[Guile-commits] branch main updated: Compile with -fexcess-precision=standard for i[3456]86 when we can

[Rob Browning]

This is an automated email from the git hooks/post-receive script.

rlb pushed a commit to branch main
in repository guile.

The following commit(s) were added to refs/heads/main by this push:
     new 9b1effb58 Compile with -fexcess-precision=standard for i[3456]86 when 
we can
9b1effb58 is described below

commit 9b1effb58579821ffa9357df808d5264e45fb30c
Author: Rob Browning <rlb@defaultvalue.org>
AuthorDate: Sat Oct 5 17:18:06 2024 -0500

    Compile with -fexcess-precision=standard for i[3456]86 when we can
    
    * configure.ac: when -fexcess-precision=standard is available and we're
    building for i[3456]86, use it.  This fixes floating point precision
    problems caused by x87 (80-bit) floating point, and detected by
    numbers.test.
    
    Closes: 43262
---
 NEWS         |  2 ++
 configure.ac | 24 ++++++++++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/NEWS b/NEWS
index b6566903b..9fd14c39d 100644
--- a/NEWS
+++ b/NEWS
@@ -24,6 +24,8 @@ Changes in 3.0.11 (since 3.0.10)
    (<https://bugs.gnu.org/71891>)
 ** Compiler no longer crashes when inlining 3.0.9 procedure with kw args
    (<https://bugs.gnu.org/72936>)
+** Guile is compiled with -fexcess-precision=standard for i[3456]86 when 
possible
+   (<https://debbugs.gnu.org/43262>)
 
 
 Changes in 3.0.10 (since 3.0.9)
diff --git a/configure.ac b/configure.ac
index 262b171f1..ce3ae0226 100644
--- a/configure.ac
+++ b/configure.ac
@@ -62,6 +62,30 @@ AC_PROG_INSTALL
 AC_PROG_CC
 gl_EARLY
 
+AC_MSG_CHECKING([whether the compiler supports -fexcess-precision=standard])
+old_CFLAGS="$CFLAGS"
+CFLAGS="$CFLAGS -fexcess-precision=standard"
+fexcess_precision=''
+AC_COMPILE_IFELSE([AC_LANG_PROGRAM([int foo;], [])], [fexcess_precision=1]])
+CFLAGS="$old_CFLAGS"
+if test -n "$fexcess_precision"; then
+  AC_MSG_RESULT([yes])
+else
+  AC_MSG_RESULT([no])
+fi
+AC_MSG_CHECKING([whether we want -fexcess-precision=standard for $target_cpu])
+case "$target_cpu" in
+    i[[3456]]86)
+        AC_MSG_RESULT([yes])
+        if test -n "$fexcess_precision"; then
+          CFLAGS="$CFLAGS -fexcess-precision=standard"
+        else
+          AC_MSG_WARN([floating-point precision may exceed C99 rules])
+        fi
+        ;;
+    *) AC_MSG_RESULT([no]) ;;
+esac
+
 AC_MSG_CHECKING([whether the compiler supports -flto])
 old_CFLAGS="$CFLAGS"
 LTO_CFLAGS="-flto"