OK, final patch. This was the most painful, because I hate autoconf quoting, way too many brackets and parentheses. But the patch silences the remaining warnings about AC_TRY_COMPILE, AC_TRY_LINK, and AC_TRY_RUN. And it is verified to be a no-op in terms of the generated configure file (except the one comment changed on purpose, and one extraneous newline character removed).
Hopefully this makes gmp future-proof for at least 10 years in terms of autoconf/automake :) The remaining warnings are: > glibtoolize: Consider adding 'AC_CONFIG_MACRO_DIRS([m4])' to configure.ac, > glibtoolize: and rerunning glibtoolize and aclocal. > glibtoolize: Consider adding '-I m4' to ACLOCAL_AMFLAGS in Makefile.am. which is perfectly safe to ignore (it’s just advice). And this: > configure.ac:4045: warning: AC_PROG_LEX without either yywrap or noyywrap is > obsolete > ./lib/autoconf/programs.m4:716: _AC_PROG_LEX is expanded from... > ./lib/autoconf/programs.m4:709: AC_PROG_LEX is expanded from... > aclocal.m4:9472: AM_PROG_LEX is expanded from... > configure.ac:4045: the top level which is also safe to ignore. I tried to silence it, but did not find a way: AM_PROG_LEX does not accept an extra argument, and switching from AM_PROG_LEX to AC_PROG_LEX would introduce a functional change. I’m not sure if the change would risk breaking some platforms for gmp with exotic flex implementations, so I did not want to risk it. I have filed the issue with automake, asking for their opinion on what I think is a automake / autoconf feature mismatch. Cheers, FX
ac_try.diff
Description: Binary data
_______________________________________________ gmp-bugs mailing list gmp-bugs@gmplib.org https://gmplib.org/mailman/listinfo/gmp-bugs