E. Rosten wrote:
There are some reasons for doing this (bigish changes
in standards compliance in g++ 4 compared to g++ 3)
Reasons, yes. Good reasons, no.
What happens when GCC 4.1 comes out, and we find that it has still more
standards restrictions, just as every point release since 3.0 has? Do
you then start checking for 4.1 vs. 4.0?
Earlier posts had it right to begin with: it is always best to test for
specific features. Testing for version numbers is effectively turning
that into a test for a set of N features (where N is the number of
entries in GCC 4.0's ChangeLog), instead of just the one feature you're
actually interested in.
_______________________________________________
Autoconf mailing list
Autoconf@xxxxxxx
http://lists.gnu.org/mailman/listinfo/autoconf