Enum OracleNumber 

pub enum OracleNumber {
    Inline {
        coefficient_le: [u8; 16],
        scale: i16,
        is_integer: bool,
    },
    Text {
        text: Box<str>,
        is_integer: bool,
    },
}

Inline, lossless decimal carrier for an Oracle NUMBER.

The common case is OracleNumber::Inline (coefficient × 10^-scale, allocation-free). Values that cannot be represented exactly inline fall back to OracleNumber::Text (a boxed canonical-text carrier).

Variants

Inline

value == coefficient × 10^-scale, with the sign carried in coefficient. scale may be negative (the value has trailing zeros to the left of the implied point). is_integer mirrors the legacy decoder’s flag — whether the canonical text contains a decimal point — so the Python int-vs-float dispatch is preserved exactly.

The coefficient is stored as its little-endian i128 bytes rather than a bare i128 field: a bare i128 forces 16-byte alignment, which rounds the enum up to 32 bytes and would blow QueryValue’s 32-byte budget once the discriminant is added. The [u8; 16] form keeps 8-byte alignment so OracleNumber is 24 bytes. Access via OracleNumber::coefficient.

Fields

coefficient_le: [u8; 16]

scale: i16

is_integer: bool

Text

Defensive fallback for values that do not fit the inline form exactly (39–40 significant digit integers that overflow i128, or the -1e126 single-byte sentinel). Boxed so the enum stays small.

Fields

text: Box<str>

is_integer: bool

Implementations

impl OracleNumber

pub fn coefficient(&self) -> Option<i128>

The inline coefficient as an i128, or None for the boxed-text fallback. value == coefficient × 10^-scale.

pub fn scale(&self) -> Option<i16>

The inline scale, or None for the boxed-text fallback.

pub fn from_wire(bytes: &[u8]) -> Result<Self>

Decode an Oracle NUMBER wire form into the inline representation, falling back to a boxed canonical-text carrier when the value cannot be represented exactly inline. The canonical text — whether produced inline or stored in the fallback — is byte-identical to the legacy decoder.

ZERO-ALLOCATION for the common inline case: the digit walk writes into a fixed stack buffer (Oracle NUMBER has at most 40 significant digits), and the inline coefficient/scale is folded directly — no scratch Vec/String is heap-allocated. Only the rare text fallback (sentinel / i128 overflow) touches the heap, and only then.

pub fn from_canonical_text(text: &str) -> Self

Construct from already-canonical decimal text (the bind / parse path). Parses the text into the inline form when it fits, else keeps it boxed. The text MUST already be canonical Oracle NUMBER text (the form the decoder emits). Integral trailing-zero values are folded into the same coefficient/negative-scale form the wire decoder emits.

pub fn from_canonical_text_with_flag(text: &str, is_integer: bool) -> Self

Like Self::from_canonical_text but with the caller-supplied is_integer flag (the borrowed fetch path already decoded it from the wire, so it is authoritative — preserve it verbatim).

pub fn as_borrowed_text(&self) -> Option<&str>

Borrow the canonical text when it is stored as boxed text (the fallback form), else None — the inline numeric form synthesizes its text on demand and has no &str to lend.

pub fn is_integer(&self) -> bool

Whether the canonical text is integral (carries no decimal point). Mirrors the legacy is_integer flag exactly.

pub fn fmt_into(&self, out: &mut String)

THE single shared canonical formatter. Appends the canonical decimal text to out. Byte-identical to super::codecs::decode_number_text_into.

pub fn to_canonical_string(&self) -> String

Canonical decimal text as an owned String.

pub fn to_canonical_cow(&self) -> Cow<'_, str>

Canonical decimal text as a Cow: borrowed for the boxed-text fallback (zero allocation), owned for the inline form (formatted once on demand).

pub fn to_i64(&self) -> Option<i64>

Exact i64 when the value is an integer that fits; else None.

pub fn to_i128(&self) -> Option<i128>

Exact i128 when the value is an integer that fits; else None.

Trait Implementations

impl Clone for OracleNumber

fn clone(&self) -> OracleNumber

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for OracleNumber

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for OracleNumber

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for OracleNumber

impl PartialEq for OracleNumber

fn eq(&self, other: &OracleNumber) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for OracleNumber