Hi Dmitry,
On Fri, Apr 04, 2025 at 04:04:21PM +0300, Dmitry V. Levin wrote:
> On Fri, Apr 04, 2025 at 01:38:01PM +0200, Mark Wielaard wrote:
> > But maybe we should just use AC_CHECK_HEADERS([stdatomic.h]) and drop
> > the AC_COMPILE_IFELSE trick?
>
> Sure. And if we want to keep AC_MSG_ERROR, then
> AC_CHECK_HEADERS([stdatomic.h], [], [AC_MSG_ERROR([stdatomic.h required])])
> should be fine.
That would indeed simplify things. Pushed the attached patch.
Cheers,
Mark
>From 3b71249d4396f742d68e774313c2a2d9caf4d668 Mon Sep 17 00:00:00 2001
From: Mark Wielaard <m...@klomp.org>
Date: Sun, 6 Apr 2025 18:04:10 +0200
Subject: [PATCH] configure: Use AC_CHECK_HEADERS to detect stdatomic.h
* configure.ac: Use AC_CHECK_HEADERS instead of
AC_COMPILE_IFELSE.
Suggested-by: Dmitry V. Levin <l...@strace.io>
Signed-off-by: Mark Wielaard <m...@klomp.org>
---
configure.ac | 6 +-----
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/configure.ac b/configure.ac
index 1679a17dd23d..27488e3fe44a 100644
--- a/configure.ac
+++ b/configure.ac
@@ -228,11 +228,7 @@ AS_IF([test "x$ac_cv_tls" != xyes],
dnl Although we test for C11 above that doesn't mean we have stdatomic.h
dnl We need at least gcc 4.9+ for that.
-AC_CACHE_CHECK([whether gcc provides stdatomic.h], ac_cv_has_stdatomic,
- [AC_COMPILE_IFELSE([AC_LANG_SOURCE([[#include <stdatomic.h>]])],
- ac_cv_has_stdatomic=yes, ac_cv_has_stdatomic=no)])
-AS_IF([test "x$ac_cv_has_stdatomic" = "xno"],
- [AC_MSG_ERROR([stdatomic.h required])])
+AC_CHECK_HEADERS([stdatomic.h], [], [AC_MSG_ERROR([stdatomic.h required])])
dnl This test must come as early as possible after the compiler configuration
dnl tests, because the choice of the file model can (in principle) affect
--
2.49.0
