On Thursday 21 September 2006 23:54, Eric Blake wrote:
= No, it is FreeBSD that is broken. If you use getopt_long to implement the
= POSIX requirements of "pr -s _", the GNU version complies whether or not
= POSIXLY_CORRECT is set (POSIX requires it to interpret -s with no [...]
Where did pr enter the picture? We are talking about gm4 here...
= > Do I have you converted yet?
=
= Nope. All you have managed to do is convince me that GNU m4 and gnulib
= should continue rejecting the current FreeBSD getopt_long implementation
= as broken.
So, respecting the POSIXLY_CORRECT makes something broken, while ignoring it
makes it correct?
Where is the standard for getopt_long documented, anyway?
-mi
