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, 13 insertions, 11 deletions

diff --git a/math/mathcalls.h b/math/mathcalls.h
index d529c1a8e2..d566020aee 100644
--- a/math/mathcalls.h
+++ b/math/mathcalls.h
@@ -147,7 +147,7 @@ __MATHCALL (cbrt,, (_Mdouble_ __x));
 __MATHCALL (ceil,, (_Mdouble_ __x));
 
 /* Absolute value of X.  */
-__MATHCALL (fabs,, (_Mdouble_ __x));
+__MATHCALLX (fabs,, (_Mdouble_ __x), (__const__));
 
 /* Largest integer not greater than X.  */
 __MATHCALL (floor,, (_Mdouble_ __x));
@@ -159,10 +159,10 @@ __MATHCALL (fmod,, (_Mdouble_ __x, _Mdouble_ __y));
 #ifdef __USE_MISC
 /* Return 0 if VALUE is finite or NaN, +1 if it
    is +Infinity, -1 if it is -Infinity.  */
-__MATHDECL (int,isinf,, (_Mdouble_ __value));
+__MATHDECLX (int,isinf,, (_Mdouble_ __value), (__const__));
 
 /* Return nonzero if VALUE is finite and not NaN.  */
-__MATHDECL (int,finite,, (_Mdouble_ __value));
+__MATHDECLX (int,finite,, (_Mdouble_ __value), (__const__));
 
 /* Deal with an infinite or NaN result.
    If ERROR is ERANGE, result is +Inf;
@@ -170,7 +170,7 @@ __MATHDECL (int,finite,, (_Mdouble_ __value));
    otherwise result is NaN.
    This will set `errno' to either ERANGE or EDOM,
    and may return an infinity or NaN, or may do something else.  */
-__MATHCALL (infnan,, (int __error));
+__MATHCALLX (infnan,, (int __error), (__const__));
 
 /* Return X times (2 to the Nth power).  */
 __MATHCALL (scalbn,, (_Mdouble_ __x, int __n));
@@ -185,18 +185,18 @@ __MATHCALL (significand,, (_Mdouble_ __x));
 
 #if defined __USE_MISC || defined __USE_ISOC9X
 /* Return X with its signed changed to Y's.  */
-__MATHCALL (copysign,, (_Mdouble_ __x, _Mdouble_ __y));
+__MATHCALLX (copysign,, (_Mdouble_ __x, _Mdouble_ __y), (__const__));
 #endif
 
 #ifdef __USE_ISOC9X
 /* Return representation of NaN for double type.  */
-__MATHCALL (nan,, (__const char *__tagb));
+__MATHCALLX (nan,, (__const char *__tagb), (__const__));
 #endif
 
 
 #if defined __USE_MISC || defined __USE_XOPEN
 /* Return nonzero if VALUE is not a number.  */
-__MATHDECL (int,isnan,, (_Mdouble_ __value));
+__MATHDECLX (int,isnan,, (_Mdouble_ __value), (__const__));
 
 /* Return the binary exponent of X, which must be nonzero.  */
 __MATHDECL (int,ilogb,, (_Mdouble_ __x));
@@ -235,7 +235,7 @@ __MATHCALL (lgamma,_r, (_Mdouble_, int *));
 __MATHCALL (rint,, (_Mdouble_ __x));
 
 /* Return X + epsilon if X < Y, X - epsilon if X > Y.  */
-__MATHCALL (nextafter,, (_Mdouble_ __x, _Mdouble_ __y));
+__MATHCALLX (nextafter,, (_Mdouble_ __x, _Mdouble_ __y), (__const__));
 
 /* Return the remainder of integer divison X / Y with infinite precision.  */
 __MATHCALL (remainder,, (_Mdouble_ __x, _Mdouble_ __y));
@@ -257,7 +257,7 @@ __MATHCALL (round,, (_Mdouble_ __x));
 
 /* Round X to the integral value in floating-point format nearest but
    not larger in magnitude.  */
-__MATHCALL (trunc,, (_Mdouble_ __x));
+__MATHCALLX (trunc,, (_Mdouble_ __x), (__const__));
 
 /* Compute remainder of X and Y and put in *QUO a value with sign of x/y
    and magnitude congruent `mod 2^n' to the magnitude of the integral
@@ -276,8 +276,10 @@ __MATHCALL (fmin,, (_Mdouble_ __x, _Mdouble_ __y));
 
 
 /* Classify given number.  */
-__MATHDECL_1 (int, __fpclassify,, (_Mdouble_ __value));
+__MATHDECL_1 (int, __fpclassify,, (_Mdouble_ __value))
+     __attribute__ ((__const__));
 
 /* Test for negative number.  */
-__MATHDECL_1 (int, __signbit,, (_Mdouble_ __value));
+__MATHDECL_1 (int, __signbit,, (_Mdouble_ __value))
+     __attribute__ ((__const__));
 #endif /* Use ISO C 9X.  */