Hello! This is with Autoconf 2.59. Consider the following: configure.ac #v+ AC_INIT([foo], [1]) AC_DEFINE([FOO], [0]) AC_DEFINE([FOO], [1]) AC_CONFIG_FILES([Makefile]) AC_OUTPUT #v- Makefile.in #v+ DEFS = @DEFS@ #v- #v+ $ autoconf $ ./configure configure: creating ./config.status config.status: creating Makefile $ cat Makefile DEFS = -DPACKAGE_NAME=\"foo\" -DPACKAGE_TARNAME=\"foo\" -DPACKAGE_VERSION=\"1\" -DPACKAGE_STRING=\"foo\ 1\" -DPACKAGE_BUGREPORT=\"\" -DFOO=0 -DFOO=1 #v- When used with GCC, this will obviously result in ``warning: "FOO" redefined'' and the _last_ definition (i.e. `FOO=1') will be used. Now. configure.ac #v+ AC_INIT([foo], [1]) AC_DEFINE([FOO], [0], [FOO 0]) AC_DEFINE([FOO], [1], [FOO 1]) AC_CONFIG_HEADER([config.h]) AC_OUTPUT #v- #v+ $ autoconf $ autoheader $ ./configure configure: creating ./config.status config.status: creating config.h $ grep FOO config.h /* FOO 1 */ #define FOO 0 #v- The _first_ definition (i.e. `FOO=0') will be used. And the commentary from the _last_ one. In both cases I would have expected that a subsequent `AC_DEFINE' on the same symbol would simply override the previous definition, like in a variable assignment. What is the expected outcome? The manual doesn't tell me. Regards, Thomas
Attachment:
signature.asc
Description: Digital signature
_______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
