That’s the UDO init code.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
> 
> from gdb, this message is printed
> 
> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
> 
> 
> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
> Steps to reproduce
> 
> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
> 
> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
> 
> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
> 
> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
> 
> I will add that equivalent code works without segfault when running directly from cs

Tested here on Windows with both MinGW and VS builds and both crash in
the same place:

  csound64.dll!useropcdset(CSOUND_ * csound, UOPCODE * p) Line 1255 C
Symbols loaded.
> csound64.dll!insert(CSOUND_ * csound, int insno, event * newevtp) Line 300 C Symbols loaded.
  csound64.dll!process_score_event(CSOUND_ * csound, event * evt, int
rtEvt) Line 777 C Symbols loaded.
  csound64.dll!process_rt_event(CSOUND_ * csound, int sensType) Line
873 C Symbols loaded.
  csound64.dll!sensevents(CSOUND_ * csound) Line 1061 C Symbols loaded.
  csound64.dll!csoundPerform(CSOUND_ * csound) Line 2147 C Symbols loaded.
  csound.exe!main(int argc, char * * argv) Line 327 C Symbols loaded.
  [External Code] Annotated Frame

Haven't found a cause for it yet.

On Thu, Oct 26, 2017 at 5:02 PM, Victor Lazzarini
 wrote:
> That’s the UDO init code.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>
>> from gdb, this message is printed
>>
>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>
>>
>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>> Steps to reproduce
>>
>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>
>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>
>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>
>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>
>> I will add that equivalent code works without segfault when running directly from csd file.
>>

I wonder if the whole code is being sent. UDP has a maximum number of characters (MTU), which this might be exceeding.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
> 
> from gdb, this message is printed
> 
> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
> 
> 
> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
> Steps to reproduce
> 
> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
> 
> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
> 
> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
> 
> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
> 
> I will add that equivalent code works without segf

What is happening is that the message is broken into two pieces and they get compiled separately and so it
doesn’t quite work. There is a problem with large bits of code, which I will try to address.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
> 
> from gdb, this message is printed
> 
> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
> 
> 
> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
> Steps to reproduce
> 
> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
> 
> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
> 
> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
> 
> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
> 
> I will add that equivalent code works without segfault when running directly fro

There's something else I'm seeing that might be a part of this:

When the program crashes, tp is NULL.  instno is 201 (I think that
must be assigned to the Chorus UDO).

If I modify the orc code to make the mixer instrument 100 instead of
10000, I don't get a crash.

If I run the code as a CSD on its own its fine.  (Attached)

On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
 wrote:
> What is happening is that the message is broken into two pieces and they get compiled separately and so it
> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>
>> from gdb, this message is printed
>>
>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>
>>
>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>> Steps to reproduce
>>
>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>
>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>
>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>
>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>
>> I will add that equivalent code works without segfault when running directly from csd file.
>>
>

I was having trouble sending large orcs with nc and so I have added support for longer
code or code broken in lines using { and }.
Anyway, I seem to be able to reproduce the segfault with chorus

========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
> 
> There's something else I'm seeing that might be a part of this:
> 
> When the program crashes, tp is NULL.  instno is 201 (I think that
> must be assigned to the Chorus UDO).
> 
> If I modify the orc code to make the mixer instrument 100 instead of
> 10000, I don't get a crash.
> 
> If I run the code as a CSD on its own its fine.  (Attached)
> 
> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>  wrote:
>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>> 
>>> from gdb, this message is printed
>>> 
>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>> 
>>> 
>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>> Steps to reproduce
>>> 
>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>> 
>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>> 
>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>> 
>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>> 
>>> I will add that equivalent code works without segfault when runni

I get 

   1252	    if (!p->ip) {
   1253	      /* search for already allocated, but not active instance */
   1254	      /* if none was found, allocate a new instance */
-> 1255	      if (!tp->act_instance)
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
> 
> There's something else I'm seeing that might be a part of this:
> 
> When the program crashes, tp is NULL.  instno is 201 (I think that
> must be assigned to the Chorus UDO).
> 
> If I modify the orc code to make the mixer instrument 100 instead of
> 10000, I don't get a crash.
> 
> If I run the code as a CSD on its own its fine.  (Attached)
> 
> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>  wrote:
>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>> 
>>> from gdb, this message is printed
>>> 
>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>> 
>>> 
>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>> Steps to reproduce
>>> 
>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>> 
>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>> 
>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>> 
>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>> 
>>> I will add that equivalent code works without segfault when running directly from csd file.
>>> 
>> 
> 

Yes, we're on the same page as that's exactly where I had the crash
with NULL tp.

On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
 wrote:
> I get
>
>    1252     if (!p->ip) {
>    1253       /* search for already allocated, but not active instance */
>    1254       /* if none was found, allocate a new instance */
> -> 1255       if (!tp->act_instance)
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>
>> There's something else I'm seeing that might be a part of this:
>>
>> When the program crashes, tp is NULL.  instno is 201 (I think that
>> must be assigned to the Chorus UDO).
>>
>> If I modify the orc code to make the mixer instrument 100 instead of
>> 10000, I don't get a crash.
>>
>> If I run the code as a CSD on its own its fine.  (Attached)
>>
>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>  wrote:
>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>> ========================
>>> Prof. Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>>
>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>
>>>> from gdb, this message is printed
>>>>
>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>
>>>>
>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>> Steps to reproduce
>>>>
>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>
>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>
>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>
>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>
>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>
>>>
>> 

It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.

========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
> 
> Yes, we're on the same page as that's exactly where I had the crash
> with NULL tp.
> 
> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>  wrote:
>> I get
>> 
>>   1252     if (!p->ip) {
>>   1253       /* search for already allocated, but not active instance */
>>   1254       /* if none was found, allocate a new instance */
>> -> 1255       if (!tp->act_instance)
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>> 
>>> There's something else I'm seeing that might be a part of this:
>>> 
>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>> must be assigned to the Chorus UDO).
>>> 
>>> If I modify the orc code to make the mixer instrument 100 instead of
>>> 10000, I don't get a crash.
>>> 
>>> If I run the code as a CSD on its own its fine.  (Attached)
>>> 
>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>  wrote:
>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>> 
>>>>> from gdb, this message is printed
>>>>> 
>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>> 
>>>>> 
>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>> Steps to reproduce
>>>>> 
>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>> 
>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>> 
>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>> 
>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>> 
>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>> 
>>>> 
>>> <

I’ve added an InitError for that case.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 27 Oct 2017, at 00:03, Victor Lazzarini  wrote:
> 
> It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.
> 
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952 
> 
>> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
>> 
>> Yes, we're on the same page as that's exactly where I had the crash
>> with NULL tp.
>> 
>> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>>  wrote:
>>> I get
>>> 
>>>  1252     if (!p->ip) {
>>>  1253       /* search for already allocated, but not active instance */
>>>  1254       /* if none was found, allocate a new instance */
>>> -> 1255       if (!tp->act_instance)
>>> ========================
>>> Prof. Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>> 
>>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>>> 
>>>> There's something else I'm seeing that might be a part of this:
>>>> 
>>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>>> must be assigned to the Chorus UDO).
>>>> 
>>>> If I modify the orc code to make the mixer instrument 100 instead of
>>>> 10000, I don't get a crash.
>>>> 
>>>> If I run the code as a CSD on its own its fine.  (Attached)
>>>> 
>>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>>  wrote:
>>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>> 
>>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>>> 
>>>>>> from gdb, this message is printed
>>>>>> 
>>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>>> 
>>>>>> 
>>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>>> Steps to reproduce
>>>>>> 
>>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>>> 
>>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>>> 
>>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>>> 
>>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>>> 
>>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>>> 
>>>>> 

But there's still a problem here I think.

If you run the code on its own outside of the context of UDP (see the
test.csd I attached in previous message in this thread) it works just
fine.

I think it's good to protect against trying to deref tp when it's
NULL, but it seems there's still a bug to fix.

On Thu, Oct 26, 2017 at 7:11 PM, Victor Lazzarini
 wrote:
> I’ve added an InitError for that case.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 27 Oct 2017, at 00:03, Victor Lazzarini  wrote:
>>
>> It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.
>>
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>>> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
>>>
>>> Yes, we're on the same page as that's exactly where I had the crash
>>> with NULL tp.
>>>
>>> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>>>  wrote:
>>>> I get
>>>>
>>>>  1252     if (!p->ip) {
>>>>  1253       /* search for already allocated, but not active instance */
>>>>  1254       /* if none was found, allocate a new instance */
>>>> -> 1255       if (!tp->act_instance)
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>>
>>>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>>>>
>>>>> There's something else I'm seeing that might be a part of this:
>>>>>
>>>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>>>> must be assigned to the Chorus UDO).
>>>>>
>>>>> If I modify the orc code to make the mixer instrument 100 instead of
>>>>> 10000, I don't get a crash.
>>>>>
>>>>> If I run the code as a CSD on its own its fine.  (Attached)
>>>>>
>>>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>>>  wrote:
>>>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>>>> ========================
>>>>>> Prof. Victor Lazzarini
>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>> Maynooth University,
>>>>>> Maynooth, Co Kildare, Ireland
>>>>>> Tel: 00 353 7086936
>>>>>> Fax: 00 353 1 7086952
>>>>>>
>>>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>>>>
>>>>>>> from gdb, this message is printed
>>>>>>>
>>>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>>>>
>>>>>>>
>>>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>>>> Steps to reproduce
>>>>>>>
>>>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>>>>
>>>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>>>>
>>>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>>>>
>>>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>>>>
>>>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>>>>
>>>>>>
>>>>> 
>>>>
>>

But there's still a problem here I think.

If you run the code on its own outside of the context of UDP (see the
test.csd I attached in previous message in this thread) it works just
fine.

I think it's good to protect against trying to deref tp when it's
NULL, but it seems there's still a bug to fix.

On Thu, Oct 26, 2017 at 7:11 PM, Victor Lazzarini
<Victor.Lazzarini@mu.ie> wrote:

I’ve added an InitError for that case.

========================

Prof. Victor Lazzarini

Dean of Arts, Celtic Studies, and Philosophy,

Maynooth University,

Maynooth, Co Kildare, Ireland

Tel: 00 353 7086936

Fax: 00 353 1 7086952

On 27 Oct 2017, at 00:03, Victor Lazzarini <Victor.Lazzarini@MU.IE> wrote:

It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.

========================

Prof. Victor Lazzarini

Dean of Arts, Celtic Studies, and Philosophy,

Maynooth University,

Maynooth, Co Kildare, Ireland

Tel: 00 353 7086936

Fax: 00 353 1 7086952

On 27 Oct 2017, at 00:02, Steven Yi <stevenyi@GMAIL.COM> wrote:

Yes, we're on the same page as that's exactly where I had the crash

with NULL tp.

On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini

<Victor.Lazzarini@mu.ie> wrote:

I get

1252     if (!p->ip) {

1253       /* search for already allocated, but not active instance */

1254       /* if none was found, allocate a new instance */

-> 1255       if (!tp->act_instance)

========================

Prof. Victor Lazzarini

Dean of Arts, Celtic Studies, and Philosophy,

Maynooth University,

Maynooth, Co Kildare, Ireland

Tel: 00 353 7086936

Fax: 00 353 1 7086952

On 26 Oct 2017, at 22:41, Steven Yi <stevenyi@gmail.com> wrote:

There's something else I'm seeing that might be a part of this:

When the program crashes, tp is NULL.  instno is 201 (I think that

must be assigned to the Chorus UDO).

If I modify the orc code to make the mixer instrument 100 instead of

10000, I don't get a crash.

If I run the code as a CSD on its own its fine.  (Attached)

On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini

<Victor.Lazzarini@mu.ie> wrote:

What is happening is that the message is broken into two pieces and they get compiled separately and so it

doesn’t quite work. There is a problem with large bits of code, which I will try to address.

========================

Prof. Victor Lazzarini

Dean of Arts, Celtic Studies, and Philosophy,

Maynooth University,

Maynooth, Co Kildare, Ireland

Tel: 00 353 7086936

Fax: 00 353 1 7086952

On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson <hlolli@gmail.com> wrote:

from gdb, this message is printed

Thread 1 "csound" received signal SIGSEGV, Segmentation fault.

0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0

On 26 October 2017 at 22:35, Hlöðver Sigurðsson <hlolli@gmail.com> wrote:

Steps to reproduce

1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)

2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`

3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.

4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).

I will add that equivalent code works without segfault when running directly from csd file.

<test.csd>

yes, I think it is to do with merging. I have not had a chance to investigate yet.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 27 Oct 2017, at 09:44, Steven Yi  wrote:
> 
> Perhaps there is an issue with merging when there is a large instrument number introduced? Thinking there is a difference between first compile pass and latter ones. (Sidenote: in cs7, we should revisit named instruments having numbers assigned, as one can easily assign a numbered instrument after a named one that redefines the assigned number)
> On Fri, Oct 27, 2017 at 1:23 AM Victor Lazzarini  wrote:
> yes, there is a bug alright.
> 
> Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy
> Maynooth University
> Ireland
> 
> On 27 Oct 2017, at 01:39, Steven Yi  wrote:
> 
>> But there's still a problem here I think.
>> 
>> If you run the code on its own outside of the context of UDP (see the
>> test.csd I attached in previous message in this thread) it works just
>> fine.
>> 
>> I think it's good to protect against trying to deref tp when it's
>> NULL, but it seems there's still a bug to fix.
>> 
>> On Thu, Oct 26, 2017 at 7:11 PM, Victor Lazzarini
>>  wrote:
>>> I’ve added an InitError for that case.
>>> ========================
>>> Prof. Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>> 
>>>> On 27 Oct 2017, at 00:03, Victor Lazzarini  wrote:
>>>> 
>>>> It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.
>>>> 
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
>>>>> 
>>>>> Yes, we're on the same page as that's exactly where I had the crash
>>>>> with NULL tp.
>>>>> 
>>>>> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>>>>>  wrote:
>>>>>> I get
>>>>>> 
>>>>>> 1252     if (!p->ip) {
>>>>>> 1253       /* search for already allocated, but not active instance */
>>>>>> 1254       /* if none was found, allocate a new instance */
>>>>>> -> 1255       if (!tp->act_instance)
>>>>>> ========================
>>>>>> Prof. Victor Lazzarini
>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>> Maynooth University,
>>>>>> Maynooth, Co Kildare, Ireland
>>>>>> Tel: 00 353 7086936
>>>>>> Fax: 00 353 1 7086952
>>>>>> 
>>>>>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>>>>>> 
>>>>>>> There's something else I'm seeing that might be a part of this:
>>>>>>> 
>>>>>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>>>>>> must be assigned to the Chorus UDO).
>>>>>>> 
>>>>>>> If I modify the orc code to make the mixer instrument 100 instead of
>>>>>>> 10000, I don't get a crash.
>>>>>>> 
>>>>>>> If I run the code as a CSD on its own its fine.  (Attached)
>>>>>>> 
>>>>>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>>>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>>>>>> ========================
>>>>>>>> Prof. Victor Lazzarini
>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>> Maynooth University,
>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>> Tel: 00 353 7086936
>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>> 
>>>>>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>>>>>> 
>>>>>>>>> from gdb, this message is printed
>>>>>>>>> 
>>>>>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>>>>>> 
>>>>>>>>> 
>>>>>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>>>>>> Steps to reproduce
>>>>>>>>> 
>>>>>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>>>>>> 
>>>>>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>>>>>> 
>>>>>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>>>>>> 
>>>>>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>>>>>> 
>>>>>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>>>>>> 
>>>>>>>> 
>>>>>>> 

I think I got this now. The insertion of opcodes was happening at the wrong time in the
compilation. This only applied to compilations after the first one, so that is why it worked
in normal running. 
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 27 Oct 2017, at 14:01, Victor Lazzarini  wrote:
> 
> yes, I think it is to do with merging. I have not had a chance to investigate yet.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952 
> 
>> On 27 Oct 2017, at 09:44, Steven Yi  wrote:
>> 
>> Perhaps there is an issue with merging when there is a large instrument number introduced? Thinking there is a difference between first compile pass and latter ones. (Sidenote: in cs7, we should revisit named instruments having numbers assigned, as one can easily assign a numbered instrument after a named one that redefines the assigned number)
>> On Fri, Oct 27, 2017 at 1:23 AM Victor Lazzarini  wrote:
>> yes, there is a bug alright.
>> 
>> Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy
>> Maynooth University
>> Ireland
>> 
>> On 27 Oct 2017, at 01:39, Steven Yi  wrote:
>> 
>>> But there's still a problem here I think.
>>> 
>>> If you run the code on its own outside of the context of UDP (see the
>>> test.csd I attached in previous message in this thread) it works just
>>> fine.
>>> 
>>> I think it's good to protect against trying to deref tp when it's
>>> NULL, but it seems there's still a bug to fix.
>>> 
>>> On Thu, Oct 26, 2017 at 7:11 PM, Victor Lazzarini
>>>  wrote:
>>>> I’ve added an InitError for that case.
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 27 Oct 2017, at 00:03, Victor Lazzarini  wrote:
>>>>> 
>>>>> It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.
>>>>> 
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>> 
>>>>>> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
>>>>>> 
>>>>>> Yes, we're on the same page as that's exactly where I had the crash
>>>>>> with NULL tp.
>>>>>> 
>>>>>> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>>>>>>  wrote:
>>>>>>> I get
>>>>>>> 
>>>>>>> 1252     if (!p->ip) {
>>>>>>> 1253       /* search for already allocated, but not active instance */
>>>>>>> 1254       /* if none was found, allocate a new instance */
>>>>>>> -> 1255       if (!tp->act_instance)
>>>>>>> ========================
>>>>>>> Prof. Victor Lazzarini
>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>> Maynooth University,
>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>> Tel: 00 353 7086936
>>>>>>> Fax: 00 353 1 7086952
>>>>>>> 
>>>>>>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>>>>>>> 
>>>>>>>> There's something else I'm seeing that might be a part of this:
>>>>>>>> 
>>>>>>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>>>>>>> must be assigned to the Chorus UDO).
>>>>>>>> 
>>>>>>>> If I modify the orc code to make the mixer instrument 100 instead of
>>>>>>>> 10000, I don't get a crash.
>>>>>>>> 
>>>>>>>> If I run the code as a CSD on its own its fine.  (Attached)
>>>>>>>> 
>>>>>>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>>>>>>  wrote:
>>>>>>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>>>>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>>>>>>> ========================
>>>>>>>>> Prof. Victor Lazzarini
>>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>>> Maynooth University,
>>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>>> Tel: 00 353 7086936
>>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>> 
>>>>>>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>>>>>>> 
>>>>>>>>>> from gdb, this message is printed
>>>>>>>>>> 
>>>>>>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>>>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>>>>>>> Steps to reproduce
>>>>>>>>>> 
>>>>>>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>>>>>>> 
>>>>>>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>>>>>>> 
>>>>>>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>>>>>>> 
>>>>>>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>>>>>>> 
>>>>>>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>>>>>>> 
>>>>>>>>> 
>>>>>>>> 
>>>>>

Tested here on Windows and it's now working great.  (Though, I had to
add aL *= 0dbfs and aR *= 0dbfs in instr 1 to hear anything ;) ).
Nice find!

On Fri, Oct 27, 2017 at 1:32 PM, Victor Lazzarini
 wrote:
> I think I got this now. The insertion of opcodes was happening at the wrong time in the
> compilation. This only applied to compilations after the first one, so that is why it worked
> in normal running.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 27 Oct 2017, at 14:01, Victor Lazzarini  wrote:
>>
>> yes, I think it is to do with merging. I have not had a chance to investigate yet.
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>>> On 27 Oct 2017, at 09:44, Steven Yi  wrote:
>>>
>>> Perhaps there is an issue with merging when there is a large instrument number introduced? Thinking there is a difference between first compile pass and latter ones. (Sidenote: in cs7, we should revisit named instruments having numbers assigned, as one can easily assign a numbered instrument after a named one that redefines the assigned number)
>>> On Fri, Oct 27, 2017 at 1:23 AM Victor Lazzarini  wrote:
>>> yes, there is a bug alright.
>>>
>>> Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy
>>> Maynooth University
>>> Ireland
>>>
>>> On 27 Oct 2017, at 01:39, Steven Yi  wrote:
>>>
>>>> But there's still a problem here I think.
>>>>
>>>> If you run the code on its own outside of the context of UDP (see the
>>>> test.csd I attached in previous message in this thread) it works just
>>>> fine.
>>>>
>>>> I think it's good to protect against trying to deref tp when it's
>>>> NULL, but it seems there's still a bug to fix.
>>>>
>>>> On Thu, Oct 26, 2017 at 7:11 PM, Victor Lazzarini
>>>>  wrote:
>>>>> I’ve added an InitError for that case.
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>>
>>>>>> On 27 Oct 2017, at 00:03, Victor Lazzarini  wrote:
>>>>>>
>>>>>> It seems it can’t find an instrument. If reduce instr 10000 to  instr 10 it doesn’t crash anymore.
>>>>>>
>>>>>> ========================
>>>>>> Prof. Victor Lazzarini
>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>> Maynooth University,
>>>>>> Maynooth, Co Kildare, Ireland
>>>>>> Tel: 00 353 7086936
>>>>>> Fax: 00 353 1 7086952
>>>>>>
>>>>>>> On 27 Oct 2017, at 00:02, Steven Yi  wrote:
>>>>>>>
>>>>>>> Yes, we're on the same page as that's exactly where I had the crash
>>>>>>> with NULL tp.
>>>>>>>
>>>>>>> On Thu, Oct 26, 2017 at 7:00 PM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> I get
>>>>>>>>
>>>>>>>> 1252     if (!p->ip) {
>>>>>>>> 1253       /* search for already allocated, but not active instance */
>>>>>>>> 1254       /* if none was found, allocate a new instance */
>>>>>>>> -> 1255       if (!tp->act_instance)
>>>>>>>> ========================
>>>>>>>> Prof. Victor Lazzarini
>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>> Maynooth University,
>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>> Tel: 00 353 7086936
>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>
>>>>>>>>> On 26 Oct 2017, at 22:41, Steven Yi  wrote:
>>>>>>>>>
>>>>>>>>> There's something else I'm seeing that might be a part of this:
>>>>>>>>>
>>>>>>>>> When the program crashes, tp is NULL.  instno is 201 (I think that
>>>>>>>>> must be assigned to the Chorus UDO).
>>>>>>>>>
>>>>>>>>> If I modify the orc code to make the mixer instrument 100 instead of
>>>>>>>>> 10000, I don't get a crash.
>>>>>>>>>
>>>>>>>>> If I run the code as a CSD on its own its fine.  (Attached)
>>>>>>>>>
>>>>>>>>> On Thu, Oct 26, 2017 at 5:37 PM, Victor Lazzarini
>>>>>>>>>  wrote:
>>>>>>>>>> What is happening is that the message is broken into two pieces and they get compiled separately and so it
>>>>>>>>>> doesn’t quite work. There is a problem with large bits of code, which I will try to address.
>>>>>>>>>> ========================
>>>>>>>>>> Prof. Victor Lazzarini
>>>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>>>> Maynooth University,
>>>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>>>> Tel: 00 353 7086936
>>>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>>>
>>>>>>>>>>> On 26 Oct 2017, at 21:38, Hlöðver Sigurðsson  wrote:
>>>>>>>>>>>
>>>>>>>>>>> from gdb, this message is printed
>>>>>>>>>>>
>>>>>>>>>>> Thread 1 "csound" received signal SIGSEGV, Segmentation fault.
>>>>>>>>>>> 0x00007ffff7899b8f in useropcdset () from /usr/local/lib/libcsound64.so.6.0
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On 26 October 2017 at 22:35, Hlöðver Sigurðsson  wrote:
>>>>>>>>>>> Steps to reproduce
>>>>>>>>>>>
>>>>>>>>>>> 1. Start Csound with `csound -odac --port=6006` (useing dev branch compiled today)
>>>>>>>>>>>
>>>>>>>>>>> 2. Useing Nodejs (Im useing v8.0) send udp messages to the csound instance with this script https://pastebin.com/2GSqXKmP notice `csound command: Segmentation fault`
>>>>>>>>>>>
>>>>>>>>>>> 3. With the same javascript script, delete or comment with semicolons line number 59 (where Chorus opcode is called) and notice that there's no segfault and everything works.
>>>>>>>>>>>
>>>>>>>>>>> 4. Uncomment line 59 and allow it to be, but delete line 63 to 76, and notice that again everything works (and seemingly Chorus had nothing to do with the segfault).
>>>>>>>>>>>
>>>>>>>>>>> I will add that equivalent code works without segfault when running directly from csd file.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>> 
>>>>>>>>
>>>>>>
>>>>>
>>