parselly 1.2.0 → 1.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e1e4e245059433130b385388d399fe29355a1c679a15327a02cb3b49e69ae23b
-  data.tar.gz: ceb1167e8a32c25543b96988f754f76eb01fe287fc279a6ceb8dd26ffc258ca9
+  metadata.gz: 5bc554d1e8b9c0bba096e513d21aa054f35589a7e52f1c0e2108c0639ec1027b
+  data.tar.gz: 55f3bec1107b38b70bdd4995375fe8cab816036fe86d6cf27e96566689cce03f
 SHA512:
-  metadata.gz: 33dab2f628019bbd51d482c53ae267f98914a5a561beb146ce57b3625f30535b661de168b2c6150faa79f16af8a49a8c15e7d6047a118577359fdd9334249a7f
-  data.tar.gz: 38803e8cc427a8eaa0b2a63f9446cf91334a368a7ea86908da80eb62cca3569a7b44f4c9536165a57f99048cf604dd948814b12dff25ac49c43ab637ec344324
+  metadata.gz: 5dc070854c73e51ac4e21dde9910cd687c788f2bdfe99c9c22870d2e73222dcfc9679cf26e54e9c81032284a201b840e0af395958347f3e0bc10967294f990d0
+  data.tar.gz: 37a0b28ade0b3d74a062fee148d18d7448f200ac04bbb107bdf036a92808b7dc5b803a179dbb4363100fb290351d14e504651a07ac74b9dd68829e5c36af3a28

data/README.md

@@ -1,30 +1,77 @@
 # Parselly [![Gem Version](https://badge.fury.io/rb/parselly.svg)](https://badge.fury.io/rb/parselly) [![CI](https://github.com/ydah/parselly/actions/workflows/test.yml/badge.svg)](https://github.com/ydah/parselly/actions/workflows/test.yml)
 
-Parselly is a module providing a simple way to parse and extract data from a css selector.
+Pure Ruby CSS selector parser.
 
 ## Installation
 
-Add this line to your application's Gemfile:
 ```ruby
 gem 'parselly'
 ```
 
-And then execute:
 ```bash
 bundle install
 ```
 
-Or install it yourself as:
+Or install it directly:
+
 ```bash
 gem install parselly
 ```
 
-## Development
+Requires Ruby 2.7 or newer.
+
+## Usage
+
+```ruby
+require 'parselly'
+
+ast = Parselly.parse('article#main.content[data-state="open"] > a:hover')
+
+ast.ids
+#=> ["main"]
+
+ast.attributes
+#=> [{ name: "data-state", operator: "=", value: "open" }]
+
+ast.pseudo_class_names
+#=> ["hover"]
+
+ast.specificity
+#=> [1, 3, 2]
+```
+
+Strict parsing raises `Parselly::LexError` or `Parselly::SyntaxError` for invalid selectors:
+
+```ruby
+Parselly.parse('div >')
+```
+
+Use tolerant mode when you want a `Parselly::ParseResult` instead:
+
+```ruby
+result = Parselly.parse('div >', tolerant: true)
+
+result.success?
+#=> false
+
+result.errors.first[:message]
+#=> "Parse error: unexpected $end '' at 1:6"
+```
 
-After checking out the repo, run `bin/setup` to install dependencies. You can also run `bin/console` for an interactive prompt that will allow you to experiment.
+Use `Parselly.sanitize` to escape text for a CSS identifier:
 
-To install this gem onto your local machine, run `bundle exec rake install`. To release a new version, update the version number in `version.rb`, and then run `bundle exec rake release`, which will create a git tag for the version, push git commits and the created tag, and push the `.gem` file to [rubygems.org](https://rubygems.org).
+```ruby
+Parselly.sanitize('1st item')
+#=> "\\31 st\\ item"
+```
+
+## Development
+
+```bash
+bin/setup
+bundle exec rake
+```
 
-## Contributing
+## License
 
-Bug reports and pull requests are welcome on GitHub at https://github.com/ydah/parselly.
+MIT

data/Rakefile

@@ -7,6 +7,16 @@ namespace 'build' do
   task :parser do
     sh 'bundle exec racc parser.y --embedded --frozen -o lib/parselly/parser.rb -t --log-file=parser.output'
   end
+
+  desc 'verify generated parser files are in sync'
+  task check_parser: :parser do
+    sh 'git diff --exit-code lib/parselly/parser.rb parser.output'
+  end
+end
+
+desc 'run parser benchmarks'
+task :benchmark do
+  ruby 'benchmark/parser_benchmark.rb'
 end
 
 require 'rspec/core/rake_task'

data/lib/parselly/lexer.rb

@@ -5,12 +5,76 @@ require 'strscan'
 module Parselly
   class Lexer
     Identifier = Struct.new(:value, :raw) do
+      attr_accessor :position
+
       def to_s
         value
       end
+
+      def ==(other)
+        other.respond_to?(:value) ? value == other.value : value == other
+      end
+    end
+
+    TokenValue = Struct.new(:value, :raw, :position, :quote, keyword_init: true) do
+      def to_s
+        value.to_s
+      end
+
+      def ==(other)
+        other.respond_to?(:value) ? value == other.value : value == other
+      end
+    end
+
+    Token = Struct.new(:type, :value, :position, keyword_init: true) do
+      def [](index)
+        to_ary[index]
+      end
+
+      def []=(index, new_value)
+        case index
+        when 0
+          self.type = new_value
+        when 1
+          self.value = new_value
+        when 2
+          self.position = new_value
+        else
+          raise IndexError, "index #{index} outside of token"
+        end
+      end
+
+      def first
+        type
+      end
+
+      def last
+        position
+      end
+
+      def to_ary
+        [type, value, position]
+      end
+
+      alias to_a to_ary
+
+      def ==(other)
+        return super unless other.respond_to?(:to_ary)
+
+        other_type, other_value, other_position = other.to_ary
+        return false unless type == other_type
+        return false unless value == other_value
+        return position == other_position unless position.is_a?(Hash) && other_position.is_a?(Hash)
+
+        other_position.all? { |key, expected| position[key] == expected }
+      end
     end
 
     TOKENS = {
+      # Namespace and column combinators
+      '|' => :PIPE,
+      '||' => :COLUMN,
+
       # Combinators
       '>' => :CHILD,
       '+' => :ADJACENT,
@@ -37,27 +101,42 @@ module Parselly
       '*=' => :SUBSTRINGMATCH
     }.freeze
 
-    # Pre-compiled regular expressions for better performance
-    MULTI_CHAR_OPERATORS = [
-      [/~=/, :INCLUDES],
-      [/\|=/, :DASHMATCH],
-      [/\^=/, :PREFIXMATCH],
-      [/\$=/, :SUFFIXMATCH],
-      [/\*=/, :SUBSTRINGMATCH]
-    ].freeze
-
-    SINGLE_CHAR_OPERATOR_REGEX = /[>+~\[\]():,.#*=-]/.freeze
-    WHITESPACE_REGEX = /[ \t\n\r]+/.freeze
-    STRING_DOUBLE_REGEX = /"([^"\\]|\\.)*"/.freeze
-    STRING_SINGLE_REGEX = /'([^'\\]|\\.)*'/.freeze
-    IDENTIFIER_REGEX = /(?:--|-?[a-zA-Z_])(?:[\w-]|\\[^\n\r\f])*/.freeze
+    MULTI_CHAR_TOKENS = {
+      '~=' => :INCLUDES,
+      '|=' => :DASHMATCH,
+      '^=' => :PREFIXMATCH,
+      '$=' => :SUFFIXMATCH,
+      '*=' => :SUBSTRINGMATCH,
+      '||' => :COLUMN
+    }.freeze
+
+    SINGLE_CHAR_OPERATOR_REGEX = /[|>+~\[\]():,.#*=-]/.freeze
+    WHITESPACE_REGEX = /[ \t\n\r\f]+/.freeze
+    COMMENT_REGEX = %r{/\*[^*]*\*+(?:[^/*][^*]*\*+)*/}.freeze
+    ESCAPE_SEQUENCE = /\\(?:[0-9a-fA-F]{1,6}[ \t\n\r\f]?|[^\n\r\f])/.freeze
+    IDENTIFIER_REGEX = /
+      (?:
+        --
+        |
+        -?(?:[a-zA-Z_]|[^\x00-\x7F]|#{ESCAPE_SEQUENCE})
+      )
+      (?:[a-zA-Z0-9_-]|[^\x00-\x7F]|#{ESCAPE_SEQUENCE})*
+    /x.freeze
     NUMBER_REGEX = /\d+(\.\d+)?/.freeze
-    ESCAPE_REGEX = /\\(.)/.freeze
+    HEX_ESCAPE_REGEX = /\\([0-9a-fA-F]{1,6})([ \t\n\r\f])?/.freeze
+    ESCAPED_NEWLINE_REGEX = /\\(?:\r\n|[\n\r\f])/.freeze
+    SIMPLE_ESCAPE_REGEX = /\\([^\n\r\f])/.freeze
+    REPLACEMENT_CHARACTER = "\uFFFD"
 
     attr_reader :line, :column
 
     def initialize(input)
-      @scanner = StringScanner.new(input)
+      unless input.valid_encoding?
+        raise_lexer_error('Invalid input encoding', { line: 1, column: 1, offset: 0 })
+      end
+
+      preprocessed_input, @offset_map = preprocess_input(input)
+      @scanner = StringScanner.new(preprocessed_input)
       @line = 1
       @column = 1
       @tokens = []
@@ -65,54 +144,99 @@ module Parselly
 
     def tokenize
       until @scanner.eos?
-        skip_whitespace
+        skip_ignored
         break if @scanner.eos?
 
-        pos = { line: @line, column: @column, offset: @scanner.pos }
+        start_position = current_position
 
-        if (token = scan_string)
-          @tokens << [:STRING, token, pos]
-        elsif (token = scan_number)
-          @tokens << [:NUMBER, token, pos]
-        elsif (token = scan_operator)
-          @tokens << [token, @scanner.matched, pos]
-        elsif (token = scan_identifier)
-          @tokens << [:IDENT, token, pos]
+        if (token = scan_string(start_position))
+          type, value = token
+          @tokens << build_token(type, value, start_position)
+        elsif (value = scan_number)
+          @tokens << build_token(:NUMBER, value, start_position)
+        elsif (type = scan_operator)
+          @tokens << build_token(type, @scanner.matched, start_position)
+        elsif (value = scan_identifier(start_position))
+          @tokens << build_token(:IDENT, value, start_position)
         else
           char = @scanner.getch
-          raise "Unexpected character: #{char} at #{pos[:line]}:#{pos[:column]} (offset #{pos[:offset]})"
+          raise_lexer_error("Unexpected character: #{char}", start_position)
         end
       end
 
-      @tokens << [false, nil, { line: @line, column: @column, offset: @scanner.pos }]
+      @tokens << Token.new(type: false, value: nil, position: eof_position)
       @tokens
     end
 
     private
 
-    def skip_whitespace
-      while @scanner.scan(WHITESPACE_REGEX)
-        matched = @scanner.matched
-        newline_count = matched.count("\n")
-        if newline_count > 0
-          @line += newline_count
-          @column = matched.size - matched.rindex("\n")
+    def preprocess_input(input)
+      output = +''
+      offset_map = { 0 => 0 }
+      chars = input.each_char.to_a
+      original_offset = 0
+      index = 0
+
+      while index < chars.length
+        char = chars[index]
+        original_start = original_offset
+        original_offset += char.bytesize
+
+        if char == "\r"
+          if chars[index + 1] == "\n"
+            index += 1
+            original_offset += chars[index].bytesize
+          end
+          append_preprocessed(output, offset_map, "\n", original_start, original_offset)
+        elsif char == "\f"
+          append_preprocessed(output, offset_map, "\n", original_start, original_offset)
+        elsif char == "\0" || surrogate_codepoint?(char)
+          append_preprocessed(output, offset_map, REPLACEMENT_CHARACTER, original_start, original_offset)
         else
-          @column += matched.size
+          append_preprocessed(output, offset_map, char, original_start, original_offset)
         end
+
+        index += 1
       end
+
+      offset_map[output.bytesize] = original_offset
+      [output, offset_map]
     end
 
-    def scan_operator
-      # Check multi-character operators first
-      MULTI_CHAR_OPERATORS.each do |regex, token|
-        if @scanner.scan(regex)
+    def append_preprocessed(output, offset_map, value, original_start, original_end)
+      offset_map[output.bytesize] = original_start
+      output << value
+      offset_map[output.bytesize] = original_end
+    end
+
+    def surrogate_codepoint?(char)
+      char.ord.between?(0xD800, 0xDFFF)
+    end
+
+    def skip_ignored
+      loop do
+        if @scanner.scan(WHITESPACE_REGEX)
+          update_position(@scanner.matched)
+        elsif @scanner.peek(2) == '/*'
+          pos = { line: @line, column: @column, offset: @scanner.pos }
+          unless @scanner.scan(COMMENT_REGEX)
+            raise_lexer_error('Unterminated comment', pos)
+          end
           update_position(@scanner.matched)
-          return token
+        else
+          break
         end
       end
+    end
+
+    def scan_operator
+      two_chars = @scanner.peek(2)
+      if (token = MULTI_CHAR_TOKENS[two_chars])
+        @scanner.pos += 2
+        update_position(two_chars)
+        return token
+      end
 
-      # Single character operators
       return unless @scanner.scan(SINGLE_CHAR_OPERATOR_REGEX)
 
       char = @scanner.matched
@@ -120,25 +244,26 @@ module Parselly
       TOKENS[char]
     end
 
-    # NOTE: Unlike identifiers (where backslash escapes are processed),
-    # escape sequences inside strings (e.g., \n, \", \', \\) are NOT processed.
-    # The raw string content is returned as-is after removing outer quotes.
-    # This is a known limitation for attribute values, as strings are treated
-    # as raw text for simplicity. Identifiers process escapes to support patterns
-    # like .hover\:bg-blue-500, but strings in attributes don't require this.
-    def scan_string
-      if @scanner.scan(STRING_DOUBLE_REGEX)
-        str = @scanner.matched
-        update_position(str)
-        str[1..-2] # Remove quotes
-      elsif @scanner.scan(STRING_SINGLE_REGEX)
-        str = @scanner.matched
-        update_position(str)
-        str[1..-2] # Remove quotes
+    def scan_string(position)
+      quote = @scanner.peek(1)
+      return unless quote == '"' || quote == "'"
+
+      @scanner.getch
+      update_position(quote)
+      raw = +''
+
+      until @scanner.eos?
+        char = @scanner.peek(1)
+        return build_string_token(:STRING, raw, position, quote) if char == quote && consume_string_char(raw)
+        return build_string_token(:BAD_STRING, raw, position, quote) if newline?(char)
+
+        consume_string_char(raw)
       end
+
+      build_string_token(:STRING, raw, position, quote)
     end
 
-    def scan_identifier
+    def scan_identifier(position)
       # Match identifiers with optional escape sequences
       # CSS allows \<any-char> as escape in identifiers (e.g., .hover\:bg-blue-500)
       #
@@ -150,9 +275,7 @@ module Parselly
 
       ident = @scanner.matched
       update_position(ident)
-      # Remove backslashes from escaped characters
-      normalized = ident.gsub(ESCAPE_REGEX, '\1')
-      Identifier.new(normalized, ident)
+      Identifier.new(unescape_css(ident), ident).tap { |identifier| identifier.position = position }
     end
 
     def scan_number
@@ -163,15 +286,102 @@ module Parselly
       num
     end
 
+    def consume_string_char(raw)
+      char = @scanner.getch
+      update_position(char)
+      return true if char == '"' || char == "'"
+
+      raw << char
+      return true unless char == '\\'
+      return true if @scanner.eos?
+
+      escaped = @scanner.getch
+      update_position(escaped)
+      raw << escaped
+      true
+    end
+
+    def build_string_token(type, raw, position, quote)
+      [type, TokenValue.new(value: unescape_css(raw), raw: raw, position: position, quote: quote)]
+    end
+
+    def newline?(char)
+      char == "\n" || char == "\r" || char == "\f"
+    end
+
     def update_position(text)
-      text.each_char do |char|
-        if char == "\n"
-          @line += 1
-          @column = 1
-        else
-          @column += 1
-        end
+      unless text.match?(/[\n\r\f]/)
+        @column += text.each_char.count
+        return
       end
+
+      lines = text.split(/\r\n|[\n\r\f]/, -1)
+      @line += lines.length - 1
+      @column = lines.last.each_char.count + 1
+    end
+
+    def current_position
+      { line: @line, column: @column, offset: original_offset(@scanner.pos) }
+    end
+
+    def original_offset(preprocessed_offset)
+      @offset_map.fetch(preprocessed_offset, preprocessed_offset)
+    end
+
+    def build_token(type, value, start_position)
+      position = start_position.merge(
+        start_line: start_position[:line],
+        start_column: start_position[:column],
+        start_offset: start_position[:offset],
+        end_line: @line,
+        end_column: @column,
+        end_offset: original_offset(@scanner.pos)
+      )
+
+      value.position = position if value.respond_to?(:position=)
+      Token.new(type: type, value: value, position: position)
+    end
+
+    def eof_position
+      current_position.merge(
+        start_line: @line,
+        start_column: @column,
+        start_offset: original_offset(@scanner.pos),
+        end_line: @line,
+        end_column: @column,
+        end_offset: original_offset(@scanner.pos)
+      )
+    end
+
+    def unescape_css(value)
+      value
+        .gsub(ESCAPED_NEWLINE_REGEX, '')
+        .gsub(HEX_ESCAPE_REGEX) { decode_hex_escape(Regexp.last_match(1)) }
+        .gsub(SIMPLE_ESCAPE_REGEX, '\1')
+    end
+
+    def decode_hex_escape(hex)
+      codepoint = hex.to_i(16)
+      return REPLACEMENT_CHARACTER if codepoint.zero? || codepoint > 0x10FFFF
+
+      codepoint.chr(Encoding::UTF_8)
+    rescue RangeError
+      REPLACEMENT_CHARACTER
+    end
+
+    def raise_lexer_error(message, position)
+      error = {
+        message: "#{message} at #{position[:line]}:#{position[:column]} (offset #{position[:offset]})",
+        line: position[:line],
+        column: position[:column],
+        offset: position[:offset]
+      }
+
+      if defined?(Parselly::LexError)
+        raise Parselly::LexError, error
+      end
+
+      raise error[:message]
     end
   end
 end