On Wednesday 19 September 2007, Stepan Kasal wrote: > On Fri, Sep 07, 2007 at 03:46:49PM -0400, Mike Frysinger wrote: > > > I suggest to add AC_REQUIRE([AC_CANONICAL_HOST]) to the beginning of > > > the macro, because that is what provides $host_cpu. But that macro > > > is already there, IIRC it is brought in AM_INIT_AUTOMAKE, so this > > > does not cause problems in practice. > > > > as you say, it shouldnt be needed in the normal case, but it certainly > > cant hurt anything and if it were split off into a sep .m4, this change > > would only be a good thing > > I agree with you here. But, forgive me, I cannot resist some > nit-picking: it does not matter whether the macro is defined in > a separate .m4 file or not. > > If, in a future release of Automake, AM_INIT_AUTOMAKE would no longer > require the macro, the AC_REQUIRE([AC_CANONICAL_HOST]) in > UTIL_CHECK_SYSCALL would suddenly become effective. > > Or if someone pasted our macro to a project which does not use > Automake, the AC_REQUIRE would ensure that the macro works as > expected. this is what i meant ... > > > + case $util_cv_syscall_$1 in #( > > > + no) AC_MSG_WARN([Unable to detect syscall $1.]) ;; > > > ... > > > +[m4_ifval([$1], > > > + [#( > > > + $1) syscall="$2" ;;dnl > > > > what are those #( for ? > > The paren matches the closing paren after the pattern, which helps > with my editor. i figured it was to make things like vim happy ... perhaps it should be: dnl #( so that the #( isnt output to the final file ... -mike
Attachment:
signature.asc
Description: This is a digitally signed message part.
