Pd-dev December 2007

pd-dev@lists.iem.at

28 participants
143 discussions

[ pure-data-Patches-1852128 ] compile with -mno-cygwin
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1852128, was opened at 2007-12-17 07:21 Message generated for change (Comment added) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1852128&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: bugfix Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) >Assigned to: Hans-Christoph Steiner (eighthave) Summary: compile with -mno-cygwin Initial Comment: This allow to compile with cygwin shell, and might not be as clean as it was done by a makefile specialist ---------------------------------------------------------------------- >Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-12-19 14:10 Message: Logged In: YES user_id=564396 Originator: NO while the original post is rather cryptic, i guess this is for the pd-extended build-system (at least, according to the file-description it is about "externals/Makefile" which is pd-extended only) i therefore assign it to hans :-) ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1852128&group_…

1 0

[ pure-data-Patches-1374659 ] MinGW Makefile for Pd
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1374659, was opened at 2005-12-06 18:42 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1374659&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: feature Status: Open Resolution: None Priority: 1 Private: No Submitted By: Hans-Christoph Steiner (eighthave) >Assigned to: Miller Puckette (millerpuckette) Summary: MinGW Makefile for Pd Initial Comment: Attached is a Makefile for building Pd on Windows using MinGW. I have successfully build the HEAD of the MAIN branch of Pd in CVS with it. It goes with these instructions: http://puredata.org/docs/developer/win For now, this is kind of a hacked up cross between makefile.nt and makefile.in. Ideally, the MinGW specifics would be incorporated into the configure.in and makefile.in system, so that on all platforms, you could build Pd by doing: ./configure && make ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-10-23 01:22 Message: Logged In: YES user_id=27104 Originator: YES updated for 0.40.3 File Added: makefile.mingw ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2006-04-12 15:54 Message: Logged In: YES user_id=564396 now what to do with that one? the patch has been accepted but (afterwards) retracted by the committer... i guess, i close it for now and whenever the "real" configure/makefile changes go in it should be opened again. ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2006-04-06 02:59 Message: Logged In: YES user_id=27104 This is not quite ready to be included in Pd. For now, its just used to build Pd-extended. Soon, it will be turned into patches against configure.in and makefile.in so that MinGW will also work with ./configure && make install ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2006-03-23 23:29 Message: Logged In: YES user_id=313747 Took it (untested... I don't have mingw.) ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2005-12-20 01:09 Message: Logged In: YES user_id=27104 this is the makefile.mingw that was used to build Pd-0.38.4-extendedRC6.exe ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2005-12-19 23:53 Message: Logged In: YES user_id=27104 This new makefile compiles everything, including the externals and pd.com, the cmd.exe version (using the regular s_entry.c, no s_entry_com.c necessary, thanks to MinGW's -mwindows). ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1374659&group_…

1 0

[ pure-data-Patches-1848356 ] streamin/streamout doc updates, bugs, and a security issue
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1848356, was opened at 2007-12-11 04:29 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1848356&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: bugfix Status: Open Resolution: None Priority: 5 Private: Yes Submitted By: Russell Bryant (russellbryant) >Assigned to: Guenter Geiger (ggeiger) Summary: streamin/streamout doc updates, bugs, and a security issue Initial Comment: This issue has been marked private as it includes details on a remotely exploitable buffer overflow vulnerability in the code for [streamin~]. A while ago, I was looking around at different options for streaming audio in and out of Pd. I spent a short time looking at the code for [streamin~] and [streamout~]. I ended up not using them for my project. However, I made this set of changes while looking at the code. Since I moved on to something else, these changes aren't exactly complete, but point out some important things, so I think they may be worth putting in as is. Here is a summary of the changes that are included. stream.h changes: 1) doxygen style documentation updates 2) added an ifdef to protect against multiple or recursive inclusion of the header 3) Documentation of some of the fields in the data structures 4) Fixed potential alignment bugs by adding the packed attribute for the structure that defines frame headers. This attribute should always be used when it is used to directly interpret data from the network or when used to prepare data to send directly over the network. This section of the gcc manual has a tidbit about the attribute: http://gcc.gnu.org/onlinedocs/gcc-4.2.2/gcc/Variable-Attributes.html#Variab… 5) Point out the fact that the code is not endianness-safe. The code should eventually be updated to respect network byte order. 6) Point out that the version field of the frame header is ignored when parsing incoming frames. It should eventually be checked, but doesn't matter much until the format gets changed such that the version number gets bumped. 7) Change a list of #defines to an enum, for the sake of better coding practice. 8) Change the use of "int" in the frame header structure to explicitly list that the field is 32-bits by using the int32_t type. streamin~.c changes: 1) Fix a remotely exploitable buffer overflow that would allow remote execution of arbitrary code. The frame header includes 4 bytes to specify the frame length. However, the code only allocates enough space for the frame header and 256 bytes of payload data. The size portion of the frame header was not checked to ensure that its value was valid. This little hack should fix it. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1848356&group_…

1 0

[ pure-data-Patches-1852385 ] Improve strtokcpy() efficiency, prevent stack overflow
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1852385, was opened at 2007-12-17 16:07 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1852385&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: bugfix Status: Open Resolution: None Priority: 5 Private: No Submitted By: Russell Bryant (russellbryant) >Assigned to: Miller Puckette (millerpuckette) Summary: Improve strtokcpy() efficiency, prevent stack overflow Initial Comment: This patch provides a revised implementation of the strtokcpy() function in s_path.c. It provides the following benefits: 1) Prevent potential overflow of a stack buffer. This function did nothing to ensure that it didn't write past the end of the destination buffer. It is possible to cause this to happen by providing certain command line arguments that are longer than MAXPDSTRING. Also, there may be other ways to trigger this bug if namelist_append_files() is used anywhere beyond the uses I reviewed, which are the ones in pd/*.c. 2) Copy bytes from the string in the same loop that looks for the delimiter. This is simply for efficiency in that the string only has to be traversed once, instead of twice (one to find the delimiter, and the second to copy up to it). 3) Document the function using doxygen style tags. ---------------------------------------------------------------------- Comment By: Russell Bryant (russellbryant) Date: 2007-12-17 16:07 Message: Logged In: YES user_id=1942915 Originator: YES File Added: strokcpy.txt ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1852385&group_…

1 0

[ pure-data-Patches-1848295 ] Fix small potential stack overflow
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1848295, was opened at 2007-12-11 01:53 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1848295&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: bugfix Status: Open Resolution: None Priority: 5 Private: No Submitted By: Russell Bryant (russellbryant) >Assigned to: Miller Puckette (millerpuckette) Summary: Fix small potential stack overflow Initial Comment: Attached is a patch to fix another potential stack buffer overflow in s_main.c. I didn't send this over in private because I don't think it can be exploited in any useful way, so it's just a small bug. ---------------------------------------------------------------------- Comment By: Russell Bryant (russellbryant) Date: 2007-12-11 03:41 Message: Logged In: YES user_id=1942915 Originator: YES Also, let me just clarify my original summary just a little bit. You can cause the overrun of the buffer by providing a really long input string to the -schedlib command line option. It's just not exploitable in such a way that it would be considered a security issue. ---------------------------------------------------------------------- Comment By: Russell Bryant (russellbryant) Date: 2007-12-11 03:38 Message: Logged In: YES user_id=1942915 Originator: YES I'm sorry for the stupid typo. That code is actually only compiled in for a Windows build (whenever MSW is defined), so I didn't notice. ---------------------------------------------------------------------- Comment By: Russell Bryant (russellbryant) Date: 2007-12-11 03:36 Message: Logged In: YES user_id=1942915 Originator: YES File Added: filename_overflow.patch2.txt ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-12-11 02:18 Message: Logged In: YES user_id=27104 Originator: NO I am guessing there is a typo in that patch, since the new line also uses sprintf() instead of snprintf(): - sprintf(filename, "%s.dll", sys_externalschedlibname); + sprintf(filename, sizeof(filename), "%s.dll", sys_externalschedlibname); ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1848295&group_…

1 0

[ pure-data-Patches-1822010 ] hardcoded sample-type in soundfile-IO
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1822010, was opened at 2007-10-29 11:24 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1822010&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: None >Status: Closed >Resolution: Accepted Priority: 5 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Miller Puckette (millerpuckette) Summary: hardcoded sample-type in soundfile-IO Initial Comment: the current implementation in d_soundfile.c assumes that both t_sample and t_float are "float". this is not necessarily true. additionally it assumes that typeof(t_sample) and typeof(t_float) are identically. this is not necessarily true. based on the assumption that (number)arrays are of type t_float and signals are of type t_sample, the attached patch provides 2 extra functions for xferin/xferout that use different types for reading/writing soundfiles. basically it provides: soundfile_xferin_sample() (reading soundfile to signal) soundfile_xferin_float() (reading soundfile to table) soundfile_xferout_sample() (writing soundfile from signal) soundfile_xferout_float() (writing soundfile from table) ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-12-18 23:56 Message: Logged In: YES user_id=313747 Originator: NO taken. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1822010&group_…

1 0

[ pure-data-Patches-1822007 ] hardcoded sample-type in jack-transport
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1822007, was opened at 2007-10-29 11:19 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1822007&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: bugfix >Status: Closed >Resolution: Accepted Priority: 7 Private: No Submitted By: IOhannes m zmölnig (zmoelnig) Assigned to: Miller Puckette (millerpuckette) Summary: hardcoded sample-type in jack-transport Initial Comment: in s_audio_jack.c the sample-type of the jack transportlayer is assumed to be "float", which is a naive assumption. the correct type is "jack_default_audio_sample_t" (which might be either "float" or "double", depending on how jack was compiled) the attached patch respects the type-definition of both jack and pd's own t_sample. when the 2 types are the same, it just does a memcpy (this is similar as before, where it is assumed that both jack and pd use "float" internally) when they are different, it copies the data manually. LATER: improve the heuristic for finding out whether the two types are equal (currently only the sizeof() is checked) ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-12-18 23:56 Message: Logged In: YES user_id=313747 Originator: NO taken. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1822007&group_…

1 0

[ pure-data-Patches-1752248 ] cmd-click manipulates GUIs on Mac OS X
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1752248, was opened at 2007-07-11 23:53 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1752248&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: feature >Status: Closed >Resolution: Accepted Priority: 5 Private: No Submitted By: Hans-Christoph Steiner (eighthave) Assigned to: Miller Puckette (millerpuckette) Summary: cmd-click manipulates GUIs on Mac OS X Initial Comment: This patch just makes Cmd-Click do exactly what Ctrl-Click does on the other platforms. It's a one liner... ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-12-18 23:41 Message: Logged In: YES user_id=313747 Originator: NO taken. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1752248&group_…

1 0

[ pure-data-Patches-1738289 ] BUG: "View list" listview broken for arrays with $0 in name
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1738289, was opened at 2007-06-16 12:06 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1738289&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: None >Status: Closed Resolution: Fixed Priority: 3 Private: No Submitted By: Frank Barknecht (fbar) Assigned to: Miller Puckette (millerpuckette) Summary: BUG: "View list" listview broken for arrays with $0 in name Initial Comment: The "View list" function in the Properties context menu of arrays is broken, if the name of the array contains a $0. This only affects visible arrays in graphs, arrays in [table] objects work correctly with $0-names. See attached test patch for a demonstration. The error message given by Tk is: can't read "0": no such variable can't read "0": no such variable while executing "array_viewlist .gfxstub82e3a78 $0-array 0" invoked from within ".gfxstub82e3a78.listview invoke" ("uplevel" body line 1) invoked from within "uplevel #0 [list $w invoke]" (procedure "tk::ButtonUp" line 22) invoked from within "tk::ButtonUp .gfxstub82e3a78.listview" (command bound to event) ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-12-18 23:39 Message: Logged In: YES user_id=313747 Originator: NO taken. ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-18 11:30 Message: Logged In: YES user_id=564396 Originator: NO moved to "patches" section, to indicate that there is a fix for the problem ---------------------------------------------------------------------- Comment By: IOhannes m zmölnig (zmoelnig) Date: 2007-09-17 14:23 Message: Logged In: YES user_id=564396 Originator: NO btw, the bug appeared with all uses of $args (e.g. $1-table is as bad as $0-array) however, the tiny attached diff (against todays cvs of 0.41) fixes this bug by removing unneeded variables. File Added: array_viewlist.diff ---------------------------------------------------------------------- Comment By: Frank Barknecht (fbar) Date: 2007-06-18 11:09 Message: Logged In: YES user_id=569446 Originator: YES Please don't use spaces in filenames here, their names get mangled by Sourceforge. I deleted the new example patch to not confuse readers. I think the first example I posted already was illustrating the bug sufficiently. ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-06-17 18:49 Message: Logged In: YES user_id=27104 Originator: NO Yup, same for me on Mac OS X 10.4.9 using Pd-0.39.2-extended-rc4 and Pd 0.40-2. File Added: array with dollar zero.pd ---------------------------------------------------------------------- Comment By: Frank Barknecht (fbar) Date: 2007-06-17 11:14 Message: Logged In: YES user_id=569446 Originator: YES It seems any version is affected. I tested with CVS HEAD (0.4x) and pd-ext 0.39 rc-2. ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-06-17 03:31 Message: Logged In: YES user_id=27104 Originator: NO Which version of Pd? I hope it's not my bug ;) The version can be set in the "Group" pulldown menu in the bug report form. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1738289&group_…

1 0

[ pure-data-Patches-1688540 ] fix makefilename segfault when wrong argument type received
by SourceForge.net 19 Dec '07

19 Dec '07

Patches item #1688540, was opened at 2007-03-26 18:33 Message generated for change (Settings changed) made by zmoelnig You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1688540&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: bugfix >Status: Closed >Resolution: Accepted Priority: 5 Private: No Submitted By: Stephen Sinclair (radarsat1) Assigned to: Miller Puckette (millerpuckette) Summary: fix makefilename segfault when wrong argument type received Initial Comment: Currently, Pd segfaults if the "makefilename" object is configured to receive a symbol ("%s") and it receives a float argument. This patch uses strstr() to check if the sprintf string is going to accept a float, like so: strstr(s->s_name, "%s")==0 The result is saved in the t_makefilename structure. It is recalculated on the "set" message. When the wrong message type is received, a 0 or "" string is substituted, respectfully. While it is unfortunate to do a string search like this, it is not very good to have Pd segfaulting, so I see no alternative. ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-12-18 23:36 Message: Logged In: YES user_id=313747 Originator: NO taken. ---------------------------------------------------------------------- Comment By: Stephen Sinclair (radarsat1) Date: 2007-07-31 03:21 Message: Logged In: YES user_id=616837 Originator: YES Deleting original patch as requested. pd-cvs-makefilename-20070326.patch is replaced by makefilename_scanformat.patch ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-07-31 03:04 Message: Logged In: YES user_id=27104 Originator: NO Does this new patch replace pd-cvs-makefilename-20070326.patch? If so, pd-cvs-makefilename-20070326.patch should be deleted from this tracker. I committed it to packages/patches so it'll be automatically included in the Pd-0.40.3-extended nightly builds. ---------------------------------------------------------------------- Comment By: Stephen Sinclair (radarsat1) Date: 2007-07-30 20:23 Message: Logged In: YES user_id=616837 Originator: YES Attached is a more complete patch scanning the symbol argument more thoroughly for sprintf-compatible formatting. Also now differentiates between float and int arguments, and supports format flags. Could probably be extended to support multiple arguments. This patch also stops makefilename from spitting out (null) symbols. Patch generated with "diff -uw x_connective.c" File Added: makefilename_scanformat.patch ---------------------------------------------------------------------- Comment By: Stephen Sinclair (radarsat1) Date: 2007-07-30 20:19 Message: Logged In: YES user_id=616837 Originator: YES File Added: makefilename_crash.pd ---------------------------------------------------------------------- Comment By: Miller Puckette (millerpuckette) Date: 2007-07-30 01:50 Message: Logged In: YES user_id=313747 Originator: NO It would take more work to do it right: check that there's only one "%f" and no "%s" or vice versa, and further check for "%%f", etc... I figure on doing this someday if nobody else does. ---------------------------------------------------------------------- Comment By: Hans-Christoph Steiner (eighthave) Date: 2007-06-14 17:11 Message: Logged In: YES user_id=27104 Originator: NO I applied this patch to the Pd-0.39.2-extended branch and it failed to compile: x_connective.c: In function 'makefilename_new': x_connective.c:1213: error: 'struct _makefilename' has no member named 'x_accept_float' x_connective.c: In function 'makefilename_float': x_connective.c:1219: error: 'struct _makefilename' has no member named 'x_accept_float' x_connective.c: In function 'makefilename_symbol': x_connective.c:1229: error: 'struct _makefilename' has no member named 'x_accept_float' x_connective.c: In function 'makefilename_set': x_connective.c:1239: error: 'struct _makefilename' has no member named 'x_accept_float' My guess is that the patch needs to be a "diff -uw" format, which is much better at dealing with different line numbers, etc. Also, an example patch to demonstrate the bug is always useful. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1688540&group_…

1 0

Jump to page:

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 December 2007