lrama 0.5.1 → 0.5.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 20a0927bedeb14281abb8a5c59daf643fdfe556b5899cbe4872878b022804da2
-  data.tar.gz: 731fdd0201266db9724a2ebcc1162427a2cfaa9d0d94543e1298b3f2683eb14f
+  metadata.gz: 32eaf486545ca4143a227a34cee2bfdb94928e002a85fcb2e098dba3dedb71c7
+  data.tar.gz: b80a7402d9f6caf9f0e8b5396e4bd6b80d84487a35c8cd4b4d711e51ffb56601
 SHA512:
-  metadata.gz: 3a6e7c0ef4e7266dae41dba15e3a5d7609cb7171b82e0dbd5b75595e8e3cf8a4a402c5b710cc0f714c54357ef9fc20fd0935e55b8c529c99cd3c09ea173efb87
-  data.tar.gz: fa867c9be43384c7f282bf22ecd2b0939cd0dca263a51914871dba42f409ddc01fb00c34d8696288c3668177fdc450ff1bc9fb5e90fdf053e2d56a5bf31bbe04
+  metadata.gz: 1d118074af8984d85e27485a0c9fb2f8c7eab9a55b771ea3eff17e26d84fe56316404f6316ba5a9c83e61591c83ff9048c6625a904529aa686417f6f295499d9
+  data.tar.gz: e3591cc5f70e6fe8a04af5070a7b5d6ee60e98718c4c6dc6db1fbef2b3be9f5239155da91bb51ba10611b5017925f1bf71aa1259fd5550b5af6d0ba458e2ab96

data/.github/workflows/test.yaml

@@ -35,6 +35,7 @@ jobs:
           ruby-version: ${{ matrix.ruby }}
           bundler-cache: true
       - run: bundle install
+      - run: bundle exec rbs collection install
       - run: bundle exec steep check
   test-ruby:
     runs-on: ubuntu-20.04
@@ -57,9 +58,7 @@ jobs:
       - run: mkdir -p tool/lrama
         working-directory: ../ruby
       - name: Copy Lrama to ruby/tool
-        run: cp -r exe lib ../ruby/tool/lrama
-        # TODO: Consider how to manage changes on ruby/ruby master and ruby/lrama
-        # run: cp -r exe lib template ../ruby/tool/lrama
+        run: cp -r exe lib template ../ruby/tool/lrama
         working-directory:
       - run: tree tool/lrama
         working-directory: ../ruby

data/.gitignore

@@ -1,4 +1,6 @@
+/.gem_rbs_collection/
 /tmp
 .irbrc
 /Gemfile.lock
 /pkg/
+coverage/

data/Gemfile

@@ -8,3 +8,4 @@ gem "stackprof"
 gem "rake"
 gem "rbs", require: false
 gem "steep", require: false
+gem "simplecov", require: false

data/LEGAL.md

@@ -5,22 +5,7 @@ mentioned below.
 
 ## GNU General Public License version 3
 
-These files are licensed under the GNU General Public License version 3. See these files for more information.
+These files are licensed under the GNU General Public License version 3 or later. See these files for more information.
 
 * template/bison/yacc.c
 * template/bison/yacc.h
-
-## Same with Ruby
-
-These files are licensed same with Ruby. See https://github.com/ruby/ruby/blob/master/COPYING for more information.
-
-* spec/fixtures/integration/ruby_3_0_5/parse.tmp.y
-* spec/fixtures/integration/ruby_3_0_5/y.tab.c
-* spec/fixtures/integration/ruby_3_0_5/y.tab.h
-* spec/fixtures/integration/ruby_3_1_0/parse.tmp.y
-* spec/fixtures/integration/ruby_3_1_0/y.tab.c
-* spec/fixtures/integration/ruby_3_1_0/y.tab.h
-* spec/fixtures/integration/ruby_3_2_0/parse.tmp.y
-* spec/fixtures/integration/ruby_3_2_0/y.tab.c
-* spec/fixtures/integration/ruby_3_2_0/y.tab.h
-

data/README.md

@@ -62,7 +62,7 @@ This branch generates "parse.c" compatible with Bison 3.8.2 for ruby 3.0, 3.1, 3
 
 1. Update `Lrama::VERSION`
 2. Release as a gem by `rake release`
-3. Update Lrama in ruby/ruby by `cp -r exe lib ruby/tool/lrama`
+3. Update Lrama in ruby/ruby by `cp -r LEGAL.md MIT exe lib ruby/tool/lrama`
 4. Create new release on [GitHub](https://github.com/ruby/lrama/releases)
 
 ## License

data/Steepfile

@@ -4,4 +4,7 @@ target :lib do
   signature "sig"
 
   check "lib/lrama/bitmap.rb"
+  check "lib/lrama/report/duration.rb"
+  check "lib/lrama/report/profile.rb"
+  check "lib/lrama/warning.rb"
 end

data/doc/TODO.md

@@ -1,7 +1,6 @@
 # TODO
 
 * command
-  * [ ] Add "--bison" option
 * lexer
   * [x] Basic functionalities
 * parser
@@ -27,7 +26,7 @@
   * [x] Table compaction
   * [x] -d option
   * yacc.c
-    * [ ] %lex-param
+    * [x] %lex-param
     * [x] %parse-param
     * [x] %printer
     * [x] Replace $, @ in user codes
@@ -46,7 +45,9 @@
   * [ ] Bison style
     * [ ] Wrap not selected reduce with "[]". See basic.output file generated by Bison.
 * Error Tolerance
-  * [x] Subset of Corchuelo et al.
+  * [x] Corchuelo et al. algorithm with N = 1 (this means the next token when error is raised)
+  * [x] Add new decl for error token semantic value initialization (%error-token)
+  * [ ] Use YYMALLOC & YYFREE
 * Lex state
 * CI
   * [x] Setup CI

data/exe/lrama

@@ -4,4 +4,4 @@
 $LOAD_PATH << File.join(__dir__, "../lib")
 require "lrama"
 
-Lrama::Command.new.run(ARGV.dup)
+Lrama::Command.new(ARGV.dup).run

data/lib/lrama/command.rb

@@ -2,94 +2,57 @@ require 'optparse'
 
 module Lrama
   class Command
-    def run(argv)
-      opt = OptionParser.new
-
-      # opt.on('-h') {|v| p v }
-      opt.on('-V', '--version') {|v| puts Lrama::VERSION ; exit 0 }
-
-      # Tuning the Parser
-      skeleton = "bison/yacc.c"
-
-      opt.on('-S', '--skeleton=FILE') {|v| skeleton = v }
-      opt.on('-t') {  } # Do nothing
-
-      # Output Files:
-      header = false
-      header_file = nil
-      report = []
-      report_file = nil
-      outfile = "y.tab.c"
-
-      opt.on('-h', '--header=[FILE]') {|v| header = true; header_file = v }
-      opt.on('-d') { header = true }
-      opt.on('-r', '--report=THINGS') {|v| report = v.split(',') }
-      opt.on('--report-file=FILE')    {|v| report_file = v }
-      opt.on('-v') {  } # Do nothing
-      opt.on('-o', '--output=FILE')   {|v| outfile = v }
-
-      # Hidden
-      trace = []
-      opt.on('--trace=THINGS') {|v| trace = v.split(',') }
-
-      # Error Recovery
-      error_recovery = false
-      opt.on('-e') {|v| error_recovery = true }
-
-      opt.parse!(argv)
-
-      trace_opts = validate_trace(trace)
-      report_opts = validate_report(report)
-
-      grammar_file = argv.shift
-
-      if !report.empty? && report_file.nil? && grammar_file
-        report_file = File.dirname(grammar_file) + "/" + File.basename(grammar_file, ".*") + ".output"
-      end
+    def initialize(argv)
+      @argv = argv
+
+      @version = nil
+      @skeleton = "bison/yacc.c"
+      @header = false
+      @header_file = nil
+      @report = []
+      @report_file = nil
+      @outfile = "y.tab.c"
+      @trace = []
+      @error_recovery = false
+      @grammar_file = nil
+      @report_file = nil
+      @trace_opts = nil
+      @report_opts = nil
+    end
 
-      if !header_file && header
-        case
-        when outfile
-          header_file = File.dirname(outfile) + "/" + File.basename(outfile, ".*") + ".h"
-        when grammar_file
-          header_file = File.dirname(grammar_file) + "/" + File.basename(grammar_file, ".*") + ".h"
-        end
-      end
+    def run
+      parse_option
 
-      if !grammar_file
-        abort "File should be specified\n"
+      if @version
+        puts Lrama::VERSION
+        exit 0
       end
 
-      Report::Duration.enable if trace_opts[:time]
+      Report::Duration.enable if @trace_opts[:time]
 
       warning = Lrama::Warning.new
-      if grammar_file == '-'
-        grammar_file = argv.shift or abort "File name for STDIN should be specified\n"
-        y = STDIN.read
-      else
-        y = File.read(grammar_file)
-      end
-      grammar = Lrama::Parser.new(y).parse
-      states = Lrama::States.new(grammar, warning, trace_state: (trace_opts[:automaton] || trace_opts[:closure]))
+      grammar = Lrama::Parser.new(@y.read).parse
+      states = Lrama::States.new(grammar, warning, trace_state: (@trace_opts[:automaton] || @trace_opts[:closure]))
       states.compute
       context = Lrama::Context.new(states)
 
-      if report_file
+      if @report_file
         reporter = Lrama::StatesReporter.new(states)
-        File.open(report_file, "w+") do |f|
-          reporter.report(f, **report_opts)
+        File.open(@report_file, "w+") do |f|
+          reporter.report(f, **@report_opts)
         end
       end
 
-      File.open(outfile, "w+") do |f|
+      File.open(@outfile, "w+") do |f|
         Lrama::Output.new(
           out: f,
-          output_file_path: outfile,
-          template_name: skeleton,
-          grammar_file_path: grammar_file,
-          header_file_path: header_file,
+          output_file_path: @outfile,
+          template_name: @skeleton,
+          grammar_file_path: @grammar_file,
+          header_file_path: @header_file,
           context: context,
           grammar: grammar,
+          error_recovery: @error_recovery,
         ).render
       end
 
@@ -144,5 +107,61 @@ module Lrama
 
       return h
     end
+
+    def parse_option
+      opt = OptionParser.new
+
+      # opt.on('-h') {|v| p v }
+      opt.on('-V', '--version') {|v| @version = true }
+
+      # Tuning the Parser
+      opt.on('-S', '--skeleton=FILE') {|v| @skeleton = v }
+      opt.on('-t') {  } # Do nothing
+
+      # Output Files:
+      opt.on('-h', '--header=[FILE]') {|v| @header = true; @header_file = v }
+      opt.on('-d') { @header = true }
+      opt.on('-r', '--report=THINGS') {|v| @report = v.split(',') }
+      opt.on('--report-file=FILE')    {|v| @report_file = v }
+      opt.on('-v') {  } # Do nothing
+      opt.on('-o', '--output=FILE')   {|v| @outfile = v }
+
+      # Hidden
+      opt.on('--trace=THINGS') {|v| @trace = v.split(',') }
+
+      # Error Recovery
+      opt.on('-e') {|v| @error_recovery = true }
+
+      opt.parse!(@argv)
+
+      @trace_opts = validate_trace(@trace)
+      @report_opts = validate_report(@report)
+
+      @grammar_file = @argv.shift
+
+      if !@grammar_file
+        abort "File should be specified\n"
+      end
+
+      if @grammar_file == '-'
+        @grammar_file = @argv.shift or abort "File name for STDIN should be specified\n"
+        @y = STDIN
+      else
+        @y = File.open(@grammar_file, 'r')
+      end
+
+      if !@report.empty? && @report_file.nil? && @grammar_file
+        @report_file = File.dirname(@grammar_file) + "/" + File.basename(@grammar_file, ".*") + ".output"
+      end
+
+      if !@header_file && @header
+        case
+        when @outfile
+          @header_file = File.dirname(@outfile) + "/" + File.basename(@outfile, ".*") + ".h"
+        when @grammar_file
+          @header_file = File.dirname(@grammar_file) + "/" + File.basename(@grammar_file, ".*") + ".h"
+        end
+      end
+    end
   end
 end

data/lib/lrama/context.rb

@@ -1,4 +1,4 @@
-require "lrama/report"
+require "lrama/report/duration"
 
 module Lrama
   # This is passed to a template
@@ -89,6 +89,16 @@ module Lrama
       return a
     end
 
+    def yytranslate_inverted
+      a = Array.new(@states.symbols.count, @states.undef_symbol.token_id)
+
+      @states.terms.each do |term|
+        a[term.number] = term.token_id
+      end
+
+      return a
+    end
+
     # Mapping from rule number to line number of the rule is defined.
     # Dummy rule is appended as the first element whose value is 0
     # because 0 means error in yydefact.

data/lib/lrama/grammar/code.rb

@@ -0,0 +1,123 @@
+require "forwardable"
+
+module Lrama
+  class Grammar
+    class Code < Struct.new(:type, :token_code, keyword_init: true)
+      extend Forwardable
+
+      def_delegators "token_code", :s_value, :line, :column, :references
+
+      # $$, $n, @$, @n is translated to C code
+      def translated_code
+        case type
+        when :user_code
+          translated_user_code
+        when :initial_action
+          translated_initial_action_code
+        end
+      end
+
+      # * ($1) error
+      # * ($$) *yyvaluep
+      # * (@1) error
+      # * (@$) *yylocationp
+      def translated_printer_code(tag)
+        t_code = s_value.dup
+
+        references.reverse.each do |ref|
+          first_column = ref.first_column
+          last_column = ref.last_column
+
+          case
+          when ref.value == "$" && ref.type == :dollar # $$
+            # Omit "<>"
+            member = tag.s_value[1..-2]
+            str = "((*yyvaluep).#{member})"
+          when ref.value == "$" && ref.type == :at # @$
+            str = "(*yylocationp)"
+          when ref.type == :dollar # $n
+            raise "$#{ref.value} can not be used in %printer."
+          when ref.type == :at # @n
+            raise "@#{ref.value} can not be used in %printer."
+          else
+            raise "Unexpected. #{self}, #{ref}"
+          end
+
+          t_code[first_column..last_column] = str
+        end
+
+        return t_code
+      end
+      alias :translated_error_token_code :translated_printer_code
+
+
+      private
+
+      # * ($1) yyvsp[i]
+      # * ($$) yyval
+      # * (@1) yylsp[i]
+      # * (@$) yyloc
+      def translated_user_code
+        t_code = s_value.dup
+
+        references.reverse.each do |ref|
+          first_column = ref.first_column
+          last_column = ref.last_column
+
+          case
+          when ref.value == "$" && ref.type == :dollar # $$
+            # Omit "<>"
+            member = ref.tag.s_value[1..-2]
+            str = "(yyval.#{member})"
+          when ref.value == "$" && ref.type == :at # @$
+            str = "(yyloc)"
+          when ref.type == :dollar # $n
+            i = -ref.position_in_rhs + ref.value
+            # Omit "<>"
+            member = ref.tag.s_value[1..-2]
+            str = "(yyvsp[#{i}].#{member})"
+          when ref.type == :at # @n
+            i = -ref.position_in_rhs + ref.value
+            str = "(yylsp[#{i}])"
+          else
+            raise "Unexpected. #{self}, #{ref}"
+          end
+
+          t_code[first_column..last_column] = str
+        end
+
+        return t_code
+      end
+
+      # * ($1) error
+      # * ($$) yylval
+      # * (@1) error
+      # * (@$) yylloc
+      def translated_initial_action_code
+        t_code = s_value.dup
+
+        references.reverse.each do |ref|
+          first_column = ref.first_column
+          last_column = ref.last_column
+
+          case
+          when ref.value == "$" && ref.type == :dollar # $$
+            str = "yylval"
+          when ref.value == "$" && ref.type == :at # @$
+            str = "yylloc"
+          when ref.type == :dollar # $n
+            raise "$#{ref.value} can not be used in initial_action."
+          when ref.type == :at # @n
+            raise "@#{ref.value} can not be used in initial_action."
+          else
+            raise "Unexpected. #{self}, #{ref}"
+          end
+
+          t_code[first_column..last_column] = str
+        end
+
+        return t_code
+      end
+    end
+  end
+end

data/lib/lrama/grammar/error_token.rb

@@ -0,0 +1,9 @@
+module Lrama
+  class Grammar
+    class ErrorToken < Struct.new(:ident_or_tags, :code, :lineno, keyword_init: true)
+      def translated_code(member)
+        code.translated_error_token_code(member)
+      end
+    end
+  end
+end

data/lib/lrama/grammar/precedence.rb

@@ -0,0 +1,11 @@
+module Lrama
+  class Grammar
+    class Precedence < Struct.new(:type, :precedence, keyword_init: true)
+      include Comparable
+
+      def <=>(other)
+        self.precedence <=> other.precedence
+      end
+    end
+  end
+end

data/lib/lrama/grammar/printer.rb

@@ -0,0 +1,9 @@
+module Lrama
+  class Grammar
+    class Printer < Struct.new(:ident_or_tags, :code, :lineno, keyword_init: true)
+      def translated_code(member)
+        code.translated_printer_code(member)
+      end
+    end
+  end
+end

data/lib/lrama/grammar/reference.rb

@@ -0,0 +1,22 @@
+# type: :dollar or :at
+# ex_tag: "$<tag>1" (Optional)
+
+module Lrama
+  class Grammar
+    class Reference < Struct.new(:type, :value, :ex_tag, :first_column, :last_column, :referring_symbol, :position_in_rhs, keyword_init: true)
+      def tag
+        if ex_tag
+          ex_tag
+        else
+          # FIXME: Remove this class check
+          if referring_symbol.is_a?(Symbol)
+            referring_symbol.tag
+          else
+            # Lrama::Lexer::Token (User_code) case
+            nil
+          end
+        end
+      end
+    end
+  end
+end

data/lib/lrama/grammar/rule.rb

@@ -0,0 +1,33 @@
+module Lrama
+  class Grammar
+    class Rule < Struct.new(:id, :lhs, :rhs, :code, :nullable, :precedence_sym, :lineno, keyword_init: true)
+      # TODO: Change this to display_name
+      def to_s
+        l = lhs.id.s_value
+        r = rhs.empty? ? "ε" : rhs.map {|r| r.id.s_value }.join(", ")
+
+        "#{l} -> #{r}"
+      end
+
+      # Used by #user_actions
+      def as_comment
+        l = lhs.id.s_value
+        r = rhs.empty? ? "%empty" : rhs.map(&:display_name).join(" ")
+
+        "#{l}: #{r}"
+      end
+
+      def precedence
+        precedence_sym&.precedence
+      end
+
+      def initial_rule?
+        id == 0
+      end
+
+      def translated_code
+        code&.translated_code
+      end
+    end
+  end
+end

data/lib/lrama/grammar/symbol.rb

@@ -0,0 +1,94 @@
+# Symbol is both of nterm and term
+# `number` is both for nterm and term
+# `token_id` is tokentype for term, internal sequence number for nterm
+#
+# TODO: Add validation for ASCII code range for Token::Char
+
+module Lrama
+  class Grammar
+    class Symbol < Struct.new(:id, :alias_name, :number, :tag, :term, :token_id, :nullable, :precedence, :printer, :error_token, keyword_init: true)
+      attr_writer :eof_symbol, :error_symbol, :undef_symbol, :accept_symbol
+
+      def term?
+        term
+      end
+
+      def nterm?
+        !term
+      end
+
+      def eof_symbol?
+        !!@eof_symbol
+      end
+
+      def error_symbol?
+        !!@error_symbol
+      end
+
+      def undef_symbol?
+        !!@undef_symbol
+      end
+
+      def accept_symbol?
+        !!@accept_symbol
+      end
+
+      def display_name
+        if alias_name
+          alias_name
+        else
+          id.s_value
+        end
+      end
+
+      # name for yysymbol_kind_t
+      #
+      # See: b4_symbol_kind_base
+      def enum_name
+        case
+        when accept_symbol?
+          name = "YYACCEPT"
+        when eof_symbol?
+          name = "YYEOF"
+        when term? && id.type == Token::Char
+          if alias_name
+            name = number.to_s + alias_name
+          else
+            name = number.to_s + id.s_value
+          end
+        when term? && id.type == Token::Ident
+          name = id.s_value
+        when nterm? && (id.s_value.include?("$") || id.s_value.include?("@"))
+          name = number.to_s + id.s_value
+        when nterm?
+          name = id.s_value
+        else
+          raise "Unexpected #{self}"
+        end
+
+        "YYSYMBOL_" + name.gsub(/[^a-zA-Z_0-9]+/, "_")
+      end
+
+      # comment for yysymbol_kind_t
+      def comment
+        case
+        when accept_symbol?
+          # YYSYMBOL_YYACCEPT
+          id.s_value
+        when eof_symbol?
+          # YYEOF
+          alias_name
+        when (term? && 0 < token_id && token_id < 128)
+          # YYSYMBOL_3_backslash_, YYSYMBOL_14_
+          alias_name || id.s_value
+        when id.s_value.include?("$") || id.s_value.include?("@")
+          # YYSYMBOL_21_1
+          id.s_value
+        else
+          # YYSYMBOL_keyword_class, YYSYMBOL_strings_1
+          alias_name || id.s_value
+        end
+      end
+    end
+  end
+end

data/lib/lrama/grammar/union.rb

@@ -0,0 +1,10 @@
+module Lrama
+  class Grammar
+    class Union < Struct.new(:code, :lineno, keyword_init: true)
+      def braces_less_code
+        # Remove braces
+        code.s_value[1..-2]
+      end
+    end
+  end
+end