Skip to content

Latest commit

 

History

History
121 lines (99 loc) · 6.57 KB

CHANGELOG.md

File metadata and controls

121 lines (99 loc) · 6.57 KB
Raw

Descriptive changelog

(All dates are DD.MM.YYYY)

0.2.2 - 10.10.2020 - Rework infinite precision division
  • Issue #127 - invalid division when using unlimited precision wasn't completely handled in the previous release. With this release that division case was completely reworked and additional tests were added.
0.2.1 - 8.10.2020 - Fix infinite precision division, kotlin version bump
  • Fixed #127 BigDecimal divide not working properly
  • Added tests to cover case from #127
  • Bumped kotlin to 1.4.10
0.2.0 - 18.8.2020 - Improvements, optimizations, bugfixes and Kotlin 1.4.0
  • Improvement #122 Add "scale" support to BigDecimal
  • Fixed #118 Rounding issue on division
  • Improvement #116 Introduce local gradle varibale to enable JVM target on different platforms
  • Fixed #112 BigDecimal divideAndRemainder bad result
  • Fixed #111 - BigDecimal narrowing
  • Fixed #104 - BigInteger.toByteArray should prefer ByteArray over Array
  • Fixed #103 - Removed coroutines
  • Fixed #90 - Incorrect Result From Negative Double
  • Fixed #88 - BigInteger.bitAt does not work correctly
  • Fixed #86 - Rounding fails when integer part is 9
  • Fixed #88 - BigInteger.bitAt() returns invalid value
  • Built with Kotlin 1.4-M3
  • Reworked to(U)ByteArray conversion methods
    • from and to conversions were not consistent, from(U)ByteArray expected a string of bytes ordered in little or big endian, while to(U)ByteArray produced Int or Long represented as byte with little endian or big endian order.
    • Replaced with a consistent to and from byte/ubyte array conversions
      • fromUByteArray always expects a big-endian ordered array of unsigned bytes
      • fromByteArray always expects a big-endian ordered array of bytes
      • toUByteArray produces unsigned byte array in big-endian order
      • toByteArray produces signed byte array in big-endian order
    • There are two helper methods that convert to and from a two's complement ByteArray, this form conforms to Java BigIntegers toByteArray
      • fromTwosComplementByteArray expects a two's complement ByteArray with at least one sign bit
      • toTwosComplementByteArrayproduces a two's complement ByteArray with at least one sign bit
  • Added secureOverwrite to BigNumber interface, with role of overwriting backing structures with zeroes. It's meant to be used by libraries that require such a functionlity (i.e. crypto). The function also breaks immutability contract of BigNumber implementations, and further operations with that instances have undefined results.
0.1.5 - 07.01.2020 - Adding linux arm targets, BigDecimal bug fixes
  • Version bump to kotlin 1.3.61
  • Gradlew wrapper bump to 6.0.1
  • Added Linux Arm 64, and Linux Arm 32 HFP build targets
  • Fixed smaller issues
0.1.4 - 10.12.2019 - BigDecimal improvements, MINGW64 and MINGW86 targets added
  • Main library now has dependancies only on the kotlin standard library (for now, coroutines will be coming back at some point in the future).
  • Renamed BigDecimal round method to roundSignificand, as it describes what it does more precisely
  • Added roundAtDigitPosition and roundAfterDecimalPoint convenience methods.
  • Use Long instead of BigInteger for BigDecimal exponent.
  • Adding MingwX64 and MingwX86 targets.
0.1.3 - 19.11.2019 - Kotlin and Gradle version bump
  • Kotlin version bump to 1.3.60 - especially important as there was a comparison performance improvement for inline classes which are heavily used by bignum library (ULong/UInt).
  • Actually bumped gradle to 5.6.1, it was mistakenly left at 5.1.1 in library versions 0.1.1 and 0.1.2
  • Cleaned up dependancies, coroutines are currently used only in test modules so they are moved there.
0.1.2 - 17.11.2019 - Performance improvements
  • Removed removeLeadingZeros and replaced with countLeadingZeros and appropriate algorithm changes.
  • Realized plural of zero is zeros not zeroes.
  • Improved exponentiation to use square-and-multiply method.
0.1.1 - 19.10.2019 - Multiplication algorithm improvements
  • Implemented Toom-Cook-3 multiplication, although still slow because of inefficient division.
  • Bumped gradle version to 5.6.1, which means that the published Gradle Metadata will be 1.0, making metadata resolution available only on Gradle >= 5.3
  • Fixed several issues related to big decimal comparison, modular integer sign, etc.
  • Added more BigDecimal extension functions.
  • Added ModularBigInteger extension functions.
  • Added Karatsuba multiplication.
  • Added copy and moveDecimalPoint methods.
  • Added fromUByteArray and toUByteArray.
0.1.0 - 31.7.2019 ByteArray support and other improvements
  • Added toByteArray and fromByteArray.
  • Added toFloat and toDouble to BigInteger and ModularBigInteger classes.
  • Added BigInteger creation from Float and Double by using tryFromFloat and tryFromDouble, with optional exact parameter to preserve precision.
  • Added BigInteger comparison with Float and Double.
  • Added BigDecimal configuration option to switch to expanded representation instead of scientific when calling toString().
  • Improved ModularBigInteger exponentiation algorithm, based on Bruce Schneier Applied Cryptography pesudocode.
0.0.9 - 11.5.2019 Adding modular integer support, changing api
  • Added modular integers - ModularBigInteger.
  • Added modInverse method to BigInteger.
  • Extracted interfaces that model big numbers better (BigNumber interface and accompanying interfaces)
  • Implemented integer reciprocal based on newton iteration (Based on paper by Yiping Cheng, Ze Lie : Refinement of a newton reciprocal algorithm for arbitrary precision numbers)
  • Implemented division by reciprocal multiplication. Not used by division at the moment as it is unoptimized and slower than basecase division in early benchmarks.
  • Fixed a bug in Int32 shift right when shift amount was an exact multiple of word size
  • Added constructor overloads
  • Added value methods (intValue, longValue...)
  • Renamed invPrecise() bigInteger method to not()
  • Renamed numberOfDigits() to numberOfDecimalDigits()
  • Introduced BigNumber and BitwiseOperations interfaces
  • Added iOS ARM 32bit support
0.0.8 - 02.04.2019 - Minor update
  • Released ios (for X64 and arm) and macos X64 artifacts
  • No functional changes
0.0.7 - 31.3.2019 Minor update
  • Added extension functions for String to BigInteger and BigDecimal
  • Fixed a couple of parsing issues
  • Added this changelog
0.0.6 - 30.3.2019 - Adding floating point support/BigDecimal
  • Added BigDecimal
  • Fixed several BigInteger bugs

0.0.5 - 20.3.2019 - Initial release/BigInteger

  • Added Big integer support