Previous: Implementation Defined Options, Up: The optional Floating-Point word set   [Contents][Index]

---

8.7.2 Ambiguous conditions ¶

df@ or df! used with an address that is not double-float aligned: ¶

Hardware-dependent. If the hardware does not support unaligned accesses, this typically results in a -23 THROW like other alignment violations.

f@ or f! used with an address that is not float aligned: ¶

Hardware-dependent. If the hardware does not support unaligned accesses, this typically results in a -23 THROW like other alignment violations.

floating-point result out of range: ¶

An FP value is returned (so nothing is out of range), but it may be an infinity or a NaN.

sf@ or sf! used with an address that is not single-float aligned: ¶

Hardware-dependent. If the hardware does not support unaligned accesses, this typically results in a -23 THROW like other alignment violations.

base is not decimal (REPRESENT, F., FE., FS.): ¶

Conversion between FP and string representations always use decimal, irrespective of base.

Both arguments are equal to zero (FATAN2): ¶

System-dependent. FATAN2 is implemented using the C library function atan2().

Using FTAN on an argument r1 where cos(r1) is zero: ¶

System-dependent. Anyway, typically the cos of r1 will not be zero because of small errors and the tan will be a very large (or very small) but finite number.

d cannot be presented precisely as a float in D>F: ¶

The result is rounded to the nearest float.

dividing by zero: ¶

On IEEE FP based systems, prouces an infinity or a NaN.

exponent too big for conversion (DF!, DF@, SF!, SF@): ¶

On IEEE-FP based systems the number is converted into an infinity.

float<1 (FACOSH): ¶

On IEEE-FP based systems this produces a NaN.

float<=-1 (FLNP1): ¶

On IEEE-FP based systems this produces a NaN (or a negative infinity for float=-1).

float<=0 (FLN, FLOG): ¶

On IEEE-FP based systems this produces a NaN (or a negative infinity for float=0).

float<0 (FASINH, FSQRT): ¶

On IEEE-FP based systems, for fsqrt this gives a NaN, for fasinh some platforms produce a NaN, others a number.

|float|>1 (FACOS, FASIN, FATANH): ¶

IEEE-FP based systems produce a NaN.

integer part of float cannot be represented by d in F>D: ¶

Platform-dependent; typically, some double number is produced and no error is reported.

string larger than pictured numeric output area (f., fe., fs.): ¶

These words do not use the pictured numeric output buffer, but a buffer big enough for the needs of these words.