Merge pull request #1829 from ksss/lex

Implement token list API

ext/rbs_extension/parser.c

@@ -2885,9 +2885,29 @@ rbsparser_parse_signature(VALUE self, VALUE buffer, VALUE end_pos)
   return rb_ensure(parse_signature_try, (VALUE)parser, ensure_free_parser, (VALUE)parser);
 }
 
+static VALUE
+rbsparser_lex(VALUE self, VALUE buffer, VALUE end_pos) {
+  lexstate *lexer = alloc_lexer(buffer, 0, FIX2INT(end_pos));
+  VALUE results = rb_ary_new();
+
+  token token = NullToken;
+  while (token.type != pEOF) {
+    token = rbsparser_next_token(lexer);
+    VALUE type = ID2SYM(rb_intern(token_type_str(token.type)));
+    VALUE location = rbs_new_location(buffer, token.range);
+    VALUE pair = rb_ary_new3(2, type, location);
+    rb_ary_push(results, pair);
+  }
+
+  free(lexer);
+
+  return results;
+}
+
 void rbs__init_parser(void) {
   RBS_Parser = rb_define_class_under(RBS, "Parser", rb_cObject);
   rb_define_singleton_method(RBS_Parser, "_parse_type", rbsparser_parse_type, 5);
   rb_define_singleton_method(RBS_Parser, "_parse_method_type", rbsparser_parse_method_type, 5);
   rb_define_singleton_method(RBS_Parser, "_parse_signature", rbsparser_parse_signature, 2);
+  rb_define_singleton_method(RBS_Parser, "_lex", rbsparser_lex, 2);
 }

ext/rbs_extension/parserstate.c

@@ -272,7 +272,7 @@ VALUE comment_to_ruby(comment *com, VALUE buffer) {
   );
 }
 
-parserstate *alloc_parser(VALUE buffer, int start_pos, int end_pos, VALUE variables) {
+lexstate *alloc_lexer(VALUE buffer, int start_pos, int end_pos) {
   VALUE string = rb_funcall(buffer, rb_intern("content"), 0);
 
   StringValue(string);
@@ -290,6 +290,11 @@ parserstate *alloc_parser(VALUE buffer, int start_pos, int end_pos, VALUE variab
   lexer->start = lexer->current;
   lexer->first_token_of_line = lexer->current.column == 0;
 
+  return lexer;
+}
+
+parserstate *alloc_parser(VALUE buffer, int start_pos, int end_pos, VALUE variables) {
+  lexstate *lexer = alloc_lexer(buffer, start_pos, end_pos);
   parserstate *parser = calloc(1, sizeof(parserstate));
   parser->lexstate = lexer;
   parser->buffer = buffer;

ext/rbs_extension/parserstate.h

@@ -93,6 +93,15 @@ void parser_insert_typevar(parserstate *state, ID id);
  * */
 bool parser_typevar_member(parserstate *state, ID id);
 
+/**
+ * Allocate new lexstate object.
+ *
+ * ```
+ * alloc_lexer(buffer, 0, 31)    // New lexstate with buffer
+ * ```
+ * */
+lexstate *alloc_lexer(VALUE buffer, int start_pos, int end_pos);
+
 /**
  * Allocate new parserstate object.
  *

lib/rbs/parser/lex_result.rb

@@ -0,0 +1,15 @@
+# frozen_string_literal: true
+
+module RBS
+  class Parser
+    class LexResult
+      attr_reader :buffer
+      attr_reader :value
+
+      def initialize(buffer:, value:)
+        @buffer = buffer
+        @value = value
+      end
+    end
+  end
+end

lib/rbs/parser/token.rb

@@ -0,0 +1,23 @@
+# frozen_string_literal: true
+
+module RBS
+  class Parser
+    class Token
+      attr_reader :type
+      attr_reader :location
+
+      def initialize(type:, location:)
+        @type = type
+        @location = location
+      end
+
+      def value
+        @location.source
+      end
+
+      def comment?
+        @type == :tCOMMENT || @type == :tLINECOMMENT
+      end
+    end
+  end
+end

lib/rbs/parser_aux.rb

@@ -1,5 +1,8 @@
 # frozen_string_literal: true
 
+require_relative "parser/lex_result"
+require_relative "parser/token"
+
 module RBS
   class Parser
     def self.parse_type(source, range: 0..., variables: [], require_eof: false)
@@ -19,6 +22,15 @@ def self.parse_signature(source)
       [buf, dirs, decls]
     end
 
+    def self.lex(source)
+      buf = buffer(source)
+      list = _lex(buf, buf.last_position)
+      value = list.map do |type, location|
+        Token.new(type: type, location: location)
+      end
+      LexResult.new(buffer: buf, value: value)
+    end
+
     def self.buffer(source)
       case source
       when String

sig/parser.rbs

@@ -1,5 +1,23 @@
 module RBS
   class Parser
+    # Result of `Parser.lex`
+    class LexResult
+      attr_reader buffer: Buffer
+      attr_reader value: Array[Token]
+
+      def initialize: (buffer: Buffer, value: Array[Token]) -> void
+    end
+
+    # Represents a token per result of `Parser.lex`.
+    class Token
+      attr_reader type: Symbol
+      attr_reader location: Location[untyped, untyped]
+
+      def initialize: (type: Symbol, location: Location[untyped, untyped]) -> void
+      def value: () -> String
+      def comment?: () -> bool
+    end
+
     # Parse a method type and return it
     #
     # When `range` keyword is specified, it starts parsing from the `begin` to the `end` of the range.
@@ -50,6 +68,14 @@ module RBS
     #
     def self.parse_signature: (Buffer | String) -> [Buffer, Array[AST::Directives::t], Array[AST::Declarations::t]]
 
+    # Parse whole RBS file and return result.
+    #
+    # ```ruby
+    # RBS::Parser.lex("# Comment\nmodule A\nend\n").value.map(&:type)
+    # # => [:tLINECOMMENT, :kMODULE, :tUIDENT, :kEND, :pEOF]
+    # ```
+    def self.lex: (Buffer | String) -> LexResult
+
     KEYWORDS: Hash[String, bot]
 
     private
@@ -62,6 +88,8 @@ module RBS
 
     def self._parse_signature: (Buffer, Integer end_pos) -> [Array[AST::Directives::t], Array[AST::Declarations::t]]
 
+    def self._lex: (Buffer, Integer end_pos) -> Array[[Symbol, Location[untyped, untyped]]]
+
     class LocatedValue
     end
   end

test/rbs/parser_test.rb

@@ -767,4 +767,26 @@ def test_proc__untyped_function_parse_error
       RBS::Parser.parse_type("^(?) { (?) -> void } -> Integer")
     end
   end
+
+  def test__lex
+    content = <<~RBS
+      # LineComment
+      class Foo[T < Integer] < Bar # Comment
+      end
+    RBS
+    tokens = RBS::Parser._lex(buffer(content), content.length)
+    assert_equal [:tLINECOMMENT, '# LineComment', 0...13], tokens[0].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:kCLASS, 'class', 14...19], tokens[1].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:tUIDENT, 'Foo', 20...23], tokens[2].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:pLBRACKET, '[', 23...24], tokens[3].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:tUIDENT, 'T', 24...25], tokens[4].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:pLT, '<', 26...27], tokens[5].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:tUIDENT, 'Integer', 28...35], tokens[6].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:pRBRACKET, ']', 35...36], tokens[7].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:pLT, '<', 37...38], tokens[8].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:tUIDENT, 'Bar', 39...42], tokens[9].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:tCOMMENT, '# Comment', 43...52], tokens[10].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:kEND, 'end', 53...56], tokens[11].then { |t| [t[0], t[1].source, t[1].range] }
+    assert_equal [:pEOF, '', 57...58], tokens[12].then { |t| [t[0], t[1].source, t[1].range] }
+  end
 end