{
"markdown": "# decimal-scaled\n\n[](https://crates.io/crates/decimal-scaled) [](https://docs.rs/decimal-scaled) [](https://blog.rust-lang.org/) [](#license) [](https://github.com/mootable/decimal-scaled/actions/workflows/docs.yml) [](https://www.bestpractices.dev/projects/12895) [](https://securityscorecards.dev/viewer/?uri=github.com/mootable/decimal-scaled) [](https://github.com/mootable/decimal-scaled/actions/workflows/cargo-audit.yml) [](https://github.com/mootable/decimal-scaled/actions/workflows/codeql.yml)\n\n**[Docs](https://mootable.github.io/decimal-scaled/)** • **[Performance](https://mootable.github.io/decimal-scaled/performance/)** • **[Comparisons](https://mootable.github.io/decimal-scaled/comparisons/)** • **[Algorithms](https://mootable.github.io/decimal-scaled/ALGORITHMS/)** • **[Roadmap](https://mootable.github.io/decimal-scaled/ROADMAP/)** • **[API reference](https://docs.rs/decimal-scaled)**\n\nA fast, precise, decimal library.\n\n- **Decimal storage** — unlike floating point, `1.1` is exactly `1.1`.\n- **Multiple rounding modes** — eight in total, `HalfToEven` by default.\n- **Up to 4 Kb numbers** — twelve widths, `D18` to `D1232`.\n- **Choose your precision at compile time** — from 0 to one less than the width. i.e. D1232 can carry a precision of 1231 digits.\n- **`no_std` friendly** — the strict, integer-only path needs no `std`.\n- **Validated by 96,159,960 value tests** — every width × scale × rounding mode.\n\n---\n\n## Install\n\n\n```toml\n[dependencies]\ndecimal-scaled = { version = \"0.5\", features = [\"macros\"] }\n```\n\n\n## First use\n\n```rust\nuse decimal_scaled::{d38, D38s12};\n\n// The most common way to make a value: a literal, scale inferred from\n// its digits.\nlet x = d38!(1.564232); // D38<6>\nassert_eq!(x.to_string(), \"1.564232\");\n\nlet price: D38s12 = \"19.99\".parse().unwrap();\nlet qty = D38s12::try_from(3i64).unwrap(); // integer, scaled by 10^SCALE (fallible)\nlet total = price * qty; // 59.97 exactly\n\nassert_eq!(total.to_string(), \"59.97\");\nassert_eq!(total, d38!(59.97, scale 12));\n\n// Transcendentals are correctly rounded to <= 0.5 ULP, integer-only,\n// and bit-identical across platforms.\nlet sqrt2 = d38!(2, scale 12).sqrt_strict();\n```\n\n## Why decimal-scaled\n\n| You need… | decimal-scaled gives you… |\n|---|---|\n| Decimal arithmetic that doesn't drift (`0.1 + 0.2 == 0.3`) | Base-10 storage; exact `+ - %`, correctly-rounded `* /`. |\n| Bit-identical results across Linux / macOS / Windows / ARM / x86 | `*_strict` transcendentals — integer-only, no platform libm. |\n| Compile-time-fixed precision with zero per-value scale byte | Const-generic `D38<19>`, `D76<35>` etc. — scale is in the type. |\n| `no_std` (or `no_std + alloc`) | Builds under `no_std + alloc` with `default-features = false`; the strict, integer-only path needs no libm. |\n| Correctly-rounded `ln` / `exp` / `sin` / `cos` / `tan` / `sqrt` / `atan` and friends — by default | Within 0.5 [ULP](https://en.wikipedia.org/wiki/Unit_in_the_last_place), `HalfToEven` by default; switch per call via `*_with(mode)` or crate-wide via the `rounding-*` features. |\n\n---\n\n## Documentation\n\nFull docs: .\n\n- [Getting started](docs/getting-started.md) — constructing values, arithmetic, formatting, parsing.\n- [The width family](docs/widths.md) — `D18` through `D1232`, scale aliases, the `Decimal` trait, picking a tier.\n- [Conversions](docs/conversions.md) — integers, floats, cross-width widening / narrowing, the float bridge.\n- [Cross-scale operations](docs/cross-scale.md) — mixed-width, mixed-`SCALE` expressions via `mul_of` / `add_of` / `cmp_of` / `clamp_of`, plus the nightly auto-inferred form.\n- [Rounding](docs/rounding.md) — `RoundingMode`, the `_with` pairs, `quantize` / `requantize`, the compile-time `rounding-*` features.\n- [Strict vs fast transcendentals](docs/strict-mode.md) — the integer-only `*_strict` path and the 0.5 ULP guarantee.\n- [The `d*!` macros](docs/macros.md) — compile-time decimal literals and scale inference.\n- [Cargo features](docs/features.md) — every feature flag and the common configurations.\n- **Bench** — per-width [Performance](docs/performance.md), the [Precision](docs/precision.md) surface, version [History](docs/history.md), the golden [Harness](docs/golden.md), and the like-for-like [Comparisons](docs/comparisons.md) against the top crates.io peers.\n- [Algorithms](ALGORITHMS.md) — every kernel with its citation (Möller–Granlund, Brent, Knuth, Karatsuba, Burnikel–Ziegler, Cody–Waite, …).\n- [Roadmap](ROADMAP.md) · [Changelog](CHANGELOG.md) · [Contributing](CONTRIBUTORS.md).\n\nAPI reference: .\n\n---\n\n## License\n\nLicensed under either of:\n\n- MIT license ([LICENSES/MIT.md](LICENSES/MIT.md))\n- Apache License, Version 2.0 ([LICENSES/Apache-2.0.md](LICENSES/Apache-2.0.md))\n\nat your option.\n\nCopyright 2026 John Moxley. Third-party code attributions are listed in\n[LICENSE-THIRD-PARTY](LICENSE-THIRD-PARTY).\n",
"bytes": 5881,
"sha": "fb28d710d8c06f289e43ad4adf237d134ec08a0d3c3608f8b956b74b1bf7d0a4",
"repo_slug": "mootable/decimal-scaled",
"fonte": "repo",
"truncated": false,
"api": "https://agentalog.com/api/listings/plg_mootable_decimal_scaled_decimal_scaled_72af776d/readme"
}