# CHANGES

## 4.1.3

* Faster gamma and lgamma calculation using Lagrange interpolation [GH-524]

  **@tompng**

* Bit-burst algorithm version of BigMath.erf and BigMath.erfc [GH-458]

  **@tompng**

* Fix Newton-Raphson division bugs: precision loss, out-of-bounds write, and missing GC guard for the divisor [GH-548] [GH-552] [GH-553]

  **@OskarEichler** **@tompng**

* Fix GC safety issues around RB_GC_GUARD and string arguments [GH-536] [GH-556]

  **@tompng**

* Honor `exception: false` for Complex and strings with null bytes [GH-549] [GH-550]

  **@OskarEichler**

* Make String#to_d accept embedded NUL bytes and reject ASCII-incompatible strings [GH-557]

  **@tompng**

* Correct nullable BigDecimal conversion RBS signature [GH-551]

  **@OskarEichler**

* Declare TypedData with RUBY_TYPED_THREAD_SAFE_FREE [GH-541]

  **@byroot**

* Fix unused-but-set-global warnings in clang-23 [GH-540]

  **@nobu**

## 4.1.2

* Fix dtoa Ractor-safety bug. Update dtoa to version from Ruby 4.0 [GH-528]

  **@jhawthorn**

* Optimize BigDecimal#to_s [GH-519]

  **@byroot**

## 4.1.1

* Make BigDecimal object embedded [GH-507]

  **@byroot**

* Multiplication with 16-decdig batch [GH-501] [GH-511]

  **@tompng**

## 4.1.0

* Drop Ruby 2.5 support [GH-505]

  **@tompng**

* Performance improvements: NTT multiplication, Newton-Raphson division, bit-burst algorithm for exp/sin, Gauss-Legendre for PI, improved log, and faster add/sub for large exponent differences [GH-407] [GH-433] [GH-434] [GH-478] [GH-484]

  **@tompng**

* Remove ENABLE_NUMERIC_STRING flag [GH-479]

  **@tompng**

* Add RBS signature and testing [GH-488] [GH-492]

  **@ksss**

* Fix erfc(x,prec) precision when x is huge [GH-502]

  **@tompng**

* Fix error compiling with ruby.wasm [GH-504]

  **@tompng**

## 4.0.1

* Fix warning [GH-475]

  **@tompng**

## 4.0.0

* `BigDecimal#divmod` return value changed to `[Integer, BigDecimal]` [GH-312]

  **@mrzasa**

* Remove `BigDecimal#precs` [GH-470]

  **@tompng**

* BigMath now supports all functions defined in Math module [GH-336] [GH-357] [GH-451] [GH-448]

  **@tompng**

* Fix incorrect exception when exponent is fractional for Infinity base [GH-453]

  **@troy-dunamu**

* Deprecate `bigdecimal/jacobian`, `bigdecimal/ludcmp` and `bigdecimal/newton` [GH-471]

  **@tompng**

## 3.3.1

* All BigMath methods converts non integer precision with to_int

  **@tompng**

## 3.3.0

* Allow calling to_d without arguments [GH-421]

  **@fsateler**

* Calculate BigMath.sin and cos in relative precision [GH-422]

  **@tompng**

* Add support for tangent function [GH-231]

  **@rhannequin**

* BigMath methods accepts numeric as an argument [GH-415]

  **@tompng**

* Round result of sqrt and BigMath methods [GH-427]

  **@tompng**

## 3.2.3

* Allow BigDecimal accept Float without precision [GH-314]

  **@mrzasa**

* Ruby implementation pow, log, exp and sqrt [GH-347] [GH-381]

  **@tompng**

* Update document [GH-348] [GH-360] [GH-365]

  **@timcraft** **@dduugg** **@mame**

* Lots of bug fixes and refactoring

## 3.2.2

* Make precision calculation in bigdecimal.div(value, 0) gc-compaction safe. [GH-340]

  **@tompng**

## 3.2.1

* Fix division precision limit. [GH-335]

  **@tompng**

## 3.2.0

* Fix division rounding. [GH-330] [GH-328]

  **@tompng**

* Fix exponential precision growth in division. [GH-329] [GH-220] [GH-222] [GH-272]

  **@tompng**

## 3.1.9

* Accept no digits in the fractional part (#302)

  **@kou**

## 3.1.8

* Add missing documents [GH-277]

  **@nobu**

* Fix memory leak in VpAlloc [GH-294] [GH-290]

  Reported by **@MaxLap**

## 3.1.7

* Only consists of CI settings and test changes.
  This release is needed for developing Ruby to run `make test-bundled-gems` with the released version of bigdecimal.

## 3.1.6

* Add LICENSE file to gem files [GH-282]

  **@oleksii-leonov**

## 3.1.5

* Add .to_s('F') digit grouping for integer part [GH-264]

  **@cryptogopher**

## 3.1.4

* Handle correctly #reminder with infinity. Fixes [GH-187] [GH-243]

  **@mrzasa**

## 3.1.3

* Adjust a local variable type to exponent. [GH-223]
* Remove checks for `struct RRational` and `struct RComplex` . [GH-233]
* Suppress macro redefinition warnings. [GH-239]

## 3.1.2

* Fix the maximum precision of the quotient. [GH-220]

  Reported by @grk

## 3.1.1

* Fix the result precision of `BigDecimal#divmod`. [GH-219]

  **Kenta Murata**

## 3.1.0

* Improve documentation [GH-209]

  **Burdette Lamar**

* Let BigDecimal#quo accept precision. [GH-214] [Bug #8826]

  Reported by Földes László

* Allow passing both float and precision in BigDecimal#div. [GH-212] [Bug #8826]

  Reported by Földes László

* Add `BigDecimal#scale` and `BigDecimal#precision_scale`

  **Kenta Murata**

* Fix a bug of `BigDecimal#precision` for the case that a BigDecimal has single internal digit [GH-205]

  **Kenta Murata**

* Fix segmentation fault due to a bug of `VpReallocReal`

  **Kenta Murata**

* Use larger precision in divide for irrational or recurring results. [GH-94] [Bug #13754]

  Reported by Lionel PERRIN

* Fix negative Bignum conversion [GH-196]

  **Jean byroot Boussier**

* Fix grammar in error messages. [GH-196]

  **Olle Jonsson**

* Improve the conversion speed of `Kernel#BigDecimal` and `to_d` methods.

  **Kenta Murata**

* Fix trailing zeros handling in `rb_uint64_convert_to_BigDecimal`. [GH-192]

  Reported by @kamipo

* Permit 0 digits in `BigDecimal(float)` and `Float#to_d`.
  It means auto-detection of the smallest number of digits to represent
  the given Float number without error. [GH-180]

  **Kenta Murata**

* Fix precision issue of Float. [GH-70] [Bug #13331]

  Reported by @casperisfine

## 3.0.2

*This version is totally same as 3.0.0.  This was released for reverting 3.0.1.*

* Revert the changes in 3.0.1 due to remaining bugs.

## 3.0.1

*This version is yanked due to the remaining bugs.*

## 3.0.0

* Deprecate `BigDecimal#precs`.

  **Kenta Murata**

* Add `BigDecimal#n_significant_digits`.

  **Kenta Murata**

* Add `BigDecimal#precision`.

  **Kenta Murata**

* Ractor support.

  **Kenta Murata**

* Fix a bug of the way to undefine `allocate` method.

  **Kenta Murata**

* Fix the default precision of `Float#to_d`.
  [Bug #13331]

  **Kenta Murata**

## 2.0.2

* Deprecate taint/trust and related methods, and make the methods no-ops

  **Jeremy Evans**

* Make BigDecimal#round with argument < 1 return Integer

  **Jeremy Evans**

* Use higher default precision for BigDecimal#power and #**

  **Jeremy Evans**
  **Kenta Murata**

## 2.0.1

* Let BigDecimal#to_s return US-ASCII string

  **Kenta Murata**

## 2.0.0

* Remove `BigDecimal.new`

  **Kenta Murata**

* Drop fat-gem support

  **Akira Matsuda**

* Do not mutate frozen BigDecimal argument in BigMath.exp

  **Jeremy Evans**

* Make Kernel#BigDecimal return argument if given correct type
  [Bug #7522]

  **Jeremy Evans**

* Undef BigDecimal#initialize_copy

  **Jeremy Evans**

* Support conversion from Complex without the imaginary part

  **Kenta Murata**

* Remove taint checking

  **Jeremy Evans**

* Code maintenance

  **Nobuyoshi Nakada**

## 1.4.4

* Fix String#to_d against the string with trailing "e" like "1e"

  **Ibrahim Awwal**

* Add BigDecimal.interpret_loosely, use it in String#to_d,
  and remove bigdecimal/util.so and rmpd_util_str_to_d

  **Kenta Murata**

## 1.4.3

* Restore subclassing support

  **Kenta Murata**

## 1.4.2

* Fix gem installation issue on mingw32.

  **Kenta Murata**

## 1.4.1

* Fix wrong packaging.

  **Ben Ford**

## 1.4.0

* Update documentation of `exception:` keyword.

  **Victor Shepelev**

* Fix VpAlloc so that '1.2.3'.to_d is 1.2

  **Nobuyoshi Nakada**

* Restore `BigDecimal.new` just for version 1.4 for very old version of Rails

  **Kenta Murata**

* Support `exception:` keyword in `BigDecimal()`

  **Kenta Murata**

* Remove `BigDecimal#initialize`

  **Kenta Murata**

* Fix the string parsing logic in `BigDecimal()` to follow `Float()`

  **Kenta Murata**

* Fix `String#to_d` to follow `String#to_f`

  **Kenta Murata**

* Update `BigDecimal#inspect` documentation

  **Dana Sherson**

* Remove `BigDecimal.ver`, `BigDecimal.allocate`, and `BigDecimal.new`

  **Kenta Murata**

* No more support Ruby < 2.3

  **Kenta Murata**

* Make BigDecimal objects frozen

  **Kenta Murata**

* Remove division by zero from the internal implementation

  **Urabe, Shyouhei**

## 1.3.5

* Add NilClass#to_d

  **Jose Ramon Camacho**

## 1.3.4

* Stop deprecation warning in dup and clone, and just return self

  **Kenta Murata**

* Improve warning message in BigDecimal.new

  **Masataka Pocke Kuwabara**

## 1.3.3

* Introduce BigDecimal::VERSION, deprecate BigDecimal.ver, and follow this version string to gem's version.

  **Kenta Murata**

* Deprecate BigDecimal.new

  **Kenta Murata**

* Deprecate BigDecimal#clone and #dup

  **Kenta Murata**

* Relax the dependent version of rake-compiler

  **yui-knk**

* Works for better windows cross-compilation support

  **SHIBATA Hiroshi**

* Use Bundler::GemHelper.install_tasks instead of bundler/gem_tasks

  **SHIBATA Hiroshi**

* Remove the old gemspec file

  **yui-knk**

* Fix for mathn removal in Ruby 2.5

  **SHIBATA Hiroshi**

* Update ruby versions in .travis.yml

  **Jun Aruga**

* Add tests for BigDecimal#truncate

  **yui-knk**

* Add tests for BigDecimal#round

  **yui-knk**

* Fix error message for precision argument

  **Marcus Stollsteimer**

## 1.3.2 and older

Omitted.