From 219dd320d69deb9068f6b2ce46034d0eb4db888a Mon Sep 17 00:00:00 2001 From: Joseph Myers Date: Fri, 4 Aug 2017 21:32:57 +0000 Subject: Reimplement min_of_type using __MATH_TG. This patch reimplements the libm-internal min_of_type macro to use __MATH_TG instead of its own local type-generic implementation, so simplifying the code and reducing the number of different type-generic implementation variants in use in glibc. Tested for x86_64. * sysdeps/generic/math_private.h (__EXPR_FLT128): Remove macro. (min_of_type_f): New macro. (min_of_type_): Likewise. (min_of_type_l): Likewise. (min_of_type_f128): Likewise. (min_of_type): Define using __MATH_TG and taking an expression argument. (math_check_force_underflow): Pass expression instead of type to min_of_type. (math_check_force_underflow_nonneg): Likewise. --- ChangeLog | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'ChangeLog') diff --git a/ChangeLog b/ChangeLog index 0909e4d77a..7cee945879 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,16 @@ +2017-08-04 Joseph Myers + + * sysdeps/generic/math_private.h (__EXPR_FLT128): Remove macro. + (min_of_type_f): New macro. + (min_of_type_): Likewise. + (min_of_type_l): Likewise. + (min_of_type_f128): Likewise. + (min_of_type): Define using __MATH_TG and taking an expression + argument. + (math_check_force_underflow): Pass expression instead of type to + min_of_type. + (math_check_force_underflow_nonneg): Likewise. + 2017-08-04 H.J. Lu * sysdeps/x86/cpu-features.h [__ASSEMBLER__] -- cgit v1.2.1