Mercurial > gnulib
view m4/ldexpf.m4 @ 40214:452ab00796c7
Fix undefined behaviour.
* lib/bitrotate.h (rotl16, rotr16, rotl8, rotr8): Case x to
'unsigned int', to avoid shift operations on 'int'.
* lib/xmemdup0.c (xmemdup0): Don't invoke memcpy with a zero size.
* tests/test-count-leading-zeros.c (main): Use a random number that has
as many bits as TYPE, not only 2*15 or 2*31 bits.
* tests/test-count-trailing-zeros.c (main): Likewise.
* tests/test-count-one-bits.c (main): Likewise.
* tests/test-memmem.c: Don't include "null-ptr.h".
(main): Use zerosize_ptr() instead of null_ptr().
* modules/memmem-tests (Files): Remove tests/null-ptr.h.
author | Bruno Haible <bruno@clisp.org> |
---|---|
date | Sat, 09 Mar 2019 20:32:25 +0100 |
parents | b06060465f09 |
children |
line wrap: on
line source
# ldexpf.m4 serial 2 dnl Copyright (C) 2011-2019 Free Software Foundation, Inc. dnl This file is free software; the Free Software Foundation dnl gives unlimited permission to copy and/or distribute it, dnl with or without modifications, as long as this notice is preserved. AC_DEFUN([gl_FUNC_LDEXPF], [ AC_REQUIRE([gl_MATH_H_DEFAULTS]) AC_REQUIRE([gl_FUNC_LDEXP]) dnl Persuade glibc <math.h> to declare ldexpf(). AC_REQUIRE([gl_USE_SYSTEM_EXTENSIONS]) dnl Test whether ldexpf() exists. We cannot assume that ldexpf(), if it dnl exists, is defined in the same library as ldexp(). This is not the case dnl on FreeBSD, NetBSD, OpenBSD. gl_MATHFUNC([ldexpf], [float], [(float, int)]) if test $gl_cv_func_ldexpf_no_libm = no \ && test $gl_cv_func_ldexpf_in_libm = no; then HAVE_LDEXPF=0 LDEXPF_LIBM="$LDEXP_LIBM" fi AC_SUBST([LDEXPF_LIBM]) ])