Pd-dev January 2018

pd-dev@lists.iem.at

12 participants
14 discussions

Re: [PD-dev] Are keyup and keyname broken?
by Dan Wilcox 30 Jan '18

30 Jan '18

Yup: https://github.com/pure-data/pure-data/issues/213 <https://github.com/pure-data/pure-data/issues/213> > On Jan 29, 2018, at 12:00 PM, pd-dev-request(a)lists.iem.at wrote: > > Message: 1 > Date: Sun, 28 Jan 2018 11:41:08 -0800 > From: Kevin Haywood <khaywood(a)ucsd.edu <mailto:khaywood@ucsd.edu>> > To: pd-dev <pd-dev(a)lists.iem.at <mailto:pd-dev@lists.iem.at>> > Subject: [PD-dev] Are keyup and keyname broken? > Message-ID: <DF767778-3443-45DF-9ECA-8C3F57DDA276(a)ucsd.edu <mailto:DF767778-3443-45DF-9ECA-8C3F57DDA276@ucsd.edu>> > Content-Type: text/plain; charset=utf-8 > > Hello, > > Apologies if this specific issue has been discussed before. I searched the archives and found related topics, but not this issue. > > I’d just like to find out if what I’m seeing on my machine is happening at large. On latest macOS (10.13.3), with Pd 0.48-1, keyup and the keyup functionality of keyname appear broken, with the following behavior: > > (The quick and easy way to observe this is to open key-help.pd, then type. If your situation is like mine, you will never see the status outlet of keyname report an output of 1.) > > Pressing a key on my keyboard produces the expected response from key, but also produces an instantaneous response from keyup and from the status outlet of keyname (even though the key has not been released). That is, if I follow the status outlet of keyname with [sel 0 1], routing the bang from 1 to the left inlet of timer, and the bang from 0 to the right inlet of timer, then pressing and holding any key returns either a time of 0 between keyname’s issuance of the values 1 and 0, or a period reflecting the blocksize (1.45125); in either case, the keyup is issued long before the OS decides to start issuing key repeats. > > Actually, complicating this issue is this: regardless of whether System Preferences > Keyboard > Key Repeat is set to Off or on (any other value), Key Repeat does not appear to /ever/ occur in any application except Pd, and yet it will /always/ occur in Pd (but only for keyboard characters which can evoke alternate characters, i.e. ‘a’, which can evoke à, á, â, ä, æ, ã, å, or ā when held during text input in most apps except Pd). > > So, is the issue (keyup firing on key press) occurring for others, and if so, is it just Mac users, or does it occur cross-platform? > > Best regards, > Kevin -------- Dan Wilcox @danomatika <http://twitter.com/danomatika> danomatika.com <http://danomatika.com/> robotcowboy.com <http://robotcowboy.com/>

2 1

porting MSP externals - sysmem_newptrclear and sysmem_freeptr
by Ed Kelly 28 Jan '18

28 Jan '18

Hi Devs, I'm attempting to port a library of MSP externals for LPC analysis and resynthesis.I've come across these elements from the Max API: t_ptr sysmem_newptrclear (long size)andvoid sysmem_freeptr (void *ptr) My head has hit a block, and I need to learn about how Pd deals with these aspects of memory / pointer allocation.Do I need to just init a pointer in the code? enclosed is the original of one of the externals (mbc.blit~.c) and my first attempt at porting it (lpc_blit~.c) Enlightnment awaits...I hope.Ed _-_-_-_-_-_-_-^-_-_-_-_-_-_-_ For Lone Shark releases, Pure Data software and published Research, go to http://sharktracks.co.uk

2 1

Are keyup and keyname broken?
by Kevin Haywood 28 Jan '18

28 Jan '18

Hello, Apologies if this specific issue has been discussed before. I searched the archives and found related topics, but not this issue. I’d just like to find out if what I’m seeing on my machine is happening at large. On latest macOS (10.13.3), with Pd 0.48-1, keyup and the keyup functionality of keyname appear broken, with the following behavior: (The quick and easy way to observe this is to open key-help.pd, then type. If your situation is like mine, you will never see the status outlet of keyname report an output of 1.) Pressing a key on my keyboard produces the expected response from key, but also produces an instantaneous response from keyup and from the status outlet of keyname (even though the key has not been released). That is, if I follow the status outlet of keyname with [sel 0 1], routing the bang from 1 to the left inlet of timer, and the bang from 0 to the right inlet of timer, then pressing and holding any key returns either a time of 0 between keyname’s issuance of the values 1 and 0, or a period reflecting the blocksize (1.45125); in either case, the keyup is issued long before the OS decides to start issuing key repeats. Actually, complicating this issue is this: regardless of whether System Preferences > Keyboard > Key Repeat is set to Off or on (any other value), Key Repeat does not appear to /ever/ occur in any application except Pd, and yet it will /always/ occur in Pd (but only for keyboard characters which can evoke alternate characters, i.e. ‘a’, which can evoke à, á, â, ä, æ, ã, å, or ā when held during text input in most apps except Pd). So, is the issue (keyup firing on key press) occurring for others, and if so, is it just Mac users, or does it occur cross-platform? Best regards, Kevin

1 0

What should I look for after "Floating point exception"?
by Ed Kelly 23 Jan '18

23 Jan '18

I'm working on a monster of a sequential polyrhythm extern. It crashes, giving "Floating point exception" in the terminal. Can anyone tell me what this means, and what I should be looking for in the code, to fix? A first feeler for why it does this!Ed _-_-_-_-_-_-_-^-_-_-_-_-_-_-_ For Lone Shark releases, Pure Data software and published Research, go to http://sharktracks.co.uk

2 2

Re: [PD-dev] [pure-data/pd-lib-builder] Target architecture detection for MinGW 32 and 64 (#38)
by IOhannes m zmölnig 21 Jan '18

21 Jan '18

(taking this to the pd-dev list, and off the pd-lib-builder issue on github) On 01/20/2018 09:55 PM, Lucas Cordiviola wrote: > @katjav: May be you know better whats happening. > > The thing started while we where working on the Autotools branch https://github.com/pure-data/pure-data/pull/275 > > While testing that we also get a working 64bit Pd. On those tests @umlaeute suggested we try https://github.com/pure-data/pure-data/pull/7. > > I never tried to build a w64 Pd without the` ./configure --with-wish=wish86.exe CPPFLAGS=-DPD_LONGINTTYPE=__int64` > > It happens that if I build an external without `-DPD_LONGINTTYPE=__int64` Pd crashes when dsp is turned on (with the external under test) > > Is still unknown what happens if w64 Pd is built without the CPPFLAG. it's pretty well known what is happening when building Pd on 64bit windows without the CPPFLAG: Pd tries to cram a 64bit pointer into a 32bit integer, dropping 32 crucial bits, leaing to a segfault when trying to access the data at the given position. gfmdsar IOhannes

1 0

Re: [PD-dev] GUI objects with libpd but without tcl/tk
by Dan Wilcox 16 Jan '18

16 Jan '18

See the PdDroidParty [loadsave] object: http://www.droidparty.net <http://www.droidparty.net/> I've also implemented it for PdParty by listening for messages sent to the specific object and then opening a browser. > On Jan 15, 2018, at 11:45 AM, pd-dev-request(a)lists.iem.at wrote: > > From: Pierre Guillot <guillotpierre6(a)gmail.com <mailto:guillotpierre6@gmail.com>> > To: pd-dev <pd-dev(a)iem.at <mailto:pd-dev@iem.at>> > Subject: Re: [PD-dev] GUI objects with libpd but without tcl/tk > Message-ID: > <CALPKkho7UPi-8WVixcSgfDFmMjLtGNhYviXg4insjMWwaHYk+w(a)mail.gmail.com <mailto:CALPKkho7UPi-8WVixcSgfDFmMjLtGNhYviXg4insjMWwaHYk+w@mail.gmail.com>> > Content-Type: text/plain; charset="utf-8" > > In fact, after activating my brain... I guess that the best solution would > be to implement a parallel mechanism where the user could send a message to > the pd instance through a specific symbol (something like "openpanel > the-default-path" to the symbol "mypdinstance") and the pd instance would > respond to the patch through another specific symbol (something like > "the-selected-path" to the symbol "openpanel"). Sorry for the noise... > > Cheers, > Pierre -------- Dan Wilcox @danomatika <http://twitter.com/danomatika> danomatika.com <http://danomatika.com/> robotcowboy.com <http://robotcowboy.com/>

2 2

configure + makefiles
by Dan Wilcox 16 Jan '18

16 Jan '18

In thinking again about Pd's build system(s) while working on the Windows build, I think there is a good solution which would allow combining both the Makefiles and autoconf. As far as I see it, the pros/cons of the various systems we have now are: * makefiles: easy to understand & fast, less adaptable / configurable * autotools: easy to use and configurable, harder to understand, lots of files I think we definitely want to pros of both: easy to understand, easy to use, configurable. Since the autotools (autoconf. automake, etc) are built around make already, it's not like these systems are mutually exclusive. The Makefile.am automaker files are used to generate Makefile.in files which are used to generate the Makefiles. I propose that we transition from using automake directly and write our own Makefile.ins. This way more of the actually logic is shifted to the Makefile.in files themselves while we still get the configuration variables from the configure script. The con for this is we lose a lot of the boilerplate stuff that automaker handles but would gain Makefiles which basically look like Miller's original makefiles. Additionally, the cross platform stuff like file lists and install rules, etc could be defined in a shared Makefile and we keep the platform specific makefiles. Last, we would be able to handle custom rules & approaches by not having to fight automake for things like generating binaries in the bin folder and making a "sourceful" release build in place with ./configure. There are a number of projects which use this approach, including Tcl/Tk. -------- Dan Wilcox @danomatika <http://twitter.com/danomatika> danomatika.com <http://danomatika.com/> robotcowboy.com <http://robotcowboy.com/>

3 6

Windows registry keys
by Dan Wilcox 16 Jan '18

16 Jan '18

As far as I can tell, the PD registry locations are: * core (src/s_file.c): HKEY_CURRENT_USER\Software\Pure-Data * gui (tcl/pd_guiprefs.tcl): HKEY_CURRENT_USER\Software\Pure-Data\org.puredata.pd.pd-gui Does this mean the gui preferences are deleted when when the core preferences are deleted since they seem to be a subkey? -------- Dan Wilcox @danomatika <http://twitter.com/danomatika> danomatika.com <http://danomatika.com/> robotcowboy.com <http://robotcowboy.com/>

1 0

dynamic clone?
by Dan Wilcox 15 Jan '18

15 Jan '18

Would it be possible to make [clone] take a create and/or destroy message so it's internals could be dynamically created? I imagine just a list like the creation args. enohp ym morf tnes ----------- Dan Wilcox danomatika.com robotcowboy.com

3 2

GUI objects with libpd but without tcl/tk
by Pierre Guillot 14 Jan '18

14 Jan '18

Hi devs, What would be the best approach to use the GUI objects from x_gui.c (openpanel, savepanel and key) using libpd without tcl/tk ? For key I only need to send messages to gensym(#key) but for the other I don't know. I don't really understand how sys_startgui works. If I can use this method to "catch" the communication that normally would be destined to tcl/tk? Another approach would be to rewrite the sys_vgui function and to catch the openpanel and savepanel messages but I would prefer not to hack the pd code. Perhaps a solution would be to add in the Pd source #ifndef PD_EXTERN_GUI around the function that allows to rewrite the function without interfering with the main code but if there is a solution that doesn't necessitate to touch the Pd sources at all it would be better. Cheers, Pierre

1 1

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 January 2018