Hi all, I just found the following on the net: http://www.alphalinux.org/archives/axp-list/March2001/0337.shtml It seems that the possible SEGV in memchr is well known if an overestimated third argument (the maximum length to be scanned) is passed in. Therefore I'd say that the bug is acutually in a subroutine of strstr that passes a too large value to memchr (I think it was str_two_way_short_needle), which is unportable code. Best, Michael
Attachment:
pgppIUhJhviTv.pgp
Description: PGP signature