Ralf Wildenhues <Ralf.Wildenhues <at> gmx.de> writes: > > m4_esyscmd([for f in foo bar; > > do echo "AC_CONFIG_FILES([$f],[chmod +x $f])"; done]) > > > > Eric, is that safe? No. The echo statement didn't provide enough [] quoting, so you are in effect invoking: AC_CONFIG_FILES(foo,chmod +x foo) AC_CONFIG_FILES(bar,chmod +x bar) Where you really wanted: AC_CONFIG_FILES([foo],[chmod +x foo]) AC_CONFIG_FILES([bar],[chmod +x bar]) > That may be safe, I haven't checked, but in any case it is very ugly. > Please do not use m4_esyscmd unless it cannot be helped at all, it's > an ugly crutch and will only lead to problems. I agree - why spawn another process to do something that m4 already does perfectly fine by itself? Especially when it is so hard to get m4_esyscmd to have the right level of quoting. > > You can use > m4_foreach_w([my_file], [autoconf autoheader ...], > [AC_CONFIG_FILES([my_file], [chmod +x my_file])]) I haven't tested this, but it seems like you are passing "my_file" as the tag rather than "autoconf", "autoheader". In other words, I think you want the my_file temporary variable expanded prior to invoking AC_CONFIG_FILES. Also, expanding my_file rather than using m4_defn([my_file]) can lead to surprises if any of your list of files contains $ or another macro name. So to be safe, I would write this as follows (also using newer m4sugar features for efficiency): m4_define([my_CONFIG_EXEC_FILE], [AC_CONFIG_FILES([$1], [chmod +x $1])]) m4_map_args([my_CONFIG_EXEC_FILE(], [)], [autoconf], [autoheader], ...) -- Eric Blake _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
