libyasm/floatnum.h File Reference

YASM floating point (IEEE) interface. More...

Go to the source code of this file.

Functions
YASM_LIB_DECL void	yasm_floatnum_initialize (void)
	Initialize floatnum internal data structures.
YASM_LIB_DECL void	yasm_floatnum_cleanup (void)
	Clean up internal floatnum allocations.
YASM_LIB_DECL yasm_floatnum *	yasm_floatnum_create (const char *str)
	Create a new floatnum from a decimal string.
YASM_LIB_DECL yasm_floatnum *	yasm_floatnum_copy (const yasm_floatnum *flt)
	Duplicate a floatnum.
YASM_LIB_DECL void	yasm_floatnum_destroy (yasm_floatnum *flt)
	Destroy (free allocated memory for) a floatnum.
YASM_LIB_DECL int	yasm_floatnum_calc (yasm_floatnum *acc, yasm_expr_op op, yasm_floatnum *operand)
	Floating point calculation function: acc = acc op operand.
YASM_LIB_DECL int	yasm_floatnum_get_int (const yasm_floatnum *flt, unsigned long *ret_val)
	Convert a floatnum to single-precision and return as 32-bit value.
YASM_LIB_DECL int	yasm_floatnum_get_sized (const yasm_floatnum *flt, unsigned char *ptr, size_t destsize, size_t valsize, size_t shift, int bigendian, int warn)
	Output a yasm_floatnum to buffer in little-endian or big-endian.
YASM_LIB_DECL int	yasm_floatnum_check_size (const yasm_floatnum *flt, size_t size)
	Basic check to see if size is valid for flt conversion (using yasm_floatnum_get_sized()).
YASM_LIB_DECL void	yasm_floatnum_print (const yasm_floatnum *flt, FILE *f)
	Print various representations of a floatnum.

---

Detailed Description

YASM floating point (IEEE) interface.

Definition in file floatnum.h.

---

Function Documentation

YASM_LIB_DECL int yasm_floatnum_calc	(	yasm_floatnum *	acc,
		yasm_expr_op	op,
		yasm_floatnum *	operand
	)

Floating point calculation function: acc = acc op operand.

Note:

Not all operations in yasm_expr_op may be supported; unsupported operations will result in an error.

Parameters:

	acc	floatnum accumulator
	op	operation
	operand	floatnum operand

Returns:

Nonzero on error.

YASM_LIB_DECL int yasm_floatnum_check_size	(	const yasm_floatnum *	flt,
		size_t	size
	)

Basic check to see if size is valid for flt conversion (using yasm_floatnum_get_sized()).

Doesn't actually check for underflow/overflow but rather checks for size=32,64,80 (at present).

Parameters:

	flt	floatnum
	size	number of bits of output space

Returns:

1 if valid size, 0 if invalid size.

YASM_LIB_DECL void yasm_floatnum_cleanup

(

void

)

Clean up internal floatnum allocations.

YASM_LIB_DECL yasm_floatnum* yasm_floatnum_copy

(

const yasm_floatnum *

flt

)

Duplicate a floatnum.

Parameters:

flt

floatnum

Returns:

Newly allocated floatnum with the same value as flt.

YASM_LIB_DECL yasm_floatnum* yasm_floatnum_create

(

const char *

str

)

Create a new floatnum from a decimal string.

The input string must be in standard C representation ([+-]123.456e[-+]789).

Parameters:

str

floating point decimal string

Returns:

Newly allocated floatnum.

YASM_LIB_DECL void yasm_floatnum_destroy

(

yasm_floatnum *

flt

)

Destroy (free allocated memory for) a floatnum.

Parameters:

flt

floatnum

YASM_LIB_DECL int yasm_floatnum_get_int	(	const yasm_floatnum *	flt,
		unsigned long *	ret_val
	)

Convert a floatnum to single-precision and return as 32-bit value.

The 32-bit value is a "standard" C value (eg, of unknown endian).

Parameters:

	flt	floatnum
	ret_val	pointer to storage for 32-bit output

Returns:

Nonzero if flt can't fit into single precision: -1 if underflow occurred, 1 if overflow occurred.

YASM_LIB_DECL int yasm_floatnum_get_sized	(	const yasm_floatnum *	flt,
		unsigned char *	ptr,
		size_t	destsize,
		size_t	valsize,
		size_t	shift,
		int	bigendian,
		int	warn
	)

Output a yasm_floatnum to buffer in little-endian or big-endian.

Puts the value into the least significant bits of the destination, or may be shifted into more significant bits by the shift parameter. The destination bits are cleared before being set. [0] should be the first byte output to the file.

Note:

Not all sizes are valid. Currently, only 32 (single-precision), 64 (double-precision), and 80 (extended-precision) are valid sizes. Use yasm_floatnum_check_size() to check for supported sizes.

Parameters:

	flt	floatnum
	ptr	pointer to storage for size bytes of output
	destsize	destination size (in bytes)
	valsize	size (in bits)
	shift	left shift (in bits)
	bigendian	endianness (nonzero=big, zero=little)
	warn	enables standard overflow/underflow warnings

Returns:

Nonzero if flt can't fit into the specified precision: -1 if underflow occurred, 1 if overflow occurred.

YASM_LIB_DECL void yasm_floatnum_initialize

(

void

)

Initialize floatnum internal data structures.

YASM_LIB_DECL void yasm_floatnum_print	(	const yasm_floatnum *	flt,
		FILE *	f
	)

Print various representations of a floatnum.

For debugging purposes only.

Parameters:

	f	file
	flt	floatnum

---