Hi folks, last night I spotted a memory leak in Edje (with the help of valgrind).
When the Edje smart object is removed from the Evas (resp. when it's deleted) it isn't free()'d correctly, it seems. The bug sits in _edje_collection_free() - it seems only parts collections are leaking, programs aren't affected. I've looked through the code but couldn't find the cause of the problem (I don't know the edje and eet code at all) so I'm sending this stuff to the list. ==12106== 8 bytes in 8 blocks are definitely lost in loss record 6 of 54 ==12106== at 0x40029A2D: malloc (in /usr/lib/valgrind/vgskin_memcheck.so) ==12106== by 0x40429D93: eet_data_get_string (eet_data.c:955) ==12106== by 0x4042A09E: eet_data_get_type (eet_data.c:1074) ==12106== by 0x4042AE24: eet_data_descriptor_decode (eet_data.c:1679) ==12106== by 0x4042AE47: eet_data_descriptor_decode (eet_data.c:1688) ==12106== by 0x4042AEFD: eet_data_descriptor_decode (eet_data.c:1757) ==12106== by 0x4042AA0F: eet_data_read (eet_data.c:1478) ==12106== by 0x402511DB: _edje_file_add (edje_load.c:402) ==12106== by 0x4025086B: edje_object_file_set (edje_load.c:31) ==12106== by 0x8048C4B: main (example.c:41) For testing purposes, I use Ben's example edje program. I enhanced it so that the various shutdown functions are actually called. Find the modified version at http://code-monkey.de/example.c The EDC I used is available at http://www.cuddletech.com/edje/edc/crossfade.edc Can someone please dig in the Edje code and try to find the bug? -- Regards, Tilman ------------------------------------------------------- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html _______________________________________________ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
