Concepts

• Floating Point Numbers (XFL)

Behaviour

• Compute an XFL (xls17) floating point from the provided exponent and mantissa

• Return that XFL as an int64_t

Definition

C

int64_t float_set (
    int32_t exponent,
    int64_t mantissa
);

Example

C

int64_t small_amount =
  float_set(-81, 1);

Parameters

🚧Caution

When setting a mantissa that is greater or fewer than 16 decimal digits the exponent will be adjusted to ensure the mantissa is exactly 16 digits. This adjustment may result in an INVALID_FLOAT in some circumstances.

📘Special case

XFL canonical 0 is also 0 in the enclosing number. Thus there is never a need to call float_set(0,0);

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Compute the addition of two XFL (xls17) floating point numbers

• Return a new XFL as an int64_t

Definition

C

int64_t float_sum (
    int64_t float1,
    int64_t float2
);

Example

C

int64_t two =
    float_sum(float_one(), float_one());

Parameters

📘Hint

To subtract two floats use float_negate on the second float then use float_sum.

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Multiply an XFL by -1

• Return a new XFL as an int64_t

Definition

C

int64_t float_negate (
    int64_t float1
);

Example

C

int64_t negative_one =
    float_negate(float_one());

📘Special case

The negation of Canonical Zero is Canonical Zero. Unlike some floating point standards (such as IEEE) there is no "negative zero" in XFL.

Parameters

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Compute the multiplication of two XFL (xls17) floating point numbers

• Return a new XFL as an int64_t

Definition

C

int64_t float_multiply (
    int64_t float1,
    int64_t float2
);

Example

C

int64_t max_vault_pusd =
    float_multiply(vault_xrp, exchange_rate);

Parameters

🚧Caution

Certain multiplications may overflow, which return with an INVALID_FLOAT error. However an underflow returns as XFL Canonical Zero (i.e. enclosing number = 0).

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Return one(1) as an XFL int64_t

Definition

C

int64_t float_one ();

Example

C

int64_t still_one =
    float_one();

Parameters

This function has no parameters.

Return Code

Concepts

• Floating Point Numbers (XFL)

• Serialized Objects

Behaviour

• Read a serialized floating point number.

• If there are more fields/data after the serialized floating pointer number then ignore them.

• Return it as an XFL enclosing number

Definition

C

int64_t float_sto_set (
    uint32_t read_ptr,
    uint32_t read_len
);

Example

C

int64_t vault_pusd = float_sto_set(vault, 8);
if (vault_pusd < 0)
  rollback("Failed to parse serialized float.", 33, 1);

Parameters

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Divide an XFL by another XFL

• Return a new XFL as an int64_t

Definition

C

int64_t float_divide (
    int64_t float1,
    int64_t float2
);

Example

C

int64_t still_one =
    float_divide(float_one(), float_one());

Parameters

Return Code

Concepts

Behaviour

• Evaluate a comparison of two XFL floating point numbers

• Return the result of the comparison as a boolean encoded in an int64_t.

Definition

C

int64_t float_compare (
    int64_t float1,
    int64_t float2,
    uint32_t mode
);

Example

C

if (float_compare(pusd_to_send, 0, COMPARE_LESS) == 1)
{
  // pusd_to_send is less than 0
}

Parameters

🚧Caution

Always verify the function returned 1 rather than non-zero, as negative error codes will be classed as non-zero.

Return Code

Concepts

Behaviour

• Compute the multiplication of an XFL (xls17) floating point number and the quotient of two integers

• Return a new XFL as an int64_t

Definition

C

int64_t float_multiply (
    int64_t float1,
    uint32_t round_up,
    uint32_t numerator,
    uint32_t denominator
);

Example

C

int64_t max_vault_pusd =
    float_mulratio(max_vault_pusd, 0,
        COLLATERALIZATION_NUMERATOR, COLLATERALIZATION_DENOMINATOR);

Parameters

🚧Caution

Certain multiplications may overflow, which return with an INVALID_FLOAT error. However an underflow returns as XFL Canonical Zero (i.e. enclosing number = 0).

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Return the mantissa part of an XFL as an unsigned integer

Definition

C

int64_t float_mantissa (
    int64_t float1
);

Example

C

int64_t exponent =
    float_exponent(float_one());

📘Hint

The mantissa of a negative XFL is positive. Use float_sign to get the sign.

Parameters

Return Code

Concepts

• Floating Point Numbers (XFL)

Behaviour

• Return 1 if the XFL is negative, otherwise return 0

Definition

C

int64_t float_sign (
    int64_t float1
);

Example

C

int64_t sign =
    float_sign(float_one());

📘Hint

The sign bit inside the XFL is the 0 when the XFL is negative, however this function follows the standard computing convention to return 1 if it is negative.

Parameters

Return Code

Concepts

Behaviour

• Read an XFL floating point number and optionally a field code and currency code

• Write a serialized amount to write_ptr according to the parameters provided

Definition

C

int64_t float_sto (
    uint32_t write_ptr,
    uint32_t write_len,
    uint32_t cread_ptr,
    uint32_t cread_len,  
    uint32_t iread_ptr,
    uint32_t iread_len,  
    int64_t float1,
    uint32_t field_code
);

Example

C

#define SBUF(str) (uint32_t)(str), sizeof(str)
uint8_t amt_out[48];
if (float_sto(SBUF(amt_out),
    SBUF(currency), SBUF(hook_accid), pusd_to_send, -1) < 0)
        rollback(SBUF("Peggy: Could not dump pusd amount into sto"), 1);

Parameters

📘Hint

To output an XAH amount prepopulate the field code in the output buffer then pass the output buffer incremented to the new start and 0 as field_code

Return Code

Concepts

Behaviour

• Compute a the decimal logarithm of an XFL number

• Return the new XFL

❗️Warning

Due to speed constraints,float_log converts the argument to an IEEE base-2 double precision floating point before applying base 10 log. Therefore the returned result will often contain less precision than expected.

Definition

C

int64_t float_log (
    int64_t float1
);

Example

C

int64_t zero =
    float_log(float_one());

🚧Warning

If a negative number is passed the function will return COMPLEX_NOT_SUPPORTED if the root is an even root.

Parameters

Return Code

Concepts

Behaviour

• Compute a the nth root of an XFL number

• Return the new XFL

❗️Warning

Due to speed constraints,float_root converts the argument to an IEEE base-2 double precision floating point before applying n-th root. Therefore the returned result will often contain less precision than expected. If you need better precision you should consider dividing your XFL into a high and a low product then individually take the square roots of those products and multiply the results together.

Definition

C

int64_t float_root (
    int64_t float1,
    uint32_t n
);

Example

C

int64_t three =
    float_root(nine, 2);

🚧Warning

If a negative number is passed the function will return COMPLEX_NOT_SUPPORTED if the root is an even root.

Parameters

Return Code

🚧 Replaced by macro

This function was replaced by a macro. Please use the macro below in your code instead. To check the validity of the XFL please use float_mantissa in conjunction with this macro.

Concepts

Behaviour

• Return the exponent part of an XFL as a signed integer

Definition

Because exponents can be negative, and because negatives are reserved for error states, exponents cannot be returned from functions. Therefore this function has become a macro as shown below.

C

#define float_exponent(f)\
        (((int32_t)(((f) >> 54U) & 0xFFU)) - 97)

Example

C

int64_t exponent =
    float_exponent(float_one());

Parameters

Concepts

Behaviour

• Divide 1 by an XFL

• Return a new XFL as an int64_t

Definition

C

int64_t float_invert (
    int64_t float1
);

Example

C

int64_t still_one =
    float_invert(float_one());

Parameters

Return Code

Concepts

Behaviour

• Left shift (multiply by 10) the XFL by the number of specified decimal places

• Convert the resulting XFL to an integer, discarding any remainder

• Return the integer

Definition

C

int64_t float_int (
    int64_t float1,
    uint32_t decimal_places,
  	uint32_t absolute
);

Example

C

int64_t drops =
    float_int(xrpbalance, 6, 0);

Parameters

📘Hint

Negative return values are reserved for error codes. Therefore if you need to execute this function against a negative XFL you should use absolute = 1

Return Code