character_set 1.8.0 → 1.9.0

data/lib/character_set/ruby_fallback/set_methods.rb

@@ -1,95 +1,258 @@
 class CharacterSet
   module RubyFallback
+    # Set-like methods, backed by a Hash of codepoints (@__hash) and a
+    # memoized, sorted Array of its keys (@__keys) that mutators reset.
     module SetMethods
-      (Enumerable.instance_methods -
-        %i[include? member? to_a] +
-        %i[empty? hash length size]).each do |mthd|
+      # Delegate Enumerable methods to the sorted keys Array. This is faster
+      # than the Enumerable implementations that go through #each, and it
+      # keeps CharacterSet::Pure free of C-implemented methods.
+      (Enumerable.instance_methods - %i[each include? member? to_a]).each do |mthd|
         class_eval <<-RUBY, __FILE__, __LINE__ + 1
-          def #{mthd}(*args, &block)
-            @__set.#{mthd}(*args, &block)
+          def #{mthd}(...)
+            sorted_keys.#{mthd}(...)
           end
         RUBY
       end
 
-      %i[< <= <=> > >= === disjoint? include? intersect? member?
-         proper_subset? proper_superset? subset? superset?].each do |mthd|
-        class_eval <<-RUBY, __FILE__, __LINE__ + 1
-          def #{mthd}(enum, &block)
-            if enum.is_a?(CharacterSet) || enum.is_a?(CharacterSet::Pure)
-              enum = enum.instance_variable_get(:@__set)
-            end
-            @__set.#{mthd}(enum, &block)
-          end
-        RUBY
+      def each(&block)
+        block or return enum_for(__method__) { size }
+        sorted_keys.each(&block)
+        self
       end
 
-      %i[<< add add? clear delete delete? delete_if each filter! keep_if
-         reject! select! subtract].each do |mthd|
-        class_eval <<-RUBY, __FILE__, __LINE__ + 1
-          def #{mthd}(*args, &block)
-            result = @__set.#{mthd}(*args, &block)
-            result.is_a?(Set) ? self : result
-          end
-        RUBY
+      def to_a(stringify = false)
+        stringify ? sorted_keys.map { |cp| cp.chr('utf-8') } : sorted_keys.dup
       end
 
-      %i[& + - ^ | difference intersection union].each do |mthd|
-        class_eval <<-RUBY, __FILE__, __LINE__ + 1
-          def #{mthd}(enum, &block)
-            if enum.respond_to?(:map)
-              enum = enum.map { |el| el.is_a?(String) ? el.ord : el }
-            end
-            self.class.new(@__set.#{mthd}(enum, &block).to_a)
-          end
-        RUBY
+      def size
+        @__hash.size
       end
+      alias length size
 
-      unless RUBY_PLATFORM[/java/i]
-        def freeze
-          @__set.to_a
-          @__set.freeze
-          super
-        end
+      def empty?
+        @__hash.empty?
+      end
+
+      def hash
+        @__hash.hash
+      end
+
+      def include?(cp)
+        @__hash.key?(cp)
+      end
+      alias member? include?
+      alias === include?
+
+      def add(cp)
+        @__hash[cp] = true
+        @__keys = nil
+        self
+      end
+      alias << add
+
+      def add?(cp)
+        assert_not_frozen
+        add(cp) unless include?(cp)
+      end
+
+      def delete(cp)
+        @__keys = nil if @__hash.delete(cp)
+        self
+      end
+
+      def delete?(cp)
+        assert_not_frozen
+        delete(cp) if include?(cp)
+      end
+
+      def clear
+        @__hash.clear
+        @__keys = nil
+        self
+      end
+
+      def keep_if
+        assert_not_frozen
+        block_given? or return enum_for(__method__) { size }
+        old_size = size
+        sorted_keys.each { |cp| @__hash.delete(cp) unless yield(cp) }
+        @__keys = nil if size != old_size
+        self
+      end
+
+      def delete_if
+        assert_not_frozen
+        block_given? or return enum_for(__method__) { size }
+        keep_if { |cp| !yield(cp) }
       end
 
       def merge(other)
+        assert_not_frozen
         raise ArgumentError, 'pass an Enumerable' unless other.respond_to?(:each)
-        # pass through #add to use the checks in SetMethodAdapters
-        other.each { |e| add(e) }
+
+        if (other_hash = codepoint_hash(other))
+          # Members of another CharacterSet are valid codepoints already,
+          # so they can go in with one Hash#update instead of n adds.
+          @__hash.update(other_hash)
+          @__keys = nil
+        else
+          # pass through #add to use the checks in SetMethodAdapters
+          other.each { |e| add(e) }
+        end
+        self
+      end
+
+      # The operands of the following methods are always CharacterSets,
+      # ensured by SetMethodAdapters - but they might be C-backed, so
+      # only their public interface is used.
+
+      def subtract(other)
+        other.each { |cp| @__hash.delete(cp) }
+        @__keys = nil
         self
       end
 
+      def |(other)
+        dup.merge(other)
+      end
+      alias + |
+      alias union |
+
+      def -(other)
+        dup.subtract(other)
+      end
+      alias difference -
+
+      def &(other)
+        result = self.class.new
+        # Fresh results have no memoized keys yet, so their Hash can be
+        # filled directly.
+        result_hash = codepoint_hash(result)
+        small, large = size < other.size ? [self, other] : [other, self]
+        large = lookup(large)
+        small.each { |cp| result_hash[cp] = true if large.include?(cp) }
+        result
+      end
+      alias intersection &
+
+      def ^(other)
+        (self | other) - (self & other)
+      end
+
+      def intersect?(other)
+        small, large = size < other.size ? [self, other] : [other, self]
+        large = lookup(large)
+        small.any? { |cp| large.include?(cp) }
+      end
+
+      def disjoint?(other)
+        !intersect?(other)
+      end
+
+      def <=>(other)
+        return unless character_set?(other)
+
+        case size <=> other.size
+        when -1 then -1 if proper_subset?(other)
+        when +1 then +1 if proper_superset?(other)
+        else 0 if self == other
+        end
+      end
+
+      def subset?(other)
+        assert_character_set(other)
+        return false if size > other.size
+
+        other = lookup(other)
+        all? { |cp| other.include?(cp) }
+      end
+      alias <= subset?
+
+      def proper_subset?(other)
+        assert_character_set(other)
+        return false if size >= other.size
+
+        other = lookup(other)
+        all? { |cp| other.include?(cp) }
+      end
+      alias < proper_subset?
+
+      def superset?(other)
+        assert_character_set(other)
+        size >= other.size && other.all? { |cp| @__hash.key?(cp) }
+      end
+      alias >= superset?
+
+      def proper_superset?(other)
+        assert_character_set(other)
+        size > other.size && other.all? { |cp| @__hash.key?(cp) }
+      end
+      alias > proper_superset?
+
       def ==(other)
         if equal?(other)
           true
-        elsif other.instance_of?(self.class)
-          @__set == other.instance_variable_get(:@__set)
-        elsif other.is_a?(CharacterSet) || other.is_a?(CharacterSet::Pure)
-          size == other.size && other.all? { |cp| @__set.include?(cp) }
+        elsif (other_hash = codepoint_hash(other))
+          @__hash == other_hash
+        elsif character_set?(other)
+          size == other.size && other.all? { |cp| @__hash.key?(cp) }
         else
           false
         end
       end
 
       def eql?(other)
-        return false unless other.is_a?(self.class)
+        other.is_a?(self.class) && @__hash.eql?(codepoint_hash(other))
+      end
 
-        @__set.eql?(other.instance_variable_get(:@__set))
+      def freeze
+        sorted_keys.freeze
+        @__hash.freeze
+        super
       end
 
       def initialize_dup(orig)
         super
-        @__set = orig.instance_variable_get(:@__set).dup
+        @__hash = @__hash.dup
+        @__keys = nil # dups are usually mutated, so don't bother copying
       end
 
-      def initialize_clone(orig)
+      def initialize_clone(orig, freeze: nil)
         super
-        @__set = orig.instance_variable_get(:@__set).clone
+        @__hash = @__hash.clone(freeze: freeze)
+        @__keys = @__keys&.clone(freeze: freeze)
       end
 
-      def to_a(stringify = false)
-        result = @__set.to_a
-        stringify ? result.map { |cp| cp.chr('utf-8') } : result
+      private
+
+      def sorted_keys
+        @__keys ||= @__hash.keys.sort!
+      end
+
+      def character_set?(obj)
+        obj.is_a?(CharacterSet) ||
+          defined?(CharacterSet::Pure) && obj.is_a?(CharacterSet::Pure)
+      end
+
+      def assert_character_set(obj)
+        character_set?(obj) or raise ArgumentError, 'value must be a set'
+      end
+
+      def assert_not_frozen
+        raise FrozenError, "can't modify frozen #{self.class}" if frozen?
+      end
+
+      # The codepoint Hash of another Ruby-backed CharacterSet, or nil if
+      # `other` is not a CharacterSet - or is C-backed and so has no Hash.
+      def codepoint_hash(other)
+        other.instance_variable_get(:@__hash) if character_set?(other)
+      end
+
+      # The cheapest object that answers #include? for the codepoints of
+      # `set`: its Hash if it is Ruby-backed, which skips the argument checks
+      # of SetMethodAdapters, otherwise the set itself.
+      def lookup(set)
+        codepoint_hash(set) || set
       end
     end
   end

data/lib/character_set/ruby_fallback.rb

@@ -2,6 +2,8 @@ require 'character_set/ruby_fallback/set_methods'
 require 'character_set/ruby_fallback/character_set_methods'
 
 class CharacterSet
+  # The RubyFallback is used by non-standard Rubies that can't use
+  # the C extension. It is tested in the specs via CharacterSet::Pure.
   module RubyFallback
     include CharacterSet::RubyFallback::SetMethods
     include CharacterSet::RubyFallback::CharacterSetMethods
@@ -11,25 +13,9 @@ class CharacterSet
     end
 
     def initialize(enum = [])
-      @__set = CharacterSet::RubyFallback::SortedSet.new
+      @__hash = {}
+      @__keys = nil
       super
     end
   end
 end
-
-if RUBY_PLATFORM[/java/i]
-  # JRuby has sorted_set in the stdlib.
-  require 'set'
-  CharacterSet::RubyFallback::Set       = ::Set
-  CharacterSet::RubyFallback::SortedSet = ::SortedSet
-else
-  # For other rubies, set/sorted_set are vendored due to dependency issues:
-  #
-  # - issues with default vs. installed gems such as [#2]
-  # - issues with the sorted_set dependency rb_tree
-  # - long-standing issues in recent versions of sorted_set
-  #
-  # The RubyFallback, and thus these set classes, are only used for testing,
-  # and for exotic rubies which use neither C nor Java.
-  require 'character_set/ruby_fallback/vendored_set_classes'
-end

data/lib/character_set/shared_methods.rb

@@ -82,6 +82,14 @@ class CharacterSet
           self - self.class.surrogate
         end
 
+        def marshal_dump
+          to_a
+        end
+
+        def marshal_load(data)
+          initialize(data)
+        end
+
         # CharacterSet-specific stringification methods
 
         def to_s(opts = {}, &block)
@@ -164,9 +172,38 @@ class CharacterSet
           each_with_object({}) { |cp, h| (h[yield(cp)] ||= self.class.new).add(cp) }
         end
 
+        # Note: Set#divide behavior with a 2-arity block changed in Ruby 4.
+        # This implementation has the new, more sensible behavior.
+        # https://github.com/ruby/ruby/blob/453a69e/spec/ruby/core/set/divide_spec.rb#L34-L48
         def divide(&func)
-          require 'character_set/ruby_fallback'
-          CharacterSet::RubyFallback::Set.new(to_a).divide(&func)
+          func or return enum_for(__method__) { size }
+
+          klass = defined?(::Set) ? ::Set : ::Array
+
+          if func.arity == 2
+            class << dig = {}
+              require 'tsort'
+              include TSort
+
+              alias tsort_each_node each_key
+              def tsort_each_child(node, &block)
+                fetch(node).each(&block)
+              end
+            end
+
+            each do |cp|
+              dig[cp] = arr = []
+              each { |cp2| arr << cp2 if cp != cp2 && func.call(cp, cp2) }
+            end
+
+            set = klass.new
+            dig.each_strongly_connected_component do |css|
+              set << self.class.new(css)
+            end
+            set
+          else
+            klass.new(classify(&func).values)
+          end
         end
 
         def join(separator = '')

data/lib/character_set/version.rb

@@ -1,3 +1,3 @@
 class CharacterSet
-  VERSION = '1.8.0'
+  VERSION = '1.9.0'
 end

data/lib/character_set/writer.rb

@@ -4,7 +4,7 @@ class CharacterSet
       def write(codepoint_ranges, opts = {}, &block)
         content = codepoint_ranges.map do |range|
           if range.size > 2 && opts[:abbreviate] != false
-            bounds = [range.min, range.max]
+            bounds = [range.min, range.max] # minmax is slow on some rubies
             bounds.map { |cp| write_codepoint(cp, opts, &block) }.join('-')
           else
             range.map { |cp| write_codepoint(cp, opts, &block) }.join

data/lib/character_set.rb

@@ -7,9 +7,14 @@ require 'character_set/shared_methods'
 require 'character_set/version'
 require 'character_set/writer'
 
+# A set of Unicode codepoints with engine-specific storage backends.
 class CharacterSet
   begin
-    require 'character_set/character_set'
+    if RUBY_ENGINE == 'jruby'
+      require 'character_set/java'
+    else
+      require 'character_set/character_set' # C extension
+    end
   rescue LoadError
     require 'character_set/ruby_fallback'
     prepend RubyFallback

metadata

@@ -1,16 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: character_set
 version: !ruby/object:Gem::Version
-  version: 1.8.0
+  version: 1.9.0
 platform: ruby
 authors:
 - Janosch Müller
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-01-07 00:00:00.000000000 Z
+date: 1980-01-02 00:00:00.000000000 Z
 dependencies: []
-description:
 email:
 - janosch84@gmail.com
 executables: []
@@ -18,33 +16,23 @@ extensions:
 - ext/character_set/extconf.rb
 extra_rdoc_files: []
 files:
-- ".gitattributes"
-- ".github/workflows/gouteur.yml"
-- ".github/workflows/lint.yml"
-- ".github/workflows/tests.yml"
-- ".gitignore"
-- ".gouteur.yml"
-- ".rspec"
 - ".rubocop.yml"
-- BENCHMARK.md
 - CHANGELOG.md
-- Gemfile
 - LICENSE.txt
 - README.md
 - Rakefile
-- bin/console
-- bin/setup
-- character_set.gemspec
 - ext/character_set/character_set.c
 - ext/character_set/extconf.rb
-- ext/character_set/unicode_casefold_table.h
-- ext/character_set/unicode_casefold_table.h.tmpl
+- ext/character_set/java/src/main/java/org/jaynetics/characterset/CharacterSetService.java
+- ext/character_set/java/src/main/java/org/jaynetics/characterset/RubyCharacterSet.java
+- ext/character_set/java/src/main/java/org/jaynetics/characterset/StringOperations.java
 - lib/character_set.rb
 - lib/character_set/character.rb
 - lib/character_set/core_ext.rb
 - lib/character_set/core_ext/regexp_ext.rb
 - lib/character_set/core_ext/string_ext.rb
 - lib/character_set/expression_converter.rb
+- lib/character_set/java.rb
 - lib/character_set/parser.rb
 - lib/character_set/predefined_sets.rb
 - lib/character_set/predefined_sets/any.cps
@@ -67,31 +55,14 @@ files:
 - lib/character_set/ruby_fallback.rb
 - lib/character_set/ruby_fallback/character_set_methods.rb
 - lib/character_set/ruby_fallback/set_methods.rb
-- lib/character_set/ruby_fallback/vendored_set_classes.rb
 - lib/character_set/set_method_adapters.rb
 - lib/character_set/shared_methods.rb
 - lib/character_set/version.rb
 - lib/character_set/writer.rb
-- tasks/benchmark.rake
-- tasks/benchmarks/count_in.rb
-- tasks/benchmarks/cover.rb
-- tasks/benchmarks/delete_in.rb
-- tasks/benchmarks/keep_in.rb
-- tasks/benchmarks/scan.rb
-- tasks/benchmarks/shared.rb
-- tasks/benchmarks/used_by.rb
-- tasks/benchmarks/z_add.rb
-- tasks/benchmarks/z_delete.rb
-- tasks/benchmarks/z_merge.rb
-- tasks/benchmarks/z_minmax.rb
-- tasks/sync_casefold_data.rake
-- tasks/sync_predefined_sets.rake
-- tasks/sync_ruby_spec.rake
 homepage: https://github.com/jaynetics/character_set
 licenses:
 - MIT
 metadata: {}
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -99,15 +70,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 2.1.0
+      version: 3.0.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.0.dev
-signing_key:
+rubygems_version: 4.0.3
 specification_version: 4
 summary: Build, read, write and compare sets of Unicode codepoints.
 test_files: []

data/.gitattributes

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

data/.github/workflows/gouteur.yml

@@ -1,20 +0,0 @@
-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: 3.3
-      - name: Prepare
-        run: |
-          bundle install --jobs 4
-          bundle exec rake compile
-      - name: Test
-        run: bundle exec gouteur

data/.github/workflows/lint.yml

@@ -1,29 +0,0 @@
-# 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: 3.3
-    - 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

@@ -1,28 +0,0 @@
-name: tests
-
-on:
-  push:
-  pull_request:
-  schedule:
-    - cron: '11 11 14 * *' # at 11:11 am on the 14th of every month
-
-jobs:
-  build:
-    runs-on: ubuntu-latest
-
-    strategy:
-      matrix:
-        ruby: [ '2.4', '2.7', '3.0', '3.1', '3.2', '3.3', '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
-      - uses: codecov/codecov-action@v3
-        if: matrix.ruby == '3.2'

data/.gitignore

@@ -1,33 +0,0 @@
-*.bundle
-*.gem
-*.iml
-*.stTheme.cache
-*.sublime-project
-*.sublime-workspace
-*.swp
-*.tmlanguage.cache
-*.tmPreferences.cache
-*~
-.byebug_history
-.DS_Store
-.idea/
-.ruby-gemset
-.ruby-version
-.tags
-.tags1
-.tool-versions
-.vscode
-bbin/
-binstubs/*
-bundler_stubs/*/.yardoc
-Gemfile.lock
-/.bundle/
-/_yardoc/
-/coverage/
-/doc/
-/pkg/
-/spec/reports/
-/tmp/
-
-# rspec failure tracking
-.rspec_status

data/.gouteur.yml

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

data/.rspec

@@ -1,3 +0,0 @@
---format documentation
---color
---require spec_helper