a-numaux__libc-x86.ads wraps around long double trig functions that
Lynx178 does not have, so use the default a-numaux.ads instead to match
what is offered on Lynx178.
This is done by removing X86_TARGET_PAIRS from the x86-lynx178elf
configuration as it is now redundant: the default a-numaux pacakge is
used, plus this target also can use the default s-atocou pacakge (since
atomic builtins are supported on the i586 compiler).
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* Makefile.rtl: Remove X86_TARGET_PAIRS for x86-lynx178elf.diff --git a/gcc/ada/Makefile.rtl b/gcc/ada/Makefile.rtl
--- a/gcc/ada/Makefile.rtl
+++ b/gcc/ada/Makefile.rtl
@@ -1911,7 +1911,6 @@ ifeq ($(strip $(filter-out lynxos178%,$(target_os))),)
endif
ifeq ($(strip $(filter-out %86, $(target_cpu))),)
LIBGNAT_TARGET_PAIRS += system.ads<libgnat/system-lynxos178-x86.ads
- LIBGNAT_TARGET_PAIRS += $(X86_TARGET_PAIRS)
endif
ifeq ($(strip $(filter-out lynxos178e,$(target_os))),)
