@decimal.Decimal

This page tracks the current repository implementation and is written as the 0.2.0 API baseline.

Representation

Finite values are stored as:

coefficient * 10^exponent10

with an attached working precision.

Constructors and Parsing

Decimal::make
Decimal::zero
Decimal::one
Decimal::inf
Decimal::nan
Decimal::from_int
Decimal::from_bigint
Decimal::from_float
Decimal::from_double
Decimal::from_string
Decimal::from_bin_float

Notes:

Finite public constructors normalize by removing removable powers of 10.
from_string accepts plain decimal and scientific notation.
Invalid strings return None.

Access, Normalization, and Comparison

classify
precision
sign
coefficient
exponent10
is_zero
normalized
with_precision
compare
min
max
clamp

Notes:

compare aborts on NaN.
clamp aborts if the bounds are unordered or NaN.

Arithmetic and Conversion

neg
abs
add
sub
mul
div
to_bin_float

Supported operators:

+
-
*
/
unary -

Conversion notes:

Decimal-to-binary conversion may be approximate for non-dyadic values.
Binary-to-decimal conversion is exact for the currently stored finite BinFloat value.

Trait Surface

Decimal currently implements:

@def.Floating
@arithmetic.ParseChecked
@arithmetic.SqrtChecked
@arithmetic.DivChecked
@arithmetic.CompareChecked
@arithmetic.PowNatChecked
@arithmetic.PowIntChecked
Eq, Add, Sub, Mul, Div, Neg, Show

Behavior note:

This pass does not expose transcendental or constant traits from Decimal.
Checked arithmetic is the intended integration surface with Luna-Flow/arithmetic.