https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7304
--- Comment #1 from Karsten Bräckelmann <guent...@rudersport.de> --- Thanks Hanno for the report. However, I believe the described issue to not be a problem. First of all, the memcmp function does not "read all bytes of both variables", but compares the source strings' raw memory byte-for-byte, up until the first mismatching byte or the maximum byte number is reached. The spamc.c longoptions[] name members are null terminated C strings. In the case of a command-line argument longer than the compared-to longopts[i].name string, that null char will definitely terminate memcmp, preventing out-of-bounds operation. (With no earlier character mismatch, that null char is guaranteed to mis-match because null cannot be included in the command-line argument.) I believe this to be not a bug. Not closing just yet, though, open for additional feedback. -- You are receiving this mail because: You are the assignee for the bug.
