Braden McDaniel <braden@xxxxxxxxxxxxx> writes: > autoconf, basically. autoconf's check for whether it can use -g is > busted, but the effect of this is just a benign warning from the > compiler. (I have a patch for this that I submitted to > autoconf-patches many months ago; but it was ignored.) Sorry, I dropped the ball on that. I didn't like your patch, but didn't have time to compose a better one at the time. I just took the time, installed the following patch instead. Please try CVS Autoconf on your compiler installation to see whether this approach works for you. 2004-04-20 Paul Eggert <eggert@xxxxxxxxxxx> * lib/autoconf/c.m4 (_AC_PROG_CC_G, _AC_PROG_CXX_G): Don't consider -g to work if it generates warnings when plain compiles don't. Problem reported by Braden McDaniel in: http://mail.gnu.org/archive/html/autoconf-patches/2003-07/msg00014.html --- c.m4.~1.185.~ 2004-04-19 10:10:00 -0700 +++ c.m4 2004-04-20 12:49:07 -0700 @@ -501,13 +501,25 @@ AC_LANG_POP(C)dnl # Check whether -g works, even if CFLAGS is set, in case the package # plays around with CFLAGS (such as to build both debugging and normal # versions of a library), tasteless as that idea is. +# Don't consider -g to work if it generates warnings when plain compiles don't. m4_define([_AC_PROG_CC_G], [ac_test_CFLAGS=${CFLAGS+set} ac_save_CFLAGS=$CFLAGS -CFLAGS="-g" AC_CACHE_CHECK(whether $CC accepts -g, ac_cv_prog_cc_g, - [_AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], [ac_cv_prog_cc_g=yes], - [ac_cv_prog_cc_g=no])]) + [ac_save_c_werror_flag=$ac_c_werror_flag + ac_c_werror_flag=yes + ac_cv_prog_cc_g=no + CFLAGS="-g" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [ac_cv_prog_cc_g=yes], + [CFLAGS="" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [], + [ac_c_werror_flag=$ac_save_c_werror_flag + CFLAGS="-g" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [ac_cv_prog_cc_g=yes])])]) + ac_c_werror_flag=$ac_save_c_werror_flag]) if test "$ac_test_CFLAGS" = set; then CFLAGS=$ac_save_CFLAGS elif test $ac_cv_prog_cc_g = yes; then @@ -710,14 +722,25 @@ AC_LANG_POP(C++)dnl # Check whether -g works, even if CXXFLAGS is set, in case the package # plays around with CXXFLAGS (such as to build both debugging and # normal versions of a library), tasteless as that idea is. +# Don't consider -g to work if it generates warnings when plain compiles don't. m4_define([_AC_PROG_CXX_G], [ac_test_CXXFLAGS=${CXXFLAGS+set} ac_save_CXXFLAGS=$CXXFLAGS -CXXFLAGS="-g" AC_CACHE_CHECK(whether $CXX accepts -g, ac_cv_prog_cxx_g, - [_AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], - [ac_cv_prog_cxx_g=yes], - [ac_cv_prog_cxx_g=no])]) + [ac_save_cxx_werror_flag=$ac_cxx_werror_flag + ac_cxx_werror_flag=yes + ac_cv_prog_cxx_g=no + CXXFLAGS="-g" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [ac_cv_prog_cxx_g=yes], + [CXXFLAGS="" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [], + [ac_cxx_werror_flag=$ac_save_cxx_werror_flag + CXXFLAGS="-g" + _AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + [ac_cv_prog_cxx_g=yes])])]) + ac_cxx_werror_flag=$ac_save_cxx_werror_flag]) if test "$ac_test_CXXFLAGS" = set; then CXXFLAGS=$ac_save_CXXFLAGS elif test $ac_cv_prog_cxx_g = yes; then
