Re: crypt and null termination

Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> It seems to say that the salt passed to crypt should be null-terminated,

Hmm. The HPUX man page for crypt() just says that
salt is a two-character string chosen from the set [a-zA-Z0-9./]
which I think is the traditional spec. Looks like BSD has adopted some
local extensions.

Note that the BSD page specifies that the extended salt format starts
with '_', which is not one of the allowed characters in the traditional
format. I bet they check that before trying to fetch more than 2 bytes.
The second paragraph you quote doesn't say anything about null
termination.

Still, it wouldn't be a bad idea to add a null byte ... couldn't hurt.

regards, tom lane