Re: [PD] [PD-announce] pd 0.43-1 test 5 released
On Fri, 2011-11-11 at 11:05 -0500, Hans-Christoph Steiner wrote: On Nov 11, 2011, at 7:58 AM, Roman Haefeli wrote: Sorry for this vague bug report. I thought it's better to mention it even when I cannot provide enough useful information yet. Can you reliably reproduce the bug? Having a complicated patch with steps to reproduce the bug every time is still helpful. If the bug is intermittent, have you confirmed that it really doesn't happen in test4? In the meanwhile I posted a proper bug report: https://sourceforge.net/tracker/index.php?func=detailaid=3437312group_id=55736atid=478070 Yeah, it happens every time and it is indeed only happening on test5, but not on test4. Luckily, the patch to trigger the crash is not so complicated anymore. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
Thanks for reporting this... I'll get to work on it. I've already got another probably similar bug report too, so having both might help me find a more general problem. cheers Miller On Sun, Nov 13, 2011 at 03:58:41PM +0100, Roman Haefeli wrote: On Fri, 2011-11-11 at 11:05 -0500, Hans-Christoph Steiner wrote: On Nov 11, 2011, at 7:58 AM, Roman Haefeli wrote: Sorry for this vague bug report. I thought it's better to mention it even when I cannot provide enough useful information yet. Can you reliably reproduce the bug? Having a complicated patch with steps to reproduce the bug every time is still helpful. If the bug is intermittent, have you confirmed that it really doesn't happen in test4? In the meanwhile I posted a proper bug report: https://sourceforge.net/tracker/index.php?func=detailaid=3437312group_id=55736atid=478070 Yeah, it happens every time and it is indeed only happening on test5, but not on test4. Luckily, the patch to trigger the crash is not so complicated anymore. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
On Sun, Nov 13, 2011 at 03:58:41PM +0100, Roman Haefeli wrote: In the meanwhile I posted a proper bug report: https://sourceforge.net/tracker/index.php?func=detailaid=3437312group_id=55736atid=478070 Is Valgrind reporting its first Invalid Write in the same spot, or not ? Can you also post Valgrind backtraces for all the Invalid Writes ? Perhaps even any related warnings (but note that Valgrind always provide a bunch of unrelated warnings). __ | Mathieu BOUCHARD - téléphone : +1.514.383.3801 - Montréal, QC___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
Hi Matju On Sun, 2011-11-13 at 12:56 -0500, Mathieu Bouchard wrote: On Sun, Nov 13, 2011 at 03:58:41PM +0100, Roman Haefeli wrote: In the meanwhile I posted a proper bug report: https://sourceforge.net/tracker/index.php?func=detailaid=3437312group_id=55736atid=478070 Is Valgrind reporting its first Invalid Write in the same spot, or not ? Why are you asking? Are you going to help fix this? Can you also post Valgrind backtraces for all the Invalid Writes ? How does it help to also have the Valgrind backtrace? Unfortunately, I'm not familiar with Valgrind. Probably you could give me a hand with it or even create the backtrace yourself (?). I think I posted all the necessary information to make Pd crash. Perhaps even any related warnings (but note that Valgrind always provide a bunch of unrelated warnings). Is it something like that it considers Pd's symbol table to be a memory leak because it gets never freed as long as Pd is running? I'm not sure, but I vaguely remember you saying something the like. Roman ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
Le 2011-11-13 à 20:45:00, Roman Haefeli a écrit : Can you also post Valgrind backtraces for all the Invalid Writes ? How does it help to also have the Valgrind backtrace? In case the bug is more subtle than GDB can find. GDB finds only the final consequences, whereas Valgrind usually finds an early cause. For some bugs, it does not matter. I think I posted all the necessary information to make Pd crash. It's often not sufficient to cause a crash. Perhaps even any related warnings (but note that Valgrind always provide a bunch of unrelated warnings). Is it something like that it considers Pd's symbol table to be a memory leak because it gets never freed as long as Pd is running? I'm not sure, but I vaguely remember you saying something the like. That's only if you enable the leak-checker, and that thing is off by default, and should be kept off while looking for the cause of a crash. Why are you asking? Are you going to help fix this? Nevermind, then. __ | Mathieu BOUCHARD - téléphone : +1.514.383.3801 - Montréal, QC___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
Hi Miller On Sun, 2011-11-06 at 22:55 -0800, Miller Puckette wrote: Another test version... I still am working on 2 bugs. The biggest change is that I updated tcl/tk to version 8.5 in the Windows compiled package. It's on the usual http://crca.ucsd.edu/~msp/software.htm or: git clone git://pure-data.git.sourceforge.net/gitroot/pure-data/pure-data It seems that test5 is crashing in certain circumstances during patch editing, where test4 is not crashing under the exact same circumstances. What happens is that when I change the arguments of a certain abstraction and the click to the empty canvas to let it re-instantiate with the new arguments, Pd segfaults. Also, when editing that said abstraction (many instances are used in the patch) and then saving it, Pd segfaults. Since this happens in a not so small project, isolating the problem and creating a small test patch could be quite time consuming. Would a gdb traceback already help to track it down? This on Ubntu 11.04 with Pd-0.43.1test5. Sorry for this vague bug report. I thought it's better to mention it even when I cannot provide enough useful information yet. Roman ___ Pd-announce mailing list pd-annou...@iem.at http://lists.puredata.info/listinfo/pd-announce ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
On Nov 11, 2011, at 7:58 AM, Roman Haefeli wrote: Hi Miller On Sun, 2011-11-06 at 22:55 -0800, Miller Puckette wrote: Another test version... I still am working on 2 bugs. The biggest change is that I updated tcl/tk to version 8.5 in the Windows compiled package. It's on the usual http://crca.ucsd.edu/~msp/software.htm or: git clone git://pure-data.git.sourceforge.net/gitroot/pure-data/pure-data It seems that test5 is crashing in certain circumstances during patch editing, where test4 is not crashing under the exact same circumstances. What happens is that when I change the arguments of a certain abstraction and the click to the empty canvas to let it re-instantiate with the new arguments, Pd segfaults. Also, when editing that said abstraction (many instances are used in the patch) and then saving it, Pd segfaults. Since this happens in a not so small project, isolating the problem and creating a small test patch could be quite time consuming. Would a gdb traceback already help to track it down? This on Ubntu 11.04 with Pd-0.43.1test5. Sorry for this vague bug report. I thought it's better to mention it even when I cannot provide enough useful information yet. Can you reliably reproduce the bug? Having a complicated patch with steps to reproduce the bug every time is still helpful. If the bug is intermittent, have you confirmed that it really doesn't happen in test4? .hc 'You people have such restrictive dress for women,’ she said, hobbling away in three inch heels and panty hose to finish out another pink-collar temp pool day. - “Hijab Scene #2, by Mohja Kahf ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [PD-announce] pd 0.43-1 test 5 released
I was hoping your fix for the (closing patches sometimes gave Tk errors) would fix this, but alas it did not. This is a crasher bug, so I marked it high priority: https://sourceforge.net/tracker/?func=detailaid=3434543group_id=55736atid=478070 Thanks for the _WIN32 change, it'll help reduce some really annoying work as more porting happens. Here is one more bugfix that should be included in 0.43.1. It fixed a key bug but has minimal risk of regression: https://sourceforge.net/tracker/?func=detailaid=3424988group_id=55736atid=478072 .hc On Nov 7, 2011, at 1:55 AM, Miller Puckette wrote: Another test version... I still am working on 2 bugs. The biggest change is that I updated tcl/tk to version 8.5 in the Windows compiled package. It's on the usual http://crca.ucsd.edu/~msp/software.htm or: git clone git://pure-data.git.sourceforge.net/gitroot/pure-data/pure-data cheers Miller ___ Pd-announce mailing list pd-annou...@iem.at http://lists.puredata.info/listinfo/pd-announce Programs should be written for people to read, and only incidentally for machines to execute. - from Structure and Interpretation of Computer Programs ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] [PD-announce] pd 0.43-1 test 5 released
Another test version... I still am working on 2 bugs. The biggest change is that I updated tcl/tk to version 8.5 in the Windows compiled package. It's on the usual http://crca.ucsd.edu/~msp/software.htm or: git clone git://pure-data.git.sourceforge.net/gitroot/pure-data/pure-data cheers Miller ___ Pd-announce mailing list pd-annou...@iem.at http://lists.puredata.info/listinfo/pd-announce ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list