[uClibc]Re: [TuxScreen-devel] arm-uclibc-c++ says it can't generate executables?

[Jeff Rush]

Alright, so uClibc/libm/fpmacros.c has a function called:

    long int __finitef ( float x )

although the comments above is name it __isfinitef(), and the 
configure for /buildroot-tux/build/gcc-3.1/libstdc++-v3/ has a
variable for HAVE_FINITEF.

The question is, is uclibc deficit for not having such a function
or is gcc-3.1 wrong for assuming it's presence?

I can't figure out the uclibc-toolchain.mk process enough to
tell where gcc-3.1 configure get's it's ideas from.  In configure.in
it seems to assume that if we are (a) cross-compiling and (b)
running under linux, then forcibly enable various items.

case "$target_alias" in *-linux*)
      # Check for available headers.
      AC_CHECK_HEADERS([nan.h ieeefp.h endian.h sys/isa_defs.h \
	machine/endian.h machine/param.h sys/machine.h sys/types.h \
	fp.h locale.h float.h inttypes.h])
      SECTION_FLAGS='-ffunction-sections -fdata-sections'
      AC_SUBST(SECTION_FLAGS)
      GLIBCPP_CHECK_LINKER_FEATURES
      GLIBCPP_CHECK_COMPLEX_MATH_SUPPORT
      GLIBCPP_CHECK_WCHAR_T_SUPPORT
      os_include_dir="config/os/gnu-linux"
      AC_DEFINE(HAVE_COPYSIGN)
      AC_DEFINE(HAVE_COPYSIGNF)
      AC_DEFINE(HAVE_FINITEF)   <--------------------------
      AC_DEFINE(HAVE_FREXPF)
      AC_DEFINE(HAVE_HYPOTF)
      AC_DEFINE(HAVE_SINCOS)
      AC_DEFINE(HAVE_SINCOSF)
      if test x"long_double_math_on_this_cpu" = x"yes"; then
	AC_DEFINE(HAVE_HYPOTL)
      fi
      ;;

-Jeff Rush