On Wed, 2007-09-05 at 19:14 +0800, Halton Huo wrote: > Hello there, > > I just upgrade sqlite from 3.4.1 to 3.5.0, and run tracker, which use > sqlite to save data. > > I got core dump everytime like following: > > [EMAIL PROTECTED] ([EMAIL PROTECTED]) signal SEGV (no mapping at the fault > address) in t_splay at > 0xfea296e2 > 0xfea296e2: t_splay+0x0022: movl 0x00000008(%ebx),%ecx > Current function is tracker_exec_proc > 1719 rc = sqlite3_step (stmt); > (dbx 2) where > current thread: [EMAIL PROTECTED] > [1] t_splay(0x84ce2b0), at 0xfea296e2 > [2] t_delete(0x84ce2b0), at 0xfea295bd > [3] realfree(0x84ce230), at 0xfea292d0 > [4] cleanfree(0x0), at 0xfea29930 > [5] _malloc_unlocked(0x78, 0x0, 0x70, 0xfece8538, 0x2f, 0x81de080), at > 0xfea28dba > [6] malloc(0x78), at 0xfea28cbf > [7] sqlite3_malloc(0x70), at 0xfecac661 > [8] sqlite3MallocZero(0x70), at 0xfecac116 > [9] allocateCursor(0x825be30, 0x0, 0x0), at 0xfecc1ad6 > [10] sqlite3VdbeExec(0x825be30), at 0xfecc63d1 > [11] sqlite3Step(0x825be30), at 0xfecc7377 > [12] sqlite3_step(0x825be30), at 0xfecc74f5 > =>[13] tracker_exec_proc(db_con = 0x81c5b78, procedure = 0x80ed1e0 > "GetServiceID", param_count = 2, ... = 0x84b9950, ...), line 1719 in > "tracker-db-sqlite.c" > [14] tracker_db_get_file_info(db_con = 0x81c5b78, info = 0x824e710), > line 180 in "tracker-db.c" > [15] process_files_thread(), line 1250 in "trackerd.c" > [16] g_thread_create_proxy(0x81c5040), at 0xfebc03fa > [17] _thr_setup(0xfe870a00), at 0xfea7d952 > [18] _lwp_start(), at 0xfea7dbb0 > > I have a page http://www.sqlite.org/34to35.html, but still do not know > where to start. Any thoughts?
This is heap corruption. Which means - problem could be pretty much anywhere :) If you can run the program under valgrind and post the results, it will quite likely shed more light. Dan. ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------