Trying to narrow down the problem further, I ran sensekey.csd in a 
single-directory installation of Windows 5.10.1, floats (WinXP, SP3), which 
contained only the following files:
sensekey.csd
csound.exe
libsndfile-1.dll
stdopcod.dll
csound32.dll.5.2
portaudio.dll
pmidi.dll
portmidi.dll
porttime.dll
pthreadGC2.dll
Runs were from the command line, and all Csound environment variables were 
eliminated. All renders were terminated by Ctrl-C, which is where the 
Segmentation Error occurs. This error occurs at every run, regardless of 
MIDI or audio input/output flags (CsOptions).

My results were identical to Oeyvind's. In particular, whenever the -M flag 
is used - whether or not the .csd is processing MIDI, the run locks up with:
pm_winmm_term called
begin closing open devices...

In addition, unfortunately, we are back to the PortAudio problem of 5.08 
(fixed in 5.09). Whenever -odac is specified, we once again get:
 *** error: unknown rtaudio module: 'PortAudio'
Failed to initialise real time audio output
Csound tidy up: Segmentation violation
pm_winmm_term called
begin closing open devices...

When neither -odac nor -M are specified, test.wav is written (with no 
audio), and at Ctrl-C, the Segmentation Error occurs - but there is a proper 
return to the command line.

Art Hunkins

----- Original Message ----- 
From: "Oeyvind Brandtsegg" 
To: 
Sent: Thursday, February 26, 2009 6:28 PM
Subject: [Csnd] Re: Re: Re: Windows (float installer) 5.10.1 bugs? (was: 
Csound 5.06 vs. 5.09 and 5.10)


There are differences, but it hangs all the same.

**Using portmidi it hangs at:
Csound tidy up: Interrupt
Csound tidy up: Segmentation violation
pm_winmm_term called
begin closing open devices...

** Using mme (rtmidi: WinMM module enabled), it hangs at:
Csound tidy up: Interrupt
Csound tidy up: Segmentation violation
inactive allocs returned to freespace
end of score.              overall amps:  10000.0  10000.0
           overall samples out of range:        0        0
0 errors in performance
Elapsed time at end of performance: real: 8.828s, CPU: 8.829s

** no difference if using portaudio or mme for rtaudio
** no difference if using -d or --displays

best
Oeyvind

2009/2/27 Andres Cabrera :
> HI,
>
> Some module is having trouble with deallocation. Can you run under gdb
> (or try different -+rtaudio and -+rtmidi options or even -d to see if
> its FLTK?)
>
> Cheers,
> Andrés
>
> On Thu, Feb 26, 2009 at 6:09 PM, Oeyvind Brandtsegg  
> wrote:
>> I can confirm that the segmentation fault (when exiting Csound with
>> Ctrl-C) happens with Csound 5.10 doubles on my system.
>>
>> If I'm using only realtime out , I get a segfault, but it appears to
>> be able to clean up and exit.
>> Csound tidy up: Interrupt
>> Csound tidy up: Segmentation violation
>> inactive allocs returned to freespace
>> end of score. overall amps: 10000.0 10000.0
>> overall samples out of range: 0 0
>> 0 errors in performance
>> Elapsed time at end of performance: real: 6.480s, CPU: 6.469s
>> ... then the command prompt appears
>>
>> If I'm enabling a midi device (-M), even though I'm not using it for
>> anything, then Csound is not able to exit after the segfault.
>> Csound tidy up: Interrupt
>> Csound tidy up: Segmentation violation
>> pm_winmm_term called
>> begin closing open devices...
>> ... here it hangs until I force quit in the task manager.
>>
>> Test file attached.
>> If I let Csound run until the score is finished (11 seconds),
>> it exits normally if no midi device is enabled,
>> but it's still not able to exit if a midi device is enabled.
>> With midi enabled, and letting the score run until finished, it hangs at
>> 0 errors in performance
>> Elapsed time at end of performance: real: 11.242s, CPU: 11.250s
>> ... until force quit in task manager
>>
>> best
>> Oeyvind
>>
>>
>>
>> 2009/2/26 Art Hunkins :
>>> This message is primarily for Mike G., but any other insight much
>>> appreciated as well.
>>>
>>> I find that working with the Windows float installer 5.10.1 from the 
>>> command
>>> line gives segmentation faults *generally* - not just in the specific 
>>> case
>>> described in my previous mail (see below). The basic problem now appears
>>> completely unrelated to MIDI input or even to audio output.
>>>
>>>
>>> I did the following, which illustrates the problem clearly (WinXP):
>>>
>>> 1) Cleared all previous Csound-related environment variables (I know 
>>> them
>>> well by now).
>>>
>>> 2) Did default install of Windows 5.10.1 float version.
>>>
>>> 3) Selected to install (continue) without Python. (Given this, I was
>>> surprised to see that PYTHONPATH nonetheless was written as an 
>>> environment
>>> variable.) Selected to install variables for all users.
>>>
>>> 4) Took sensekey.com (from the install package), copying it to
>>> \csound5.10.1\bin.
>>>
>>> 5) Navigated, on the commandline to the above directory.
>>>
>>> 5) Ran csound sensekey.csd, pressing several ASCII keys. The complete
>>> console output is copied below.
>>>
>>> 6) During the run, the Python popup window appeared; after closing it, 
>>> the
>>> run continued normally. (I thought all our problems with Python install 
>>> and
>>> these popup windows were long since behind us - ugh.) I gather this is 
>>> an
>>> unrelated issue to the segmentation error however.
>>>
>>> 7) Otherwise, the run is fine until I press Ctrl-C. Then comes the
>>> segmentation error. Probably as a consequence(?), there is no clean 
>>> ending
>>> back to the command line (no ASCII input is accepted, though the cursor 
>>> is
>>> blinking [which was not the case in my previous example]).
>>>
>>> 8) The relevence (significance) of both WARNING messages completely 
>>> escapes
>>> me. I'd guess that both indicate minor bugs of some sort. (Please take a
>>> close look. Perhaps they do relate somehow to the segmentation error.) 
>>> The
>>> WARNING about csnd.dll not being found is indeed perplexing because the
>>> (identical) csnd.dll is indeed located both in \bin and \plugins.
>>>
>>> It is important to me that any (float) version of (Windows) Csound I 
>>> use,
>>> run fully from the command line. Currently, 5.10.1 doesn't. 5.06 and 
>>> 5.07
>>> do. (5.08 had the PortAudio problem, and 5.09 as well as 5.10.1 have 
>>> this
>>> segmentation error issue. I feel somewhat stuck in the dark ages.
>>>
>>> I'd appreciate any insight, especially from you, Mike.
>>>
>>> Appreciatively,
>>>
>>> Art Hunkins
>>>
>>> -------------------------
>>> Microsoft Windows XP [Version 5.1.2600]
>>> (C) Copyright 1985-2001 Microsoft Corp.
>>>
>>> C:\Documents and Settings\Art Hunkins>cd \csound5.10.1\bin
>>>
>>> C:\Csound5.10.1\bin>csound sensekey.csd
>>> time resolution is 279.365 ns
>>> PortAudio real-time audio module for Csound
>>> PortMIDI real time MIDI plugin for Csound
>>> virtual_keyboard real time MIDI plugin for Csound
>>> 0dBFS level = 32768.0
>>> Csound version 5.10 beta (float samples) Jan 8 2009
>>> libsndfile-1.0.18pre24
>>> Reading options from $CSOUNDRC: C:\Csound5.10.1\.csoundrc
>>> WARNING: could not open library 'C:\Csound5.10.1\plugins\csnd.dll' (-1)
>>> UnifiedCSD: sensekey.csd
>>> STARTING FILE
>>> Creating options
>>> Creating orchestra
>>> Creating score
>>> orchname: C:\DOCUME~1\ARTHUN~1\LOCALS~1\Temp\cs90.orc
>>> scorename: C:\DOCUME~1\ARTHUN~1\LOCALS~1\Temp\cs91.sco
>>> rtmidi: PortMIDI module enabled
>>> rtaudio: PortAudio module enabled ... using callback interface
>>> orch compiler:
>>> 14 lines read
>>> instr 1
>>> Elapsed time at end of orchestra compile: real: 0.142s, CPU: 0.140s
>>> sorting score ...
>>> ... done
>>> Elapsed time at end of score sort: real: 0.149s, CPU: 0.140s
>>> Csound version 5.10 beta (float samples) Jan 8 2009
>>> displays suppressed
>>> 0dBFS level = 32768.0
>>> orch now loaded
>>> audio buffered in 128 sample-frame blocks
>>> reading 256-byte blks of shorts from adc (RAW)
>>> PortAudio V19-devel (built Jan 6 2009 21:14:14)
>>> PortAudio: available input devices:
>>> 0: Microsoft Sound Mapper - Input
>>> 1: Avance Sound
>>> 2: VIA AC'97 Audio (WAVE)
>>> 3: Bluetooth AV Audio
>>> 4: Bluetooth SCO Audio
>>> 5: Primary Sound Capture Driver
>>> 6: Avance Sound
>>> 7: VIA AC'97 Audio (WAVE)
>>> 8: Bluetooth AV Audio
>>> 9: Bluetooth SCO Audio
>>> PortAudio: selected input device 'Avance Sound'
>>> PortAudio: available output devices:
>>> 0: Microsoft Sound Mapper - Output
>>> 1: Avance Sound
>>> 2: VIA AC'97 Audio (WAVE)
>>> 3: Bluetooth AV Audio
>>> 4: Bluetooth SCO Audio
>>> 5: Primary Sound Driver
>>> 6: Avance Sound
>>> 7: VIA AC'97 Audio (WAVE)
>>> 8: Bluetooth AV Audio
>>> 9: Bluetooth SCO Audio
>>> 10: GIGAPort AG ASIO driver
>>> PortAudio: selected output device 'Avance Sound'
>>> WARNING: buffer size should be an integer multiple of ksmps in 
>>> full-duplex
>>> mode
>>> writing 256-byte blks of shorts to dac
>>> SECTION 1:
>>> new alloc for instr 1:
>>> i1 -1.00000
>>> i1 49.00000
>>> i1 -1.00000
>>> i1 50.00000
>>> i1 -1.00000
>>> Csound tidy up: Interrupt
>>> Csound tidy up: Segmentation violation
>>> inactive allocs returned to freespace
>>> end of score. overall amps: 0.0
>>> overall samples out of range: 0
>>> 0 errors in performance
>>> Elapsed time at end of performance: real: 8.467s, CPU: 8.468s
>>>
>>>
>>> ----- Original Message ----- From: "Art Hunkins" 
>>> To: 
>>> Sent: Tuesday, February 24, 2009 7:44 PM
>>> Subject: Re: Csound 5.06 vs. 5.09 and 5.10
>>>
>>>
>>>> Followup to previous post:
>>>>
>>>> 5.07 works correctly, just like 5.06.
>>>>
>>>> 5.08 doesn't run at all (neither does the Lettuce standalone) due to 
>>>> the
>>>> PortAudio issue in that version, which was fixed by Mike G. in 5.09.
>>>> (Perhaps his fix is related to my issue, though that seems unlikely.)
>>>>
>>>> Still mystified and seeking illumination -
>>>>
>>>> Art Hunkins
>>>>
>>>> ----- Original Message ----- From: "Art Hunkins" 
>>>> To: 
>>>> Sent: Tuesday, February 24, 2009 7:18 PM
>>>> Subject: Csound 5.06 vs. 5.09 and 5.10
>>>>
>>>>
>>>>> This question concerns WinXP and realtime, MIDI-controlled performance
>>>>> .csd's.
>>>>>
>>>>> I've three separate folders, one each containing the following files 
>>>>> from
>>>>> 5.06, 5.09 and 5.10:
>>>>>
>>>>> csound32.dll.5.1 (csound32.dll.5.2 in the case of 5.10)
>>>>> stdopcod.dll
>>>>> csound.exe
>>>>> libsndfile-1.dll
>>>>> pmidi.dll
>>>>> portaudio.dll
>>>>> portmidi.dll
>>>>> porttime.dll
>>>>> pthreadGC2.dll
>>>>> rtpa.dll
>>>>> rtwinmm.dll
>>>>>
>>>>> These are the only Csound files required to render my compositions. 
>>>>> *No
>>>>> environment variables are set* and .csd's are run from within these 
>>>>> folders.
>>>>> (Please note that the way I've set up things, multiple versions of 
>>>>> Csound
>>>>> can easily be run from unique directories on the same computer - with 
>>>>> no
>>>>> environment variables set. Just put everything for one version in a 
>>>>> single
>>>>> directory. You may need no more files than the above dozen or so, plus 
>>>>> your
>>>>> .csd's.
>>>>>
>>>>> Problem: Whereas standalone Lettuce executables run fine in each of 
>>>>> the
>>>>> three folders, only 5.06 runs *from the command line* correctly. 5.09 
>>>>> and
>>>>> 5.10 issue the following error when Ctrl-C is pressed:
>>>>> Csound tidy up: Interrupt
>>>>> Csound tidy up: Segmentation violation (as always - *so* helpful!)
>>>>> pm_winmm_term called
>>>>> begin closing open devices...
>>>>>
>>>>> and the system hangs.
>>>>>
>>>>> Note there is no .csoundrc file, and that I'm defaulting to PortMIDI. 
>>>>> My
>>>>> CsOptions include -M11 (my MIDI device). All installations use files 
>>>>> from
>>>>> the float Windows installers.
>>>>>
>>>>> Everything worked so well in 5.06.
>>>>>
>>>>> Why should 5.09 and 5.10 fail (at the command line only)?
>>>>>
>>>>> Art Hunkins
>>>>
>>>
>>>
>>>
>>> Send bugs reports to this list.
>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>>> csound"
>>>
>>
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
>> csound"
>
>
>
> --
>
>
> Andrés
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
> csound"


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
csound"= 

Art, 

Are you seeing this message at the beginning of every run?

  PortAudio real-time audio module for Csound

I suspect that you are not seeing it because you do not have the file
rtpa.dll in your directory.  That is the PortAudio plugin for Csound and if
it is missing, it would explain this message:

  *** error: unknown rtaudio module: 'PortAudio'

It is also possible that the rtpa.dll will require other files to be present
that don't seem logical.  I don't know about Windows, but on OS X , the
PortAudio plugin is also linked to the FLTK, liblo, libpng, and zlib
libraries and will not load unless they are present (even though they are
technically not required).

Thus, it is more difficult to help track down the real bugs here if you are
not running the standard and complete distribution of Csound.

Anthony Kozar
mailing-lists-1001 AT anthonykozar DOT net
http://anthonykozar.net/

Art Hunkins wrote on 2/27/09 3:45 PM:

> Trying to narrow down the problem further, I ran sensekey.csd in a
> single-directory installation of Windows 5.10.1, floats (WinXP, SP3), which
> contained only the following files:
> sensekey.csd
> csound.exe
> libsndfile-1.dll
> stdopcod.dll
> csound32.dll.5.2
> portaudio.dll
> pmidi.dll
> portmidi.dll
> porttime.dll
> pthreadGC2.dll
> Runs were from the command line, and all Csound environment variables were
> eliminated. All renders were terminated by Ctrl-C, which is where the
> Segmentation Error occurs. This error occurs at every run, regardless of
> MIDI or audio input/output flags (CsOptions).
> [...]
> In addition, unfortunately, we are back to the PortAudio problem of 5.08
> (fixed in 5.09). Whenever -odac is specified, we once again get:
> *** error: unknown rtaudio module: 'PortAudio'
> Failed to initialise real time audio output
> Csound tidy up: Segmentation violation
> pm_winmm_term called
> begin closing open devices...