Print

The library provides consistent, human-readable formatting across decimal32 · decimal64 · decimal128 Display is available via: std::string conversion , operator<< (ostream) std::format and low-level BID helpers (display_bidXX, print)

math::decimal_t<uint32_t> a{"123.45"};

std::string s = static_cast<std::string>(a); // "123.45"
std::cout << a;                              // "123.45"
std::format("value={}", a);                  // "value=123.45"

Low-level formatting operates directly on raw BID values:

math::bid::display_bid32(uint32_t)
math::bid::display_bid64(uint64_t)
math::bid::display_bid128(BID_UINT128)

Example:

auto raw = math::bid::uint32_to_bid32(12345, -2);
math::bid::display_bid32(raw);  // "123.45"

Numbers are rendered in normalized decimal form:

Internal Output
(12345, -2) "123.45"
(5, -3) "0.005"
(42, 3) "4200"
  • Negative values include -
  • No scientific notation for decimal32 / decimal64
  • decimal128 is normalized to fixed notation

Zero is canonically normalized:

0, -0, 0E±k → "0"
- Sign and exponent are ignored - All representations collapse to "0"

Value Output
NaN "nan"
+∞ "+inf"
-∞ "-inf"

Example:

math::bid::display_bid32(0x78000000u);  // "+inf"
math::bid::display_bid32(0xf8000000u);  // "-inf"
math::bid::display_bid32(0x7c000000u);  // "nan"

Internally, decimal128 may use scientific form:

+125E-2

Display is normalized:

"1.25" → "1.25"

"3.141592653589793238462643383279502"
→ "3.141592653589793238462643383279502"
  • No E notation in output
  • Precision: 34 significant digits
  • Inputs exceeding precision are rounded
"3.1415926535897932384626433832795029"
→ "3.141592653589793238462643383279503"


decimal32{0.1} → "0.1000000"
decimal64{0.1} → "0.1000000000000000"
This reflects: \(\text{binary float} \neq \text{exact decimal value}\)

  • Full decimal expansion is preserved
  • No “pretty-printing” heuristics

Internal helper used by all display functions:

math::bid::print(sign, mantissa, exponent, buffer, size)

Behavior:

  • Places decimal point according to exponent
  • Pads with zeros
  • Preserves trailing zeros 
print(false, 123, -2) → "1.23"
print(false, 5, -3)   → "0.005"
print(false, 42, 3)   → "4200"
Area Note
Float inputs May expand (e.g. "0.1000000")
Zero normalization Always rendered as "0"
Invalid input Always mapped to NaN
Scientific Notation Exponent is off by 1, see: test/thirdparty.cpp