On Mar 01, 2005, at 13:21, Tomasz Kojm wrote:
On Tue, 1 Mar 2005 12:52:29 -0500 Dale Walsh <[EMAIL PROTECTED]> wrote:
Well Tomasz, it just crashed, I'm sitting at the gdb prompt, anything
you want to examine from here?
Program received signal EXC_BAD_ACCESS, Could not access memory. cli_bm_addpatt (root=0x205c80, pattern=0x21dd890) at matcher-bm.c:62 62 matcher-bm.c: No such file or directory. in matcher-bm.c (gdb) bt #0 cli_bm_addpatt (root=0x205c80, pattern=0x21dd890) at matcher-bm.c:62 #1 0x037eac80 in cli_parse_add (root=0x205c80, virname=0xbfffc240 "Worm.JS.Redlof.A (Clam)", hexsig=0xbfffc258 "272f74765f706f7075702e68746d272c272f7669657765722e63666d3f2a7069643d
Is the problem reproducible with clamscan or clamd?
In matcher-bm.c:62 change unsigned int size = DHASH(256, 256, 256); to unsigned int size = 131072;
Does it help?
So far, the FC1 looks like it doesn't hang any more and the Mac hasn't crashed since I've implemented the change (made an additional change as well).
What I don't understand is how increasing the value of "size" in "cli_bm_init()" from 62.25k (127488 - DHASH( 211 *256 + 37 * 256 + 256)) to 128k (131072) is going to make that much of a difference but perhaps I'm missing something that is transpiring somewhere else that has a dramatic affect based on this value.
Also, shouldn't I also adjust cli_bm_free() to be equivalent? (I did this already because it seemed to be the logical thing to do)
-- Dale
_______________________________________________ http://lurker.clamav.net/list/clamav-users.html