Re: (P)review: sigset_t for more than 32 signals

[Garrett Wollman]

<<On Mon, 6 Sep 1999 10:51:41 -0400 (EDT), Peter Dufault <[EMAIL PROTECTED]> said:

> /* Now just insert the macros to make this work...
>  */

I'd be inclined to make sigemptyset() and sigfillset(), at a minimum,
functions only (not macros).

I'd define sigaddset something like this (function version):

int
sigaddset(sigset_t *ss, int sig)
{
        int index, bit;

        /*
         * Since most programs ignore the return value of sigaddset(),
         * we really do want to abort here rather than simply returning
         * an error.  This assertion ensures that we never act on an
         * uninitialized sigset_t.
         */
        assert(ss->sigset_size == sizeof *ss);

        if (sig < 1 || sig > NSIG) {
                errno = EINVAL;
                return 1;
        }
        sig--;
        index = sig / (CHAR_BIT * sizeof ss->sigset_bits[0]);
        bit = sig % (CHAR_BIT * sizeof ss->sigset_bits[0]);
        ss->sigset_bits[index] |= (1 << bit);
        return 0;
}

I actually don't see any reason to have sigaddset be defined as a
macro at all -- any program which calls it frequently enough to make a
difference is doing something wrong.

-GAWollman

--
Garrett A. Wollman   | O Siem / We are all family / O Siem / We're all the same
[EMAIL PROTECTED]  | O Siem / The fires of freedom 
Opinions not those of| Dance in the burning flame
MIT, LCS, CRS, or NSA|                     - Susan Aglukark and Chad Irschick


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message