Improve MatchData#dup

CHANGELOG.md

@@ -19,6 +19,7 @@ Bug fixes:
 Compatibility:
 
 * Implemented `String#start_with?(Regexp)` (#1771, @zhublik).
+* Implemented `MatchData#dup` (#1792, @XrXr).
 
 Performance:
 

spec/ruby/core/matchdata/begin_spec.rb

@@ -101,4 +101,11 @@
       match_data.begin(:æ).should == 1
     end
   end
+
+  context "when uninitialized" do
+    it "raises TypeError" do
+      match_data = MatchData.allocate
+      -> { match_data.begin(0) }.should raise_error(TypeError)
+    end
+  end
 end

spec/ruby/core/matchdata/captures_spec.rb

@@ -1,6 +1,11 @@
 require_relative '../../spec_helper'
 
 describe "MatchData#captures" do
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.captures }.should raise_error(TypeError)
+  end
+
   it "returns an array of the match captures" do
     /(.)(.)(\d+)(\d)/.match("THX1138.").captures.should == ["H","X","113","8"]
   end

spec/ruby/core/matchdata/dup_spec.rb

@@ -0,0 +1,14 @@
+require_relative '../../spec_helper'
+
+describe "MatchData#dup" do
+  it "duplicates the match data" do
+    original = /ll/.match("hello")
+    original.instance_variable_set(:@custom_ivar, 42)
+    duplicate = original.dup
+
+    duplicate.instance_variable_get(:@custom_ivar).should == 42
+    original.regexp.should == duplicate.regexp
+    original.string.should == duplicate.string
+    original.offset(0).should == duplicate.offset(0)
+  end
+end

spec/ruby/core/matchdata/element_reference_spec.rb

@@ -1,6 +1,11 @@
 require_relative '../../spec_helper'
 
 describe "MatchData#[]" do
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data[1] }.should raise_error(TypeError)
+  end
+
   it "acts as normal array indexing [index]" do
     md = /(.)(.)(\d+)(\d)/.match("THX1138.")
 

spec/ruby/core/matchdata/end_spec.rb

@@ -101,4 +101,11 @@
       match_data.end(:æ).should == 2
     end
   end
+
+  context "when uninitialized" do
+    it "raises TypeError" do
+      match_data = MatchData.allocate
+      -> { match_data.end(0) }.should raise_error(TypeError)
+    end
+  end
 end

spec/ruby/core/matchdata/inspect_spec.rb

@@ -5,6 +5,11 @@
     @match_data = /(.)(.)(\d+)(\d)/.match("THX1138.")
   end
 
+  it "returns a String when receiver is uninitialized" do
+    match_data = MatchData.allocate
+    match_data.inspect.should be_kind_of(String)
+  end
+
   it "returns a String" do
     @match_data.inspect.should be_kind_of(String)
   end

spec/ruby/core/matchdata/named_captures_spec.rb

@@ -12,4 +12,9 @@
   it 'returns the latest matched capture, even if a later one that does not match exists' do
     /\A(?<a>.)(?<b>.)(?<b>.)(?<a>.)?\z/.match('012').named_captures.should == { 'a' => '0', 'b' => '2' }
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.named_captures }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/names_spec.rb

@@ -30,4 +30,9 @@
     r = /(?<hay>hay)(?<dot>.)(?<hay>tack)/
     'haystack'.match(r).names.should == r.names
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.names }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/offset_spec.rb

@@ -27,4 +27,9 @@
       match_data.offset(4).should == [6, 7]
     end
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.offset(0) }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/post_match_spec.rb

@@ -31,4 +31,9 @@
     str = "abc".force_encoding Encoding::ISO_8859_1
     str.match(/c/).post_match.encoding.should equal(Encoding::ISO_8859_1)
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.post_match }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/pre_match_spec.rb

@@ -31,4 +31,9 @@
     str = "abc".force_encoding Encoding::ISO_8859_1
     str.match(/a/).pre_match.encoding.should equal(Encoding::ISO_8859_1)
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.pre_match }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/regexp_spec.rb

@@ -15,4 +15,9 @@
     'he[[o'.gsub('[', ']')
     $~.regexp.should == /\[/
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.regexp }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/shared/eql.rb

@@ -23,4 +23,15 @@
     a = 'haystack'.match(/hay/)
     a.send(@method, Object.new).should be_false
   end
+
+  it "returns false if arguments are different and both are uninitialized" do
+    a = MatchData.allocate
+    b = MatchData.allocate
+    a.send(@method, b).should be_false
+  end
+
+  it "returns true if arguments are the identical and uninitialized" do
+    match_data = MatchData.allocate
+    match_data.send(@method, match_data).should be_true
+  end
 end

spec/ruby/core/matchdata/shared/length.rb

@@ -2,4 +2,9 @@
   it "length should return the number of elements in the match array" do
     /(.)(.)(\d+)(\d)/.match("THX1138.").send(@method).should == 5
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.send(@method) }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/string_spec.rb

@@ -22,4 +22,9 @@
     $~.string.should == 'he[[o'
     $~.string.frozen?.should == true
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.string }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/to_a_spec.rb

@@ -4,4 +4,9 @@
   it "returns an array of matches" do
     /(.)(.)(\d+)(\d)/.match("THX1138.").to_a.should == ["HX1138", "H", "X", "113", "8"]
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.to_a }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/to_s_spec.rb

@@ -4,4 +4,9 @@
   it "returns the entire matched string" do
     /(.)(.)(\d+)(\d)/.match("THX1138.").to_s.should == "HX1138"
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.to_s }.should raise_error(TypeError)
+  end
 end

spec/ruby/core/matchdata/values_at_spec.rb

@@ -18,4 +18,9 @@
   it 'takes names and indices' do
     /\A(?<a>.)(?<b>.)\z/.match('01').values_at(0, 1, 2, :a, :b).should == ['01', '0', '1', '0', '1']
   end
+
+  it "raises TypeError when uninitialized" do
+    match_data = MatchData.allocate
+    -> { match_data.values_at(0) }.should raise_error(TypeError)
+  end
 end

src/main/java/org/truffleruby/core/regexp/MatchDataGuards.java

@@ -0,0 +1,20 @@
+/*
+ * Copyright (c) 2015, 2019 Oracle and/or its affiliates. All rights reserved. This
+ * code is released under a tri EPL/GPL/LGPL license. You can use it,
+ * redistribute it and/or modify it under the terms of the:
+ *
+ * Eclipse Public License version 2.0, or
+ * GNU General Public License version 2, or
+ * GNU Lesser General Public License version 2.1.
+ */
+
+package org.truffleruby.core.regexp;
+
+import com.oracle.truffle.api.object.DynamicObject;
+import org.truffleruby.Layouts;
+
+public class MatchDataGuards {
+    public static boolean isInitialized(DynamicObject matchData) {
+        return Layouts.MATCH_DATA.getSource(matchData) != null;
+    }
+}

src/main/java/org/truffleruby/core/regexp/MatchDataLayout.java

@@ -23,21 +23,26 @@ public interface MatchDataLayout extends BasicObjectLayout {
     DynamicObjectFactory createMatchDataShape(DynamicObject logicalClass,
             DynamicObject metaClass);
 
-    Object[] build(DynamicObject source, DynamicObject regexp, Region region, @Nullable Region charOffsets);
+    Object[] build(@Nullable DynamicObject source, @Nullable DynamicObject regexp, @Nullable Region region,
+            @Nullable Region charOffsets);
 
     boolean isMatchData(DynamicObject object);
 
     boolean isMatchData(Object object);
 
     DynamicObject getSource(DynamicObject object);
 
+    void setSource(DynamicObject object, DynamicObject value);
+
     /** Either a Regexp or a String for the case of String#gsub(String) */
     DynamicObject getRegexp(DynamicObject object);
 
     void setRegexp(DynamicObject object, DynamicObject value);
 
     Region getRegion(DynamicObject object);
 
+    void setRegion(DynamicObject object, Region value);
+
     Region getCharOffsets(DynamicObject object);
 
     void setCharOffsets(DynamicObject object, Region value);