James Sutherland jas88 at
Fri Mar 10 11:27:20 GMT 2000

On Fri, 10 Mar 2000, Peter Samuelson wrote:
> [James Sutherland]
> > Obviously, if crypt() CAN return NULL, for whatever reason, then it
> > must be properly handled and checked for.
> True -- unless the only time it will return NULL is for invalid input.
> Then it is sufficient to verify that it will never be given invalid
> input.  In the case of crypt(), that shouldn't be too hard.

Assuming that IS the problem here, then yes. I'll take a look at the UFC
source later today, and see what could cause it to return NULL.


More information about the samba-technical mailing list