eichin@mit.edu (Mark W. Eichin) (06/15/91)
I built perl-4.010 with gcc-1.40, all options defaulted (in particular, -O). make test worked fine, but h2ph core dumps on reloc.h (* and sys/* were otherwise fine): (gdb) where #0 0x500f016 in nxzonefreenolock () #1 0x500ee48 in free () #2 0x2c134 in safefree () #3 0x235c0 in str_replace () #4 0x9b32 in do_subst () #5 0x14a9e in eval () #6 0x3c86 in cmd_exec () #7 0x4080 in cmd_exec () #8 0x4314 in cmd_exec () #9 0x4314 in cmd_exec () #10 0x1d28c in main () Running in perl -d leads to a core dump within the last two or three lines of the input file (I can't tell but I don't think the -d is affecting it.) -w only pointed out an example of $t not being initialized, which isn't really surprising given the layout of h2ph. Unfortunately, as is I can't get any more data with the native version of gdb (which is based on gdb-3.1, but they *haven't* released source for it... sigh...) And because of the wierd shared library implementation on the NeXT, I can't drop in the GNU malloc, or the version I cooked up with "paranoid" enhancements (range checking, scribble checking, and a few other things.) _Mark_ <eichin@athena.mit.edu> MIT Student Information Processing Board Watchmaker Computing <eichin@watch.com>