update from main archive 970304 cvs/libc-970307 cvs/libc-970306 cvs/libc-970305 cvs/libc-970304

1997-03-04 04:31  Ulrich Drepper  <drepper@cygnus.com>

	* Makerules: Add rules to handle versioning.
	* config.h.in (DO_VERSIONING): New macro.
	* config.make.in (versioning): New variable.
	* configure.in: Add checks for .symver directive in gas and
	--version-script option to ld.  Define DO_VERSIONING and
	versioning if appropriate.

	* math/Makefile (routines): Add s_signbit, s_fpclassify, s_fmax,
	s_fmin, and s_fdim.
	* math/math.h: Define ISO C 9X constants, macros and functions.
	* math/mathcalls.h: Likewise.
	* sysdeps/libm-ieee754/s_fdim.c: New file.
	* sysdeps/libm-ieee754/s_fdimf.c: New file.
	* sysdeps/libm-ieee754/s_fdiml.c: New file.
	* sysdeps/libm-ieee754/s_fmax.c: New file.
	* sysdeps/libm-ieee754/s_fmaxf.c: New file.
	* sysdeps/libm-ieee754/s_fmaxl.c: New file.
	* sysdeps/libm-ieee754/s_fmin.c: New file.
	* sysdeps/libm-ieee754/s_fminf.c: New file.
	* sysdeps/libm-ieee754/s_fminl.c: New file.
	* sysdeps/libm-ieee754/s_fpclassify.c: New file.
	* sysdeps/libm-ieee754/s_fpclassifyf.c: New file.
	* sysdeps/libm-ieee754/s_fpclassifyl.c: New file.
	* sysdeps/libm-ieee754/s_signbit.c: New file.
	* sysdeps/libm-ieee754/s_signbitf.c: New file.
	* sysdeps/libm-ieee754/s_signbitl.c: New file.

	* stdio-common/printf_fphex.c: Correct printing of decimal point
	character.
	Simplify conversion of mantissa to string.
	* stdio-common/vfscanf.c: Handle %A format.
	Optimize termination of floating-point scanning.
	* stdio-common/tstscanf.c (main): Add new test to scanf to test
	scanning float values with given width.
	* stdlib/strtod.c: Add handling of floating-point numbers in
	hexadecimal notation.

	* stdlib/stdlib.h: Use __USE_ISOC9X feature macro for new long long
	functions.
	Pretty print #if directives.
	* string/string.h: Pretty print #if directives.

	* sysdeps/ieee754/dbl2mpn.c: Update copyright.
	* sysdeps/ieee754/ldbl2mpn.c: Likewise.
	* sysdeps/ieee754/mpn2dbl.c: Likewise.
	* sysdeps/ieee754/mpn2flt.c: Likewise.
	* sysdeps/ieee754/mpn2ldbl.c: Likewise.

	* sysdeps/unix/sysv/linux/poll.c: Implement poll function by
	falling back to select-based implementation if syscall isn't
	available.
	* sysdeps/unix/sysv/linux/syscalls.list: Add s_poll.

	* time/leapseconds: Update from tzdata1997b.
	* time/zic.c: Update from tzcode1997b.

1997-03-01 15:08  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* time/Makefile $(tzfiles:%=$(objpfx)z.%): Remove unneeded
	depedencies between installed $(tzlinks) and $(tzbases) files.

1997-03-01 14:27  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* math/math.h: Make compatible with traditional preprocessor;
	requires carefull placement of whitespace in macro arguments.
	Use __CONCAT instead of ##.
	Declare long double functions only if __STDC__ or __GNUC__.
	* math/mathcall.h: Avoid whitespace before argument of macro call
	that is used as function name.

	* sysdeps/m68k/fpu/__math.h: Use __CONCAT instead of ##.
	(__m81_u, __m81_inline): Depend on __LIBC_M81_MATH_INLINES instead
	of __NO_M81_MATH_INLINES.
	[!__LIBC_M81_MATH_INLINES]: Don't define internal functions
	starting with __ieee754.
	[!__NO_MATH_INLINES && __OPTIMIZE__]: Define user visible
	functions as inlines.
	(__m81_defun): Put __attribute__ between return type and function
	name.
	* math/math.h: Include <__math.h> also if __LIBC_M81_MATH_INLINES
	is defined.
	* sysdeps/m68k/fpu/e_acos.c: Define __LIBC_M81_MATH_INLINES
	instead of __NO_M81_MATH_INLINES.
	* sysdeps/m68k/fpu/e_fmod.c: Likewise.
	* sysdeps/m68k/fpu/k_cos.c: Likewise.
	* sysdeps/m68k/fpu/k_sin.c: Likewise.
	* sysdeps/m68k/fpu/k_tan.c: Likewise.
	* sysdeps/m68k/fpu/s_atan.c: Likewise. De-ANSI-declify.
	* sysdeps/m68k/fpu/s_frexp.c: Likewise.
	* sysdeps/m68k/fpu/s_ilogb.c: Likewise.
	* sysdeps/m68k/fpu/s_isinf.c: Likewise.
	* sysdeps/m68k/fpu/s_modf.c: Likewise.
	* sysdeps/m68k/fpu/s_scalbn.c: Likewise.

1997-02-27 21:51  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* Makefile (tests): Cope with $PATH not including the current
	directory.

1997-02-27 18:04  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/unix/Makefile ($(common-objpfx)mk-local_lim): Use
	$(common-objdir-compile).
	($(common-objpfx)make-ioctls): Likewise.
	(mk-local_lim-CFLAGS): Set this instead of local_lim-CFLAGS.
	($(common-objpfx)sys/param.h): Use $(make-target-directory).
	($(addprefix $(common-objpfx),$(sys/param.h-includes))):
	Likewise.
	($(common-objpfx)sys/syscall.h): Likewise.
	($(common-objpfx)local_lim.h): Let make deal with command
	failure.
	($(common-objpfx)param.h.dep): Use temporary file and update
	target atomically.
	($(common-objpfx)errnos): Avoid the Useless Use of cat Award.
	(include $(common-objpfx)param.h.dep): Ignore error.

	* sysdeps/posix/Makefile ($(common-objpfx)mk-stdiolim): Use
	$(common-objdir-compile).
	(mk-stdiolim-CFLAGS): Renamed from cded-objdir-includes, use
	$(shell pwd) instead of $$cwd.

	* sysdeps/generic/Makefile ($(common-objpfx)det_endian): Use
	$(common-objdir-compile).
	($(objpfx)make_siglist): Use $(native-compile).
	(make_siglist-CFLAGS): New variable.

	* Makerules (ALL_BUILD_CFLAGS): Renamed from BUILD_CFLAGS, leaving
	the old name for the user to pass additional flags to the host
	compiler.  Fix reference to config header.
	(native-compile, common-objdir-compile): Rewritten to make more
	generally usable.
	* sysdeps/unix/sysv/sysv4/solaris2/Makefile: Set ALL_BUILD_CFLAGS
	instead of BUILD_CFLAGS.

	* sysvips/sys/ipc.h: Warn if needed feature select macro are not
	defined.

1997-02-27 17:11  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sunrpc/Makefile ($(objpfx)rpc-proto.d, $(objpfx)rpc-proto.c):
	New rules to generate dependencies for the RPC service objects.

1997-02-27 16:26  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* argp/argp-parse.c (parser_finalize): Always set *END_INDEX if
	supplied.

1997-02-28 03:27  Ulrich Drepper  <drepper@cygnus.com>

	* stdlib/strtod.c (STRTOF): Make sure return value is large enough
	so that clearing second word is necessary.

	* sysdeps/unix/sysv/linux/netinet/in_systm.h: Don't use kernel
	header since it is wrong for 64 bit systems.
	Patch by a sun <asun@zoology.washington.edu>.

1997-02-27 10:34:11  Richard Henderson  <rth@tamu.edu>

	* sysdeps/unix/sysv/linux/alpha/brk.S: Support both the Linux/i386
	and OSF/1 style brk syscalls.  We may want to change Linux/Alpha
	for the benefit of running foreign binaries.

1997-03-01 20:21  Miles Bader  <miles@gnu.ai.mit.edu>
1997-02-25 19:42  Miles Bader  <miles@gnu.ai.mit.edu>

1 files changed, 36 insertions, 0 deletions

diff --git a/sysdeps/libm-ieee754/s_fdimf.c b/sysdeps/libm-ieee754/s_fdimf.c
new file mode 100644
index 0000000000..a8d8192c0f
--- /dev/null
+++ b/sysdeps/libm-ieee754/s_fdimf.c
@@ -0,0 +1,36 @@
+/* Return positive difference between arguments.
+   Copyright (C) 1997 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If not,
+   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+#include <math.h>
+
+float
+__fdimf (float x, float y)
+{
+  int clsx = fpclassify (x);
+  int clsy = fpclassify (y);
+
+  if (clsx == FP_NAN || clsy == FP_NAN
+      || (y < 0 && clsx == FP_INFINITE && clsy == FP_INFINITE))
+    /* Raise invalid flag.  */
+    return x - y;
+
+  return x < y ? 0 : x - y;
+}
+weak_alias (__fdimf, fdimf)