Update.

1997-03-25 02:15  Ulrich Drepper  <drepper@cygnus.com>

	* gmon/gmon.c: Optimize a bit by using more sets of records to
	write in a single writev call.

	* math/math.h: Add definitions of macros __MATHCALLX and __MATHDECLX.
	* math/mathcalls.h: Use __MATHCALLX for fabs, infnan, copysign,
	nan, isnan, nextafter, trunc, __fpclassify and signbit to mark as
	`const'.
	Use __MATHDECLX for isinf and finite.

	* sysdeps/generic/setenv.c [_LIBC]: Define __clearenv and make
	clearenv a weak alias.

	Implement complex exponential function.
	* sysdeps/libm-ieee754/s_cexp.c: New file.
	* sysdeps/libm-ieee754/s_cexpf.c: New file.
	* sysdeps/libm-ieee754/s_cexpl.c: New file.

	* sysdeps/libm-ieee754/s_nan.c: Define function as __nan and make
	nan a weak alias.
	* sysdeps/libm-ieee754/s_nanf.c: Likewise for nanf.
	* sysdeps/libm-ieee754/s_nanl.c: Likewise for nanl.

	* sysdeps/unix/sysv/linux/iovec.h: Don't use kernel header because
	of type clashes.  Add all definitions here.

	* sysdeps/unix/sysv/linux/shmat.c: Likewise.  Correct types according
	to XPG4.2.
	* sysdeps/unix/sysv/linux/shmdt.c: Likewise.
	* sysdeps/unix/sysv/linux/msgrcv.c: Likewise.

	* sysvipc/sys/shm.h (shmat, shmdt): Correct types.

	* sysdeps/unix/sysv/linux/sys/kd.h: Define _LINUX_TYPES_H to avoid
	use of kernel types.

1997-03-25 00:00  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/unix/sysv/linux/netinet/ip.h (struct ip_timestamp):
	Correct typos.  Reported by a sun <asun@zoology.washington.edu>.

1997-03-20 21:58  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/libm-ieee754/s_asinhl.c: Fix sign of result.

1997-03-20 16:20  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/libm-ieee754/e_atan2l.c: Fix typo.

1997-03-20 14:23  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* sysdeps/m68k/fpu/__math.h (__logb, logb): Don't define.
	* sysdeps/m68k/fpu/s_logb.c: Removed.
	* sysdeps/m68k/fpu/s_logbf.c: Removed.
	* sysdeps/m68k/fpu/s_logbl.c: Removed.

	* sysdeps/m68k/fpu/__math.h (__ieee754_atan2, __ieee754_pow):
	Don't define here.
	* sysdeps/m68k/fpu/e_atan2.c: Rewritten.
	* sysdeps/m68k/fpu/e_pow.c: Likewise.
	* sysdeps/m68k/fpu/e_atan2f.c: Based on e_atan2.c.
	* sysdeps/m68k/fpu/e_atan2l.c: Likewise.
	* sysdeps/m68k/fpu/e_powf.c: Based on e_pow.c.
	* sysdeps/m68k/fpu/e_powl.c: Likewise.

	* sysdeps/m68k/fpu/s_log2.c: New file.
	* sysdeps/m68k/fpu/s_log2f.c: New file.
	* sysdeps/m68k/fpu/s_log2l.c: New file.
	* sysdeps/m68k/fpu/s_exp2.c: New file.
	* sysdeps/m68k/fpu/s_exp2f.c: New file.
	* sysdeps/m68k/fpu/s_exp2l.c: New file.

1997-03-20 14:46  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* Makeconfig (all): Make sure this is always the default goal.

1997-03-20 11:09  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* math/libm-test.c: Increase epsilon in many places.  Fix many
	typos.  Allow optinal argument also for short option.

1997-03-20 11:09  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* math/Makefile ($(objpfx)$(tests)): Link against libm.

1997-03-24 23:14  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/libm-ieee754/s_fpclassifyl.c: Correct recognition of
	denormalized numbers.

	* sysdeps/i386/huge_val.h: Remove references to byte order macros.
	Don't include <endian.h>.

1997-03-19 15:18  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>
	* sysdeps/m68k/huge_val.h: Remove references to byte order macros.
	Don't include <endian.h>.

1997-03-24 23:09  Ulrich Drepper  <drepper@cygnus.com>

	* inet/rcmd.c (iruserok): Use access instead of euidaccess.

1997-03-15 18:08  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>

	* math/cmathcalls.h: Remove whitespace before function name in use
	of __MATHDECL macro, to make it compilable with a traditional
	preprocessor.

1997-03-24 15:31  Ulrich Drepper  <drepper@cygnus.com>

	* configure.in: Use AC_PROG_CC instead of AC_PROG_TOOL to find
	compiler.

1997-03-24 02:34  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/libm-ieee754/s_finitel.c (__finitel): Shift return value
	by 31 positions to get 0/1 result.
	Patch by Joe Keane <jgk@jgk.org>.

1997-03-23 12:15  H.J. Lu  <hjl@gnu.ai.mit.edu>

	* posix/getopt.c (__getopt_nonoption_flags): Make it extern
	to prevent from ld linking in getopt.o even if there is another
	incompatible one.

	* posix/getopt_init.c (__getopt_nonoption_flags): Remove extern.

1997-03-23 23:30   Thorsten Kukuk  <kukuk@vt.uni-paderborn.de>

	* nis/nis_call.c (__do_niscall): Print message if cold start file
	does not exist.
	* nis/nis_file.c: Don't print error messages.
	* nis/nis_local_name.c (nis_local_host): Fix pointer errors.
	* nis/rpcsvc/ypclnt.h: Fix copyright and prototypes.
	* nis/rpcsvc/nis.h: Likewise.
	* nis/ypclnt.c: Fix prototypes.

1997-03-24 01:36  Ulrich Drepper  <drepper@cygnus.com>

	* sysdeps/libm-i387/e_scalb.S: Pop additional result before
	returning.
	* sysdeps/libm-i387/e_scalbf.S: Likewise.
	* sysdeps/libm-i387/e_scalbl.S: Likewise.
	Reported by Andreas Jaeger <aj@arthur.pfalz.de>.

	* elf/rtld.c (dl_main): Set main_map->l_opencount to 1 also if
	dynamic linker was called implicitly by the kernel.
	Reported by Eric Youngdale <eric@andante.jic.com>.

	* math/Makefile (CFLAGS-test-float.c, CFLAGS-test-double.c,
	CFLAGS-test-longdouble.c): New variables.  Set to -fno-inline to
	prevent clever optimizations which corrupt the tests.

1997-03-23 21:33  Andreas Jaeger  <aj@arthur.pfalz.de>

	* math/libm-test.c (scalb_test): New function.
	(sqrt_test): New function.
	(scalbn_test): New function.
	(ilogb_test): New function.
	(main): Added calls for new test functions.
	(ldexp_test): Add another test for ldexp(x,0).

1997-03-23 12:35  Andreas Jaeger  <aj@arthur.pfalz.de>

	* math/libm-test.c (remquo_test): Correct messages so that they
	match the tests.
	(copysign_test): Likewise.

1997-03-23 16:28  Ulrich Drepper  <drepper@cygnus.com>

	* stdio-common/Makefile (routines): Add printf_size.

	* stdio-common/printf.h: Add declarations for printf_size and
	printf_size_info.

	* stdio-common/printf-parse.h (__printf_arginfo_table): Correct
	declaration.
	(parse_one_spec): Test whether __printf_function_table is not NULL
	before using registered handlers and don't test
	__printf_arginfo_table.  Update nargs also when handler is
	registered.

	* stdio-common/vfprintf.c (vfprintf): Count number of specifiers
	processed in fast loop.

	* stdio-common/printf_size: New file.

1997-03-22 04:53  Ulgmon_out_readrich Drepper  <drepper@cygnus.com>

1 files changed, 63 insertions, 0 deletions

diff --git a/sysdeps/libm-ieee754/s_cexpl.c b/sysdeps/libm-ieee754/s_cexpl.c
new file mode 100644
index 0000000000..6b3d409158
--- /dev/null
+++ b/sysdeps/libm-ieee754/s_cexpl.c
@@ -0,0 +1,63 @@
+/* Return value of complex exponential function for long double complex value.
+   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 <complex.h>
+#include <math.h>
+
+
+__complex__ long double
+__cexpl (__complex__ long double x)
+{
+  __complex__ long double retval;
+
+  if (isfinite (__real__ x))
+    {
+      if (isfinite (__imag__ x))
+	{
+	  retval = __expl (__real__ x) * (__cosl (__imag__ x)
+					  + 1i * __sinl (__imag__ x));
+	}
+      else
+	/* If the imaginary part is +-inf or NaN and the real part is
+	   not +-inf the result is NaN + iNan.  */
+	retval = __nanl ("") + 1.0i * __nanl ("");
+    }
+  else if (__isinfl (__real__ x))
+    {
+      if (isfinite (__imag x))
+	{
+	  if (signbit (__real__ x) == 0 && __imag__ x == 0.0)
+	    retval = HUGE_VALL;
+	  else
+	    retval = ((signbit (__real__ x) ? 0.0 : HUGE_VALL)
+		      * (__cosl (__imag__ x) + 1i * __sinl (__imag__ x)));
+	}
+      else if (signbit (__real__ x))
+	retval = HUGE_VALL + 1.0i * __nanl ("");
+      else
+	retval = 0.0;
+    }
+  else
+    /* If the real part is NaN the result is NaN + iNan.  */
+    retval = __nanl ("") + 1.0i * __nanl ("");
+
+  return retval;
+}
+weak_alias (__cexpl, cexpl)