@EEight said:

@lacuna said:
I just see this flag on linux:
-nosleep -- spin, don't sleep (may lower latency on multi-CPUs)

Oh yes and there are startup flags for loading a different scheduler (I corrected this in my first post now)
https://puredata.info/docs/faq/commandline

  -rt or -realtime -- use real-time priority
  -nrt             -- don't use real-time priority
  -sleep           -- sleep when idle, don't spin (true by default)
  -nosleep         -- spin, don't sleep (may lower latency on multi-CPUs)
  -schedlib <file> -- plug in external scheduler
  -extraflags <s>  -- string argument to send schedlib
  -batch           -- run off-line as a batch process
  -nobatch         -- run interactively (true by default) 

Not sure, would be interested to know too.

@EEight output of full text search of sleep with grep in pd's source folder -nri flags set:
https://pastebin.com/3mBw6Mnj

/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/u_pdsend.c:86:            sleep (nretry < 5 ? 1 : 5);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_stuff.h:65:extern int sys_sleepgrain;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_stuff.h:163:EXTERN void sys_microsleep(int microsec);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_stuff.h:369:EXTERN int* get_sys_sleepgrain(void);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:58:int sys_nosleep = 0;  /* skip all "sleep" calls and spin instead */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:115:int* get_sys_sleepgrain() { return &sys_sleepgrain; }
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:424:"-sleepgrain <n>  -- specify number of milliseconds to sleep when idle\n",
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:520:"-sleep           -- sleep when idle, don't spin (true by default)\n",
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:521:"-nosleep         -- spin, don't sleep (may lower latency on multi-CPUs)\n",
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:732:        else if (!strcmp(*argv, "-sleepgrain"))
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:737:            sys_sleepgrain = 1000 * atof(argv[1]);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:1242:        else if (!strcmp(*argv, "-sleep"))
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:1244:            sys_nosleep = 0;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:1247:        else if (!strcmp(*argv, "-nosleep"))
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_main.c:1249:            sys_nosleep = 1;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:202:extern int sys_nosleep;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:204:/* sleep (but cancel the sleeping if pollem is set and any file descriptors are
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:207:sleep. */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:208:static int sys_domicrosleep(int microsec, int pollem)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:226:          perror("microsleep select");
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:242:        Sleep(microsec/1000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:244:        usleep(microsec);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:251:    /* sleep (but if any incoming or to-gui sending to do, do that instead.)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:253:void sys_microsleep(int microsec)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:256:    sys_domicrosleep(microsec, 1);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_inter.c:909:    int didsomething = sys_domicrosleep(0, 1);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:16:    correct thread synchronization (by defining THREADSIGNAL) or just sleeping
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:21:    switch to usleep in s_inter.c
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:91:#include <windows.h>    /* for Sleep() */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:549:            sys_microsleep(sys_sleepgrain);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:550:            if (!pa_stream)     /* sys_microsleep() may have closed device */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:591:            sys_microsleep(sys_sleepgrain);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_pa.c:592:            if (!pa_stream)     /* sys_microsleep() may have closed device */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_oss.c:672:            sys_microsleep(2000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:105:/* if more than this sleep detected, should be more than periodsize/samplerate ??? */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:106:static double sleep_time;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:340:  sleep_time =  (float) alsamm_period_size/ (float) alsamm_sr;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:798:      sleep(1);       /* wait until the suspend flag is released */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:1336:  if ((timenow = sys_getrealtime()) > (timelast + sleep_time))
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsamm.c:1342:             timenow,timelast,sleep_time,(timelast + sleep_time));
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/s_audio_alsa.c:691:        sys_microsleep(5000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:23:int sys_usecsincelastsleep(void);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:24:int sys_sleepgrain;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:442:will now sleep. */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:454:    if (sys_sleepgrain < 100)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:455:        sys_sleepgrain = sys_schedadvance/4;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:456:    if (sys_sleepgrain < 100)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:457:        sys_sleepgrain = 100;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:458:    else if (sys_sleepgrain > 5000)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:459:        sys_sleepgrain = 5000;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:477:               the machine sleeps.  */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:539:                    /* if even that had nothing to do, sleep. */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:541:                    sys_microsleep(sys_sleepgrain);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:574:        Sleep(1000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/src/m_sched.c:576:        sleep(1);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptwinmm.c:67:PMEXPORT void Pt_Sleep(int32_t duration)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptwinmm.c:69:    Sleep(duration);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptmacosx_mach.c:128:void Pt_Sleep(int32_t duration)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptmacosx_mach.c:130:    usleep(duration * 1000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptmacosx_cf.c:137:void Pt_Sleep(int32_t duration)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptmacosx_cf.c:139:    usleep(duration * 1000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptlinux.c:14:of sleeping when realtime threads request a sleep of <=2ms (as a way 
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptlinux.c:132:void Pt_Sleep(int32_t duration)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/ptlinux.c:134:    usleep(duration * 1000);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/porttime.h:82:    Pt_Sleep() pauses, allowing other threads to run.
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/porttime/porttime.h:88:PMEXPORT void Pt_Sleep(int32_t duration);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/portmidi/pm_mac/pmmacosxcm.c:492:            usleep((useconds_t) 
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portmidi/patches/mac_limit_rate_override.patch:52:             usleep((useconds_t) 
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/win/pa_win_util.c:102:void Pa_Sleep( long msec )
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/win/pa_win_util.c:104:    Sleep( msec );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:108:void Pa_Sleep( long msec )
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:110:#ifdef HAVE_NANOSLEEP
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:116:    nanosleep(&req, &rem);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:117:    /* XXX: Try sleeping the remaining time (contained in rem) if interrupted by a signal? */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:120:        {                   /* to usleep must be < 1000000.   */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:121:        usleep( 999000 );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:124:    usleep( msec * 1000 );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:599:        /* Test before and after in case whatever underlying sleep call isn't interrupted by pthread_cancel */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:601:        Pa_Sleep( intervalMsec );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:644:            PA_DEBUG(( "%s: Watchdog sleeping for %lu msecs before unthrottling\n", __FUNCTION__, th->throttledSleepTime ));
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:645:            Pa_Sleep( th->throttledSleepTime );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/os/unix/pa_unix_util.c:704:        Pa_Sleep( intervalMsec );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/wmme/pa_win_wmme.c:2185:    unsigned long throttledSleepMsecs;
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/wmme/pa_win_wmme.c:2565:    /* time to sleep when throttling due to >100% cpu usage.
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/wmme/pa_win_wmme.c:2567:    stream->throttledSleepMsecs =
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/wmme/pa_win_wmme.c:3194:                            /* sleep to give other processes a go */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/wmme/pa_win_wmme.c:3195:                            Sleep( stream->throttledSleepMsecs );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_utilities.c:395:      /* No match yet, so let's sleep and try again. */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_utilities.c:396:      Pa_Sleep( 100 );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_blocking.h:69:#define PA_MAC_BLIO_BUSY_WAIT_SLEEP_INTERVAL (5)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_blocking.c:454:             Pa_Sleep( PA_MAC_BLIO_BUSY_WAIT_SLEEP_INTERVAL );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_blocking.c:535:             Pa_Sleep( PA_MAC_BLIO_BUSY_WAIT_SLEEP_INTERVAL );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core_blocking.c:607:            Pa_Sleep( msecPerBuffer );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/coreaudio/pa_mac_core.c:2722:       Pa_Sleep( 100 );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/asio/pa_asio.cpp:3401:        Sleep(1);
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/alsa/pa_linux_alsa.c:1129:        Pa_Sleep( 10 );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/alsa/pa_linux_alsa.c:2759:        /* self->threading.throttledSleepTime = (unsigned long) (minFramesPerHostBuffer / sampleRate / 4 * 1000); */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/alsa/pa_linux_alsa.c:3831:                Pa_Sleep( 1 ); /* avoid hot loop */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/alsa/pa_linux_alsa.c:3849:            if( timeouts > 1 ) /* sometimes device times out, but normally once, so we do not sleep any time */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/hostapi/alsa/pa_linux_alsa.c:3851:                Pa_Sleep( 1 ); /* avoid hot loop */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/src/common/pa_util.h:152:/* void Pa_Sleep( long msec );  must also be implemented in per-platform .c file */
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/include/portaudio.h:1211:/** Put the caller to sleep for at least 'msec' milliseconds. This function is
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/include/portaudio.h:1215: The function may sleep longer than requested so don't rely on this for accurate
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/include/portaudio.h:1218:void Pa_Sleep( long msec );
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/include/pa_win_wmme.h:64:    to THREAD_PRIORITY_NORMAL and sleeps the thread if the CPU load exceeds 100%
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/portaudio/portaudio/include/pa_linux_alsa.h:91:/** Set the maximum number of times to retry opening busy device (sleeping for a
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x5.htm:599:<P> In linux, a "-nosleep" flag causes Pd to poll instead of sleeping as it
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x5.htm:798:<P> Fixed a thread-safety problem in sys_microsleep().
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x5.htm:1224:the controlling parameter for MIDI jitter is "-sleepgrain", which specifies
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x5.htm:1225:the interval of time Pd sleeps when it believes it's idle.
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x5.htm:1373:<P> -sleepgrain: if you aren't using audio I/O, this can reduce time jitter in
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:488:scheduling; "-sleepgrain 1" sets the sleep grain to 1 (see under MIDI below),
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:489:and typing "-rt -sleepgrain 1" does both.
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:527:-sleepgrain &lt;n&gt;  -- specify number of milliseconds to sleep when idle
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:573:-nosleep         -- never relinquish CPU (only for multiprocessors!)
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:610:<H4> MIDI and sleepgrain</H4>
/pd-0.50-2.src.tar.gz.extracted/pd-0.50-2/doc/1.manual/x3.htm:619:<P> The "sleepgrain" controls how long (in milliseconds) Pd sleeps between

good night sweet dreams

Purr Data 2.10.0 is now available:

https://github.com/jonwwilkes/purr-data/releases/tag/2.10.0

Changes:

• iem_spec2/spec2_tabreceive_enable~: fix array error handler and set sane default array name value
• fix partconv crashers in bsaylor lib and add perfroutine for array errors
• adaptive/nlms3~: fix typo that caused a double free
• fix lyonpotpourri crashers in dsp, perform and constructor routines
• at least keep the inoperable streamout13~ and streamin13~ from crashing when instantiating
• use some sane default values in ekext/lpreson~ to prevent segfault
• quick fixes to keep cxc/mean~ from crashing when dsp is turned on
• greatly reduce undefined behavior in all dsp objects
• fix hex2dec so that it actually does something useful
• fix #523: crash with manual width adjustment on subpatch
• add ability to change makefile flags for Gem from toplevel makefile
• fix stray bugs detected by obs
• unauthorized/cooled~: increase string buffer size to accommodate the terminating nul character
• unauthorized/cooled~: fix memory access bug trying to concatenate into a string constant
• iemmatrix/mtx_dispersive_dline: add missing void return type
• allow make options like -j8 to be passed to the Gem compilation, which takes awfully long on a single cpu.
• cxc/cxc_split: fix use of un-initialized pointer
• ggee/serial_bird: fix undefined behavior with the ++ operator
• ext13/scramble~: fix header for scramble~
• jasch_lib/detox/detox: reformat for sanity's sake, fix array overflow, undefined behavior
• linux desktop: remove the -rt -audiobuf options from the desktop files.
• linux desktop: change DEFAULTADVANCE to 20 ms for Linux.
• linux desktop: remove leftover TargetEnvironment=Unity lines in menu entries for Purr Data
• linux desktop: add some comments and a few more useful desktop action examples to the main desktop file, so that the user understands how to adjust these if needed.
• linux desktop: replace pd-gui -> nw in the ForceQuit actions, which is the proper name of the GUI program on Linux
• linux desktop: remove useless %U arguments from desktop actions.
• linux desktop: invoke desktop actions via /bin/sh.
• linux desktop: migrate the desktop actions from the ancient Unity syntax to the current freedesktop.org standard
• linux desktop: remove sticky options from the desktop files. For now, keep -rt -audiobuf 20.
• Gem: sync with https://github.com/umlaeute/Gem, QT4L and startup issues have been fixed
• linux: fix the Debian control files once again, since the dependency auto-detection needs a Depends line in there.
• debuild: Support for ARM (e.g., Raspbian)
• update nw-update to nw.js 0.24.4 to fix font issues under Linux
• backport 'add-to-path' from vanilla rev. c917dd19, to make Gem happy.
• usability improvements in the documentation browser.
• switch Gem to the latest from upstream.
• add missing dlls for fluid~ on Windows. Fixes #540.
• Debian packaging: Demote python and fluid-soundfont dependencies, as suggested in #540.
• polish the externals/Makefile clean targets, and delete redundant files in repo
• fix compile options for Xcode 10 - fftease and lyonpotpourri externals.
• update pd-lua to latest upstream.
• fix compile options for Xcode 10 - externals and abstractions.
• fix compile options for Xcode 10.
• ios header needs to be included before base64.h, to avoid compile errors on macOS 10.14.
• fix improper string access in pd_getdirname on Mac.
• fix list cat crasher, update help patch, add missing test abstractions
• get rid of obsolete and unneeded unicap and sndobj dependencies on Linux.
• mark some globals as extern to fix compilation if g_canvas.h is included more than once

Please report bugs here:

https://git.purrdata.net/jwilkes/purr-data/issues

If @jancsika has a moment maybe he can explain where things stand.

In Purr Data you can do [floatsize(---[pdinfo] to tell whether floats are single or double precision. The current possible outputs from that message would be "32" for single precision or "64" for double precision.

Pd and the externals you use have to be compiled for one or the other-- you can't mix the two floatsizes. Currently Purr Data is able to build itself and most externals with floatsize=64, but some libraries need to be revised so that they work. So atm we're shipping floatsize=64 Purr Data with just the core and a few convenience libraries. I named these builds "purr-double-trouble" since they don't yet ship with all the externals.

That is all a separate issue from the arch for which a piece of software is compiled. "Arch" means chip architecture and uses nicknames like "i32", "x86_64", "arm64". For general purpose computers these generally divide into two broad categories-- the ones which the hardware shuttles data around in groups of 64 bits, and hardware that shuttles data in groups of 32 bits. Some 64-bit platforms even let you run old stuff built for a 32-bit hardware in a special compatibility mode.

So to clarify: you can have:

• single-precision Pd built for a 32-bit arch
• single-precision Pd built for a 64-bit arch
• double-precision Pd built for a 32-bit arch
• double-precision Pd built for a 64-bit arch

Hi,

not sure if this is a bug. Don't wanna noobishly spam to the devs' bug reports pile but this needs to be clarified. Is it a deken issue or on the repository or what?

Some searches for externals in the 64-bit version of PD vanilla for windows (I'm on windows 10) the deken search (in the menu: Help -> Find externals) don't return results while the same externals (in some cases) can found on the 32-bit version. Some externals are neither found in the 32- nor 64-bit version. Additionally - as a marker that something is wrong here - PD does not output the line

[deken]: No matching externals found. Try using the full name e.g. 'freeverb'.

which is the case for truly non-existing names (like searching "sdfsdfsdf").

Example of external found as expected on 32- and 64-version:
cyclone

Example of external found as expected on 32- but NOT on 64-bit version:
moonlib

Example of external found neither on 32- or 64-bit version:
apple

What I'm most eager to find out is whether, say, the moonlib external is just not compiled for 64-bit windows and therefore is not found or whether it should be found and just isn't (which hints more towards a bug).

Does anyone have an idea? I'm not completely new to PD but not up to date at all with its development. It seems the current version 0.49 vanilla is the first to have a 64-bit windows version.

kind regards

Thanks for the info @Pierre-Guillot!
@Pierre-Guillot said:

Hi @alfonso.santimone

I argue that adding external is just a matter of incorporating them in libpd.dll. Am I right?

Yes but the libpd used by Camomile is the static library (.lib/.so/.a) that why libpd is embedded in Camomile.

1. In fact, I guess that some of the external libraries are already compatible with multiinstance support because it depends on what parts of the Pd API the libraries use. But to ensure the multiinstance support you must use the C flags -DPDINSTANCE=1 (and -DPDTHREAD=1) while compiling the library or libpd and avoid all the static variables (except if they use thread local storage and you ensure to init them well on each thread). Here is one of the 1st posts by Miller about it: https://lists.puredata.info/pipermail/pd-dev/2017-04/020980.html

So i guess i have to remove directly from the external code all the static variable, so basically remove the "static" keyword from the variable declarations.
How can i compile libpd as a .lib/.so/.a and not .dll?
Should i use those flags in the various makefiles or just as a fleg in the command line? (assuming working with msys2 on Win64)

2. The same way you did it for the "standard" approach.

So it's enough to use a x64 compiler and a x64 target?

3. Including an external library with libpd is pretty straightforward (as long as, in this case, the library has the multiinstance support). See this: https://github.com/libpd/libpd/wiki/Adding-Pure-Data-external-libraries-to-your-project. And if a library is included in libpd it will also be included in Camomile.

So you also have to compile libpd then Camomile (all the instructions are on the Github repository - I updated it a week ago https://github.com/pierreguillot/Camomile/tree/dev/v1.0.6-lv2#compilation - or just do the same steps as on Travis and Appveyor and I can help if needed).

Cheers

I don't really know what Travis and Appeveyor are. I have to check them out.
How to include the static pthread lib for Win compilation? You mention it the Camomile build process readme.

PS: Another approach would have been to ship the libpd dynamic library next to each plugin. i need to try this but I afraid that, on Windows and Linux, the DAW think that the libpd library is another plugin try to load it.

I guess that most hosts have an avoid list for specific .dll. At least Reaper have a config file where is possible to remove a specific .dll from being scanned as a plugin.

thanks!

@jancsika

Some report after testing Purr Data compiled for Win64.

1. I have a strange bug related to MIDI device. I'll have to check if it exists in PD vanilla 32 bit. I have the same bug in PD vanilla for 64bit compiled by Lucas. I don't think it is a multiclient midi driver issue. But i'll check it better. BTW if i choose a MIDI device and apply so it goes saved for the next start of Purr Data..well Purr Data doesn't start anymore and i have to clean the prefs via regedit because it seems that a command line pd.exe -nomidi does not work for Purr Data as it does with Vanilla

2. zexy library does not work even if the .dll externals files are compiled. I have to check the build log because i remember there were some red warnings about specific zexy externals. So every zexy help file i tried shows red outlined externals because the zexy stuff can't be instantiated.

3. there are minor naming errors in some help files that i guess are in the official Purr Data 32bit release (i.e. linearpan~ in pan lib should be linear_pan~

4. timestretch~ just crash Purr Data and the program quits.

5. moocow folder is empty

6. pixelTango does not work but i guess it has something to do with Gem externals dependencies

7. SMLib help file is not present

8. pmpd loads but does not seem to work

9. Gem and fluidsynth~ don't compile for win64 target (so as you know i removed those from the building process)

@cuinjune I tried to compile the lissa seq patch. but when i open the executable it opens only a small empty window.
i also tried to compile a help patch for testing, with the same result.
but your example works fine(Win32Example).

Hello, I am able to load a couple of PD externals. I am using Pd 0.46.7 on Linux Mint 18.2 Cinnamon 64-bit. I am using the find externals command under the help menu. I see the directories in my ~pd-externals folder. But when I try and use objects like [counter] that are supposed to be in these externals I get a can"t create error?
I get this message when I start PD
sys_midiapi 1
input channels = 2, output channels = 2
priority 6 scheduling failed; running at normal priority
tried /home/lulujoe13/pd-externals/zexy.l_ia64 and failed
tried /home/lulujoe13/pd-externals/zexy.pd_linux and failed
tried /home/lulujoe13/pd-externals/zexy/zexy.l_ia64 and failed
tried /home/lulujoe13/pd-externals/zexy/zexy.pd_linux and failed
/home/lulujoe13/pd-externals/zexy: can't load library
tried /home/lulujoe13/pd-externals/cyclone.l_ia64 and failed
tried /home/lulujoe13/pd-externals/cyclone.pd_linux and failed
tried /home/lulujoe13/pd-externals/cyclone/cyclone.l_ia64 and failed
tried /home/lulujoe13/pd-externals/cyclone/cyclone.pd_linux and succeeded

Cyclone 0.3 beta 2; Released: June 19th, 2017
This is a sub library containing the objects:
[!-], [!-~], [!/], [!/~], [!=~], [%~], [+=~], [<=~], [<~], [==~], [>=~] and [>~]

tried /home/lulujoe13/pd-externals/upp.l_ia64 and failed
tried /home/lulujoe13/pd-externals/upp.pd_linux and failed
tried /home/lulujoe13/pd-externals/upp/upp.l_ia64 and failed
tried /home/lulujoe13/pd-externals/upp/upp.pd_linux and failed
/home/lulujoe13/pd-externals/upp: can't load library
tried /home/lulujoe13/pd-externals/Gem.l_ia64 and failed
tried /home/lulujoe13/pd-externals/Gem.pd_linux and failed
tried /home/lulujoe13/pd-externals/Gem/Gem.l_ia64 and failed
tried /home/lulujoe13/pd-externals/Gem/Gem.pd_linux and succeeded
/home/lulujoe13/pd-externals/Gem/Gem.pd_linux: libftgl.so.2: cannot open shared object file: No such file or directory
/home/lulujoe13/pd-externals/Gem: can't load library
deken-plugin.tcl (Pd externals search) in /home/lulujoe13/pd-externals/deken-plugin loaded.
Platform detected: Linux-x86_64-64bit
'deken-plugin.tcl' already loaded, ignoring: '/home/lulujoe13/pd-externals/deken-plugin/deken-plugin.tcl'
Opened Alsa Client 128 in:0 out:2
counter 1
... couldn't create

Can anyone offer suggestions and help? Thanks. JoeF.

Purr Data 2.3.1 is now available:

https://github.com/jonwwilkes/purr-data/releases/tag/2.3.1

Bug fixes in 2.3.1:

• fixed some display bugs in GUI
• fixed display bug with preset_hub
• fixed undefined behavior on x64 systems with the dollar arg parser
• strengthen the testing system

New in 2.3.0:

• added first draft of an external test suite to make sure a majority of external classes can load and instantiate correctly. Previously, the CI runners could register a "successful" build even if build errors kept one or more libraries from building correctly. Now there is a minimum number of objects that must instantiate in the tests or the build will fail. We can build on this to guarantee an exact number of creators for each library, making it much easier (and less risky) to imiprove the build system.
• ported tof/imagebang and allow it to instantiate with no arguments
• added "arch", "platform", and "libdir" methods to [pdinfo] and update help patch
• default loader now uses the "hexmunge" code to find filenames with hex encodings that accommodate object names with special characters

Bugfixes in 2.3.0:

• fixed up external arg types, use static declarations to protect against namespace pollution, fix allocation errors, fix crashers, header problems, fix buggy aliases for externals, fix some makefiles, fix various buffer overflows
• added missing help files for some external aliases (don't think they are used, but they are required by the Makefile)
• handle special case of trailing "/" or "/~" in legacy external alias classnames
• updated zexy, markex, iemmatrix alias files to use the simplified hexmunger
• removed external hexloader loader from default loaded libs now that we have rudimentary hexloading in the main loader
• removed arbitrary limit of 128 characters for classnames that are absolute paths
• fixed harmgen interface to use proper A_GIMME args
• fixed invalid reads in iemmatrix
• bump lyonpotpourri to head to fix missing object name argument
• make [declare -lib] handle absolute paths (and namespace-prefixed paths) consistent with object boxes
• removed references to helpers in fluid~ that got removed
• allow iemlib, unauthorized, lyonpotpourri iem_spec2, bin_ambi, iem_ambi, iem_adaptifilt and mrpeach objects to instantiate without arguments. Try to set sane defaults for these situations, while also outputting a warning. (Might consider changing the warning to an error...)
• allow class_addcreator to register an additional creator with the namespace prefix if one was used
• triage iem16's lack of shared lib with statically declared copy/pasta functions
• remove a bunch of state files from iem16 that somehow got added to the repo at the outset
• remove unnecessary pd files from fluid~
• make all cyclone classes instantiate when no arguments are provided
• switch lyonpotpourri submodule to gitlab mirror. This is needed to make some quick fixes found by the test system (which we can later request to merge upstream)

Report issues here:

https://git.purrdata.net/jwilkes/purr-data

Hi there,

Thanks for taking the time. I am getting a little bit frustrated here, so I guess it’s best to ask the community:

How do I get MrPeach to work on 0.47.1? I am using a Raspberry Pi 3 (running Linux version 4.9.20-v7+) and Pure Data 0.47.1.

I installed mrpeach using the “find externals” functionality. Of course I edited the path to “/home/pi/pd-externals” and created a “mrpeach” entry in the startup preferences.

Some background information: I want my Python code to “communicate” with pd but actually I do not care how they do this. OSC-Messages seem to be the easiest solution.
Well, seemed. I am a total beginner to pd

Help much appreciated

Daniel

By the way here is the log:

tried ./osc/routeOSC.l_arm and failed
tried ./osc/routeOSC.pd_linux and failed
tried ./osc/routeOSC/routeOSC.l_arm and failed
tried ./osc/routeOSC/routeOSC.pd_linux and failed
tried ./osc/routeOSC.pd and failed
tried ./osc/routeOSC.pat and failed
tried ./osc/routeOSC/osc/routeOSC.pd and failed
tried /home/pi/pd-externals/osc/routeOSC.l_arm and failed
tried /home/pi/pd-externals/osc/routeOSC.pd_linux and succeeded
tried ./mrpeach.l_arm and failed
tried ./mrpeach.pd_linux and failed
tried ./mrpeach/mrpeach.l_arm and failed
tried ./mrpeach/mrpeach.pd_linux and failed
tried ./mrpeach.pd and failed
tried ./mrpeach.pat and failed
tried ./mrpeach/mrpeach.pd and failed
tried /home/pi/pd-externals/mrpeach.l_arm and failed
tried /home/pi/pd-externals/mrpeach.pd_linux and failed
tried /home/pi/pd-externals/mrpeach/mrpeach.l_arm and failed
tried /home/pi/pd-externals/mrpeach/mrpeach.pd_linux and failed
tried /home/pi/pd-externals/mrpeach.pd and failed
tried /home/pi/pd-externals/mrpeach.pat and failed
tried /home/pi/pd-externals/mrpeach/mrpeach.pd and failed
mrpeach: can't load library