[Python-Dev] [Fwd: segfault in sre on 64-bit plats] (original) (raw)

Trent Mick trentm@ActiveState.com
Thu, 17 Aug 2000 14:41:04 -0700

On Thu, Aug 17, 2000 at 09:17:42AM +0800, Mark Favas wrote:

[Trent] > This test on Win32 and Linux32 hits the recursion limit check of 10000 in > SREMATCH(). However, on Linux64 the segfault occurs at a recursion depth of > 7500. I don't want to just willy-nilly drop the recursion limit down to make > the problem go away. >

Sorry for the delay - yes, I had these segfaults due to exceeding the stack size on Tru64 Unix (which, by default, is 2048 kbytes) before Fredrick introduced the recusrion limit of 10000 in sre.c. You'd expect a 64-bit OS to use a bit more bytes of the stack when handling recursive calls, but your 7500 down from 10000 sounds a bit much - unless the

Actually with pointers being twice the size the stack will presumably get comsumed more quickly (right?), so all other things being equal the earlier stack overflow is expected.

stack size limit you're using on Linux64 is smaller than that for Linux32 - what are they?

------------------- snip --------- snip ---------------------- #include <sys/time.h> #include <sys/resource.h> #include <unistd.h>

int main(void) { struct rlimit lims; if (getrlimit(RLIMIT_STACK, &lims) != 0) { printf("error in getrlimit\n"); exit(1); } printf("cur stack limit = %d, max stack limit = %d\n", lims.rlim_cur, lims.rlim_max); return 0; } ------------------- snip --------- snip ----------------------

On Linux32:

cur stack limit = 8388608, max stack limit = 2147483647

On Linux64:

cur stack limit = 8388608, max stack limit = -1

Trent

-- Trent Mick TrentM@ActiveState.com