Lines Matching +refs:clang +refs:format +refs:fallback +refs:style
305 dnl For the binary64 format, the smallest normal number is 2^(-1022).
306 dnl For the binary32 format, the smallest normal number is 2^(-126).
520 dnl Try to determine the format of double
532 AC_MSG_WARN([format of `double' unsupported or not recognized: $mpfr_cv_c_double_format])
536 dnl Now try to determine the format of long double
558 AC_MSG_WARN([This format is known on GCC/PowerPC platforms,])
564 AC_MSG_WARN([This format is known on GCC/PowerPC platforms,])
572 …AC_MSG_WARN([format of `long double' unsupported or not recognized: $mpfr_cv_c_long_double_format])
652 dnl 1. Use the hex format for the double constants (this format should be
664 AC_MSG_CHECKING(decimal float format)
802 AC_MSG_CHECKING(if __float128 can be used as a fallback)
812 [Build float128 functions with float128 fallback])
813 AC_DEFINE([_Float128],[__float128],[__float128 fallback])],
1128 dnl Determine the format of a real floating type (first argument),
1141 dnl Concerning long double: The 10-byte IEEE extended format is generally
1150 dnl clang's LTO (-flto). If we just compile with LTO, the generated
1163 AC_CACHE_CHECK([format of floating-point type `$1'],
1323 # format found on m68k
1390 # format used on HP 9000/785 under HP-UX
1435 # format used on 32-bit PowerPC (Mac OS X and Debian GNU/Linux)
1458 # format used on ppc64le
1588 dnl as a fallback.
1598 [Define to 1 if the compiler accepts gcc style __attribute__ ((mode (XX)))])
1622 precision, as -Werror=format yields an error, even though this