NAME

t/native_pbc/number.t - Floating-Point Numbers

SYNOPSIS

% prove t/native_pbc/number.t

DESCRIPTION

Tests word-size/float-type/endian-ness for different architectures.

These tests usually only work on updated native pbc test files. See tools/dev/mk_native_pbc to create the platform-specific native pbcs.

PLATFORMS

_1   (8_le) i386 32 bit opcode_t, 4 byte intval, 8 byte double
     (linux-gcc-i386, freebsd-gcc, cygwin, ...)

_2   (12_le) i386 32 bit opcode_t, 4 byte intval, 12 byte long double
     (linux-gcc-i386 or cygwin with --floatval="long double")

_3   (8_be) PPC BE 32 bit opcode_t, 4 byte intval, 8 byte double
     (darwin-ppc or sparc32)

_4   (8_le) x86_64 64 bit opcode_t, 8 byte intval, 8 byte double
     (linux-gcc-x86_64 resp. amd64, solaris-cc-64int)

_5   (16_le) x86_64 64 bit opcode_t, 8 byte intval, 16 byte long double
     (linux-gcc-x86_64, solaris-cc-64int --floatval="long double")

_6   (8_be) big-endian 64 bit opcode_t, 8 byte intval, 8 byte double
     (Sparc64/Solaris, MIPS irix or similar)

_7   (16_be) big-endian 64 bit opcode_t, 8 byte intval, 16 byte long double
     (Sparc64/Solaris --floatval="long double")

_8   (4_le) i386 32 bit opcode_t, 4 byte intval, 4 byte single float
     (linux-gcc-i386 or cygwin with --floatval="float")

_9   (4_be) big-endian 32 bit opcode_t, 4 byte intval, 4 byte single float
     (darwin/ppc with --floatval="float")