[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf
From: |
Bruno Haible |
Subject: |
Re: [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf.h. |
Date: |
Sun, 13 Jul 2008 01:08:09 +0200 |
User-agent: |
KMail/1.5.4 |
> With those changes, I pushed this to the repository.
At a second glance, I could see a small improvement: When, say, the system's
isnsn(double) works and is in libc, but isnan(long double) does not work and
thus needs the replacement, the gnulib isnan will use entirely replaced and
therefore does not need libm. I'm applying this:
2008-07-12 Bruno Haible <address@hidden>
* m4/isnan.m4 (gl_ISNAN): When the replacement is used, set ISNAN_LIBM
to empty.
--- m4/isnan.m4.orig 2008-07-13 01:03:57.000000000 +0200
+++ m4/isnan.m4 2008-07-13 01:01:34.000000000 +0200
@@ -1,4 +1,4 @@
-# isnan.m4 serial 1
+# isnan.m4 serial 2
dnl Copyright (C) 2007-2008 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
@@ -11,9 +11,6 @@
AC_REQUIRE([gl_FUNC_ISNAND])
AC_REQUIRE([gl_FUNC_ISNANL])
- ISNAN_LIBM="$ISNANF_LIBM $ISNAND_LIBM $ISNANL_LIBM"
- AC_SUBST([ISNAN_LIBM])
-
# If we replaced any of the underlying isnan* functions, replace
# the isnan macro; it undoubtedly suffers from the same flaws.
AC_MSG_CHECKING([whether isnan macro works])
@@ -21,12 +18,15 @@
&& test $gl_func_isnand = yes \
&& test $gl_func_isnanl = yes; then
AC_MSG_RESULT([yes])
+ ISNAN_LIBM="$ISNANF_LIBM $ISNAND_LIBM $ISNANL_LIBM"
else
+ AC_MSG_RESULT([no])
+ REPLACE_ISNAN=1
# Make sure the rpl_isnan[fdl] functions get built.
gl_BUILD_ISNANF
gl_BUILD_ISNAND
gl_BUILD_ISNANL
- REPLACE_ISNAN=1
- AC_MSG_RESULT([no])
+ ISNAN_LIBM=
fi
+ AC_SUBST([ISNAN_LIBM])
])
--- m4/isnand.m4.orig 2008-07-13 01:03:57.000000000 +0200
+++ m4/isnand.m4 2008-07-13 00:59:53.000000000 +0200
@@ -41,7 +41,7 @@
fi
])
-dnl Pull in replacement isnand definition.
+dnl Pull in replacement isnand definition. It does not need -lm.
AC_DEFUN([gl_BUILD_ISNAND],
[
AC_LIBOBJ([isnand])
--- m4/isnanf.m4.orig 2008-07-13 01:03:57.000000000 +0200
+++ m4/isnanf.m4 2008-07-13 00:59:15.000000000 +0200
@@ -59,7 +59,7 @@
fi
])
-dnl Pull in replacement isnanf definition.
+dnl Pull in replacement isnanf definition. It does not need -lm.
AC_DEFUN([gl_BUILD_ISNANF],
[
AC_LIBOBJ([isnanf])
--- m4/isnanl.m4.orig 2008-07-13 01:03:57.000000000 +0200
+++ m4/isnanl.m4 2008-07-13 00:59:25.000000000 +0200
@@ -55,7 +55,7 @@
fi
])
-dnl Pull in replacement isnanl definition.
+dnl Pull in replacement isnanl definition. It does not need -lm.
AC_DEFUN([gl_BUILD_ISNANL],
[
AC_LIBOBJ([isnanl])
- Re: new modules 'isnan', 'isnanf', 'isnand', (continued)
- Re: new modules 'isnan', 'isnanf', 'isnand', Simon Josefsson, 2008/07/08
- Re: new modules 'isnan', 'isnanf', 'isnand', Bruno Haible, 2008/07/08
- Re: new modules 'isnan', 'isnanf', 'isnand', Bruno Haible, 2008/07/08
- [patch 2/4] Add isnand module., Ben Pfaff, 2008/07/11
- [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf.h., Ben Pfaff, 2008/07/11
- Re: [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf.h., Bruno Haible, 2008/07/11
- Re: [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf.h., Ben Pfaff, 2008/07/12
- Re: [patch 1/4] Rename isnand.h to isnand-nolibm.h, similarly for isnanf.h.,
Bruno Haible <=
- [patch 4/4] Add isnan module., Ben Pfaff, 2008/07/11
- [patch 3/4] Add isnanf module., Ben Pfaff, 2008/07/11