[Issue migrated from SourceForge | ID: 1837775 | Submitted by 'sciss']
[http://sourceforge.net/support/tracker.php?aid=1837775]

fillRect and fillOval (and maybe others?) are using the strokeColor, not the fillColor:

GUI.cocoa;
(
var width= 640, height= 480, w, drawFunc;
w= GUI.window.new("trail test3", Rect(128, 64, width, height), false);
drawFunc= {|v|
GUI.pen.fillColor_(Color.blue);
GUI.pen.strokeColor_(Color.green);
// these should be blue, but in fact they are green
GUI.pen.fillRect(Rect.aboutPoint(Point(100,100), 60, 40));
GUI.pen.fillOval(Rect.aboutPoint(Point(300,100), 60, 40));
};
x = GUI.userView.new(w, Rect(10, 10, width-20,
height-20)).drawFunc_(drawFunc);
w.front;
)

[Issue migrated from SourceForge | ID: 572166 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=572166]

bug in the drag primitive

• dragging from range changes the current setting in an
  SCRangeSlider

[Issue migrated from SourceForge | ID: 1869861 | Submitted by 'timblech']
[http://sourceforge.net/support/tracker.php?aid=1869861]

string constants should be handled as const char* instead of char*, as the syntax is deprecated by gcc-4.2.

the attached patch fixes this issue ...

[Issue migrated from SourceForge | ID: 1887774 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=1887774]

SuperCollider currently does not receive successfully SysEx messages.
It's hard to give an example since it is dependant of the hardware you use. And using IACDriver for this purpose is not a good idea since it would feed itself again and again for large messages.

Most trials i have done were with a Waldorf MicroQ, from wich i was sending full program dumps of 392bytes each. the MIDIIn.doSysexAction was never triggered and occasionally it crashed.

I made 3 propositions after working on the code (in SC_CoreMidi.cpp file).
After those modifications i was able to receive successfully SysEx messages.
You can find those modifications/suggestions in the dev-list :

http://lists.create.ucsb.edu/pipermail/sc-dev/2008-February/011903.html

Anyway more tests with different machines should be better.

charles a picasso
[email protected]

[Issue migrated from SourceForge | ID: 1870544 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1870544]

If you set an environment variable in Windows

SC_SYNTHDEF_DIR = D:\SuperCollider\PsyCollider\synthdefs

scsynth.exe fails to read with this error:

localhost
booting 57110
Using vector unit: no
*** ERROR: open directory failed 'D'

[Issue migrated from SourceForge | ID: 1873963 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1873963]

o = OSCresponderNode(s.addr, '/b_info', { |t, r, m|
m.postln
}).add;

b = b.read("sounds/a11wlk01.wav");
--> [ /b_info, 18, 0, 0, 0 ]

... where it should print [ /b_info, 18, 188893, 1, 44100 ]

b = b.read("dummkopf.aif");
--> [ /b_info, 18, 0, 0, 0 ]

... where the server should print an error that the file could not be opened.

It isn't clear to me from these tests whether the buffer is actually populated or not.

[Issue migrated from SourceForge | ID: 1897486 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1897486]

[0, 1, 2, 3, 0, 1, 2, 3].removeAll([1, 2])
--> [ 0, 3, 0, 1, 2, 3 ]

Only the first occurrence of each given item is removed.

[Issue migrated from SourceForge | ID: 1813039 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1813039]

OSX incorrectly saves .quark files as RTF.

To reproduce: Write any text into an SC code window. Then save using the quark format from the dropdown.

In terminal, cat the new file and you will see something like:

{\rtf1\mac\ansicpg10000\cocoartf824\cocoasubrtf420
{\fonttbl\f0\fswiss\fcharset77 Helvetica;}
{\colortbl;\red255\green255\blue255;}
\pard\tx560\tx1120\tx1680\tx2240\tx2800\tx3360\tx3920\tx4480\tx5040\tx5600\tx6160\tx6720\ql\qnatural\pardirnatural

\f0\fs24 \cf0
(test: "test quark")}

[Issue migrated from SourceForge | ID: 1834737 | Submitted by 'sciss']
[http://sourceforge.net/support/tracker.php?aid=1834737]

an immediate n_free, g_freeAll or g_deepFree crashes scsynth if the node was paused. tested with latest SVN build (6650) on macbook pro intel, mac os x 10.4.10. to reproduce:

s.boot;

// now either of these:

(
s.sendBundle( nil,
[ "/g_new", 1007, 0, 0 ],
[ "/s_new", "default", 1011, 1, 1007 ],
[ "/n_run", 1011, 0 ]// REQUIRED TO PRODUCE THE CRASH
);
s.sendMsg( "/g_deepFree", 1007 );// BOOOM!
)

(
s.sendBundle( nil,
[ "/g_new", 1007, 0, 0 ],
[ "/s_new", "default", 1011, 1, 1007 ],
[ "/n_run", 1011, 0 ]// REQUIRED TO PRODUCE THE CRASH
);
s.sendMsg( "/g_freeAll", 1007 );// BOOM!
)

(
s.sendBundle( nil,
[ "/g_new", 1007, 0, 0 ],
[ "/s_new", "default", 1011, 1, 1007 ],
[ "/n_run", 1011, 0 ]// REQUIRED TO PRODUCE THE CRASH
);
s.sendMsg( "/n_free", 1011 );
)

[Issue migrated from SourceForge | ID: 1121530 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=1121530]

in the latest build, the exponentiation operator ** is wasted! It
returns negative results when squaring negatives!
i.e. neg(5)**2 returns -25....

wasted all my programs

[email protected]

[Issue migrated from SourceForge | ID: 1865867 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1865867]

This seems to have been fixed in Median but it's still broken in Pitch.

e.g.:

SynthDef(\kaboom, { |outbus|
varsig = In.ar(outbus, 1),
pitch, hasFreq, trig, mnote;

pitch, hasFreq = Pitch.kr(sig, median: 11);

mnote = pitch.cpsmidi.round;
// send a trigger on change of freq or hasFreq on/off
trig = (HPZ1.kr(mnote).abs > 0) + (HPZ1.kr(hasFreq).abs > 0);
SendTrig.kr(trig, hasFreq, pitch);
}).memStore;

[Issue migrated from SourceForge | ID: 1896121 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1896121]

At some point in the past, string literals were immutable objects, but that changed somewhere along the way.

That produces the following deadly side effect:

f = {
varstringLiteral = "xyz";
stringLiteral.debug("stringLiteral before extension");
stringLiteral.extend(10, $.);
stringLiteral.debug("stringLiteral after extension");
};

f.def.prototypeFrame;
[ xyz ]

f.value;
stringLiteral before extension: xyz
stringLiteral after extension: xyz.......

f.def.prototypeFrame;
[ xyz....... ]

f.value;
stringLiteral before extension: xyz.......
stringLiteral after extension: xyz.......

An operation within a function has changed the function definition.

[Issue migrated from SourceForge | ID: 1324323 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=1324323]

i downloladed the latest full version (20050916), and when i try to
boot the localhost server it prints "result=5" in the comments
windows, and 30 seconds later it says "server failed to boot".

when i try to boot the internal server the application unexpectedly
quits

for the localhost server i tried both "s = Server.local; s.boot" as well
as clicking on the boot button on the localhost gui window. for the
internal server i only tried clickiong on the boot button on the internal
server gui window.

thanks

bob bellerue
[email protected]

[Issue migrated from SourceForge | ID: 1083902 | Submitted by 'bwillkie']
[http://sourceforge.net/support/tracker.php?aid=1083902]

Hello,

Just installed SuperCollider3_bin_20041125.tar.gz on 4 flat pannel
imacs running OS 10.3.6. I had no problems running as an admin
user but under any other user I got an error on start up:

ERROR: Primitive '_FilePutString' failed

under the SuperCollider_f/synthdef directory executing:

%sudo chmod 666 ./*

fixed the problem.

-Brian

[Issue migrated from SourceForge | ID: 1896281 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1896281]

I see this at the beginning of my library compilation transcript. (This is with dewdrop_lib installed.)

compiling dir: 'C:\Program Files\PsyCollider\SCClassLibrary'
in file '\Extensions\dewdrop_lib\ddwPeakMonitor\SequenceableColl-asBus.sc' line 8 char 53

Curious, it tells me something is in there but doesn't tell me what.

Is it a warning that was removed, but some remnant remains?

Not urgent for 3.2. Do not delay release of 3.2 for this.

[Issue migrated from SourceForge | ID: 1897484 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1897484]

Helpfile is wrong on takeAt.

takeAt(index)

Same as removeAt, but reverses the order of the items following those that which was taken.

y = [ 1, 2, 3, 4 ];
y.takeAt(1);
y.postln;

"reverses the order" is not correct.

[Issue migrated from SourceForge | ID: 1797829 | Submitted by 'dpunktnpunkt']
[http://sourceforge.net/support/tracker.php?aid=1797829]

When building SuperCollider, I tried to use a PREFIX that's in my homedirectory, like so:

PREFIX = '/home/daniel/lib/supercollider'

Turns out that the binaries that are copied to $PREFIX/bin aren't linked in a way to find the libraries in $PREFIX/lib. The attached patch to SConstruct has the binaries look in the right place when they're run.

[Issue migrated from SourceForge | ID: 1874838 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1874838]

If SuperCollider is not the frontmost application when Document.new is called, then the return value from the call is nil.

To reproduce:

d = Document.new("oops"); d.postln;
--> a CocoaDocument

// after running this line, cmd-tab to another application
{ d = Document.new("oops"); d.postln }.defer(5.0);

// "a Function" prints immediately, then after 5 seconds:
--> nil

[Issue migrated from SourceForge | ID: 1866474 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1866474]

This is fine:

a = Synth(\default);

a.set(\gate, 0);

s.quit;

This is not:

(
s.makeBundle(1.0, {
a = Synth(\default);
});
)

a.set(\gate, 0);

s.quit;
---> RESULT = -1073741819

And a popup "scsynth.exe has generated errors."

Not sure if that's a huge deal since it is on shutdown anyway, but it suggests something is being left behind in some data structure somewhere.

[Issue migrated from SourceForge | ID: 1821651 | Submitted by 'redfrik']
[http://sourceforge.net/support/tracker.php?aid=1821651]

if i disable editable before setting the string, nothing gets show. if i disable after then it's ok.
swingosc works fine and there the order doesn't matter.

GUI.cocoa
(
w= GUI.window.new;
GUI.textView.new(w, Rect(10, 10, 100, 20))
.editable_(false)
.string_("asdf");//not displayed
w.front;
)
(
w= GUI.window.new;
GUI.textView.new(w, Rect(10, 10, 100, 20))
.string_("asdf")
.editable_(false);
w.front;
)

_f

[Issue migrated from SourceForge | ID: 572167 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=572167]

SCSliderBase, knobColor bug

the knobColor property can't be set

[Issue migrated from SourceForge | ID: 1870000 | Submitted by 'sciss']
[http://sourceforge.net/support/tracker.php?aid=1870000]

Pen.strokeOval ignores the Pen.width settings
(SVN rev. 6944)

[Issue migrated from SourceForge | ID: 1873965 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1873965]

In Windows:

f = SoundFile.openRead("sounds/a11wlk01.wav");
f.dump;

---> prints:

Instance of SoundFile { (04260E10, gc=B4, fmt=00, flg=00, set=03)
instance variables [7]
fileptr : RawPointer BC60048
headerFormat : " "
sampleFormat : "int16"
numFrames : Integer 188893
numChannels : Integer 44100
sampleRate : Integer 0
path : "sounds/a11wlk01.wav"
}

but it should print:

Instance of SoundFile { (04260E10, gc=B4, fmt=00, flg=00, set=03)
instance variables [7]
fileptr : RawPointer BC60048
headerFormat : "aiff"
sampleFormat : "int16"
numFrames : Integer 188893
numChannels : Integer 1
sampleRate : Integer 44100
path : "sounds/a11wlk01.wav"
}

headerFormat, numChannels and sampleRate are not populated correctly.

[Issue migrated from SourceForge | ID: 1876621 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1876621]

w = SCWindow.new;

c = SCCompositeView(w,Rect(50,0,300,300)).relativeOrigin_(false);

a = SC2DSlider(c,Rect(50,0,100,100));
b = SC2DSlider(c,Rect(150,100,100,100));

c.background = Gradient(Color.rand,Color.rand);

w.front;

TO DO:

Regression-test ALL SCViews (ugh) -- I'll do what I can, don't have much time during the workweek, though

[Issue migrated from SourceForge | ID: 786733 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=786733]

building on linux gives me an error

g++ -DHAVE_CONFIG_H -I. -I. -I../..
-DSC_AUDIO_API=SC_AUDIO_API_JACK
-DSC_PLUGIN_DIR="/usr/local/lib/SuperCollider/plugins"
-DSC_PLUGIN_EXT=".so" -DSC_PLUGIN_LOAD_SYM="load"
-I../../headers/common -I../../headers/plugin_interface
-I../../headers/server -Wall -DSC_LINUX -D_REENTRANT
-DNDEBUG -Wp,-MD,.deps/SC_CoreAudio.pp -c
SC_CoreAudio.cpp -fPIC -DPIC -o .libs/SC_CoreAudio.lo
SC_CoreAudio.cpp: In member function bool SC_CoreAudioDriver::DriverStart()': SC_CoreAudio.cpp:1408: error:assert' undeclared
(first use this function)
SC_CoreAudio.cpp:1408: error: (Each undeclared
identifier is reported only once
for each function it appears in.)
make[3]: *** [SC_CoreAudio.lo] Error 1
make[3]: Leaving directory
/opt/cvs/SuperCollider3/source/server' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory/opt/cvs/SuperCollider3/source'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/opt/cvs/SuperCollider3'
make: *** [all-recursive-am] Error 2

[Issue migrated from SourceForge | ID: 1870525 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1870525]

If I put a shortcut to a folder into the userExtensionsDir in Windows, .sc files referenced by the shortcut are not compiled.

Among other problems, if there is ever an effort to make Quarks windows-compatible, this shortcoming would prevent Quarks from working altogether because installing a quark would create a shortcut in Extensions to the real quark directory.

[Issue migrated from SourceForge | ID: 572164 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=572164]

typo in SCRangeSlider.sc

SCRangeSlider
receiveDrag {
// change to .x, .y instead of lo, hi
this.lo = currentDrag.x;
this.hi = currentDrag.y;
currentDrag = nil;
}

[Issue migrated from SourceForge | ID: 1853824 | Submitted by 'timblech']
[http://sourceforge.net/support/tracker.php?aid=1853824]

the documentation of Server.bootSync says:

condition is an optional instance of Condition used for evaluating this.

in the implementation, the condition argument is not optional, though, but mandatory ... the attached patch solves that issue ...

[Issue migrated from SourceForge | ID: 1897485 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1897485]

a = [0, 1, 2, 3];
a.clipPut(6, 10)
a

'a' is unchanged - clipPut failed silently with no error but no effect either.

[Issue migrated from SourceForge | ID: 1804647 | Submitted by 'danstowell']
[http://sourceforge.net/support/tracker.php?aid=1804647]

Various users have reported that occasional audio dropouts can occur when starting a synth (brief, e.g. 10 ms). This has only arisen fairly recently so may be due to a change in SC or to a change in OSX (e.g. its memory management). It seems to be triggered by using UGens that have not been used so far in the SC session, meaning the plugin code is not yet paged-in to memory.

Main discussion thread:
http://www.create.ucsb.edu/pipermail/sc-users/2007-August/035837.html

Dan found that he could cause this to recur if he closed SC, ran memory-hungry apps, then restarted SC. (Didn't happen without running the apps.) (PPC PowerBook, OSX 10.4.10, SC updated from SVN on 2007-09-15.) Seems to confirm it's a paging issue. James H found it happening on his MacBook Pro, 1.83 GHz, 1.5 GB RAM. So it seems to happen on both Intel & PPC. (See email lists for other reports)

During the initial thread of discussion, James McC suggested "a ugen to call upon loading, a function to read the first word of each of its _next function addresses". Florian suggested mlocking scsynth's memory. The issue was discussed at SC Symposium 2007 but not sure if there was a definitive solution.

[Issue migrated from SourceForge | ID: 1804666 | Submitted by 'danstowell']
[http://sourceforge.net/support/tracker.php?aid=1804666]

.....!.......

[Issue migrated from SourceForge | ID: 1843401 | Submitted by 'timblech']
[http://sourceforge.net/support/tracker.php?aid=1843401]

the following invalid code crashes sclang:

v = {};
v();

[Issue migrated from SourceForge | ID: 738882 | Submitted by 'clements']
[http://sourceforge.net/support/tracker.php?aid=738882]

The command-line build of SuperCollider3 that I get from running
pbxbuild at the command line fails on UnaryOpGen.cpp because
vfloat32 is undefined. Detective work shows that this typedef
is conditional in SC_Types.h on the #define'dness of VEC,
which is only enabled when -faltivec is a command-line option
to the compiler.

I'm much more familiar with 'make' than I am with 'pbxbuild', but
I noticed the following code in the build file:

            F50D04F902E68E8201CA2799 = {
                    buildRules = (
                    );
                    buildSettings = {
                            COPY_PHASE_STRIP = NO;
                            OPTIMIZATION_CFLAGS = "-O0 -faltivec";
                    };
                    isa = PBXBuildStyle;
                    name = Development;
            };

This suggests to me that "-faltivec" is set only when optimization
is enabled. Since the code won't compile without it, I suggest that
this be moved from optimization cflags to all-the-time cflags.

Finally, I'd be grateful to know how to enable the optimization
at the command line. Something like pbxbuild OPT=yes ...?

[Issue migrated from SourceForge | ID: 1807381 | Submitted by 'timblech']
[http://sourceforge.net/support/tracker.php?aid=1807381]

i am running into memory issues, when running a more or less trivial supercollider script, which can be found at http://klingt.org/~tim/midi2osc.sc and which is basically a very simple midi to osc converter.

i am running the script in the command line:
sclang midi2osc.sc

but it keeps allocating memory, so after a some time, it requires several hundred mb in the ram. however, when running it the same script from within emacs, it works fine, without continuously allocating memory.

memory debugging results:
i had a quick look at it with the totalview memory debugger ... it basically looks fine, the only reported memory leaks seem to be calls to pthread and alsa api functions, which is basically fine ...

examining the heap, it seems, that the memory is allocated from the following piece of code:

[quoted text muted]
Supercollider3/Source/lang/LangSource/InitAlloc.cpp#40]
1 AllocPool::NewArea PC=0xb7e476c3, FP=0xbf9e7cf8 [/home/tim/workspace/
Supercollider3/Source/common/SC_AllocPool.cpp#202]
2 AllocPool::Alloc PC=0xb7e47bb1, FP=0xbf9e7d68 [/home/tim/workspace/
Supercollider3/Source/common/SC_AllocPool.cpp#399]
3 AdvancingAllocPool::AddChunk PC=0xb7dc0a7e, FP=0xbf9e7d98 [/home/tim/
workspace/Supercollider3/Source/lang/LangSource/AdvancingAllocPool.cpp#56]
4 AdvancingAllocPool::Alloc PC=0xb7dc0b34, FP=0xbf9e7dc8 [/home/tim/
workspace/Supercollider3/Source/lang/LangSource/AdvancingAllocPool.cpp#71]
5 PyrGC::NewFrame PC=0xb7dc6208, FP=0xbf9e7e38 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/GC.cpp#427]
6 executeMethod PC=0xb7ddb475, FP=0xbf9e7e98 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/PyrMessage.cpp#1059]
7 Interpret PC=0xb7dcd092, FP=0xbf9e7f68 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/PyrInterpreter3.cpp#2056]
8 runInterpreter PC=0xb7dcdf8a, FP=0xbf9e7f88 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/PyrInterpreter3.cpp#120]
9 runLibrary PC=0xb7dcef5c, FP=0xbf9e7fc8 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/PyrLexer.cpp#2230]
10 SC_LanguageClient::tick PC=0xb7dfa3eb, FP=0xbf9e7fd8 [/home/tim/
workspace/Supercollider3/Source/lang/LangSource/SC_LanguageClient.cpp#287]
11 SC_TerminalClient::daemonLoop PC=0xb7dfc076, FP=0xbf9e8008 [/home/
tim/workspace/Supercollider3/Source/lang/LangSource/
SC_TerminalClient.cpp#329]
12 SC_TerminalClient::run PC=0xb7dfc8a6, FP=0xbf9e8038 [/home/tim/
workspace/Supercollider3/Source/lang/LangSource/SC_TerminalClient.cpp#233]
13 main PC=0x0804875e, FP=0xbf9e80a8 [/home/tim/workspace/
Supercollider3/Source/lang/LangSource/cmdLineFuncs.cpp#36]
14 __libc_start_main PC=0xb7b4404d, FP=0xbf9e8118 [/lib/i686/cmov/
libc.so.6]

the memory chunks are still referred to, though ...

it might be of interest, that pyr_new_area allocates memory of about 6MB from in a call of SC_LanguageClient::compileLibrary(), and an aditional
about 2MB from SC_LanguageClient::initRuntime() ... all other memory chunks are allocated in the backtrace above, having a size of about 256KB ...

[Issue migrated from SourceForge | ID: 1856972 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=1856972]

When fasterBetter is true (default), 0.2 hangs the primitive for SimpleNumber:asFraction. Somehow this particular number (1/5) never gets out of the Farey approximation loop.

My solution up to now has been to use fasterBetter = false

J.S.Lach:
[email protected]

[Issue migrated from SourceForge | ID: 928864 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=928864]

Running OS 10.2.8 on a 17" G4 iMac.

Renaming the folder 'SuperCollider_f' to 'SuperCollider
[option-f = italicized f]' prevented the introspection
menu items ('Open Class Def', 'Implementations of',
'References to') from being able to open the *.sc files.

IMHO a low priority bug, but a ReadMe file warning may
be appropriate.

[email protected]

[Issue migrated from SourceForge | ID: 1816592 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1816592]

Self-explanatory - an Instr's path variable should have the path to the file containing the Instr definition. For the last several months, the path is always nil.

Assigning to myself for an initial look. If I can't fix it, it will go to felix.

[Issue migrated from SourceForge | ID: 1875947 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1875947]

As reported by Alberto on the mailing list, messages sent to the server without a timestamp (nil latency) are quantized to the next hardware buffer block boundary -- where I think most of us typically expect things to happen on control block boundaries. Using the default settings, 64 sample quantization is significantly better than 512 samples.

Not a regression issue -- the same issue is observable in two-year-old builds.

Not critical for 3.2, since a workaround is to reduce the hardware buffer size. Just logging it here so the issue doesn't get forgotten.

Server.default = s = Server.internal;
s.boot;
s.scope;

SynthDef(\tick, { OffsetOut.ar(0, XLine.ar(0.2, 0.01, 0.0003,
doneAction: 2)) }).memStore;

Tdef(\tic, { loop { s.sendBundle(s.latency, ["s_new", "tick", -1, 0,
0]); 0.005.wait; } }).play;

s.latency = 0.2;// ok, 200 Hz
s.latency = nil;// quantized to hardwarebuffersize! (512 on my machine)

Tdef(\tic).stop;

[Issue migrated from SourceForge | ID: 1865286 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1865286]

[].lace // kaboom

Probably some other primitives break also.

High priority because it's easy to do this innocently.

[Issue migrated from SourceForge | ID: 731909 | Submitted by 'cruxxial']
[http://sourceforge.net/support/tracker.php?aid=731909]

if the SCWindow was initially created with a smaller size and then
resized by the user, views placed outside of that range disappear
when clicked on.

(mainly testing the bug tracker here)
-felix

[Issue migrated from SourceForge | ID: 1886415 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1886415]

SynthDef(\sin, { |freq = 440, amp = 0.2, out = 0|
Out.ar(out, (SinOsc.ar(freq, 0, amp) * EnvGen.kr(Env.perc(0.01, 0.1),
doneAction: 2)) ! 2);
}).memStore;

// sound
a = Synth(\sin);

// no sound
s.makeBundle(0.2, { a = Synth(\sin); });

// sound
s.latency = nil;
p = Pbind(\freq, Pexprand(200, 800, inf), \amp, Pwhite(0.1, 0.3, inf),
\delta, 0.125, \instrument, \sin).play;

p.stop;

// no sound
s.latency = 0.1;
p = Pbind(\freq, Pexprand(200, 800, inf), \amp, Pwhite(0.1, 0.3, inf),
\delta, 0.125, \instrument, \sin).play;

p.stop;

[Issue migrated from SourceForge | ID: 1870523 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1870523]

Was just trying to play with psycollider on a laptop with a weak wifi connection. It would finish compiling and then freeze - uninstall and reinstall didn't fix it.

Turning off the wifi connection instantly makes it respond.

So:

• strong (wired) wifi - works
• no internet - works
• weak internet - hangs

[Issue migrated from SourceForge | ID: 1204875 | Submitted by 'danstowell']
[http://sourceforge.net/support/tracker.php?aid=1204875]

The LPF UGen fails if its cutoff is set to precisely
zero. This is because the line

float C = 1.f / tan(pfreq);

results in a NaN when fed a value of zero, and it looks
like the NaN value pollutes the whole server's output
(since the output seems to "max out" and nothing can be
heard until forcibly stopping the output with [Cmd+.]).

Recommend checking for a zero value, and perhaps
adjusting pfreq to a very very small but nonzero value?
Or perhaps checking whether C is NaN just after the
line I've quoted.

[Issue migrated from SourceForge | ID: 1804665 | Submitted by 'danstowell']
[http://sourceforge.net/support/tracker.php?aid=1804665]

Please ignore (test item)

[Issue migrated from SourceForge | ID: 1880224 | Submitted by 'timblech']
[http://sourceforge.net/support/tracker.php?aid=1880224]

in the current svn checkout, running the GrainBuf ugen causes scsynth to segfault.

it can be reproduced using the help file of GrainBuf.

[Issue migrated from SourceForge | ID: 918069 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=918069]

building from cvs on debian linux (unstable) for ppc I
get :

g++ -DHAVE_CONFIG_H -I. -I. -I../..
-I../../headers/common -I../../headers/plugin_interface
-I../../headers/server -Wall -DSC_LINUX -D_REENTRANT
-DNDEBUG -fsigned-char -maltivec -g -O2
-Wp,-MD,.deps/MacUGens.pp -c MacUGens.cpp -fPIC -DPIC
-o .libs/MacUGens.o
MacUGens.cpp: In function void cmdDemoFunc(World*, void*, sc_msg_iter*, void*) ': MacUGens.cpp:406: error: invalid conversion fromint
()(World, void_)' to bool (_)(World_, void_)' MacUGens.cpp:406: error: invalid conversion fromint
()(World, void*)' tobool (_)(World_, void_)' MacUGens.cpp:406: error: invalid conversion fromint
()(World, void_)' to bool (*)(World*, void*)' make[3]: *** [MacUGens.lo] Error 1 make[3]: Leaving directory/home/marco/src/SuperCollider3/source/plugins'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory
/home/marco/src/SuperCollider3/source' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory/home/marco/src/SuperCollider3'
make: *** [all-recursive-am] Error 2

[Issue migrated from SourceForge | ID: 1259873 | Submitted by 'yeeking']
[http://sourceforge.net/support/tracker.php?aid=1259873]

Running late July 2005 CVS supercollider under Debian
GNU/Linux i686

When running the example code found in
Non-Realtime-Synthesis.rtf, i.e.:

c = [ 0.2, [\s_new, \NRT-sine, 1001, 0, 0]].asRawOSC;

I get the error:

ERROR: Variable 'sine' not defined.
in file 'selected text'
line 1 char 29 :

c = [ 0.2, [\s_new, \NRT-sine¥, 1001, 0, 0]].asRawOSC;

If i edit the code to:

c = [ 0.2, [\s_new, "NRT-sine", 1001, 0, 0]].asRawOSC;

It executes fine.

[Issue migrated from SourceForge | ID: 1637723 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=1637723]

I just installed the latest binary build available from Weslyan. While reading through 'First Steps.rtf', I've come to the part of the page where it describes how to bring the post window to top using a key chord.

After correctly describing the key chord as the apple key plus a '' character, the next sentence incorrectly says that this is often referred to as "Cmd-/" when it should say "Cmd-".

Yes, I know, this is incredibly minor, but that means it's also incredibly easy to fix. :-)

[Issue migrated from SourceForge | ID: 572165 | Submitted by 'nobody']
[http://sourceforge.net/support/tracker.php?aid=572165]

typo in SCView.sc

SCView
getPropertyList {
this.properties.collect({ arg name;
[name, this.perform(name)]
});
}

doesn't return
should be ^this.properties.collect ...

[Issue migrated from SourceForge | ID: 1815773 | Submitted by 'jamshark70']
[http://sourceforge.net/support/tracker.php?aid=1815773]

scsynth prints the output properties for the input device, and the input properties for the output device.

The ultralite has a 2-ch main out, 8 individual outs, headphones and SPDIF. So the output under "input device" looks like the output.

"MOTU UltraLite" Input Device
Streams: 4
0 channels 2
1 channels 8
2 channels 2
3 channels 2

"MOTU UltraLite" Output Device
Streams: 3
0 channels 8
1 channels 2
2 channels 2

And look what happens if I switch to "built-in mic" for input. "Built-in Microphone" has no output capability, hence 0 streams. But it should have a stream for input!

"Built-in Microphone" Input Device
Streams: 0

"Built-in Microphone" Output Device
Streams: 3
0 channels 8
1 channels 2
2 channels 2