Pd-dev August 2002

pd-dev@lists.iem.at

25 participants
32 discussions

Pd and Intel IA32 compiler
by Olaf Matthes 13 Aug '02

13 Aug '02

Hi developers, has anybody ever tried to use the Intel C++ compiler for Pd under Win? I just tried but noticed one anomaly: When compiling I get the following warning for every EXTERN_STRUCT definition: warning #1128: a storage class may not be specified here Changing the definition in m_pd.h seems to help (I don't get the warning any longer): /* and depending on the compiler, hidden data structures are declared differently: */ #ifdef __GNUC__ #define EXTERN_STRUCT struct #else #ifdef IA32 /* Intel IA32 compiler */ #define EXTERN_STRUCT struct #else /* non-GNU compiler */ #define EXTERN_STRUCT extern struct #endif #endif It's also necessary to define "IA32" for the compiler to get the above code working... (maybe the compiler identifies itselfe by it's own but I couldn't find any hints) In the end it seems that this compiler reduces CPU usage a bit (on my P4 with compiler options G7 and QaxW set). More tests have to show if it's worth the effort. I should probably recompile all libraries as well... I could upload a version of Pd0.35 compiled with this compiler in case someone is interested or is willing to test the amount of optimization one gets in real 'on-stage' situations. Olaf

3 4

Tcl Bug
by JaReK 13 Aug '02

13 Aug '02

Howdy. I got the following bug when running PD 35.1 on Win2k sp2 can't read "tkPriv(x)": no such element in array while executing "if {(283 != $tkPriv(x)) || (182 != $tkPriv(y))} { set tkPriv(mouseMoved) 1 }" invoked from within "if {!$tk_strictMotif} { if {(283 != $tkPriv(x)) || (182 != $tkPriv(y))} { set tkPriv(mouseMoved) 1 } if {$tkPriv(mouseMoved)} { .bgerrorT..." (command bound to event) Bug started occurring after I closed all patches I was editing and all that was present on the monitor was the "main" pd window and the cmd window Regards ... Jarek

1 0

alsa rme pd again
by martin pi 12 Aug '02

12 Aug '02

hi, sine i installed the alsa drivers i get strange sound from pd and all other software. this is a bit off topic, yet someone might help me, since i know the card is rather popular in here. actually the sound seems to be about 2 octaves lower than it has to be. i am no sound engineer, so i do not know, how this happens technically. thanks for help martin martin pi http://attacksyour.net/pi johann strauss gasse 32|7 1040 vienna www.machfeld.net

3 4

float & symbol in second inlet
by Olaf Matthes 12 Aug '02

12 Aug '02

Hi, is it possible to create an object with a second inlet that accepts floats and symbols? That's no problem for the first (leftmost) inlet but what with additional inlets? Any ideas? thanks, Olaf

2 1

Re: [PD-dev] pure-data CVS
by g�nter geiger 09 Aug '02

09 Aug '02

On Thu, 8 Aug 2002, [ISO-8859-1] Juha Vehvil�inen wrote: > Hi, > > If i can help with the build-system, let me know. yes, always, first problem I have is that I want to structure the externals somehow, but I didn't come up with a good classification yet. Maybe we should discuss this on the list, so I have CC'd this. There has been some discussion about different classes of externals which as far as I can remember didn't come to a conclusion, maybe we should talk about that again. > > With autofetching, I though of pd instead of saying > "couldn't create foo" would see if foo can be downloaded > (and built?, and installed) from the cvs. Don't know > if this could work even reasonably well though. ah, I understand. In the first place this would mean a change in pd's code, something like a message ; pd loadfailed <externalname> that gets issued by the loader, then it would be possible to implement, but surely only for linux. But actually its a good idea to have a build system this way, where you can start a small script/program and this will download and build the external that you requested. Guenter

3 5

strange segfaults
by martin pi 08 Aug '02

08 Aug '02

hi, i compiled a usually working library (linux) on a different computer yesterday and got strange segmentation faults occuring 10 minutes after running it. gdb revealed, that pd tries to gensym (dogensym() or dogen() as far as i recall). the calling statement was the setup method of an external i did not use in the patch (yet of the same lib). i did not save the backtrace and i cannor reconstruct the behaviour on my development computer. any thoughts are welcome martin martin pi attacksyour.net/pi johann strauss gasse 32 | 7 1040 vienna ++43 699 10 44 37 42 sil.at

2 1

BorlandC++
by Thomas Grill 07 Aug '02

07 Aug '02

Hi Miller, could you please change m_pd.h / line 29 to #if defined(__GNUC__) || defined(__BORLANDC__) so that PD will instantly compile with the free BorlandC++ compiler as well greetings, Thomas

1 0

Re: [PD-dev] object creation bug?
by Thomas Grill 03 Aug '02

03 Aug '02

Hi Miller, > well, it's intentional: you get all the pointer&integer arguments, THEN > all the flost arguments. It's easier to implement message-passing > efficiently this way. (I should have written this down somewhere!) even if it's intentional, i dont like it very much. I find it better to sacrifice two or three cpu cycles for more flexibility. Of course, one can do it with gimme arguments, but it's for sure even slower then..... The reason why i stumbled across this issue is that it works in Max/MSP. greetings, Thomas

2 2

object creation bug?
by Thomas Grill 03 Aug '02

03 Aug '02

Hi Miller, hi all, i'm stuck with the following codelet: -- snip ----------------------------------------------------------------------- ----- #include "m_pd.h" struct simple_t { t_object obj; }; static struct _class * simple1_class; _class* class_simple1 (float arg1, t_symbol *arg2) { /* arg1 and arg2 are invalid!! */ post("arguments: %f %s",arg1,arg2->s_name); /* seg fault!! */ return 0; } extern "C" __declspec(dllexport) void simple1_setup() { simple1_class = class_new( gensym("simple1"), (t_newmethod)class_simple1,NULL,sizeof(simple_t), 0, A_FLOAT, A_SYMBOL, A_NULL); } -- snip ----------------------------------------------------------------------- ----- It's simple as can be. I defined a object having a float and a symbol a creation arguments. The new method "class_simple1" gets invalid values for arg1 and arg2, though. It's the same for 0.34-4 and 0.35 If it's a PD bug then it's strange that noone noticed it before..... greetings, Thomas

2 1

Re: [PD-dev] STL & C++ & platforms - a question
by jdavison＠speakeasy.org 02 Aug '02

02 Aug '02

andy(a)e-molecule.com writes: >I am currently using STLPort quite heavily in a large application. It runs on >both linux, windows and sgi with good performance. It is threadsafe, in fact >fully reentrant, and I'd say its quite indispensible for any serious work in >C++. I second that. I started working with STLport with the Sun WorkShop C++ 4.2 compiler in 1999 and very quickly concluded that STLport is an absolute necessity when working with a compiler whose ISO/IEC 14882:1998 Standard Containers Library and Standard Algorithms Library is deficient or absent. (I've also had good results with the ObjectSpace C++<ToolKit>, but it is neither Free nor open-source.) -- John Davison, jdavison at-sign speakeasy period org

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 August 2002