|RDP on Minix-386 email@example.com (1997-01-26)|
|Re: RDP on Minix-386 aswNS@hamp.hampshire.edu (1997-01-29)|
|Re: RDP on Minix-386 firstname.lastname@example.org (1997-01-30)|
|From:||aswNS@hamp.hampshire.edu (Albert S Woodhull)|
|Date:||29 Jan 1997 11:47:30 -0500|
|Organization:||Hampshire College, Amherst MA|
> One of the porting challenges was when the ACK compiler protested that
> some functions were called with more arguments than its limit.
> Although I have gotten past that problem, I am curious to know if I am
> correct in the assumption that the ANSI spec. explicitly permits a
> limit for the number of arguments to a variadic function? If so, I
> can report the problem as a legitimate "bug" which needs my fix :-).
This is an interesting question, to which I assumed the answer would
be yes, of course there is a limit. However, in tracking down all
references to argument lists and variable argument lists in K&R 2nd
ed. (1988), Plauger and Brodie's Standard C Reference (1996), Lewine's
POSIX Programmer's Guide (1992) and the Posix standard itself I can't
find any discussion of a limit to the number of arguments.
However, POSIX refers to the ARG_MAX macro, which defines the number
of bytes available for the combined argument and environment lists for
an exec(). POSIX calls for this to have a minumum size of 4096. Larger
values are permitted, and the sysconf() call can be used to find the
> I am also curious to know if there is any interest within this group
> in the results. In particular, I could use some help in fleshing out
> the C grammar to the point where it might serve as a "lint" program or
I certainly think a lint for Minix would be of interest.
Albert S. Woodhull
Hampshire College, Amherst, MA
Return to the
Search the comp.compilers archives again.