Compare/Check/Repair Float

Detailed Description

Functions
Bool	CompareFloatTolerant (Float32 a, Float32 b)
Bool	CompareFloatTolerant (Float64 a, Float64 b)
Bool	CheckFloat (Float32 r)
Bool	CheckFloat (Float64 r)
Float32	RepairFloat (Float32 r)
Float64	RepairFloat (Float64 r)

Function Documentation

CompareFloatTolerant() [1/2]

Bool CompareFloatTolerant	(	Float32	a,
		Float32	b
	)

Checks if two floats are close to each other on a bit basis (rather than a fixed epsilon).
Single precision version.

Parameters

[in]	a	A float value.
[in]	b	A float value.

Returns

true if a and b are sufficiently close to each other, otherwise false.

CompareFloatTolerant() [2/2]

Bool CompareFloatTolerant	(	Float64	a,
		Float64	b
	)

Checks if two floats are close to each other on a bit basis (rather than a fixed epsilon).
Double precision version.

Parameters

[in]	a	A float value.
[in]	b	A float value.

Returns

true if a and b are sufficiently close to each other, otherwise false.

CheckFloat() [1/2]

Bool CheckFloat

(

Float32

r

)

Checks if r is a valid floating point number.
Single precision version.

Parameters

[in]

r

The value to check.

Returns

true if r is a valid floating point number, otherwise false.
true for FP_NORMAL, FP_SUBNORMAL and FP_ZERO.
false for FP_NAN, FP_INFINITE.

CheckFloat() [2/2]

Bool CheckFloat

(

Float64

r

)

Checks if r is a valid floating point number.
Double precision version.

Parameters

[in]

r

The value to check.

Returns

true if r is a valid floating point number, otherwise false.
true for FP_NORMAL, FP_SUBNORMAL and FP_ZERO.
false for FP_NAN, FP_INFINITE.

RepairFloat() [1/2]

Float32 RepairFloat

(

Float32

r

)

Repairs floating point values.
Single precision version.

Warning

Only NaN and Infinity are corrected. Denormalized numbers are not.

Note

Useful when importing other formats where the floating point data is not always valid.

Parameters

[in]

r

The float value to repair.

Returns

The repaired value.

RepairFloat() [2/2]

Float64 RepairFloat

(

Float64

r

)

Repairs floating point values.
Double precision version.

Warning

Only NaNs and Infinity are corrected. Denormalized numbers are not.

Note

Useful when importing other formats where the floating point data is not always valid.

Parameters

[in]

r

The float value to repair.

Returns

The repaired value.