[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[bug-gawk] [PATCH 2/2] use AC_CANONICAL_HOST instead of uname
From: |
Mike Frysinger |
Subject: |
[bug-gawk] [PATCH 2/2] use AC_CANONICAL_HOST instead of uname |
Date: |
Sat, 7 Dec 2013 01:54:23 -0500 |
The autoconf macro is designed for parsing the target and works in all the
various scenarios (cross-compile/native/etc...) while `uname` only works
when doing a native build.
---
configure.ac | 27 +++++++++++++++------------
1 file changed, 15 insertions(+), 12 deletions(-)
diff --git a/configure.ac b/configure.ac
index 0619a75..b5e1ac5 100644
--- a/configure.ac
+++ b/configure.ac
@@ -59,6 +59,7 @@ AC_ARG_ENABLE([lint], [ --disable-lint Disable gawk
lint checking],
fi
)
+AC_CANONICAL_HOST
AC_USE_SYSTEM_EXTENSIONS
dnl checks for programs
@@ -119,8 +120,8 @@ dnl need -D_SYSV3 for ISC
fi
dnl check for systems where libc is borked for regex handling
-case `uname` in
-MirBSD | MirOS)
+case $host_os in
+mirbsd*)
AC_DEFINE([LIBC_IS_BORKED], 1, [libc is broken for regex handling])
;;
esac
@@ -263,8 +264,8 @@ AC_CHECK_LIB(m, fmod)
AC_CHECK_LIB(m, isinf)
AC_CHECK_LIB(m, ismod)
dnl Don't look for libsigsegv on OSF/1, gives us severe headaches
-case `uname` in
-OSF1) : ;;
+case $host_os in
+osf1) : ;;
*)
gl_LIBSIGSEGV
;;
@@ -289,8 +290,8 @@ if test "x$enable_extensions" != "xno"; then
extensions_supported=no
dnl On MirBSD (and probably other systems), don't even try.
- case `uname` in
- MirBSD | MirOS)
+ case $host_os in
+ mirbsd*)
: ;;
*)
AC_CHECK_HEADER(dlfcn.h,
@@ -306,10 +307,12 @@ if test "x$enable_extensions" != "xno"; then
extensions_supported=yes
AC_DEFINE([DYNAMIC], 1, [dynamic loading is possible])
# Add -export-dynamic for old extensions. Only works
for GCC
- if test "$GCC" = yes &&
- uname | $EGREP -i 'linux|freebsd' > /dev/null
- then
- LDFLAGS="$LDFLAGS -export-dynamic"
+ if test "$GCC" = yes; then
+ case $host_os in
+ linux*|freebsd*)
+ LDFLAGS="$LDFLAGS -export-dynamic"
+ ;;
+ esac
fi
fi
])
@@ -325,8 +328,8 @@ fi
dnl check for how to use getpgrp
dnl have to hardwire it for VMS POSIX. Sigh.
dnl ditto for BeOS, OS/2, and MS-DOS.
-case `(uname) 2> /dev/null` in
-*VMS*|*BeOS*|*OS/2*|*MS-DOS*)
+case $host_os in
+vms*|beos*|os2*|msdos)
AC_DEFINE(GETPGRP_VOID, 1,
[Define to 1 if the getpgrp function requires zero arguments.])
;;
--
1.8.4.3