Pd-dev September 2007

pd-dev@lists.iem.at

36 participants
77 discussions

[ pure-data-Bugs-1788587 ] unpackOSC doesn't like typed floats
by SourceForge.net 06 Sep '07

06 Sep '07

Bugs item #1788587, was opened at 2007-09-05 18:01 Message generated for change (Comment added) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: externals Group: None >Status: Closed >Resolution: Accepted Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Martin Peach (mrpeach) Summary: unpackOSC doesn't like typed floats Initial Comment: unpackOSC does not decode an osc-message correctly that contains a floating-point number and typetags. (see attached patch to trigger the bug) the problem only occurs if compiled with strict-aliasing enabled (default on gcc>3.0 with optimization levels 2, 3, s), because unpackOSC uses pointer-casts for re-interpreting the data. for information on this see: http://mail.opensolaris.org/pipermail/tools-gcc/2005-August/000047.html one way to prevent this problem is by adding the "-fno-strict-aliasing" to the CFLAGS. unfortunately, osc comes without build-system so you cannot add this flag... tested with: pd-0.40.1, pd-0.41test6 (linux x86) ---------------------------------------------------------------------- >Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-06 10:18 Message: Logged In: YES user_id=564396 Originator: YES fixed (btw, mrpeach, can you close an open issue yourself, or does this require administrator privileges?) ---------------------------------------------------------------------- Comment By: Martin Peach (mrpeach) Date: 2007-09-05 21:48 Message: Logged In: YES user_id=919007 Originator: NO Should be fixed now... ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 18:04 Message: Logged In: YES user_id=564396 Originator: YES one more remark: neither the version in pd-extended nor matt wright's original "dumpOSC" application usually exhibit this problems, as they use different compiler flags. ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 18:04 Message: Logged In: YES user_id=564396 Originator: YES a more generic solution for this problem is to use unions (as described in the above-mentioned article) instead of pointer-casts. the attached patch does this and it seems to fix the problem, it would be nice if this could make it into osc (for the one single person who does not use the pd-extended build-system) File Added: unpackOSC.c.diff ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_…

1 0

glist_grab not grabbing arrows
by federico 06 Sep '07

06 Sep '07

I searched thru the externals... didn't found one using arrow keys. anyone can give an example of capturing arrow keys? here's my code calling glist_grab: static int grid_click(t_gobj* z, t_glist* glist, int xpix, int ypix, int shift, int alt, int dbl, int doit) { t_grid* x = (t_grid*)z; t_canvas* canvas = glist_getcanvas(glist); if(doit) { glist_grab(glist, &x->x_obj.te_g, (t_glistmotionfn)grid_motion, grid_key, xpix, ypix); ... ty Federico

3 5

sys_idlehook and audio API interaction
by Claude Heiland-Allen 05 Sep '07

05 Sep '07

Hi, I made an external to let patches do things when Pd is idle: [idlebang] -- output a bang when Pd is idle but it seems to be heavily dependent on the audio API. With OSS, I get around 200 bangs per second. With ALSA, I get around 70000 bangs per second. With JACK, I get around 1 bang per second. With audio off, I get around 160000 bangs per second. I haven't tried PortAudio yet. Any idea why this might be? Source code tarball is here: https://devel.goto10.org/dl.php?repname=maximus&path=%2Fidlebang%2F&rev=0&i… Thanks for any insights, Claude -- http://claudiusmaximus.goto10.org

2 3

[ pure-data-Bugs-1788587 ] unpackOSC doesn't like typed floats
by SourceForge.net 05 Sep '07

05 Sep '07

Bugs item #1788587, was opened at 2007-09-05 12:01 Message generated for change (Comment added) made by mrpeach You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: externals Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Martin Peach (mrpeach) Summary: unpackOSC doesn't like typed floats Initial Comment: unpackOSC does not decode an osc-message correctly that contains a floating-point number and typetags. (see attached patch to trigger the bug) the problem only occurs if compiled with strict-aliasing enabled (default on gcc>3.0 with optimization levels 2, 3, s), because unpackOSC uses pointer-casts for re-interpreting the data. for information on this see: http://mail.opensolaris.org/pipermail/tools-gcc/2005-August/000047.html one way to prevent this problem is by adding the "-fno-strict-aliasing" to the CFLAGS. unfortunately, osc comes without build-system so you cannot add this flag... tested with: pd-0.40.1, pd-0.41test6 (linux x86) ---------------------------------------------------------------------- >Comment By: Martin Peach (mrpeach) Date: 2007-09-05 15:48 Message: Logged In: YES user_id=919007 Originator: NO Should be fixed now... ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 12:04 Message: Logged In: YES user_id=564396 Originator: YES one more remark: neither the version in pd-extended nor matt wright's original "dumpOSC" application usually exhibit this problems, as they use different compiler flags. ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 12:04 Message: Logged In: YES user_id=564396 Originator: YES a more generic solution for this problem is to use unions (as described in the above-mentioned article) instead of pointer-casts. the attached patch does this and it seems to fix the problem, it would be nice if this could make it into osc (for the one single person who does not use the pd-extended build-system) File Added: unpackOSC.c.diff ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_…

1 0

new external... please help
by Thomas O Fredericks 05 Sep '07

05 Sep '07

I do not know if I should post here or on the main list. Anyways, I made a new external called folderpanel (I know it is possible to reproduce this with tot, by that solution proved buggy with cyclone and Gem or something, I don't remember the details...) that lets you select a directory. It is based on parts of cyclone and pd's savepanel. It did not work before, but then I added a space or a \\; in in the tk code somewhere and it works now. I do not know much c, so please bear with me. Here are a few questions: - how do I guarantee that the tk code, lines 78 to 83, only run once? - what does pd_bind do? - do you see any other weirdness in my code ( I know there are useless pieces of code, like all references to s and s->s_name) ? what I would like to add: - starts browsing in the path received. how would you implement this? I have seen http://www.tcl.tk/man/tcl8.3/TkCmd/chooseDirectory.htm but I am unsure of the implementation with pd. Source code, help file and i386 linux binary attached (to compile: make pd_linux) Thanks -- thomas ouellet fredericks, tof(a)danslchamp.org, montreal, canada

1 0

[ pure-data-Bugs-1759014 ] random tcpserver crashes
by SourceForge.net 05 Sep '07

05 Sep '07

Bugs item #1759014, was opened at 2007-07-23 18:17 Message generated for change (Comment added) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1759014&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: externals Group: None >Status: Closed >Resolution: Accepted Priority: 7 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Martin Peach (mrpeach) Summary: random tcpserver crashes Initial Comment: tcpserver easily crashes, when deleting one of several [tcpserver] objects. i was only able to reproduce this reliably on x86_64 platform by the following way (sounds a bit more complicated than it needs to be): 1. create [tcpserver 5600] (ok) 2. create [tcpserver 5800] (ok) 3. create [tcpserver 0] (ok) 4. create [tcpserver 0] (ok) 5. rename (3) to [tcpserver 5600] (the object will fail to create) 6. rename (4) to [tcpserver 5800] (pd will crash) i had no problem with the same procedure on i686. nevertheless, after a look at the code i wonder why it doesn't crash more often... tcpserver uses a static binbuf, which is destroyed without checks in the free method. since the binbuf is static, the next tcpserver that is destroyed will try to free the binbuf again, leading to a segfault. the good news is, that the binbuf is really nowhere needed in the code (it only gets created and deleted). the attached patch removes it, leading to a no more crashes on x86_84 ---------------------------------------------------------------------- >Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 18:05 Message: Logged In: YES user_id=564396 Originator: YES let's assume it is really fixed :-) ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-07-23 18:49 Message: Logged In: NO OK, it's fixed. Thanks for finding it! ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1759014&group_…

1 0

[ pure-data-Bugs-1788587 ] unpackOSC doesn't like typed floats
by SourceForge.net 05 Sep '07

05 Sep '07

Bugs item #1788587, was opened at 2007-09-05 18:01 Message generated for change (Comment added) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: externals Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Martin Peach (mrpeach) Summary: unpackOSC doesn't like typed floats Initial Comment: unpackOSC does not decode an osc-message correctly that contains a floating-point number and typetags. (see attached patch to trigger the bug) the problem only occurs if compiled with strict-aliasing enabled (default on gcc>3.0 with optimization levels 2, 3, s), because unpackOSC uses pointer-casts for re-interpreting the data. for information on this see: http://mail.opensolaris.org/pipermail/tools-gcc/2005-August/000047.html one way to prevent this problem is by adding the "-fno-strict-aliasing" to the CFLAGS. unfortunately, osc comes without build-system so you cannot add this flag... tested with: pd-0.40.1, pd-0.41test6 (linux x86) ---------------------------------------------------------------------- >Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 18:04 Message: Logged In: YES user_id=564396 Originator: YES one more remark: neither the version in pd-extended nor matt wright's original "dumpOSC" application usually exhibit this problems, as they use different compiler flags. ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-05 18:04 Message: Logged In: YES user_id=564396 Originator: YES a more generic solution for this problem is to use unions (as described in the above-mentioned article) instead of pointer-casts. the attached patch does this and it seems to fix the problem, it would be nice if this could make it into osc (for the one single person who does not use the pd-extended build-system) File Added: unpackOSC.c.diff ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_…

1 0

[ pure-data-Bugs-1788587 ] unpackOSC doesn't like typed floats
by SourceForge.net 05 Sep '07

05 Sep '07

1 0

[ pure-data-Bugs-1788587 ] unpackOSC doesn't like typed floats
by SourceForge.net 05 Sep '07

05 Sep '07

Bugs item #1788587, was opened at 2007-09-05 18:01 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: externals Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Martin Peach (mrpeach) Summary: unpackOSC doesn't like typed floats Initial Comment: unpackOSC does not decode an osc-message correctly that contains a floating-point number and typetags. (see attached patch to trigger the bug) the problem only occurs if compiled with strict-aliasing enabled (default on gcc>3.0 with optimization levels 2, 3, s), because unpackOSC uses pointer-casts for re-interpreting the data. for information on this see: http://mail.opensolaris.org/pipermail/tools-gcc/2005-August/000047.html one way to prevent this problem is by adding the "-fno-strict-aliasing" to the CFLAGS. unfortunately, osc comes without build-system so you cannot add this flag... tested with: pd-0.40.1, pd-0.41test6 (linux x86) ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788587&group_…

1 0

[ pure-data-Bugs-1788567 ] [pack symbol] accepts non-symbols
by SourceForge.net 05 Sep '07

05 Sep '07

Bugs item #1788567, was opened at 2007-09-05 11:30 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788567&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: puredata Group: v0.40.2 Status: Open Resolution: None Priority: 5 Private: No Submitted By: Hans-Christoph Steiner (eighthave) Assigned to: Miller Puckette (millerpuckette) Summary: [pack symbol] accepts non-symbols Initial Comment: If [pack symbol ...] receives a message with just a selector, then it treats it as if it was a symbol. I.e. this packs a symbol: [test( <--this works | | [test( <-- this throws an error | | [pack symbol symbol] Tested on Mac OS X: 0.40-2 and 0.39.3-extended-rc5. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1788567&group_…

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Pd-dev September 2007