_BAD_ACCESS, Could not access memory.
>> Reason: 13 at address: 0x0000000000000000
>> 0x00007fff86acb812 in strlen ()
>> (gdb) bt
>> #0  0x00007fff86acb812 in strlen ()
>> #1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
>> str=0x732f73726573552f 
) at
>> csound_orc_semantics.c:65
>> #2  0x000000010008a864 in _PyEval_SliceIndex ()
>> #3  0x00000001000874d4 in PyEval_EvalFrameEx ()
>> #4  0x0000000100084093 in PyEval_EvalCodeEx ()
>> #5  0x0000000100083a24 in PyEval_EvalCode ()
>> #6  0x00000001000a2c2c in PyParser_ASTFromFile ()
>> #7  0x00000001000a2cd3 in PyRun_FileExFlags ()
>> #8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
>> #9  0x00000001000b3363 in Py_Main ()
>> #10 0x00007fff938bb5fd in start ()
>> 
>> I'm not sure yet how the problem should be solved.  To note, the API
>> related Java code in Blue had not changed from what I was using in
>> CS5, so this behavior is new in CS6.
>> 
>> Thanks!
>> steven
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT 
>> organizations don't have a clear picture of how application performance 
>> affects their revenue. With AppDynamics, you get 100% visibility into your 
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk_______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> 
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT 
> organizations don't have a clear picture of how application performance 
> affects their revenue. With AppDynamics, you get 100% visibility into your 
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi Victor,

(Happy Holidays! :) )

Thanks for looking into it.  I'll try to find out why this works when
setChannel is called after the first chnexport is called but not
before.  I assume there's just some branch of code that wasn't updated
to use STRINGDAT as you mentioned.

Thanks!
steven

On Wed, Dec 25, 2013 at 11:30 AM, Victor Lazzarini
 wrote:
> My impression is that this line is dodgy  (bus.c:1171)
>
>     if(type == CSOUND_STRING_CHANNEL)
>       chn->data = (MYFLT *) ((STRINGDAT *)p->arg)->data;
>
> because when a new string is set, chn->data is liable to point elsewhere and
> ((STRINGDAT *)p->arg)->data  does not point to the contents of the channel anymore.
>
> I think the solution might be to turn “chn->data” into STRINGDAT pointers (and update
> all of the code relating to string channels). At the moment, chn->data is carrying a pointer
> to char* for string channels and that is not going to work for chnexport.
>
> Victor
>
> On 25 Dec 2013, at 16:16, Victor Lazzarini  wrote:
>
>> The crash is in printksset_S (prints opcode), but the bug is in chn_export. If you use chnget instead, it does not crash.
>> On 24 Dec 2013, at 23:07, Steven Yi  wrote:
>>
>>> Hi All,
>>>
>>> I was debugging an issue found with Blue and the image opcodes.  The
>>> bug turned out not to be related to the image opcodes, but rather had
>>> to do with string channels.  I'm finding this behavior:
>>>
>>> 1. If I compile a CSD, performKsmps once (so that instr 0 is
>>> executed), then call SetChannel, all is well
>>> 2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.
>>>
>>> I've attached a python test file that illustrates this.  The crash
>>> comes from strlen:
>>>
>>> SECTION 1:
>>> new alloc for instr 1:
>>> Program received signal EXC_BAD_ACCESS, Could not access memory.
>>> Reason: 13 at address: 0x0000000000000000
>>> 0x00007fff86acb812 in strlen ()
>>> (gdb) bt
>>> #0  0x00007fff86acb812 in strlen ()
>>> #1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
>>> str=0x732f73726573552f 
) at
>>> csound_orc_semantics.c:65
>>> #2  0x000000010008a864 in _PyEval_SliceIndex ()
>>> #3  0x00000001000874d4 in PyEval_EvalFrameEx ()
>>> #4  0x0000000100084093 in PyEval_EvalCodeEx ()
>>> #5  0x0000000100083a24 in PyEval_EvalCode ()
>>> #6  0x00000001000a2c2c in PyParser_ASTFromFile ()
>>> #7  0x00000001000a2cd3 in PyRun_FileExFlags ()
>>> #8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
>>> #9  0x00000001000b3363 in Py_Main ()
>>> #10 0x00007fff938bb5fd in start ()
>>>
>>> I'm not sure yet how the problem should be solved.  To note, the API
>>> related Java code in Blue had not changed from what I was using in
>>> CS5, so this behavior is new in CS6.
>>>
>>> Thanks!
>>> steven
>>> ------------------------------------------------------------------------------
>>> Rapidly troubleshoot problems before they affect your business. Most IT
>>> organizations don't have a clear picture of how application performance
>>> affects their revenue. With AppDynamics, you get 100% visibility into your
>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk_______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT
>> organizations don't have a clear picture of how application performance
>> affects their revenue. With AppDynamics, you get 100% visibility into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT
> organizations don't have a clear picture of how application performance
> affects their revenue. With AppDynamics, you get 100% visibility into your
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi Steven,

I think it was updated, but not correctly. The correct way is make the channel carry the STRINGDAT rather than
the char pointer.

Regards

Victor
On 26 Dec 2013, at 21:45, Steven Yi  wrote:

> Hi Victor,
> 
> (Happy Holidays! :) )
> 
> Thanks for looking into it.  I'll try to find out why this works when
> setChannel is called after the first chnexport is called but not
> before.  I assume there's just some branch of code that wasn't updated
> to use STRINGDAT as you mentioned.
> 
> Thanks!
> steven
> 
> On Wed, Dec 25, 2013 at 11:30 AM, Victor Lazzarini
>  wrote:
>> My impression is that this line is dodgy  (bus.c:1171)
>> 
>>    if(type == CSOUND_STRING_CHANNEL)
>>      chn->data = (MYFLT *) ((STRINGDAT *)p->arg)->data;
>> 
>> because when a new string is set, chn->data is liable to point elsewhere and
>> ((STRINGDAT *)p->arg)->data  does not point to the contents of the channel anymore.
>> 
>> I think the solution might be to turn “chn->data” into STRINGDAT pointers (and update
>> all of the code relating to string channels). At the moment, chn->data is carrying a pointer
>> to char* for string channels and that is not going to work for chnexport.
>> 
>> Victor
>> 
>> On 25 Dec 2013, at 16:16, Victor Lazzarini  wrote:
>> 
>>> The crash is in printksset_S (prints opcode), but the bug is in chn_export. If you use chnget instead, it does not crash.
>>> On 24 Dec 2013, at 23:07, Steven Yi  wrote:
>>> 
>>>> Hi All,
>>>> 
>>>> I was debugging an issue found with Blue and the image opcodes.  The
>>>> bug turned out not to be related to the image opcodes, but rather had
>>>> to do with string channels.  I'm finding this behavior:
>>>> 
>>>> 1. If I compile a CSD, performKsmps once (so that instr 0 is
>>>> executed), then call SetChannel, all is well
>>>> 2. If I compile a CSD, run SetChannel, then run performKsmps, I get a crash.
>>>> 
>>>> I've attached a python test file that illustrates this.  The crash
>>>> comes from strlen:
>>>> 
>>>> SECTION 1:
>>>> new alloc for instr 1:
>>>> Program received signal EXC_BAD_ACCESS, Could not access memory.
>>>> Reason: 13 at address: 0x0000000000000000
>>>> 0x00007fff86acb812 in strlen ()
>>>> (gdb) bt
>>>> #0  0x00007fff86acb812 in strlen ()
>>>> #1  0x0000000101ae7505 in cs_strdup (csound=0x101090e00,
>>>> str=0x732f73726573552f 
) at
>>>> csound_orc_semantics.c:65
>>>> #2  0x000000010008a864 in _PyEval_SliceIndex ()
>>>> #3  0x00000001000874d4 in PyEval_EvalFrameEx ()
>>>> #4  0x0000000100084093 in PyEval_EvalCodeEx ()
>>>> #5  0x0000000100083a24 in PyEval_EvalCode ()
>>>> #6  0x00000001000a2c2c in PyParser_ASTFromFile ()
>>>> #7  0x00000001000a2cd3 in PyRun_FileExFlags ()
>>>> #8  0x00000001000a2821 in PyRun_SimpleFileExFlags ()
>>>> #9  0x00000001000b3363 in Py_Main ()
>>>> #10 0x00007fff938bb5fd in start ()
>>>> 
>>>> I'm not sure yet how the problem should be solved.  To note, the API
>>>> related Java code in Blue had not changed from what I was using in
>>>> CS5, so this behavior is new in CS6.
>>>> 
>>>> Thanks!
>>>> steven
>>>> ------------------------------------------------------------------------------
>>>> Rapidly troubleshoot problems before they affect your business. Most IT
>>>> organizations don't have a clear picture of how application performance
>>>> affects their revenue. With AppDynamics, you get 100% visibility into your
>>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk_______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> Rapidly troubleshoot problems before they affect your business. Most IT
>>> organizations don't have a clear picture of how application performance
>>> affects their revenue. With AppDynamics, you get 100% visibility into your
>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> 
>> 
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT
>> organizations don't have a clear picture of how application performance
>> affects their revenue. With AppDynamics, you get 100% visibility into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT 
> organizations don't have a clear picture of how application performance 
> affects their revenue. With AppDynamics, you get 100% visibility into your 
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net