Common compiled languages keep a visible separation between integer and floating point numbers. The purpose of this package is to allow performing mathematical operations without having to do type conversions:

• where possible / reasonable, floating point numbers are converted to integers;
• operations that require converting integers to floating point are deferred until a native numerical value is requested.

NOTE: Support for radicals is very experimental.

For the reasons outlined at https://forums.swift.org/t/why-does-swift-not-support-rational-numbers/62725, this kind of abstraction isn't feasible in a general-purpose useful way. Noticeably, the performance penalty and the increased frequency of overflow issues make it applicable only in a subset of circumstances. You should kee that in mind if choosing this package for some specific use.

Goals of this package:

• ease of use
• correctness
• preserving integer math

Not goals:

• performance
• completeness of mathematical API

    let m: Nuggle = 2.5
    let n: Nuggle = 1.5
    let p: Nuggle = 1

    let ex: Nuggle = 2.5 + 1.5 + 1

    XCTAssertEqual(m + n + p, ex)

    XCTAssertEqual(m + n, 4)
    XCTAssertEqual(m + n + p, 5)
    XCTAssertEqual(m + n + p, 4 + 1.6/4.8 + 2/3)

    XCTAssertEqual(m / n, 5 / 3)
    XCTAssertEqual((m / n).description, "5/3")

    XCTAssertEqual(ex.exactInt(), 5)

    XCTAssertEqual((m * 3).double(), 7.5)

Except where/if otherwise specified, all the files in this package are copyright of the package contributors mentioned in the NOTICE file and licensed under the Apache 2.0 License, which is permissive for business use.