Start work on String#byteindex

include/natalie/env.hpp

@@ -63,6 +63,16 @@ class Env : public Cell {
     [[noreturn]] void raise_name_error(StringObject *name, String);
     [[noreturn]] void raise_not_comparable_error(Value lhs, Value rhs);
 
+    // Old error message style, e.g.:
+    // - no implicit conversion from nil to string
+    // - no implicit conversion of Integer into String
+    [[noreturn]] void raise_type_error(const Object *obj, const char *expected);
+
+    // New error message style, e.g.:
+    // - no implicit conversion of nil into String
+    // - no implicit conversion of Integer into String
+    [[noreturn]] void raise_type_error2(const Object *obj, const char *expected);
+
     template <typename... Args>
     [[noreturn]] void raise_name_error(SymbolObject *name, const char *format, Args... args) {
         auto message = String::format(format, args...);

include/natalie/object.hpp

@@ -376,8 +376,17 @@ class Object : public Cell {
     IoObject *to_io(Env *env);
     IntegerObject *to_int(Env *env);
     StringObject *to_s(Env *env);
+
+    // Old error message style, e.g.:
+    // - no implicit conversion from nil to string
+    // - no implicit conversion of Integer into String
     StringObject *to_str(Env *env);
 
+    // New error message style, e.g.:
+    // - no implicit conversion of nil into String
+    // - no implicit conversion of Integer into String
+    StringObject *to_str2(Env *env);
+
 protected:
     ClassObject *m_klass { nullptr };
 

include/natalie/string_object.hpp

@@ -276,6 +276,8 @@ class StringObject : public Object {
     StringObject *successive(Env *);
     StringObject *successive_in_place(Env *);
 
+    Value byteindex(Env *, Value, Value = nullptr) const;
+
     Value index(Env *, Value, Value);
     Value index(Env *, Value, size_t start);
     nat_int_t index_int(Env *, Value, size_t byte_start);
@@ -336,7 +338,7 @@ class StringObject : public Object {
     Value delete_suffix(Env *, Value);
     Value delete_suffix_in_place(Env *, Value);
     StringObject *downcase(Env *, Value, Value);
-    Value downcase_in_place(Env *, Value, Value);
+    Value downcase_in_place(Env *, Value = nullptr, Value = nullptr);
     Value dump(Env *);
     Value each_byte(Env *, Block *);
     Value encode(Env *, Value = nullptr, Value = nullptr, HashObject * = nullptr);

lib/natalie/compiler/binding_gen.rb

@@ -1241,6 +1241,7 @@ def generate_name
 gen.binding('String', 'ascii_only?', 'StringObject', 'is_ascii_only', argc: 0, pass_env: false, pass_block: false, return_type: :bool)
 gen.binding('String', 'b', 'StringObject', 'b', argc: 0, pass_env: true, pass_block: false, return_type: :Object)
 gen.binding('String', 'bytes', 'StringObject', 'bytes', argc: 0, pass_env: true, pass_block: true, return_type: :Object)
+gen.binding('String', 'byteindex', 'StringObject', 'byteindex', argc: 1..2, pass_env: true, pass_block: false, return_type: :Object)
 gen.binding('String', 'byteslice', 'StringObject', 'byteslice', argc: 1..2, pass_env: true, pass_block: false, return_type: :Object)
 gen.binding('String', 'bytesize', 'StringObject', 'bytesize', argc: 0, pass_env: false, pass_block: false, return_type: :size_t)
 gen.binding('String', 'capitalize', 'StringObject', 'capitalize', argc: 0..2, pass_env: true, pass_block: false, return_type: :Object)

spec/core/string/byteindex_spec.rb

@@ -0,0 +1,324 @@
+# -*- encoding: utf-8 -*-
+require_relative '../../spec_helper'
+require_relative 'fixtures/classes'
+require_relative 'shared/byte_index_common.rb'
+
+describe "String#byteindex" do
+  ruby_version_is "3.2" do
+    it "calls #to_str to convert the first argument" do
+      char = mock("string index char")
+      char.should_receive(:to_str).and_return("b")
+      "abc".byteindex(char).should == 1
+    end
+
+    it "calls #to_int to convert the second argument" do
+      offset = mock("string index offset")
+      offset.should_receive(:to_int).and_return(1)
+      "abc".byteindex("c", offset).should == 2
+    end
+
+    it "does not raise IndexError when byte offset is correct or on string boundary" do
+      "わ".byteindex("").should == 0
+      "わ".byteindex("", 0).should == 0
+      "わ".byteindex("", 3).should == 3
+    end
+
+    it_behaves_like :byte_index_common, :byteindex
+  end
+end
+
+describe "String#byteindex with String" do
+  ruby_version_is "3.2" do
+    it "behaves the same as String#byteindex(char) for one-character strings" do
+      "blablabla hello cruel world...!".split("").uniq.each do |str|
+        chr = str[0]
+        str.byteindex(str).should == str.byteindex(chr)
+
+        0.upto(str.size + 1) do |start|
+          str.byteindex(str, start).should == str.byteindex(chr, start)
+        end
+
+        (-str.size - 1).upto(-1) do |start|
+          str.byteindex(str, start).should == str.byteindex(chr, start)
+        end
+      end
+    end
+
+    it "returns the byteindex of the first occurrence of the given substring" do
+      "blablabla".byteindex("").should == 0
+      "blablabla".byteindex("b").should == 0
+      "blablabla".byteindex("bla").should == 0
+      "blablabla".byteindex("blabla").should == 0
+      "blablabla".byteindex("blablabla").should == 0
+
+      "blablabla".byteindex("l").should == 1
+      "blablabla".byteindex("la").should == 1
+      "blablabla".byteindex("labla").should == 1
+      "blablabla".byteindex("lablabla").should == 1
+
+      "blablabla".byteindex("a").should == 2
+      "blablabla".byteindex("abla").should == 2
+      "blablabla".byteindex("ablabla").should == 2
+    end
+
+    it "treats the offset as a byteindex" do
+      "aaaaa".byteindex("a", 0).should == 0
+      "aaaaa".byteindex("a", 2).should == 2
+      "aaaaa".byteindex("a", 4).should == 4
+    end
+
+    it "ignores string subclasses" do
+      "blablabla".byteindex(StringSpecs::MyString.new("bla")).should == 0
+      StringSpecs::MyString.new("blablabla").byteindex("bla").should == 0
+      StringSpecs::MyString.new("blablabla").byteindex(StringSpecs::MyString.new("bla")).should == 0
+    end
+
+    it "starts the search at the given offset" do
+      "blablabla".byteindex("bl", 0).should == 0
+      "blablabla".byteindex("bl", 1).should == 3
+      "blablabla".byteindex("bl", 2).should == 3
+      "blablabla".byteindex("bl", 3).should == 3
+
+      "blablabla".byteindex("bla", 0).should == 0
+      "blablabla".byteindex("bla", 1).should == 3
+      "blablabla".byteindex("bla", 2).should == 3
+      "blablabla".byteindex("bla", 3).should == 3
+
+      "blablabla".byteindex("blab", 0).should == 0
+      "blablabla".byteindex("blab", 1).should == 3
+      "blablabla".byteindex("blab", 2).should == 3
+      "blablabla".byteindex("blab", 3).should == 3
+
+      "blablabla".byteindex("la", 1).should == 1
+      "blablabla".byteindex("la", 2).should == 4
+      "blablabla".byteindex("la", 3).should == 4
+      "blablabla".byteindex("la", 4).should == 4
+
+      "blablabla".byteindex("lab", 1).should == 1
+      "blablabla".byteindex("lab", 2).should == 4
+      "blablabla".byteindex("lab", 3).should == 4
+      "blablabla".byteindex("lab", 4).should == 4
+
+      "blablabla".byteindex("ab", 2).should == 2
+      "blablabla".byteindex("ab", 3).should == 5
+      "blablabla".byteindex("ab", 4).should == 5
+      "blablabla".byteindex("ab", 5).should == 5
+
+      "blablabla".byteindex("", 0).should == 0
+      "blablabla".byteindex("", 1).should == 1
+      "blablabla".byteindex("", 2).should == 2
+      "blablabla".byteindex("", 7).should == 7
+      "blablabla".byteindex("", 8).should == 8
+      "blablabla".byteindex("", 9).should == 9
+    end
+
+    it "starts the search at offset + self.length if offset is negative" do
+      str = "blablabla"
+
+      ["bl", "bla", "blab", "la", "lab", "ab", ""].each do |needle|
+        (-str.length .. -1).each do |offset|
+          str.byteindex(needle, offset).should ==
+          str.byteindex(needle, offset + str.length)
+        end
+      end
+    end
+
+    it "returns nil if the substring isn't found" do
+      "blablabla".byteindex("B").should == nil
+      "blablabla".byteindex("z").should == nil
+      "blablabla".byteindex("BLA").should == nil
+      "blablabla".byteindex("blablablabla").should == nil
+      "blablabla".byteindex("", 10).should == nil
+
+      "hello".byteindex("he", 1).should == nil
+      "hello".byteindex("he", 2).should == nil
+      "I’ve got a multibyte character.\n".byteindex("\n\n").should == nil
+    end
+
+    it "returns the character byteindex of a multibyte character" do
+      "ありがとう".byteindex("が").should == 6
+    end
+
+    it "returns the character byteindex after offset" do
+      "われわれ".byteindex("わ", 3).should == 6
+      "ありがとうありがとう".byteindex("が", 9).should == 21
+    end
+
+    it "returns the character byteindex after a partial first match" do
+      "</</h".byteindex("</h").should == 2
+    end
+
+    it "raises an Encoding::CompatibilityError if the encodings are incompatible" do
+      char = "れ".encode Encoding::EUC_JP
+      -> do
+        "あれ".byteindex(char)
+      end.should raise_error(Encoding::CompatibilityError)
+    end
+
+    it "handles a substring in a superset encoding" do
+      'abc'.dup.force_encoding(Encoding::US_ASCII).byteindex('é').should == nil
+    end
+
+    it "handles a substring in a subset encoding" do
+      'été'.byteindex('t'.dup.force_encoding(Encoding::US_ASCII)).should == 2
+    end
+  end
+end
+
+describe "String#byteindex with Regexp" do
+  ruby_version_is "3.2" do
+    it "behaves the same as String#byteindex(string) for escaped string regexps" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        ["blablabla", "hello cruel world...!"].each do |str|
+          ["", "b", "bla", "lab", "o c", "d."].each do |needle|
+            regexp = Regexp.new(Regexp.escape(needle))
+            str.byteindex(regexp).should == str.byteindex(needle)
+
+            0.upto(str.size + 1) do |start|
+              str.byteindex(regexp, start).should == str.byteindex(needle, start)
+            end
+
+            (-str.size - 1).upto(-1) do |start|
+              str.byteindex(regexp, start).should == str.byteindex(needle, start)
+            end
+          end
+        end
+      end
+    end
+
+    it "returns the byteindex of the first match of regexp" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "blablabla".byteindex(/bla/).should == 0
+        "blablabla".byteindex(/BLA/i).should == 0
+
+        "blablabla".byteindex(/.{0}/).should == 0
+        "blablabla".byteindex(/.{6}/).should == 0
+        "blablabla".byteindex(/.{9}/).should == 0
+
+        "blablabla".byteindex(/.*/).should == 0
+        "blablabla".byteindex(/.+/).should == 0
+
+        "blablabla".byteindex(/lab|b/).should == 0
+
+        not_supported_on :opal do
+          "blablabla".byteindex(/\A/).should == 0
+          "blablabla".byteindex(/\Z/).should == 9
+          "blablabla".byteindex(/\z/).should == 9
+          "blablabla\n".byteindex(/\Z/).should == 9
+          "blablabla\n".byteindex(/\z/).should == 10
+        end
+
+        "blablabla".byteindex(/^/).should == 0
+        "\nblablabla".byteindex(/^/).should == 0
+        "b\nablabla".byteindex(/$/).should == 1
+        "bl\nablabla".byteindex(/$/).should == 2
+
+        "blablabla".byteindex(/.l./).should == 0
+      end
+    end
+
+    it "starts the search at the given offset" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "blablabla".byteindex(/.{0}/, 5).should == 5
+        "blablabla".byteindex(/.{1}/, 5).should == 5
+        "blablabla".byteindex(/.{2}/, 5).should == 5
+        "blablabla".byteindex(/.{3}/, 5).should == 5
+        "blablabla".byteindex(/.{4}/, 5).should == 5
+
+        "blablabla".byteindex(/.{0}/, 3).should == 3
+        "blablabla".byteindex(/.{1}/, 3).should == 3
+        "blablabla".byteindex(/.{2}/, 3).should == 3
+        "blablabla".byteindex(/.{5}/, 3).should == 3
+        "blablabla".byteindex(/.{6}/, 3).should == 3
+
+        "blablabla".byteindex(/.l./, 0).should == 0
+        "blablabla".byteindex(/.l./, 1).should == 3
+        "blablabla".byteindex(/.l./, 2).should == 3
+        "blablabla".byteindex(/.l./, 3).should == 3
+
+        "xblaxbla".byteindex(/x./, 0).should == 0
+        "xblaxbla".byteindex(/x./, 1).should == 4
+        "xblaxbla".byteindex(/x./, 2).should == 4
+
+        not_supported_on :opal do
+          "blablabla\n".byteindex(/\Z/, 9).should == 9
+        end
+      end
+    end
+
+    it "starts the search at offset + self.length if offset is negative" do
+      str = "blablabla"
+
+      ["bl", "bla", "blab", "la", "lab", "ab", ""].each do |needle|
+        (-str.length .. -1).each do |offset|
+          str.byteindex(needle, offset).should ==
+          str.byteindex(needle, offset + str.length)
+        end
+      end
+    end
+
+    it "returns nil if the substring isn't found" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "blablabla".byteindex(/BLA/).should == nil
+
+        "blablabla".byteindex(/.{10}/).should == nil
+        "blaxbla".byteindex(/.x/, 3).should == nil
+        "blaxbla".byteindex(/..x/, 2).should == nil
+      end
+    end
+
+    it "returns nil if the Regexp matches the empty string and the offset is out of range" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "ruby".byteindex(//, 12).should be_nil
+      end
+    end
+
+    it "supports \\G which matches at the given start offset" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "helloYOU.".byteindex(/\GYOU/, 5).should == 5
+        "helloYOU.".byteindex(/\GYOU/).should == nil
+
+        re = /\G.+YOU/
+        # The # marks where \G will match.
+        [
+          ["#hi!YOUall.", 0],
+          ["h#i!YOUall.", 1],
+          ["hi#!YOUall.", 2],
+          ["hi!#YOUall.", nil]
+        ].each do |spec|
+
+          start = spec[0].byteindex("#")
+          str = spec[0].delete("#")
+
+          str.byteindex(re, start).should == spec[1]
+        end
+      end
+    end
+
+    it "converts start_offset to an integer via to_int" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        obj = mock('1')
+        obj.should_receive(:to_int).and_return(1)
+        "RWOARW".byteindex(/R./, obj).should == 4
+      end
+    end
+
+    it "returns the character byteindex of a multibyte character" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "ありがとう".byteindex(/が/).should == 6
+      end
+    end
+
+    it "returns the character byteindex after offset" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "われわれ".byteindex(/わ/, 3).should == 6
+      end
+    end
+
+    it "treats the offset as a byteindex" do
+      NATFIXME 'Support Regexp', exception: TypeError do
+        "われわわれ".byteindex(/わ/, 6).should == 6
+      end
+    end
+  end
+end