Re: CVS commit: src

[Jukka Ruohonen <jruohonen%iki.fi@localhost>]

On Sun, May 01, 2011 at 02:50:55PM +0000, Christos Zoulas wrote:
> I am not sure this is valid. It could return ENOTTY too for this case
> or even EINVAL. The opengroup manual page is not clear. Do you have a
> more authoritative source of what it should return in this case?

Indeed I think the underlying idea of the submitter of this PR was only that
EINVAL is not right. SUSv4 (2008 ed.) agrees with the manual page. In other
words, as getpid() is passed, the submitter in my opinion rightly concludes
that

        /*
         * NetBSD fails here, returning EINVAL which is
         * clearly wrong, since we passed to tcsetpgrp(3)
         * a value returned from getpid(3). So it definitely
         * was a "valid pgid value".
         * Only Solaris gets all of it right.
         */

But as I've written before, I wouldn't waste too many cycles with these
issues (as long as there is a note in documentation, a PR, or a test).

- Jukka.