character_set 1.2.0-java → 1.5.0-java

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 438cb1162ff31a5413f0441ac9c976c1697b9bc8a817e9a1ed78ce08d8f63d56
-  data.tar.gz: 55988d116eaa4b2e5db705703d918409b65971bc193f7d11ef2d9a0a8a93e649
+  metadata.gz: 52823f4f35cdec44378c3828b4b38eba1f9f5bce402a70962eae1fb786132d8c
+  data.tar.gz: b1d6419575a3614675c194cbfde8530be02195cc73365a8c6ca446dd6ea909e6
 SHA512:
-  metadata.gz: eeabb985966b1b60a5264b1cc8692f496472cb96c29649dc40b60321727bb1c98921c87ad0b229e711e9ae49850840a52f7ee9a317cd54262531569d0267b320
-  data.tar.gz: 9c02685dd05d2e3563f83c04879f8e6ed0f0c4cbfeef6259135d313f847d76ceda53a1b64ba34cc121a34ab12231303dc539322d49f68c62bfc657161a1cd0ad
+  metadata.gz: c021975f912100174a5274454cfc6099a0955262e7e5fac619989a0a2aa5d624e048fe8b5f68b167157aca425c771df1bf137be12924b609b8d30dff1608142e
+  data.tar.gz: 1b702ea538bc5a5209c3544c88c9b38d328080db52640bb4a5780454d296970d8c2557ebe9c3cdd014e3a20af254c77fe694a9f56f09f3f29c039aef81dc381f

data/.gitattributes

@@ -0,0 +1,3 @@
+*.cps linguist-detectable=false
+benchmarks/* linguist-detectable=false
+spec/ruby-spec/* linguist-vendored

data/.github/workflows/gouteur.yml

@@ -0,0 +1,20 @@
+name: gouteur
+
+on: [push, pull_request]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Ruby
+        uses: ruby/setup-ruby@v1
+        with:
+          ruby-version: 2.7
+      - name: Prepare
+        run: |
+          bundle install --jobs 4
+          bundle exec rake compile
+      - name: Test
+        run: bundle exec gouteur

data/.github/workflows/lint.yml

@@ -0,0 +1,29 @@
+# based on https://github.com/rails/rails/blob/4a78dcb/.github/workflows/rubocop.yml
+
+name: rubocop linting
+
+on: [push, pull_request]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: 2.7
+    - name: Cache gems
+      uses: actions/cache@v1
+      with:
+        path: vendor/bundle
+        key: ${{ runner.os }}-rubocop-${{ hashFiles('**/Gemfile.lock') }}
+        restore-keys: |
+          ${{ runner.os }}-rubocop-
+    - name: Install gems
+      run: |
+        bundle config path vendor/bundle
+        bundle install --jobs 4 --retry 3
+    - name: Run rubocop
+      run: bundle exec rubocop --lint

data/.github/workflows/tests.yml

@@ -0,0 +1,22 @@
+name: tests
+
+on: [push, pull_request]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    strategy:
+      matrix:
+        ruby: [ '2.2', '2.7', '3.0', 'ruby-head', 'jruby-head' ]
+
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Ruby ${{ matrix.ruby }}
+        uses: ruby/setup-ruby@v1
+        with:
+          ruby-version: ${{ matrix.ruby }}
+      - name: Install dependencies
+        run: bundle install --jobs 4
+      - name: Test with Rake
+        run: bundle exec rake

data/.gitignore

@@ -15,6 +15,7 @@
 .ruby-version
 .tags
 .tags1
+.tool-versions
 .vscode
 bbin/
 binstubs/*

data/.gouteur.yml

@@ -0,0 +1,2 @@
+repos:
+  - uri: https://github.com/jaynetics/js_regex

data/.rubocop.yml

@@ -0,0 +1,17 @@
+AllCops:
+  Exclude:
+    - '**/doc/*'
+    - '**/pkg/*'
+    - '**/spec/ruby-spec/**/*'
+    - '**/vendor/**/*' # vendored dependencies
+  NewCops: enable
+  RubyInterpreters:
+    - ruby
+    - rake
+  TargetRubyVersion: 2.5 # really 2.1, but 2.5 is lowest supported by rubocop
+
+Lint/AmbiguousOperatorPrecedence:
+  Enabled: false
+
+Lint/AmbiguousRegexpLiteral:
+  Enabled: false

data/BENCHMARK.md

@@ -1,50 +1,86 @@
-Results of `rake:benchmark` on ruby 2.6.0preview1 (2018-02-24 trunk 62554) [x86_64-darwin17]
+Results of `rake:benchmark` on ruby 3.0.0p0 (2020-12-25 revision 95aff21468) [x86_64-darwin19]
 
+```
+Counting non-letters
+
+CharacterSet#count_in:  9472902.2 i/s
+        String#count:  2221799.9 i/s - 4.26x slower
+```
 ```
 Detecting non-whitespace
 
- CharacterSet#cover?: 13244577.7 i/s
-       Regexp#match?:  8027017.5 i/s - 1.65x  slower
+ CharacterSet#cover?: 12388427.2 i/s
+       Regexp#match?:  7901676.8 i/s - 1.57x slower
 ```
 ```
 Detecting non-letters
 
- CharacterSet#cover?: 13082940.8 i/s
-       Regexp#match?:  5372589.2 i/s - 2.44x  slower
+ CharacterSet#cover?: 12263689.1 i/s
+       Regexp#match?:  4940889.9 i/s - 2.48x slower
 ```
 ```
 Removing whitespace
 
-CharacterSet#delete_in:   389315.6 i/s
-         String#gsub:   223773.5 i/s - 1.74x  slower
+CharacterSet#delete_in:  2406722.6 i/s
+         String#gsub:   235760.3 i/s - 10.21x slower
 ```
 ```
 Removing whitespace, emoji and umlauts
 
-CharacterSet#delete_in:   470239.3 i/s
-         String#gsub:   278679.4 i/s - 1.69x  slower
+CharacterSet#delete_in:  1653607.6 i/s
+         String#gsub:   272782.9 i/s - 6.06x slower
 ```
 ```
 Removing non-whitespace
 
-CharacterSet#keep_in:  1138461.0 i/s
-         String#gsub:   235287.4 i/s - 4.84x  slower
+CharacterSet#keep_in:  2671038.2 i/s
+         String#gsub:   242551.0 i/s - 11.01x slower
 ```
 ```
 Extracting emoji
 
-CharacterSet#keep_in:  1474472.0 i/s
-         String#gsub:   212269.6 i/s - 6.95x  slower
+CharacterSet#keep_in:  1726496.5 i/s
+         String#gsub:   215609.2 i/s - 8.01x slower
+```
+```
+Extracting emoji to an Array
+
+   CharacterSet#scan:  2373856.1 i/s
+         String#scan:   480000.5 i/s - 4.95x slower
 ```
 ```
 Detecting whitespace
 
-CharacterSet#used_by?: 13063108.7 i/s
-       Regexp#match?:  7215075.0 i/s - 1.81x  slower
+CharacterSet#used_by?: 11988328.7 i/s
+       Regexp#match?:  6758146.8 i/s - 1.77x slower
 ```
 ```
 Detecting emoji in a large string
 
-CharacterSet#used_by?:   246527.7 i/s
-       Regexp#match?:    92956.5 i/s - 2.65x  slower
+CharacterSet#used_by?:   288223.3 i/s
+       Regexp#match?:   102384.2 i/s - 2.82x slower
+```
+```
+Adding entries
+
+    CharacterSet#add:  2538251.2 i/s
+       SortedSet#add:   443925.9 i/s - 5.72x slower
+```
+```
+Removing entries
+
+ CharacterSet#delete:  2487620.8 i/s
+    SortedSet#delete:   628816.1 i/s - 3.96x slower
+```
+```
+Merging entries
+
+  CharacterSet#merge:      551.6 i/s
+     SortedSet#merge:        1.4 i/s - 393.59x slower
+```
+```
+Getting the min and max
+
+ CharacterSet#minmax:   636890.7 i/s
+    SortedSet#minmax:      254.1 i/s - 2506.20x slower
 ```

data/CHANGELOG.md

@@ -4,6 +4,60 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## [1.5.0] - 2021-12-05
+
+### Added
+
+- new codepoints for `::assigned` and `::emoji` predefined sets, as in Ruby 3.1.0
+- latest unicode case-folding data (for `#case_insensitive`)
+- support for passing any Enumerable to `#disjoint?`, `#intersect?`
+  - this matches recent broadening of these methods in `ruby/set`
+- new instance method `#secure_token` (see README)
+- class method `::of` now accepts more than one `String`
+- `CharacterSet::ExpressionConverter` can now build output of any Set-like class
+
+### Fixed
+
+- `CharacterSet::Pure::of_expression` now returns a `CharacterSet::Pure`
+  - it used to return a regular `CharacterSet`
+
+## [1.4.1] - 2020-01-10
+
+### Fixed
+- multiple fixes for Ruby 3
+  - fixed segfault for some `String` manipulation cases
+  - added `sorted_set` as dependency, so `CharacterSet::Pure` (non-C fallback) works
+- fixed error when parsing a `Regexp` with an empty intersection (e.g. `/[a&&]/`)
+
+## [1.4.0] - 2019-06-07
+
+### Added
+- `#to_s_with_surrogate_ranges` / `Writer::write_surrogate_ranges`
+  - allows for much shorter astral plane representations e.g. in JavaScript
+  - thanks to https://github.com/singpolyma for the suggestion and groundwork (#1)
+- improved performance for `#to_s` / `Writer` by avoiding bugged `Range#minmax`
+
+### Fixed
+- '/' is now escaped by default when stringifying so as to work with //-regexp syntax
+
+## [1.3.0] - 2019-04-26
+
+### Added
+- improved `String` manipulation speed
+- improved initialization and `#merge` speed when passing a large `Range`
+- reduced memory consumption by > 90% for most use cases via dynamic resizing
+  - before, every set instance required 136 KB for codepoints
+  - now, 16 bytes for a CharacterSet in ASCII space, 8 KB for one in BMP space etc.
+- `#count_in` and `#scan_in` methods for `String` interaction
+- new predefined sets `::any`/`::all`, `::assigned`, `::surrogate`
+- conversion methods `#assigned_part`, `#valid_part`
+- sectioning methods `#ascii_part`, `#plane(n)`
+- section test methods `#ascii_part?`, `#ascii_ratio`, `#ascii_only?`, `#astral_only?`
+
+### Fixed
+- `#count` now supports passing an argument or block as usual
+- `CharacterSet::Pure#keep_in`, `#delete_in` now preserve the original encoding
+
 ## [1.2.0] - 2019-04-02
 
 ### Added

data/README.md

@@ -1,12 +1,17 @@
 # CharacterSet
 
 [![Gem Version](https://badge.fury.io/rb/character_set.svg)](http://badge.fury.io/rb/character_set)
-[![Build Status](https://travis-ci.org/jaynetics/character_set.svg?branch=master)](https://travis-ci.org/jaynetics/character_set)
+[![Build Status](https://github.com/jaynetics/character_set/workflows/tests/badge.svg)](https://github.com/jaynetics/character_set/actions)
+[![Build Status](https://github.com/jaynetics/character_set/workflows/gouteur/badge.svg)](https://github.com/jaynetics/character_set/actions)
+[![codecov](https://codecov.io/gh/jaynetics/character_set/branch/master/graph/badge.svg)](https://codecov.io/gh/jaynetics/character_set)
 
-A gem to build, read, write and compare sets of Unicode codepoints.
+This is a C-extended Ruby gem to work with sets of Unicode codepoints. It can read and write these sets in various formats and implements the stdlib `Set` interface for them.
+
+It also offers an alternate paradigm of `String` processing which grants much better performance than `Regexp` and `String` methods from the stdlib where applicable (see [benchmarks](./BENCHMARK.md)).
 
 Many parts can be used independently, e.g.:
 - `CharacterSet::Character`
+- `CharacterSet::ExpressionConverter`
 - `CharacterSet::Parser`
 - `CharacterSet::Writer`
 - [`RangeCompressor`](https://github.com/jaynetics/range_compressor)
@@ -37,7 +42,7 @@ CharacterSet.parse('[a-c]')
 CharacterSet.parse('\U00000061-\U00000063')
 ```
 
-If the gems [`regexp_parser`](https://github.com/ammar/regexp_parser) and [`regexp_property_values`](https://github.com/jaynetics/regexp_property_values) are installed, `::of_regexp` and `::of_property` can also be used. `::of_regexp` can handle intersections, negations, and set nesting.
+If the gems [`regexp_parser`](https://github.com/ammar/regexp_parser) and [`regexp_property_values`](https://github.com/jaynetics/regexp_property_values) are installed, `::of_regexp` and `::of_property` can also be used. `::of_regexp` can handle intersections, negations, and set nesting. Regexp's `i`-flag is ignored; call `#case_insensitive` on the result if needed.
 
 ```ruby
 CharacterSet.of_property('Thai') # => #<CharacterSet (size: 86)>
@@ -49,7 +54,7 @@ require 'character_set/core_ext/regexp_ext'
 
 ### Predefined utility sets
 
-`ascii`, `ascii_alnum`, `ascii_letters`, `bmp`, `crypt`, `emoji`, `newline`, `unicode`, `url_fragment`, `url_host`, `url_path`, `url_query`, `whitespace`
+`ascii`, `ascii_alnum`, `ascii_letter`, `assigned`, `bmp`, `crypt`, `emoji`, `newline`, `surrogate`, `unicode`, `url_fragment`, `url_host`, `url_path`, `url_query`, `whitespace`
 
 ```ruby
 CharacterSet.ascii # => #<CharacterSet (size: 128)>
@@ -60,7 +65,7 @@ CharacterSet.non_ascii
 
 ### Interact with Strings
 
-CharacterSet can replace some `Regexp` actions on Strings, at better speed (see [benchmarks](./BENCHMARK.md)).
+`CharacterSet` can replace some types of `String` handling with better performance than the stdlib.
 
 `#used_by?` and `#cover?` can replace some `Regexp#match?` calls:
 
@@ -71,6 +76,7 @@ CharacterSet.ascii.cover?('Tr') # => true
 ```
 
 `#delete_in(!)` and `#keep_in(!)` can replace `String#gsub(!)` and the like:
+
 ```ruby
 string = 'Tüür'
 
@@ -84,6 +90,13 @@ CharacterSet.ascii.keep_in!(string) # => ''
 string # => ''
 ```
 
+`#count_in` and `#scan` can replace `String#count` and `String#scan`:
+
+```ruby
+CharacterSet.non_ascii.count_in('Tüür') # => 2
+CharacterSet.non_ascii.scan_in('Tüür') # => ['ü', 'ü']
+```
+
 There is also a core extension for String interaction.
 ```ruby
 require 'character_set/core_ext/string_ext'
@@ -100,7 +113,7 @@ require 'character_set/core_ext/string_ext'
 
 ### Manipulate
 
-Use any [Ruby Set method](https://ruby-doc.org/stdlib-2.5.1/libdoc/set/rdoc/Set.html), e.g. `#+`, `#-`, `#&`, `#^`, `#intersect?`, `#<`, `#>` etc. to interact with other sets. Use `#add`, `#delete`, `#include?` etc. to change or check for members.
+Use [any Ruby Set method](https://ruby-doc.org/stdlib-2.5.1/libdoc/set/rdoc/Set.html), e.g. `#+`, `#-`, `#&`, `#^`, `#intersect?`, `#<`, `#>` etc. to interact with other sets. Use `#add`, `#delete`, `#include?` etc. to change or check for members.
 
 Where appropriate, methods take both chars and codepoints, e.g.:
 
@@ -122,13 +135,13 @@ non_a.include?('ü') # => true
 
 # surrogate pair halves are not included by default
 CharacterSet['a'].inversion(include_surrogates: true)
-# => #<CharacterSet (size: 1114111)>
+# => #<CharacterSet (size: 1114112)>
 ```
 
 `#case_insensitive` can be used to create a `CharacterSet` where upper/lower case codepoints are supplemented:
 
 ```ruby
-CharacterSet['1', 'a'].case_insensitive # => CharacterSet['1', 'A', 'a']
+CharacterSet['1', 'A'].case_insensitive # => CharacterSet['1', 'A', 'a']
 ```
 
 ### Write
@@ -156,18 +169,44 @@ set.to_s(escape_all: true) { |c| "<#{c.hex}>" } # => "<61>-<63><258><1F929>"
 # disable abbreviation (grouping of codepoints in ranges)
 set.to_s(abbreviate: false) # => "abc\u0258\u{1F929}"
 
-# for full js regex compatibility in case of astral members:
-set.to_s_with_surrogate_alternation # => '(?:[\u0258]|\ud83e\udd29)'
+# astral members require some trickery if we want to target environments
+# that are based on UTF-16 or "UCS-2 with surrogates", such as JavaScript.
+set = CharacterSet['a', 'b', '🤩', '🤪', '🤫']
+
+# Use #to_s_with_surrogate_ranges e.g. for JavaScript:
+set.to_s_with_surrogate_ranges
+# => '(?:[ab]|\uD83E[\uDD29-\uDD2B])'
+
+# Or use #to_s_with_surrogate_alternation if such surrogate set pairs
+# don't work in your target environment:
+set.to_s_with_surrogate_alternation
+# => '(?:[ab]|\uD83E\uDD29|\uD83E\uDD2A|\uD83E\uDD2B)'
+```
+
+### Other features
+
+#### Secure tokens
+
+Generate secure random strings of characters from a set:
+
+```ruby
+CharacterSet.new('a'..'z').secure_token(8) # => "ugwpujmt"
+CharacterSet.crypt.secure_token # => "8.1w7aBT737/pMfcMoO4y2y8/=0xtmo:"
 ```
 
-### Unicode plane methods
+#### Unicode planes
 
-There are some methods to check for planes and to handle [BMP](https://en.wikipedia.org/wiki/Plane_%28Unicode%29#Basic_Multilingual_Plane) and astral parts:
+There are some methods to check for planes and to handle ASCII, [BMP](https://en.wikipedia.org/wiki/Plane_%28Unicode%29#Basic_Multilingual_Plane) and astral parts:
 ```Ruby
+CharacterSet['a', 'ü', '🤩'].ascii_part # => CharacterSet['a']
+CharacterSet['a', 'ü', '🤩'].ascii_part? # => true
+CharacterSet['a', 'ü', '🤩'].ascii_only? # => false
+CharacterSet['a', 'ü', '🤩'].ascii_ratio # => 0.3333333
 CharacterSet['a', 'ü', '🤩'].bmp_part # => CharacterSet['a', 'ü']
 CharacterSet['a', 'ü', '🤩'].astral_part # => CharacterSet['🤩']
 CharacterSet['a', 'ü', '🤩'].bmp_ratio # => 0.6666666
 CharacterSet['a', 'ü', '🤩'].planes # => [0, 1]
+CharacterSet['a', 'ü', '🤩'].plane(1) # => CharacterSet['🤩']
 CharacterSet['a', 'ü', '🤩'].member_in_plane?(7) # => false
 CharacterSet::Character.new('a').plane # => 0
 ```

data/Rakefile

@@ -7,6 +7,13 @@ RSpec::Core::RakeTask.new(:spec)
 
 task default: :spec
 
+namespace :spec do
+  task :quick do
+    ENV['SKIP_MEMSAFETY_SPECS'] = '1'
+    Rake::Task[:spec].invoke
+  end
+end
+
 Rake::ExtensionTask.new('character_set') do |ext|
   ext.lib_dir = 'lib/character_set'
 end
@@ -106,27 +113,22 @@ task :sync_casefold_data do
     hash[from] = to if type == 'C'
   end.sort
 
-  File.open(dst_path, 'w') do |f|
-    f.puts <<-C
-// THIS FILE IS GENERATED BY $ rake sync_casefold_data - DO NOT EDIT
-// -*-C-*-
-
-typedef struct casefold_mapping {
-  unsigned long from;
-  unsigned long to;
-} casefold_mapping;
-
-#define CASEFOLD_COUNT #{mapping.size}
+  content = File.read(dst_path + '.tmpl')
+    .sub(/(CASEFOLD_COUNT )0/, "\\1#{mapping.count}")
+    .sub('{}', ['{', mapping.map { |a, b| "{0x#{a},0x#{b}}," }, '}'].join("\n"))
 
-static const casefold_mapping unicode_casefold_table[CASEFOLD_COUNT] = {
-    C
-
-    mapping.each { |from, to| f.puts "{0x#{from},0x#{to}}," }
+  File.write(dst_path, content)
+  File.unlink(src_path)
+end
 
-    f.puts '};'
+desc 'Update codepoint data for predefined sets, based on Onigmo'
+task :sync_predefined_sets do
+  %w[assigned emoji whitespace].each do |prop|
+    require 'regexp_property_values'
+    ranges = RegexpPropertyValues[prop].matched_ranges
+    str = ranges.map { |r| "#{r.min.to_s(16)},#{r.max.to_s(16)}\n" }.join.upcase
+    File.write("./lib/character_set/predefined_sets/#{prop}.cps", str, mode: 'w')
   end
-
-  File.unlink(src_path)
 end
 
 desc 'Run all IPS benchmarks'

data/benchmarks/count_in.rb

@@ -0,0 +1,13 @@
+require_relative './shared'
+
+str = 'Lorem ipsum et dolorem'
+tr = '^A-Za-z'
+cs = CharacterSet.non_ascii_letter
+
+benchmark(
+  caption: 'Counting non-letters',
+  cases: {
+    'String#count'          => -> { str.count(tr) },
+    'CharacterSet#count_in' => -> { cs.count_in(str) },
+  }
+)

data/benchmarks/delete_in.rb

@@ -14,7 +14,7 @@ benchmark(
 
 str = 'Lörem ipsüm ⛷ et dölörem'
 rx = /[\s\p{emoji}äüö]/
-cs = CharacterSet.whitespace + CharacterSet.emoji + CS['ä', 'ü', 'ö']
+cs = CharacterSet.whitespace + CharacterSet.emoji + CharacterSet['ä', 'ö', 'ü']
 
 benchmark(
   caption: 'Removing whitespace, emoji and umlauts',

data/benchmarks/scan.rb

@@ -0,0 +1,13 @@
+require_relative './shared'
+
+str = 'Lorem ipsum ⛷ et dolorem'
+rx = /\p{emoji}/
+cs = CharacterSet.emoji
+
+benchmark(
+  caption: 'Extracting emoji to an Array',
+  cases: {
+    'String#scan'       => -> { str.scan(rx) },
+    'CharacterSet#scan' => -> { cs.scan(str) },
+  }
+)

data/benchmarks/shared.rb

@@ -3,6 +3,11 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 
 require 'benchmark/ips'
 require 'character_set'
+if RUBY_VERSION.to_f >= 3.0 && !RUBY_PLATFORM[/java/i]
+  require 'sorted_set'
+else
+  require 'set'
+end
 
 def benchmark(caption: nil, cases: {})
   puts caption

data/benchmarks/z_add.rb

@@ -0,0 +1,12 @@
+require_relative './shared'
+
+cs = CharacterSet[]
+ss = SortedSet[]
+
+benchmark(
+  caption: 'Adding entries',
+  cases: {
+    'CharacterSet#add' => -> { cs.add(rand(0x10FFFF)) },
+    'SortedSet#add'    => -> { ss.add(rand(0x10FFFF)) },
+  }
+)

data/benchmarks/z_delete.rb

@@ -0,0 +1,12 @@
+require_relative './shared'
+
+cs = CharacterSet.new(0..0x10FFFF)
+ss = SortedSet.new(0..0x10FFFF)
+
+benchmark(
+  caption: 'Removing entries',
+  cases: {
+    'CharacterSet#delete' => -> { cs.delete(rand(0x10FFFF)) },
+    'SortedSet#delete'    => -> { ss.delete(rand(0x10FFFF)) },
+  }
+)

data/benchmarks/z_merge.rb

@@ -0,0 +1,15 @@
+require_relative './shared'
+
+cs1 = CharacterSet.new(0...0x88000)
+cs2 = CharacterSet.new(0x88000..0x10FFFF)
+
+ss1 = SortedSet.new(0...0x88000)
+ss2 = SortedSet.new(0x88000..0x10FFFF)
+
+benchmark(
+  caption: 'Merging entries',
+  cases: {
+    'CharacterSet#merge' => -> { cs1.merge(cs2) },
+    'SortedSet#merge'    => -> { ss1.merge(ss2) },
+  }
+)

data/benchmarks/z_minmax.rb

@@ -0,0 +1,12 @@
+require_relative './shared'
+
+cs = CharacterSet.new(0..0xFFFF)
+ss = SortedSet.new(0..0xFFFF)
+
+benchmark(
+  caption: 'Getting the min and max',
+  cases: {
+    'CharacterSet#minmax' => -> { cs.minmax },
+    'SortedSet#minmax'    => -> { ss.minmax },
+  }
+)

data/bin/console

@@ -2,6 +2,8 @@
 
 require 'bundler/setup'
 
+`bundle exec rake compile`
+
 require 'character_set'
 require 'character_set/core_ext'
 require 'character_set/pure'

data/character_set.gemspec

@@ -22,11 +22,24 @@ Gem::Specification.new do |s|
 
   s.required_ruby_version = '>= 2.1.0'
 
+  # SortedSet, needed for RubyFallback, was moved to a gem in Ruby 3.
+  # This dependency is only used if the C extension is unavailable.
+  # JRuby has it in the stdlib.
+  if RUBY_VERSION.to_f >= 3.0 && !RUBY_PLATFORM[/java/i]
+    s.add_dependency 'sorted_set', '~> 1.0'
+  end
+
   s.add_development_dependency 'benchmark-ips', '~> 2.7'
-  s.add_development_dependency 'rake', '~> 12.0'
-  s.add_development_dependency 'rake-compiler', '~> 1.0'
+  s.add_development_dependency 'get_process_mem', '~> 0.2.3'
+  s.add_development_dependency 'rake', '~> 13.0'
+  s.add_development_dependency 'rake-compiler', '~> 1.1'
   s.add_development_dependency 'range_compressor', '~> 1.0'
-  s.add_development_dependency 'regexp_parser', '~> 1.3'
-  s.add_development_dependency 'regexp_property_values', '~> 0.3.5'
+  s.add_development_dependency 'regexp_parser', '~> 2.1'
+  s.add_development_dependency 'regexp_property_values', '~> 1.0'
   s.add_development_dependency 'rspec', '~> 3.8'
+  if RUBY_VERSION.to_f >= 2.7
+    s.add_development_dependency 'codecov', '~> 0.2.12'
+    s.add_development_dependency 'gouteur', '~> 1.0.0'
+    s.add_development_dependency 'rubocop', '~> 1.8'
+  end
 end