Bruno Haible <[email protected]> writes: > Yes, this was precisely the topic of the two threads that started at > <http://lists.gnu.org/archive/html/bug-gnulib/2011-06/msg00241.html> > and <http://lists.gnu.org/archive/html/bug-gnulib/2011-06/msg00289.html>. > > The root of the problem is that we have two ways to define PATH_MAX: > - <limits.h> defines it only if it can be a literal constant, and keeps > it undefined on the Hurd, > - "pathmax.h" used to provide a definition always. > And these two semantics get messed up. > > There was some, but not complete agreement, that we should change the > semantics the "pathmax.h" so that it matches what <limits.h> does in > POSIX. See the last two mails at > <http://lists.gnu.org/archive/html/bug-gnulib/2011-06/msg00328.html> > <http://lists.gnu.org/archive/html/bug-gnulib/2011-06/msg00329.html>. > > > So here's now the proposed patch once again.
I like this approach. After applying it, getcwd builds and passes selftests [1], so please push it. /Simon [1] http://autobuild.josefsson.org/gnulib/log-201108042248625702000.txt
