[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: Committed now. trunk: r69376, r69377 release26-maint: r69378 py3k: r69380 release30-maint: r69381 (hand-merged, as you may notice I forgot to split the patch in two again in py3k) -- resolution: - fixed status: open - closed ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Robert Hancock hanc...@sedsystems.ca added the comment: That's not really a meaningful difference, though.. if the application uses this code continuously then the conditions will pile up in memory until it fills up. -- nosy: +robhancock1 -robhancock ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: That's not really a meaningful difference, though.. if the application uses this code continuously then the conditions will pile up in memory until it fills up. I'm not trying to discourage you to do the complete fix (for these two problems), but this API was made public only in tcl 8.3.1 as can be seen at http://tcl.cvs.sourceforge.net/viewvc/tcl/tcl/ChangeLog?revision=1.226view=markuppathrev=core-8-3-1 Still, the call fix should be backported to python 2.6 as it doesn't depend on changing supported tcl versions and is also more severe. ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Martin v. Löwis mar...@v.loewis.de added the comment: Still, the call fix should be backported to python 2.6 as it doesn't depend on changing supported tcl versions and is also more severe. That sounds like a good approach. If you can split the patch into two, backporting only one of them - go ahead! ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Martin v. Löwis mar...@v.loewis.de added the comment: I think the patch is correct. gpolo, what's your opinion? ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: Do we still want to say _tkinter.c supports tcl/tk from version 8.2 and newer ? If yes, then we should add a no-op Tcl_ConditionFinalize when using tcl older than 8.3, because that was when this function got added. Besides that, the patch looks fine to me too (except I prefer Tcl_Condition *cond over Tcl_Condition* cond). ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Martin v. Löwis mar...@v.loewis.de added the comment: Do we still want to say _tkinter.c supports tcl/tk from version 8.2 and newer ? If yes, then we should add a no-op Tcl_ConditionFinalize when using tcl older than 8.3, because that was when this function got added. Good point. If we want to fix this in 2.6, we *have* to say that. If we ignore 2.6, and only fix 2.7, we could require a newer Tcl version. ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: It is much more important to fix the memory leak anyway, it is possible to notice real problems using test_tkleak1.py and watching the process with ps or something else provided by the platform. Besides, not calling ConditionFinalize doesn't introduce memory leaks, so it is not in par with issue's title. But I'm unsure about the existence of systems using Python as the main language which are also using tkinter and at the same time depend on tcl/tk 8.2 for some reason. So, to me, it would seem fine to start requiring tcl/tk 8.3 (or even newer!) starting with 2.7 and 3.1 ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Martin v. Löwis mar...@v.loewis.de added the comment: Besides, not calling ConditionFinalize doesn't introduce memory leaks Are you sure about this? On Unix, TclpFinalizeCondition does pthread_cond_destroy(pcondPtr); ckfree((char *) pcondPtr); both of which release memory (IIUC). Likewise for Windows, except that it calls DeleteCriticalSection. ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: Besides, not calling ConditionFinalize doesn't introduce memory leaks Are you sure about this? On Unix, TclpFinalizeCondition does pthread_cond_destroy(pcondPtr); ckfree((char *) pcondPtr); both of which release memory (IIUC). Likewise for Windows, except that it calls DeleteCriticalSection. I think I meant something more like it doesn't result in memory definitely lost because tcl will free these conditions when tcl finishes, but it does introduce memory leak in the sense that the memory isn't freed right after it is not longer used. ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Guilherme Polo ggp...@gmail.com added the comment: Sample attached for demonstrating the leak by missing a call to Tkapp_CallDeallocArgs. -- nosy: +gpolo Added file: http://bugs.python.org/file12373/test_tkleak1.py ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue1731706 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1731706] tkinter memory leak problem
Changes by Christian Heimes: -- versions: +Python 2.6 _ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue1731706 _ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com