For the last couple of days I've been getting some pretty bad segfaults. I can reproduce it every time, by doing certain things to one of the shelves (adding modules, moving it to a different layer). When debugging, most of the time, the stack trace passes through the winselector module. Upon further examination, I saw that the obtained data from an ecore (e) event, which in this case was the instance of the gadget, was garbled. Most of the time, the bd_icon Evas_Object member pointed to an unallocated memory.
I don't think the winselector code itself overwrites that bit of memory, or does some excessive free'ing, since there's equal amount of evas_object_del and e_border_icon_add. So I took to valgrind to figure out what was going on. Too bad I've no idea how to read valgrind's output. So I'm posting to the list, in hope that someone with more experience in valgrind can help me out on this. http://pastie.caboo.se/189585 ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
