sloane@noscvax.UUCP (04/16/84)
Thanks much for the many responses to my query about the ctime()
argument types.
I received several responses:
1. Lint is wrong, ignore it and type the argument as a long
2. Lint is right, the argument should be declared as a time_t
and on a VAX it should be an int
3. Lint is right and sys/types.h is wrong, the argument should
be declared as a time_t and a time_t should be a long in
sys/types.h
Answer 3 (change sys/types.h to type time_t as a long) is the
correct one, which now brings up another question (which in my
profound ignorance I will ask of you ladies and gents): but first
some background.
The most cogent reponse to my ctime() question was that the
time_t typedef is designed to make calls to ctime() portable to
other machines; that is, by using a time_t declaration and
#including <sys/types.h> I would be doing things 'properly', as
the types.h file is modified to be appropriate to the machine
architecture it appears on. I had not heard of a time_t (or of
sys/types.h, I told you I was ignorant) until I started
scrounging around trying to solve my original ctime() vs. lint
problem.
So now my question: Just where in all that 4.2bsd
'documentation' is the proper use of types.h mentioned? I assume
that there are SEVERAL places where I should routinely use
types.h if I want to do things correctly, but is the only way to
find out by inspection? There was no mention of types.h in the
ctime() man page, I was just lucky to stumble onto it... And are
there other #include files that I should know about? I'd
appreciate pointers...
Send 'em to me via e-mail and if I get enough, maybe I'll post it
to the net.unix people...
"I must be blind, 'cause I sure as hell can't C..."
Gary K. Sloane
c/o Naval Ocean Systems Center
COTD Building 1 Room B205
San Diego, California 92152
MILNET: sloane@nosc
UUCP: ...{sdcsvax}!noscvax!sloane
DDD: (619) 225-8401 x391