On Thursday 21 September 2006 08:33, Eric Blake wrote:
= > getopt(c, v, "r::");
=
= ...even though the "::" in this line is what makes this test program leave
= the realm of POSIX-specified behavior (and why GNU and BSD differ on
= opinion on what should happen).
It would seem to me, that it is gm4, that should change to rely on
Posix-specified behavior of standard functions (like getopt) instead of
expecting the platform's getopt to be "extended" or building its own
replacement for the standard function(s).
There was a point in (recent) history, when the concept of "Embrace and
Extend" was rather frowned upon...
-mi
