On Mon Jul 02 23:49:33 2007, [EMAIL PROTECTED] wrote: > I don't have a good test case for this, but I've triggered the problem > with > some code that makes TGE sick.4 > > Program received signal SIGSEGV, Segmentation fault. > [Switching to Thread -1212335920 (LWP 8882)] > 0xb7dd7e4b in key_hash_STRING (interp=0x804e008, value=0x0, seed=3793) > at src/hash.c:110 > 110 if (s->hashval) { > (gdb) bt > #0 0xb7dd7e4b in key_hash_STRING (interp=0x804e008, value=0x0, > seed=3793) > at src/hash.c:110 > #1 0xb7dd8c3e in parrot_hash_get_bucket (interp=0x804e008, > hash=0x82ff5b8, > key=0x0) at src/hash.c:796 > #2 0xb7eb87d3 in Parrot_Hash_exists_keyed (interp=0x804e008, > pmc=0x828b534, > key=0x0) at ./src/pmc/hash.pmc:900 > #3 0xb7ce471b in Parrot_exists_i_p_kc (cur_opcode=0xb7be1230, > interp=0x804e008) at src/ops/pmc.ops:327 > #4 0xb7dd2fbd in runops_slow_core (interp=0x804e008, pc=0xb7be1230) > at src/runops_cores.c:184 > #5 0xb7dbd580 in runops_int (interp=0x804e008, offset=58) > at src/interpreter.c:769 > > Here's a patch; I'm not positive it's the right solution, but if you > can > somehow pass in a Key that doesn't do the right thing, you can trigger > the > problem. > > I hope this isn't too vague for someone to look over my shoulder. > > -- c >
chromatic - this patch doesn't apply any more (and seems to also have a lot of pheme in it for something regarding hashes.) Can you take a look at this see if your original concerns are still relevant? -- Will "Coke" Coleda