Class: Ibex::Normalizer
- Inherits:
-
Object
- Object
- Ibex::Normalizer
- Includes:
- NormalizeDeclarations, NormalizeDiagnostics, NormalizeExpander, NormalizeInlineExpansion, NormalizeInlineValidation, NormalizeLexer, NormalizeNamedReferences, NormalizeNodes, NormalizeParameterEbnfLowering, NormalizeParameterLowering, NormalizeParameterSubstitution, NormalizeParameterValidation, NormalizeParameters, NormalizeRecoveryDeclarations
- Defined in:
- lib/ibex/normalize.rb,
sig/ibex/normalize.rbs
Overview
Converts a frontend AST into immutable Grammar IR.
Constant Summary collapse
- RESERVED_NAMES =
%w[result val _values].freeze
- RUBY_KEYWORDS =
%w[ __ENCODING__ __FILE__ __LINE__ alias and begin break case class def defined do else elsif end ensure false for if in module next nil not or redo rescue retry return self super then true undef unless until when while yield ].freeze
- DEFAULT_MAX_PARAMETER_SPECIALIZATIONS =
1_000- DEFAULT_MAX_INLINE_EXPANSIONS =
10_000
Instance Method Summary collapse
- #fail_at(location, message) ⇒ bot
- #fail_hash(location, message) ⇒ bot
-
#initialize(input, mode: :default, max_parameter_specializations: DEFAULT_MAX_PARAMETER_SPECIALIZATIONS, max_inline_expansions: DEFAULT_MAX_INLINE_EXPANSIONS) ⇒ Normalizer
constructor
A new instance of Normalizer.
- #intern(name, kind, reserved: false, location: nil, documentation: nil, metadata_name: nil) ⇒ IR::GrammarSymbol
- #intern_declared_terminals ⇒ void
- #intern_reserved_symbols ⇒ void
- #intern_user_nonterminal(rule) ⇒ void
- #intern_user_nonterminals ⇒ void
- #nonterminal_name?(name) ⇒ Boolean
- #normalize ⇒ IR::Grammar
- #normalized_start_names ⇒ Array[String]
- #normalized_user_code ⇒ Hash[String, String]
- #normalized_user_code_chunks ⇒ IR::user_code_chunks
- #required_symbol(name) ⇒ IR::GrammarSymbol
- #resolved_expansion ⇒ IR::production_expansion?
- #start_rule_candidate?(rule) ⇒ Boolean
- #symbol(name) ⇒ IR::GrammarSymbol?
- #symbol_for_reference(reference) ⇒ IR::GrammarSymbol
- #undefined_nonterminal(reference) ⇒ bot
- #validate_positive_limit!(name, value) ⇒ void
- #validated_rule_documentation ⇒ Hash[String, String]
- #warn_undeclared_terminal(reference) ⇒ void
Methods included from NormalizeDiagnostics
#productive_terminal?, #productive_terminal_ids, #reachable_symbol_ids, #validate_grammar, #validate_symbol_metadata, #warn_duplicate_productions, #warn_empty_language, #warn_unreachable_nonterminals, #warn_unreachable_terminals, #warn_unused_precedence, #warn_unused_terminals
Methods included from NormalizeInlineExpansion
#advance_inline_expansion_frame?, #append_inline_array, #append_inline_symbol, #charge_inline_expansion!, #clone_inline_expansion_frames, #combine_inline_variant, #combined_inline_precedence, #combined_inline_steps, #complete_inline_expansion_frame?, #composed_inline_action, #composed_inline_production, #composition_plan_step, #copy_inline_production, #drain_inline_expansion_frames, #empty_inline_variant, #enforce_inline_choice_limit!, #expand_inline_production, #expand_inline_rules, #finish_inline_choice, #finish_inline_precedence, #inline_action_step, #inline_expansion_frame, #inline_implicit_code, #inline_production_expansion, #inline_source_provenance, #rebuild_inline_symbols, #remap_inline_production, #remap_inline_reference, #remap_inline_steps, #remap_inline_symbol, #resolve_inline_reference, #schedule_inline_choice_branches
Methods included from NormalizeExpander
#add_group_production, #add_production, #build_optional, #build_plus, #build_separated_list, #build_star, #expand_group, #expand_inline_action, #expand_optional, #expand_plus, #expand_separated_list, #expand_star, #group_value_expression, #new_helper, #normalize_action, #normalize_alternative, #normalize_item, #normalize_user_productions, #normalized_alternative_items, #synthetic_action, #synthetic_origin
Methods included from NormalizeNodes
#normalize_node_annotation, #validate_node_fields
Methods included from NormalizeNamedReferences
#add_named_reference, #named_reference_in, #reject_group_named_references, #unwrap_reference
Methods included from NormalizeParameters
#clear_parameter_alternative, #drain_parameter_worklist, #enforce_parameter_limits!, #growing_parameter_arguments?, #new_parameter_helper, #parameter_argument_contains?, #parameter_frame, #prepare_parameter_alternative?, #prepare_parameter_alternative_entry, #schedule_parameter_specialization, #specialize_parameterized_reference
Methods included from NormalizeParameterLowering
#advance_parameter_alternative, #finish_parameter_alternative, #finish_parameter_item, #lower_parameter_call, #lower_parameter_item, #parameter_items_remaining?, #process_parameter_operation, #start_parameter_item, #with_parameter_frame_context
Methods included from NormalizeParameterEbnfLowering
#finish_parameter_group_alternative, #finish_parameter_separated_list, #finish_parameter_suffix, #lower_parameter_group, #lower_parameter_separated_list, #lower_parameter_suffix, #start_parameter_group_alternative
Methods included from NormalizeParameterSubstitution
#clone_parameter_call, #clone_parameter_group, #clone_parameter_item, #clone_parameter_separated_list, #clone_parameter_suffix, #clone_parameter_symbol, #substitute_formal_reference, #substitute_parameter_alternative, #substitute_parameter_item, #substitute_parameter_precedence
Methods included from NormalizeInlineValidation
#collect_inline_edges, #collect_live_formals, #gather_inline_rules, #parameter_formal_liveness, #validate_inline_cycles, #validate_inline_start, #validate_inline_terminal_collisions, #visit_inline_cycle
Methods included from NormalizeParameterValidation
#gather_parameter_templates, #gather_parameterized_rule, #parameter_template?, #parameterized_rule?, #reject_argument_named_reference, #reject_mixed_parameterized_rule, #validate_parameter_call, #validate_parameter_formals, #validate_parameter_item, #validate_parameter_references, #validate_parameter_terminal_collisions, #validate_parameterized_item
Methods included from NormalizeRecoveryDeclarations
#read_grammar_test, #read_on_error_reduce_declaration, #read_parser_control_declaration, #read_recovery_declaration, #validate_recovery_declarations
Methods included from NormalizeLexer
#decode_lexer_literal, #normalize_lexer, #normalize_lexer_pattern, #normalize_lexer_rule, #normalize_lexer_state, #risky_lexer_pattern?, #validate_lexer_action, #validate_lexer_regexp, #validate_lexer_token
Methods included from NormalizeDeclarations
#read_conversions, #read_declaration, #read_declarations, #read_extended_declaration, #read_option, #read_options, #read_parser_parameter, #read_precedence, #read_start_declaration, #read_symbol_metadata, #read_symbol_metadata_declaration, #read_tokens, #read_value_printer, #validate_value_printers
Constructor Details
#initialize(input, mode: :default, max_parameter_specializations: DEFAULT_MAX_PARAMETER_SPECIALIZATIONS, max_inline_expansions: DEFAULT_MAX_INLINE_EXPANSIONS) ⇒ Normalizer
Returns a new instance of Normalizer.
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 |
# File 'lib/ibex/normalize.rb', line 94 def initialize(input, mode: :default, max_parameter_specializations: DEFAULT_MAX_PARAMETER_SPECIALIZATIONS, max_inline_expansions: DEFAULT_MAX_INLINE_EXPANSIONS) validate_positive_limit!(:max_parameter_specializations, max_parameter_specializations) validate_positive_limit!(:max_inline_expansions, max_inline_expansions) normalized_mode = mode.to_sym raise ArgumentError, "mode must be :default or :extended" unless %i[default extended].include?(normalized_mode) @resolution = input if input.is_a?(Frontend::Resolution) ast = input.is_a?(Frontend::Resolution) ? input.root : input normalized_mode = :extended if ast.is_a?(Frontend::AST::Root) && ast.extended fail_at(ast.loc, "fragments must be resolved before normalization") if ast.is_a?(Frontend::AST::Fragment) unresolved = ast.declarations.find { |declaration| declaration.is_a?(Frontend::AST::Include) } fail_at(unresolved.loc, "includes must be resolved before normalization") if unresolved @ast = ast @mode = normalized_mode #: IR::grammar_mode @symbols = [] #: Array[IR::GrammarSymbol] @symbols_by_name = {} #: Hash[String, IR::GrammarSymbol] @productions = [] #: Array[IR::Production] @warnings = [] #: Array[IR::grammar_warning] @helper_sequence = 0 @current_include_chain = [] #: Array[IR::source_provenance] @parameter_specializations = {} #: Hash[[String, Array[String]], String] @parameter_worklist = [] #: Array[Hash[Symbol, untyped]] @parameter_worklist_active = false @max_parameter_specializations = max_parameter_specializations @max_inline_expansions = max_inline_expansions @inline_expansion_count = 0 @inline_symbol_ids = Set.new #: Set[Integer] @inline_rule_by_symbol = {} #: Hash[Integer, String] end |
Instance Method Details
#fail_at(location, message) ⇒ bot
314 315 316 |
# File 'lib/ibex/normalize.rb', line 314 def fail_at(location, ) raise Ibex::Error, "#{location}: #{}" end |
#fail_hash(location, message) ⇒ bot
319 320 321 322 |
# File 'lib/ibex/normalize.rb', line 319 def fail_hash(location, ) rendered = location ? "#{location[:file]}:#{location[:line]}:#{location[:column]}" : "(grammar):1:1" raise Ibex::Error, "#{rendered}: #{}" end |
#intern(name, kind, reserved: false, location: nil, documentation: nil, metadata_name: nil) ⇒ IR::GrammarSymbol
236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 |
# File 'lib/ibex/normalize.rb', line 236 def intern(name, kind, reserved: false, location: nil, documentation: nil, metadata_name: nil) existing = symbol(name) if existing fail_hash(location, "symbol #{name} is both terminal and nonterminal") if existing.kind != kind return existing end precedence = @precedence[name] = || name definition = IR::GrammarSymbol.new(id: @symbols.length, name: name, kind: kind, reserved: reserved, precedence: precedence, location: location, display_name: @display_names[], semantic_type: @semantic_types[], documentation: documentation) @symbols << definition @symbols_by_name[name] = definition definition end |
#intern_declared_terminals ⇒ void
This method returns an undefined value.
174 175 176 177 |
# File 'lib/ibex/normalize.rb', line 174 def intern_declared_terminals @declared_tokens.each { |name, loc| intern(name, :terminal, location: loc) } @precedence.each_key { |name| intern(name, :terminal, location: @precedence_locations[name]) } end |
#intern_reserved_symbols ⇒ void
This method returns an undefined value.
168 169 170 171 |
# File 'lib/ibex/normalize.rb', line 168 def intern_reserved_symbols intern("$eof", :terminal, reserved: true) intern("error", :terminal, reserved: true) end |
#intern_user_nonterminal(rule) ⇒ void
This method returns an undefined value.
209 210 211 212 213 214 215 216 217 |
# File 'lib/ibex/normalize.rb', line 209 def intern_user_nonterminal(rule) definition = intern( rule.lhs, :nonterminal, location: rule.loc.to_h, documentation: @rule_documentation[rule.lhs] ) return unless rule.inline @inline_symbol_ids << definition.id @inline_rule_by_symbol[definition.id] = rule.lhs end |
#intern_user_nonterminals ⇒ void
This method returns an undefined value.
180 181 182 183 184 185 186 187 188 189 190 191 192 |
# File 'lib/ibex/normalize.rb', line 180 def intern_user_nonterminals @ast.rules.reject { |rule| parameterized_rule?(rule) }.each do |rule| intern_user_nonterminal(rule) end @start_names = normalized_start_names fail_at(@ast.loc, "grammar has no start rule") if @start_names.empty? @start_name = @start_names.fetch(0) @start_names.each do |name| next if symbol(name)&.nonterminal? fail_at(@start_location || @ast.loc, "undefined start symbol #{name}") end end |
#nonterminal_name?(name) ⇒ Boolean
278 279 280 |
# File 'lib/ibex/normalize.rb', line 278 def nonterminal_name?(name) name.match?(/\A[a-z_]/) && name != "error" end |
#normalize ⇒ IR::Grammar
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 |
# File 'lib/ibex/normalize.rb', line 127 def normalize read_declarations gather_parameter_templates gather_inline_rules intern_reserved_symbols intern_declared_terminals intern_user_nonterminals normalize_user_productions validate_value_printers validate_recovery_declarations validate_grammar IR::Grammar.new(class_name: @ast.class_name, superclass: @ast.superclass, start: @start_name, expect: @expected_conflicts, options: @options, symbols: @symbols, mode: @mode, starts: @start_names, expect_rr: @expected_rr_conflicts, parser_parameters: @parser_parameters, value_printers: @value_printers.values, grammar_tests: @grammar_tests, lexer: normalize_lexer, recovery: { sync_tokens: @recovery_sync_tokens, on_error_reduce: @on_error_reduce_groups }, productions: @productions, user_code: normalized_user_code, conversions: @conversions, warnings: @warnings, user_code_chunks: normalized_user_code_chunks, source_provenance: { file: @ast.loc.file, root: @resolution&.root_directory, byte_span: nil }) end |
#normalized_start_names ⇒ Array[String]
195 196 197 198 199 200 201 |
# File 'lib/ibex/normalize.rb', line 195 def normalized_start_names explicit = @explicit_starts return explicit if explicit candidate = @ast.rules.find { |rule| start_rule_candidate?(rule) }&.lhs candidate ? [candidate] : [] end |
#normalized_user_code ⇒ Hash[String, String]
295 296 297 298 299 |
# File 'lib/ibex/normalize.rb', line 295 def normalized_user_code %w[header inner footer].to_h do |name| [name, @ast.user_code.fetch(name, Array.new(0)).map(&:code).join] end end |
#normalized_user_code_chunks ⇒ IR::user_code_chunks
302 303 304 305 306 307 308 309 310 311 |
# File 'lib/ibex/normalize.rb', line 302 def normalized_user_code_chunks chunks_by_name = %w[header inner footer].to_h do |name| chunks = @ast.user_code.fetch(name, Array.new(0)).map do |block| location = block.loc.to_h.merge(line: block.loc.line + 1, column: 1) IR::UserCodeChunk.new(code: block.code, location: location) end [name, chunks] end chunks_by_name.reject { |_name, chunks| chunks.empty? } end |
#required_symbol(name) ⇒ IR::GrammarSymbol
260 261 262 |
# File 'lib/ibex/normalize.rb', line 260 def required_symbol(name) symbol(name) || raise(Ibex::Error, "missing normalized symbol #{name}") end |
#resolved_expansion ⇒ IR::production_expansion?
325 326 327 328 329 |
# File 'lib/ibex/normalize.rb', line 325 def resolved_expansion return unless @resolution || @current_parameter_expansion { parameter: @current_parameter_expansion, inline: nil, include_chain: @current_include_chain } end |
#start_rule_candidate?(rule) ⇒ Boolean
204 205 206 |
# File 'lib/ibex/normalize.rb', line 204 def start_rule_candidate?(rule) !parameterized_rule?(rule) && !rule.inline end |
#symbol(name) ⇒ IR::GrammarSymbol?
255 256 257 |
# File 'lib/ibex/normalize.rb', line 255 def symbol(name) @symbols_by_name[name] end |
#symbol_for_reference(reference) ⇒ IR::GrammarSymbol
265 266 267 268 269 270 271 272 273 274 275 |
# File 'lib/ibex/normalize.rb', line 265 def symbol_for_reference(reference) existing = symbol(reference.name) return existing if existing fail_at(reference.loc, "parameterized rule #{reference.name} requires arguments") if parameter_template?(reference.name) return undefined_nonterminal(reference) if nonterminal_name?(reference.name) warn_undeclared_terminal(reference) intern(reference.name, :terminal, location: reference.loc.to_h) end |
#undefined_nonterminal(reference) ⇒ bot
283 284 285 |
# File 'lib/ibex/normalize.rb', line 283 def undefined_nonterminal(reference) fail_at(reference.loc, "undefined nonterminal #{reference.name}") end |
#validate_positive_limit!(name, value) ⇒ void
This method returns an undefined value.
161 162 163 164 165 |
# File 'lib/ibex/normalize.rb', line 161 def validate_positive_limit!(name, value) return if value.is_a?(Integer) && value.positive? raise ArgumentError, "#{name} must be a positive Integer" end |
#validated_rule_documentation ⇒ Hash[String, String]
220 221 222 223 224 225 226 227 228 229 230 231 232 |
# File 'lib/ibex/normalize.rb', line 220 def validated_rule_documentation documentation = {} #: Hash[String, String] @ast.rules.each do |rule| text = rule.documentation next unless text existing = documentation[rule.lhs] fail_at(rule.loc, "conflicting documentation for rule #{rule.lhs}") if existing && existing != text documentation[rule.lhs] ||= text end documentation end |
#warn_undeclared_terminal(reference) ⇒ void
This method returns an undefined value.
288 289 290 291 292 |
# File 'lib/ibex/normalize.rb', line 288 def warn_undeclared_terminal(reference) return if @declared_tokens.empty? || reference.name.start_with?("'", '"') @warnings << { type: :undeclared_terminal, symbol: reference.name, loc: reference.loc.to_h } end |