Ralf Wildenhues <address@hidden> writes:
Except now, HAVE_MALLOC, HAVE_REALLOC, and the results of AC_CHECK_DECLS
(HAVE_DECL_*) are special cases. They seem to be documented correctly
as such in the manual, but I think a mention of these in the NEWS item
would be helpful, given that you intend to keep them that way.
(And no, I don't know the reason for each of these special cases; at
least the AC_CHECK_DECLS one looks like it's very much intended so.)
Yes, the results of AC_CHECK_DECLS have always been a special case dating
back to when this was a GCC macro (I think that's where it started),
because it's expressing trinary information.