OLD | NEW |
(Empty) | |
| 1 \input texinfo @c -*-texinfo-*- |
| 2 @comment %**start of header |
| 3 @setfilename bison.info |
| 4 @include version.texi |
| 5 @settitle Bison @value{VERSION} |
| 6 @setchapternewpage odd |
| 7 |
| 8 @finalout |
| 9 |
| 10 @c SMALL BOOK version |
| 11 @c This edition has been formatted so that you can format and print it in |
| 12 @c the smallbook format. |
| 13 @c @smallbook |
| 14 |
| 15 @c Set following if you want to document %default-prec and %no-default-prec. |
| 16 @c This feature is experimental and may change in future Bison versions. |
| 17 @c @set defaultprec |
| 18 |
| 19 @ifnotinfo |
| 20 @syncodeindex fn cp |
| 21 @syncodeindex vr cp |
| 22 @syncodeindex tp cp |
| 23 @end ifnotinfo |
| 24 @ifinfo |
| 25 @synindex fn cp |
| 26 @synindex vr cp |
| 27 @synindex tp cp |
| 28 @end ifinfo |
| 29 @comment %**end of header |
| 30 |
| 31 @copying |
| 32 |
| 33 This manual (@value{UPDATED}) is for @acronym{GNU} Bison (version |
| 34 @value{VERSION}), the @acronym{GNU} parser generator. |
| 35 |
| 36 Copyright @copyright{} 1988, 1989, 1990, 1991, 1992, 1993, 1995, 1998, |
| 37 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software |
| 38 Foundation, Inc. |
| 39 |
| 40 @quotation |
| 41 Permission is granted to copy, distribute and/or modify this document |
| 42 under the terms of the @acronym{GNU} Free Documentation License, |
| 43 Version 1.2 or any later version published by the Free Software |
| 44 Foundation; with no Invariant Sections, with the Front-Cover texts |
| 45 being ``A @acronym{GNU} Manual,'' and with the Back-Cover Texts as in |
| 46 (a) below. A copy of the license is included in the section entitled |
| 47 ``@acronym{GNU} Free Documentation License.'' |
| 48 |
| 49 (a) The FSF's Back-Cover Text is: ``You have the freedom to copy and |
| 50 modify this @acronym{GNU} manual. Buying copies from the @acronym{FSF} |
| 51 supports it in developing @acronym{GNU} and promoting software |
| 52 freedom.'' |
| 53 @end quotation |
| 54 @end copying |
| 55 |
| 56 @dircategory Software development |
| 57 @direntry |
| 58 * bison: (bison). @acronym{GNU} parser generator (Yacc replacement). |
| 59 @end direntry |
| 60 |
| 61 @titlepage |
| 62 @title Bison |
| 63 @subtitle The Yacc-compatible Parser Generator |
| 64 @subtitle @value{UPDATED}, Bison Version @value{VERSION} |
| 65 |
| 66 @author by Charles Donnelly and Richard Stallman |
| 67 |
| 68 @page |
| 69 @vskip 0pt plus 1filll |
| 70 @insertcopying |
| 71 @sp 2 |
| 72 Published by the Free Software Foundation @* |
| 73 51 Franklin Street, Fifth Floor @* |
| 74 Boston, MA 02110-1301 USA @* |
| 75 Printed copies are available from the Free Software Foundation.@* |
| 76 @acronym{ISBN} 1-882114-44-2 |
| 77 @sp 2 |
| 78 Cover art by Etienne Suvasa. |
| 79 @end titlepage |
| 80 |
| 81 @contents |
| 82 |
| 83 @ifnottex |
| 84 @node Top |
| 85 @top Bison |
| 86 @insertcopying |
| 87 @end ifnottex |
| 88 |
| 89 @menu |
| 90 * Introduction:: |
| 91 * Conditions:: |
| 92 * Copying:: The @acronym{GNU} General Public License says |
| 93 how you can copy and share Bison. |
| 94 |
| 95 Tutorial sections: |
| 96 * Concepts:: Basic concepts for understanding Bison. |
| 97 * Examples:: Three simple explained examples of using Bison. |
| 98 |
| 99 Reference sections: |
| 100 * Grammar File:: Writing Bison declarations and rules. |
| 101 * Interface:: C-language interface to the parser function @code{yypars
e}. |
| 102 * Algorithm:: How the Bison parser works at run-time. |
| 103 * Error Recovery:: Writing rules for error recovery. |
| 104 * Context Dependency:: What to do if your language syntax is too |
| 105 messy for Bison to handle straightforwardly. |
| 106 * Debugging:: Understanding or debugging Bison parsers. |
| 107 * Invocation:: How to run Bison (to produce the parser source file). |
| 108 * Other Languages:: Creating C++ and Java parsers. |
| 109 * FAQ:: Frequently Asked Questions |
| 110 * Table of Symbols:: All the keywords of the Bison language are explained. |
| 111 * Glossary:: Basic concepts are explained. |
| 112 * Copying This Manual:: License for copying this manual. |
| 113 * Index:: Cross-references to the text. |
| 114 |
| 115 @detailmenu |
| 116 --- The Detailed Node Listing --- |
| 117 |
| 118 The Concepts of Bison |
| 119 |
| 120 * Language and Grammar:: Languages and context-free grammars, |
| 121 as mathematical ideas. |
| 122 * Grammar in Bison:: How we represent grammars for Bison's sake. |
| 123 * Semantic Values:: Each token or syntactic grouping can have |
| 124 a semantic value (the value of an integer, |
| 125 the name of an identifier, etc.). |
| 126 * Semantic Actions:: Each rule can have an action containing C code. |
| 127 * GLR Parsers:: Writing parsers for general context-free languages. |
| 128 * Locations Overview:: Tracking Locations. |
| 129 * Bison Parser:: What are Bison's input and output, |
| 130 how is the output used? |
| 131 * Stages:: Stages in writing and running Bison grammars. |
| 132 * Grammar Layout:: Overall structure of a Bison grammar file. |
| 133 |
| 134 Writing @acronym{GLR} Parsers |
| 135 |
| 136 * Simple GLR Parsers:: Using @acronym{GLR} parsers on unambiguous grammars. |
| 137 * Merging GLR Parses:: Using @acronym{GLR} parsers to resolve ambiguities. |
| 138 * GLR Semantic Actions:: Deferred semantic actions have special concerns. |
| 139 * Compiler Requirements:: @acronym{GLR} parsers require a modern C compiler. |
| 140 |
| 141 Examples |
| 142 |
| 143 * RPN Calc:: Reverse polish notation calculator; |
| 144 a first example with no operator precedence. |
| 145 * Infix Calc:: Infix (algebraic) notation calculator. |
| 146 Operator precedence is introduced. |
| 147 * Simple Error Recovery:: Continuing after syntax errors. |
| 148 * Location Tracking Calc:: Demonstrating the use of @@@var{n} and @@$. |
| 149 * Multi-function Calc:: Calculator with memory and trig functions. |
| 150 It uses multiple data-types for semantic values. |
| 151 * Exercises:: Ideas for improving the multi-function calculator. |
| 152 |
| 153 Reverse Polish Notation Calculator |
| 154 |
| 155 * Rpcalc Declarations:: Prologue (declarations) for rpcalc. |
| 156 * Rpcalc Rules:: Grammar Rules for rpcalc, with explanation. |
| 157 * Rpcalc Lexer:: The lexical analyzer. |
| 158 * Rpcalc Main:: The controlling function. |
| 159 * Rpcalc Error:: The error reporting function. |
| 160 * Rpcalc Generate:: Running Bison on the grammar file. |
| 161 * Rpcalc Compile:: Run the C compiler on the output code. |
| 162 |
| 163 Grammar Rules for @code{rpcalc} |
| 164 |
| 165 * Rpcalc Input:: |
| 166 * Rpcalc Line:: |
| 167 * Rpcalc Expr:: |
| 168 |
| 169 Location Tracking Calculator: @code{ltcalc} |
| 170 |
| 171 * Ltcalc Declarations:: Bison and C declarations for ltcalc. |
| 172 * Ltcalc Rules:: Grammar rules for ltcalc, with explanations. |
| 173 * Ltcalc Lexer:: The lexical analyzer. |
| 174 |
| 175 Multi-Function Calculator: @code{mfcalc} |
| 176 |
| 177 * Mfcalc Declarations:: Bison declarations for multi-function calculator. |
| 178 * Mfcalc Rules:: Grammar rules for the calculator. |
| 179 * Mfcalc Symbol Table:: Symbol table management subroutines. |
| 180 |
| 181 Bison Grammar Files |
| 182 |
| 183 * Grammar Outline:: Overall layout of the grammar file. |
| 184 * Symbols:: Terminal and nonterminal symbols. |
| 185 * Rules:: How to write grammar rules. |
| 186 * Recursion:: Writing recursive rules. |
| 187 * Semantics:: Semantic values and actions. |
| 188 * Locations:: Locations and actions. |
| 189 * Declarations:: All kinds of Bison declarations are described here. |
| 190 * Multiple Parsers:: Putting more than one Bison parser in one program. |
| 191 |
| 192 Outline of a Bison Grammar |
| 193 |
| 194 * Prologue:: Syntax and usage of the prologue. |
| 195 * Prologue Alternatives:: Syntax and usage of alternatives to the prologue. |
| 196 * Bison Declarations:: Syntax and usage of the Bison declarations section. |
| 197 * Grammar Rules:: Syntax and usage of the grammar rules section. |
| 198 * Epilogue:: Syntax and usage of the epilogue. |
| 199 |
| 200 Defining Language Semantics |
| 201 |
| 202 * Value Type:: Specifying one data type for all semantic values. |
| 203 * Multiple Types:: Specifying several alternative data types. |
| 204 * Actions:: An action is the semantic definition of a grammar rule. |
| 205 * Action Types:: Specifying data types for actions to operate on. |
| 206 * Mid-Rule Actions:: Most actions go at the end of a rule. |
| 207 This says when, why and how to use the exceptional |
| 208 action in the middle of a rule. |
| 209 |
| 210 Tracking Locations |
| 211 |
| 212 * Location Type:: Specifying a data type for locations. |
| 213 * Actions and Locations:: Using locations in actions. |
| 214 * Location Default Action:: Defining a general way to compute locations. |
| 215 |
| 216 Bison Declarations |
| 217 |
| 218 * Require Decl:: Requiring a Bison version. |
| 219 * Token Decl:: Declaring terminal symbols. |
| 220 * Precedence Decl:: Declaring terminals with precedence and associativity. |
| 221 * Union Decl:: Declaring the set of all semantic value types. |
| 222 * Type Decl:: Declaring the choice of type for a nonterminal symbol. |
| 223 * Initial Action Decl:: Code run before parsing starts. |
| 224 * Destructor Decl:: Declaring how symbols are freed. |
| 225 * Expect Decl:: Suppressing warnings about parsing conflicts. |
| 226 * Start Decl:: Specifying the start symbol. |
| 227 * Pure Decl:: Requesting a reentrant parser. |
| 228 * Push Decl:: Requesting a push parser. |
| 229 * Decl Summary:: Table of all Bison declarations. |
| 230 |
| 231 Parser C-Language Interface |
| 232 |
| 233 * Parser Function:: How to call @code{yyparse} and what it returns. |
| 234 * Push Parser Function:: How to call @code{yypush_parse} and what it returns. |
| 235 * Pull Parser Function:: How to call @code{yypull_parse} and what it returns. |
| 236 * Parser Create Function:: How to call @code{yypstate_new} and what it returns. |
| 237 * Parser Delete Function:: How to call @code{yypstate_delete} and what it retur
ns. |
| 238 * Lexical:: You must supply a function @code{yylex} |
| 239 which reads tokens. |
| 240 * Error Reporting:: You must supply a function @code{yyerror}. |
| 241 * Action Features:: Special features for use in actions. |
| 242 * Internationalization:: How to let the parser speak in the user's |
| 243 native language. |
| 244 |
| 245 The Lexical Analyzer Function @code{yylex} |
| 246 |
| 247 * Calling Convention:: How @code{yyparse} calls @code{yylex}. |
| 248 * Token Values:: How @code{yylex} must return the semantic value |
| 249 of the token it has read. |
| 250 * Token Locations:: How @code{yylex} must return the text location |
| 251 (line number, etc.) of the token, if the |
| 252 actions want that. |
| 253 * Pure Calling:: How the calling convention differs in a pure parser |
| 254 (@pxref{Pure Decl, ,A Pure (Reentrant) Parser}). |
| 255 |
| 256 The Bison Parser Algorithm |
| 257 |
| 258 * Lookahead:: Parser looks one token ahead when deciding what to do. |
| 259 * Shift/Reduce:: Conflicts: when either shifting or reduction is valid. |
| 260 * Precedence:: Operator precedence works by resolving conflicts. |
| 261 * Contextual Precedence:: When an operator's precedence depends on context. |
| 262 * Parser States:: The parser is a finite-state-machine with stack. |
| 263 * Reduce/Reduce:: When two rules are applicable in the same situation. |
| 264 * Mystery Conflicts:: Reduce/reduce conflicts that look unjustified. |
| 265 * Generalized LR Parsing:: Parsing arbitrary context-free grammars. |
| 266 * Memory Management:: What happens when memory is exhausted. How to avoid it. |
| 267 |
| 268 Operator Precedence |
| 269 |
| 270 * Why Precedence:: An example showing why precedence is needed. |
| 271 * Using Precedence:: How to specify precedence in Bison grammars. |
| 272 * Precedence Examples:: How these features are used in the previous example. |
| 273 * How Precedence:: How they work. |
| 274 |
| 275 Handling Context Dependencies |
| 276 |
| 277 * Semantic Tokens:: Token parsing can depend on the semantic context. |
| 278 * Lexical Tie-ins:: Token parsing can depend on the syntactic context. |
| 279 * Tie-in Recovery:: Lexical tie-ins have implications for how |
| 280 error recovery rules must be written. |
| 281 |
| 282 Debugging Your Parser |
| 283 |
| 284 * Understanding:: Understanding the structure of your parser. |
| 285 * Tracing:: Tracing the execution of your parser. |
| 286 |
| 287 Invoking Bison |
| 288 |
| 289 * Bison Options:: All the options described in detail, |
| 290 in alphabetical order by short options. |
| 291 * Option Cross Key:: Alphabetical list of long options. |
| 292 * Yacc Library:: Yacc-compatible @code{yylex} and @code{main}. |
| 293 |
| 294 Parsers Written In Other Languages |
| 295 |
| 296 * C++ Parsers:: The interface to generate C++ parser classes |
| 297 * Java Parsers:: The interface to generate Java parser classes |
| 298 |
| 299 C++ Parsers |
| 300 |
| 301 * C++ Bison Interface:: Asking for C++ parser generation |
| 302 * C++ Semantic Values:: %union vs. C++ |
| 303 * C++ Location Values:: The position and location classes |
| 304 * C++ Parser Interface:: Instantiating and running the parser |
| 305 * C++ Scanner Interface:: Exchanges between yylex and parse |
| 306 * A Complete C++ Example:: Demonstrating their use |
| 307 |
| 308 A Complete C++ Example |
| 309 |
| 310 * Calc++ --- C++ Calculator:: The specifications |
| 311 * Calc++ Parsing Driver:: An active parsing context |
| 312 * Calc++ Parser:: A parser class |
| 313 * Calc++ Scanner:: A pure C++ Flex scanner |
| 314 * Calc++ Top Level:: Conducting the band |
| 315 |
| 316 Java Parsers |
| 317 |
| 318 * Java Bison Interface:: Asking for Java parser generation |
| 319 * Java Semantic Values:: %type and %token vs. Java |
| 320 * Java Location Values:: The position and location classes |
| 321 * Java Parser Interface:: Instantiating and running the parser |
| 322 * Java Scanner Interface:: Specifying the scanner for the parser |
| 323 * Java Action Features:: Special features for use in actions |
| 324 * Java Differences:: Differences between C/C++ and Java Grammars |
| 325 * Java Declarations Summary:: List of Bison declarations used with Java |
| 326 |
| 327 Frequently Asked Questions |
| 328 |
| 329 * Memory Exhausted:: Breaking the Stack Limits |
| 330 * How Can I Reset the Parser:: @code{yyparse} Keeps some State |
| 331 * Strings are Destroyed:: @code{yylval} Loses Track of Strings |
| 332 * Implementing Gotos/Loops:: Control Flow in the Calculator |
| 333 * Multiple start-symbols:: Factoring closely related grammars |
| 334 * Secure? Conform?:: Is Bison @acronym{POSIX} safe? |
| 335 * I can't build Bison:: Troubleshooting |
| 336 * Where can I find help?:: Troubleshouting |
| 337 * Bug Reports:: Troublereporting |
| 338 * More Languages:: Parsers in C++, Java, and so on |
| 339 * Beta Testing:: Experimenting development versions |
| 340 * Mailing Lists:: Meeting other Bison users |
| 341 |
| 342 Copying This Manual |
| 343 |
| 344 * Copying This Manual:: License for copying this manual. |
| 345 |
| 346 @end detailmenu |
| 347 @end menu |
| 348 |
| 349 @node Introduction |
| 350 @unnumbered Introduction |
| 351 @cindex introduction |
| 352 |
| 353 @dfn{Bison} is a general-purpose parser generator that converts an |
| 354 annotated context-free grammar into an @acronym{LALR}(1) or |
| 355 @acronym{GLR} parser for that grammar. Once you are proficient with |
| 356 Bison, you can use it to develop a wide range of language parsers, from those |
| 357 used in simple desk calculators to complex programming languages. |
| 358 |
| 359 Bison is upward compatible with Yacc: all properly-written Yacc grammars |
| 360 ought to work with Bison with no change. Anyone familiar with Yacc |
| 361 should be able to use Bison with little trouble. You need to be fluent in |
| 362 C or C++ programming in order to use Bison or to understand this manual. |
| 363 |
| 364 We begin with tutorial chapters that explain the basic concepts of using |
| 365 Bison and show three explained examples, each building on the last. If you |
| 366 don't know Bison or Yacc, start by reading these chapters. Reference |
| 367 chapters follow which describe specific aspects of Bison in detail. |
| 368 |
| 369 Bison was written primarily by Robert Corbett; Richard Stallman made it |
| 370 Yacc-compatible. Wilfred Hansen of Carnegie Mellon University added |
| 371 multi-character string literals and other features. |
| 372 |
| 373 This edition corresponds to version @value{VERSION} of Bison. |
| 374 |
| 375 @node Conditions |
| 376 @unnumbered Conditions for Using Bison |
| 377 |
| 378 The distribution terms for Bison-generated parsers permit using the |
| 379 parsers in nonfree programs. Before Bison version 2.2, these extra |
| 380 permissions applied only when Bison was generating @acronym{LALR}(1) |
| 381 parsers in C@. And before Bison version 1.24, Bison-generated |
| 382 parsers could be used only in programs that were free software. |
| 383 |
| 384 The other @acronym{GNU} programming tools, such as the @acronym{GNU} C |
| 385 compiler, have never |
| 386 had such a requirement. They could always be used for nonfree |
| 387 software. The reason Bison was different was not due to a special |
| 388 policy decision; it resulted from applying the usual General Public |
| 389 License to all of the Bison source code. |
| 390 |
| 391 The output of the Bison utility---the Bison parser file---contains a |
| 392 verbatim copy of a sizable piece of Bison, which is the code for the |
| 393 parser's implementation. (The actions from your grammar are inserted |
| 394 into this implementation at one point, but most of the rest of the |
| 395 implementation is not changed.) When we applied the @acronym{GPL} |
| 396 terms to the skeleton code for the parser's implementation, |
| 397 the effect was to restrict the use of Bison output to free software. |
| 398 |
| 399 We didn't change the terms because of sympathy for people who want to |
| 400 make software proprietary. @strong{Software should be free.} But we |
| 401 concluded that limiting Bison's use to free software was doing little to |
| 402 encourage people to make other software free. So we decided to make the |
| 403 practical conditions for using Bison match the practical conditions for |
| 404 using the other @acronym{GNU} tools. |
| 405 |
| 406 This exception applies when Bison is generating code for a parser. |
| 407 You can tell whether the exception applies to a Bison output file by |
| 408 inspecting the file for text beginning with ``As a special |
| 409 exception@dots{}''. The text spells out the exact terms of the |
| 410 exception. |
| 411 |
| 412 @node Copying |
| 413 @unnumbered GNU GENERAL PUBLIC LICENSE |
| 414 @include gpl-3.0.texi |
| 415 |
| 416 @node Concepts |
| 417 @chapter The Concepts of Bison |
| 418 |
| 419 This chapter introduces many of the basic concepts without which the |
| 420 details of Bison will not make sense. If you do not already know how to |
| 421 use Bison or Yacc, we suggest you start by reading this chapter carefully. |
| 422 |
| 423 @menu |
| 424 * Language and Grammar:: Languages and context-free grammars, |
| 425 as mathematical ideas. |
| 426 * Grammar in Bison:: How we represent grammars for Bison's sake. |
| 427 * Semantic Values:: Each token or syntactic grouping can have |
| 428 a semantic value (the value of an integer, |
| 429 the name of an identifier, etc.). |
| 430 * Semantic Actions:: Each rule can have an action containing C code. |
| 431 * GLR Parsers:: Writing parsers for general context-free languages. |
| 432 * Locations Overview:: Tracking Locations. |
| 433 * Bison Parser:: What are Bison's input and output, |
| 434 how is the output used? |
| 435 * Stages:: Stages in writing and running Bison grammars. |
| 436 * Grammar Layout:: Overall structure of a Bison grammar file. |
| 437 @end menu |
| 438 |
| 439 @node Language and Grammar |
| 440 @section Languages and Context-Free Grammars |
| 441 |
| 442 @cindex context-free grammar |
| 443 @cindex grammar, context-free |
| 444 In order for Bison to parse a language, it must be described by a |
| 445 @dfn{context-free grammar}. This means that you specify one or more |
| 446 @dfn{syntactic groupings} and give rules for constructing them from their |
| 447 parts. For example, in the C language, one kind of grouping is called an |
| 448 `expression'. One rule for making an expression might be, ``An expression |
| 449 can be made of a minus sign and another expression''. Another would be, |
| 450 ``An expression can be an integer''. As you can see, rules are often |
| 451 recursive, but there must be at least one rule which leads out of the |
| 452 recursion. |
| 453 |
| 454 @cindex @acronym{BNF} |
| 455 @cindex Backus-Naur form |
| 456 The most common formal system for presenting such rules for humans to read |
| 457 is @dfn{Backus-Naur Form} or ``@acronym{BNF}'', which was developed in |
| 458 order to specify the language Algol 60. Any grammar expressed in |
| 459 @acronym{BNF} is a context-free grammar. The input to Bison is |
| 460 essentially machine-readable @acronym{BNF}. |
| 461 |
| 462 @cindex @acronym{LALR}(1) grammars |
| 463 @cindex @acronym{LR}(1) grammars |
| 464 There are various important subclasses of context-free grammar. Although it |
| 465 can handle almost all context-free grammars, Bison is optimized for what |
| 466 are called @acronym{LALR}(1) grammars. |
| 467 In brief, in these grammars, it must be possible to |
| 468 tell how to parse any portion of an input string with just a single |
| 469 token of lookahead. Strictly speaking, that is a description of an |
| 470 @acronym{LR}(1) grammar, and @acronym{LALR}(1) involves additional |
| 471 restrictions that are |
| 472 hard to explain simply; but it is rare in actual practice to find an |
| 473 @acronym{LR}(1) grammar that fails to be @acronym{LALR}(1). |
| 474 @xref{Mystery Conflicts, ,Mysterious Reduce/Reduce Conflicts}, for |
| 475 more information on this. |
| 476 |
| 477 @cindex @acronym{GLR} parsing |
| 478 @cindex generalized @acronym{LR} (@acronym{GLR}) parsing |
| 479 @cindex ambiguous grammars |
| 480 @cindex nondeterministic parsing |
| 481 |
| 482 Parsers for @acronym{LALR}(1) grammars are @dfn{deterministic}, meaning |
| 483 roughly that the next grammar rule to apply at any point in the input is |
| 484 uniquely determined by the preceding input and a fixed, finite portion |
| 485 (called a @dfn{lookahead}) of the remaining input. A context-free |
| 486 grammar can be @dfn{ambiguous}, meaning that there are multiple ways to |
| 487 apply the grammar rules to get the same inputs. Even unambiguous |
| 488 grammars can be @dfn{nondeterministic}, meaning that no fixed |
| 489 lookahead always suffices to determine the next grammar rule to apply. |
| 490 With the proper declarations, Bison is also able to parse these more |
| 491 general context-free grammars, using a technique known as @acronym{GLR} |
| 492 parsing (for Generalized @acronym{LR}). Bison's @acronym{GLR} parsers |
| 493 are able to handle any context-free grammar for which the number of |
| 494 possible parses of any given string is finite. |
| 495 |
| 496 @cindex symbols (abstract) |
| 497 @cindex token |
| 498 @cindex syntactic grouping |
| 499 @cindex grouping, syntactic |
| 500 In the formal grammatical rules for a language, each kind of syntactic |
| 501 unit or grouping is named by a @dfn{symbol}. Those which are built by |
| 502 grouping smaller constructs according to grammatical rules are called |
| 503 @dfn{nonterminal symbols}; those which can't be subdivided are called |
| 504 @dfn{terminal symbols} or @dfn{token types}. We call a piece of input |
| 505 corresponding to a single terminal symbol a @dfn{token}, and a piece |
| 506 corresponding to a single nonterminal symbol a @dfn{grouping}. |
| 507 |
| 508 We can use the C language as an example of what symbols, terminal and |
| 509 nonterminal, mean. The tokens of C are identifiers, constants (numeric |
| 510 and string), and the various keywords, arithmetic operators and |
| 511 punctuation marks. So the terminal symbols of a grammar for C include |
| 512 `identifier', `number', `string', plus one symbol for each keyword, |
| 513 operator or punctuation mark: `if', `return', `const', `static', `int', |
| 514 `char', `plus-sign', `open-brace', `close-brace', `comma' and many more. |
| 515 (These tokens can be subdivided into characters, but that is a matter of |
| 516 lexicography, not grammar.) |
| 517 |
| 518 Here is a simple C function subdivided into tokens: |
| 519 |
| 520 @ifinfo |
| 521 @example |
| 522 int /* @r{keyword `int'} */ |
| 523 square (int x) /* @r{identifier, open-paren, keyword `int',} |
| 524 @r{identifier, close-paren} */ |
| 525 @{ /* @r{open-brace} */ |
| 526 return x * x; /* @r{keyword `return', identifier, asterisk,} |
| 527 @r{identifier, semicolon} */ |
| 528 @} /* @r{close-brace} */ |
| 529 @end example |
| 530 @end ifinfo |
| 531 @ifnotinfo |
| 532 @example |
| 533 int /* @r{keyword `int'} */ |
| 534 square (int x) /* @r{identifier, open-paren, keyword `int', identifier, close-p
aren} */ |
| 535 @{ /* @r{open-brace} */ |
| 536 return x * x; /* @r{keyword `return', identifier, asterisk, identifier, semico
lon} */ |
| 537 @} /* @r{close-brace} */ |
| 538 @end example |
| 539 @end ifnotinfo |
| 540 |
| 541 The syntactic groupings of C include the expression, the statement, the |
| 542 declaration, and the function definition. These are represented in the |
| 543 grammar of C by nonterminal symbols `expression', `statement', |
| 544 `declaration' and `function definition'. The full grammar uses dozens of |
| 545 additional language constructs, each with its own nonterminal symbol, in |
| 546 order to express the meanings of these four. The example above is a |
| 547 function definition; it contains one declaration, and one statement. In |
| 548 the statement, each @samp{x} is an expression and so is @samp{x * x}. |
| 549 |
| 550 Each nonterminal symbol must have grammatical rules showing how it is made |
| 551 out of simpler constructs. For example, one kind of C statement is the |
| 552 @code{return} statement; this would be described with a grammar rule which |
| 553 reads informally as follows: |
| 554 |
| 555 @quotation |
| 556 A `statement' can be made of a `return' keyword, an `expression' and a |
| 557 `semicolon'. |
| 558 @end quotation |
| 559 |
| 560 @noindent |
| 561 There would be many other rules for `statement', one for each kind of |
| 562 statement in C. |
| 563 |
| 564 @cindex start symbol |
| 565 One nonterminal symbol must be distinguished as the special one which |
| 566 defines a complete utterance in the language. It is called the @dfn{start |
| 567 symbol}. In a compiler, this means a complete input program. In the C |
| 568 language, the nonterminal symbol `sequence of definitions and declarations' |
| 569 plays this role. |
| 570 |
| 571 For example, @samp{1 + 2} is a valid C expression---a valid part of a C |
| 572 program---but it is not valid as an @emph{entire} C program. In the |
| 573 context-free grammar of C, this follows from the fact that `expression' is |
| 574 not the start symbol. |
| 575 |
| 576 The Bison parser reads a sequence of tokens as its input, and groups the |
| 577 tokens using the grammar rules. If the input is valid, the end result is |
| 578 that the entire token sequence reduces to a single grouping whose symbol is |
| 579 the grammar's start symbol. If we use a grammar for C, the entire input |
| 580 must be a `sequence of definitions and declarations'. If not, the parser |
| 581 reports a syntax error. |
| 582 |
| 583 @node Grammar in Bison |
| 584 @section From Formal Rules to Bison Input |
| 585 @cindex Bison grammar |
| 586 @cindex grammar, Bison |
| 587 @cindex formal grammar |
| 588 |
| 589 A formal grammar is a mathematical construct. To define the language |
| 590 for Bison, you must write a file expressing the grammar in Bison syntax: |
| 591 a @dfn{Bison grammar} file. @xref{Grammar File, ,Bison Grammar Files}. |
| 592 |
| 593 A nonterminal symbol in the formal grammar is represented in Bison input |
| 594 as an identifier, like an identifier in C@. By convention, it should be |
| 595 in lower case, such as @code{expr}, @code{stmt} or @code{declaration}. |
| 596 |
| 597 The Bison representation for a terminal symbol is also called a @dfn{token |
| 598 type}. Token types as well can be represented as C-like identifiers. By |
| 599 convention, these identifiers should be upper case to distinguish them from |
| 600 nonterminals: for example, @code{INTEGER}, @code{IDENTIFIER}, @code{IF} or |
| 601 @code{RETURN}. A terminal symbol that stands for a particular keyword in |
| 602 the language should be named after that keyword converted to upper case. |
| 603 The terminal symbol @code{error} is reserved for error recovery. |
| 604 @xref{Symbols}. |
| 605 |
| 606 A terminal symbol can also be represented as a character literal, just like |
| 607 a C character constant. You should do this whenever a token is just a |
| 608 single character (parenthesis, plus-sign, etc.): use that same character in |
| 609 a literal as the terminal symbol for that token. |
| 610 |
| 611 A third way to represent a terminal symbol is with a C string constant |
| 612 containing several characters. @xref{Symbols}, for more information. |
| 613 |
| 614 The grammar rules also have an expression in Bison syntax. For example, |
| 615 here is the Bison rule for a C @code{return} statement. The semicolon in |
| 616 quotes is a literal character token, representing part of the C syntax for |
| 617 the statement; the naked semicolon, and the colon, are Bison punctuation |
| 618 used in every rule. |
| 619 |
| 620 @example |
| 621 stmt: RETURN expr ';' |
| 622 ; |
| 623 @end example |
| 624 |
| 625 @noindent |
| 626 @xref{Rules, ,Syntax of Grammar Rules}. |
| 627 |
| 628 @node Semantic Values |
| 629 @section Semantic Values |
| 630 @cindex semantic value |
| 631 @cindex value, semantic |
| 632 |
| 633 A formal grammar selects tokens only by their classifications: for example, |
| 634 if a rule mentions the terminal symbol `integer constant', it means that |
| 635 @emph{any} integer constant is grammatically valid in that position. The |
| 636 precise value of the constant is irrelevant to how to parse the input: if |
| 637 @samp{x+4} is grammatical then @samp{x+1} or @samp{x+3989} is equally |
| 638 grammatical. |
| 639 |
| 640 But the precise value is very important for what the input means once it is |
| 641 parsed. A compiler is useless if it fails to distinguish between 4, 1 and |
| 642 3989 as constants in the program! Therefore, each token in a Bison grammar |
| 643 has both a token type and a @dfn{semantic value}. @xref{Semantics, |
| 644 ,Defining Language Semantics}, |
| 645 for details. |
| 646 |
| 647 The token type is a terminal symbol defined in the grammar, such as |
| 648 @code{INTEGER}, @code{IDENTIFIER} or @code{','}. It tells everything |
| 649 you need to know to decide where the token may validly appear and how to |
| 650 group it with other tokens. The grammar rules know nothing about tokens |
| 651 except their types. |
| 652 |
| 653 The semantic value has all the rest of the information about the |
| 654 meaning of the token, such as the value of an integer, or the name of an |
| 655 identifier. (A token such as @code{','} which is just punctuation doesn't |
| 656 need to have any semantic value.) |
| 657 |
| 658 For example, an input token might be classified as token type |
| 659 @code{INTEGER} and have the semantic value 4. Another input token might |
| 660 have the same token type @code{INTEGER} but value 3989. When a grammar |
| 661 rule says that @code{INTEGER} is allowed, either of these tokens is |
| 662 acceptable because each is an @code{INTEGER}. When the parser accepts the |
| 663 token, it keeps track of the token's semantic value. |
| 664 |
| 665 Each grouping can also have a semantic value as well as its nonterminal |
| 666 symbol. For example, in a calculator, an expression typically has a |
| 667 semantic value that is a number. In a compiler for a programming |
| 668 language, an expression typically has a semantic value that is a tree |
| 669 structure describing the meaning of the expression. |
| 670 |
| 671 @node Semantic Actions |
| 672 @section Semantic Actions |
| 673 @cindex semantic actions |
| 674 @cindex actions, semantic |
| 675 |
| 676 In order to be useful, a program must do more than parse input; it must |
| 677 also produce some output based on the input. In a Bison grammar, a grammar |
| 678 rule can have an @dfn{action} made up of C statements. Each time the |
| 679 parser recognizes a match for that rule, the action is executed. |
| 680 @xref{Actions}. |
| 681 |
| 682 Most of the time, the purpose of an action is to compute the semantic value |
| 683 of the whole construct from the semantic values of its parts. For example, |
| 684 suppose we have a rule which says an expression can be the sum of two |
| 685 expressions. When the parser recognizes such a sum, each of the |
| 686 subexpressions has a semantic value which describes how it was built up. |
| 687 The action for this rule should create a similar sort of value for the |
| 688 newly recognized larger expression. |
| 689 |
| 690 For example, here is a rule that says an expression can be the sum of |
| 691 two subexpressions: |
| 692 |
| 693 @example |
| 694 expr: expr '+' expr @{ $$ = $1 + $3; @} |
| 695 ; |
| 696 @end example |
| 697 |
| 698 @noindent |
| 699 The action says how to produce the semantic value of the sum expression |
| 700 from the values of the two subexpressions. |
| 701 |
| 702 @node GLR Parsers |
| 703 @section Writing @acronym{GLR} Parsers |
| 704 @cindex @acronym{GLR} parsing |
| 705 @cindex generalized @acronym{LR} (@acronym{GLR}) parsing |
| 706 @findex %glr-parser |
| 707 @cindex conflicts |
| 708 @cindex shift/reduce conflicts |
| 709 @cindex reduce/reduce conflicts |
| 710 |
| 711 In some grammars, Bison's standard |
| 712 @acronym{LALR}(1) parsing algorithm cannot decide whether to apply a |
| 713 certain grammar rule at a given point. That is, it may not be able to |
| 714 decide (on the basis of the input read so far) which of two possible |
| 715 reductions (applications of a grammar rule) applies, or whether to apply |
| 716 a reduction or read more of the input and apply a reduction later in the |
| 717 input. These are known respectively as @dfn{reduce/reduce} conflicts |
| 718 (@pxref{Reduce/Reduce}), and @dfn{shift/reduce} conflicts |
| 719 (@pxref{Shift/Reduce}). |
| 720 |
| 721 To use a grammar that is not easily modified to be @acronym{LALR}(1), a |
| 722 more general parsing algorithm is sometimes necessary. If you include |
| 723 @code{%glr-parser} among the Bison declarations in your file |
| 724 (@pxref{Grammar Outline}), the result is a Generalized @acronym{LR} |
| 725 (@acronym{GLR}) parser. These parsers handle Bison grammars that |
| 726 contain no unresolved conflicts (i.e., after applying precedence |
| 727 declarations) identically to @acronym{LALR}(1) parsers. However, when |
| 728 faced with unresolved shift/reduce and reduce/reduce conflicts, |
| 729 @acronym{GLR} parsers use the simple expedient of doing both, |
| 730 effectively cloning the parser to follow both possibilities. Each of |
| 731 the resulting parsers can again split, so that at any given time, there |
| 732 can be any number of possible parses being explored. The parsers |
| 733 proceed in lockstep; that is, all of them consume (shift) a given input |
| 734 symbol before any of them proceed to the next. Each of the cloned |
| 735 parsers eventually meets one of two possible fates: either it runs into |
| 736 a parsing error, in which case it simply vanishes, or it merges with |
| 737 another parser, because the two of them have reduced the input to an |
| 738 identical set of symbols. |
| 739 |
| 740 During the time that there are multiple parsers, semantic actions are |
| 741 recorded, but not performed. When a parser disappears, its recorded |
| 742 semantic actions disappear as well, and are never performed. When a |
| 743 reduction makes two parsers identical, causing them to merge, Bison |
| 744 records both sets of semantic actions. Whenever the last two parsers |
| 745 merge, reverting to the single-parser case, Bison resolves all the |
| 746 outstanding actions either by precedences given to the grammar rules |
| 747 involved, or by performing both actions, and then calling a designated |
| 748 user-defined function on the resulting values to produce an arbitrary |
| 749 merged result. |
| 750 |
| 751 @menu |
| 752 * Simple GLR Parsers:: Using @acronym{GLR} parsers on unambiguous grammars. |
| 753 * Merging GLR Parses:: Using @acronym{GLR} parsers to resolve ambiguities. |
| 754 * GLR Semantic Actions:: Deferred semantic actions have special concerns. |
| 755 * Compiler Requirements:: @acronym{GLR} parsers require a modern C compiler. |
| 756 @end menu |
| 757 |
| 758 @node Simple GLR Parsers |
| 759 @subsection Using @acronym{GLR} on Unambiguous Grammars |
| 760 @cindex @acronym{GLR} parsing, unambiguous grammars |
| 761 @cindex generalized @acronym{LR} (@acronym{GLR}) parsing, unambiguous grammars |
| 762 @findex %glr-parser |
| 763 @findex %expect-rr |
| 764 @cindex conflicts |
| 765 @cindex reduce/reduce conflicts |
| 766 @cindex shift/reduce conflicts |
| 767 |
| 768 In the simplest cases, you can use the @acronym{GLR} algorithm |
| 769 to parse grammars that are unambiguous, but fail to be @acronym{LALR}(1). |
| 770 Such grammars typically require more than one symbol of lookahead, |
| 771 or (in rare cases) fall into the category of grammars in which the |
| 772 @acronym{LALR}(1) algorithm throws away too much information (they are in |
| 773 @acronym{LR}(1), but not @acronym{LALR}(1), @ref{Mystery Conflicts}). |
| 774 |
| 775 Consider a problem that |
| 776 arises in the declaration of enumerated and subrange types in the |
| 777 programming language Pascal. Here are some examples: |
| 778 |
| 779 @example |
| 780 type subrange = lo .. hi; |
| 781 type enum = (a, b, c); |
| 782 @end example |
| 783 |
| 784 @noindent |
| 785 The original language standard allows only numeric |
| 786 literals and constant identifiers for the subrange bounds (@samp{lo} |
| 787 and @samp{hi}), but Extended Pascal (@acronym{ISO}/@acronym{IEC} |
| 788 10206) and many other |
| 789 Pascal implementations allow arbitrary expressions there. This gives |
| 790 rise to the following situation, containing a superfluous pair of |
| 791 parentheses: |
| 792 |
| 793 @example |
| 794 type subrange = (a) .. b; |
| 795 @end example |
| 796 |
| 797 @noindent |
| 798 Compare this to the following declaration of an enumerated |
| 799 type with only one value: |
| 800 |
| 801 @example |
| 802 type enum = (a); |
| 803 @end example |
| 804 |
| 805 @noindent |
| 806 (These declarations are contrived, but they are syntactically |
| 807 valid, and more-complicated cases can come up in practical programs.) |
| 808 |
| 809 These two declarations look identical until the @samp{..} token. |
| 810 With normal @acronym{LALR}(1) one-token lookahead it is not |
| 811 possible to decide between the two forms when the identifier |
| 812 @samp{a} is parsed. It is, however, desirable |
| 813 for a parser to decide this, since in the latter case |
| 814 @samp{a} must become a new identifier to represent the enumeration |
| 815 value, while in the former case @samp{a} must be evaluated with its |
| 816 current meaning, which may be a constant or even a function call. |
| 817 |
| 818 You could parse @samp{(a)} as an ``unspecified identifier in parentheses'', |
| 819 to be resolved later, but this typically requires substantial |
| 820 contortions in both semantic actions and large parts of the |
| 821 grammar, where the parentheses are nested in the recursive rules for |
| 822 expressions. |
| 823 |
| 824 You might think of using the lexer to distinguish between the two |
| 825 forms by returning different tokens for currently defined and |
| 826 undefined identifiers. But if these declarations occur in a local |
| 827 scope, and @samp{a} is defined in an outer scope, then both forms |
| 828 are possible---either locally redefining @samp{a}, or using the |
| 829 value of @samp{a} from the outer scope. So this approach cannot |
| 830 work. |
| 831 |
| 832 A simple solution to this problem is to declare the parser to |
| 833 use the @acronym{GLR} algorithm. |
| 834 When the @acronym{GLR} parser reaches the critical state, it |
| 835 merely splits into two branches and pursues both syntax rules |
| 836 simultaneously. Sooner or later, one of them runs into a parsing |
| 837 error. If there is a @samp{..} token before the next |
| 838 @samp{;}, the rule for enumerated types fails since it cannot |
| 839 accept @samp{..} anywhere; otherwise, the subrange type rule |
| 840 fails since it requires a @samp{..} token. So one of the branches |
| 841 fails silently, and the other one continues normally, performing |
| 842 all the intermediate actions that were postponed during the split. |
| 843 |
| 844 If the input is syntactically incorrect, both branches fail and the parser |
| 845 reports a syntax error as usual. |
| 846 |
| 847 The effect of all this is that the parser seems to ``guess'' the |
| 848 correct branch to take, or in other words, it seems to use more |
| 849 lookahead than the underlying @acronym{LALR}(1) algorithm actually allows |
| 850 for. In this example, @acronym{LALR}(2) would suffice, but also some cases |
| 851 that are not @acronym{LALR}(@math{k}) for any @math{k} can be handled this way. |
| 852 |
| 853 In general, a @acronym{GLR} parser can take quadratic or cubic worst-case time, |
| 854 and the current Bison parser even takes exponential time and space |
| 855 for some grammars. In practice, this rarely happens, and for many |
| 856 grammars it is possible to prove that it cannot happen. |
| 857 The present example contains only one conflict between two |
| 858 rules, and the type-declaration context containing the conflict |
| 859 cannot be nested. So the number of |
| 860 branches that can exist at any time is limited by the constant 2, |
| 861 and the parsing time is still linear. |
| 862 |
| 863 Here is a Bison grammar corresponding to the example above. It |
| 864 parses a vastly simplified form of Pascal type declarations. |
| 865 |
| 866 @example |
| 867 %token TYPE DOTDOT ID |
| 868 |
| 869 @group |
| 870 %left '+' '-' |
| 871 %left '*' '/' |
| 872 @end group |
| 873 |
| 874 %% |
| 875 |
| 876 @group |
| 877 type_decl : TYPE ID '=' type ';' |
| 878 ; |
| 879 @end group |
| 880 |
| 881 @group |
| 882 type : '(' id_list ')' |
| 883 | expr DOTDOT expr |
| 884 ; |
| 885 @end group |
| 886 |
| 887 @group |
| 888 id_list : ID |
| 889 | id_list ',' ID |
| 890 ; |
| 891 @end group |
| 892 |
| 893 @group |
| 894 expr : '(' expr ')' |
| 895 | expr '+' expr |
| 896 | expr '-' expr |
| 897 | expr '*' expr |
| 898 | expr '/' expr |
| 899 | ID |
| 900 ; |
| 901 @end group |
| 902 @end example |
| 903 |
| 904 When used as a normal @acronym{LALR}(1) grammar, Bison correctly complains |
| 905 about one reduce/reduce conflict. In the conflicting situation the |
| 906 parser chooses one of the alternatives, arbitrarily the one |
| 907 declared first. Therefore the following correct input is not |
| 908 recognized: |
| 909 |
| 910 @example |
| 911 type t = (a) .. b; |
| 912 @end example |
| 913 |
| 914 The parser can be turned into a @acronym{GLR} parser, while also telling Bison |
| 915 to be silent about the one known reduce/reduce conflict, by |
| 916 adding these two declarations to the Bison input file (before the first |
| 917 @samp{%%}): |
| 918 |
| 919 @example |
| 920 %glr-parser |
| 921 %expect-rr 1 |
| 922 @end example |
| 923 |
| 924 @noindent |
| 925 No change in the grammar itself is required. Now the |
| 926 parser recognizes all valid declarations, according to the |
| 927 limited syntax above, transparently. In fact, the user does not even |
| 928 notice when the parser splits. |
| 929 |
| 930 So here we have a case where we can use the benefits of @acronym{GLR}, |
| 931 almost without disadvantages. Even in simple cases like this, however, |
| 932 there are at least two potential problems to beware. First, always |
| 933 analyze the conflicts reported by Bison to make sure that @acronym{GLR} |
| 934 splitting is only done where it is intended. A @acronym{GLR} parser |
| 935 splitting inadvertently may cause problems less obvious than an |
| 936 @acronym{LALR} parser statically choosing the wrong alternative in a |
| 937 conflict. Second, consider interactions with the lexer (@pxref{Semantic |
| 938 Tokens}) with great care. Since a split parser consumes tokens without |
| 939 performing any actions during the split, the lexer cannot obtain |
| 940 information via parser actions. Some cases of lexer interactions can be |
| 941 eliminated by using @acronym{GLR} to shift the complications from the |
| 942 lexer to the parser. You must check the remaining cases for |
| 943 correctness. |
| 944 |
| 945 In our example, it would be safe for the lexer to return tokens based on |
| 946 their current meanings in some symbol table, because no new symbols are |
| 947 defined in the middle of a type declaration. Though it is possible for |
| 948 a parser to define the enumeration constants as they are parsed, before |
| 949 the type declaration is completed, it actually makes no difference since |
| 950 they cannot be used within the same enumerated type declaration. |
| 951 |
| 952 @node Merging GLR Parses |
| 953 @subsection Using @acronym{GLR} to Resolve Ambiguities |
| 954 @cindex @acronym{GLR} parsing, ambiguous grammars |
| 955 @cindex generalized @acronym{LR} (@acronym{GLR}) parsing, ambiguous grammars |
| 956 @findex %dprec |
| 957 @findex %merge |
| 958 @cindex conflicts |
| 959 @cindex reduce/reduce conflicts |
| 960 |
| 961 Let's consider an example, vastly simplified from a C++ grammar. |
| 962 |
| 963 @example |
| 964 %@{ |
| 965 #include <stdio.h> |
| 966 #define YYSTYPE char const * |
| 967 int yylex (void); |
| 968 void yyerror (char const *); |
| 969 %@} |
| 970 |
| 971 %token TYPENAME ID |
| 972 |
| 973 %right '=' |
| 974 %left '+' |
| 975 |
| 976 %glr-parser |
| 977 |
| 978 %% |
| 979 |
| 980 prog : |
| 981 | prog stmt @{ printf ("\n"); @} |
| 982 ; |
| 983 |
| 984 stmt : expr ';' %dprec 1 |
| 985 | decl %dprec 2 |
| 986 ; |
| 987 |
| 988 expr : ID @{ printf ("%s ", $$); @} |
| 989 | TYPENAME '(' expr ')' |
| 990 @{ printf ("%s <cast> ", $1); @} |
| 991 | expr '+' expr @{ printf ("+ "); @} |
| 992 | expr '=' expr @{ printf ("= "); @} |
| 993 ; |
| 994 |
| 995 decl : TYPENAME declarator ';' |
| 996 @{ printf ("%s <declare> ", $1); @} |
| 997 | TYPENAME declarator '=' expr ';' |
| 998 @{ printf ("%s <init-declare> ", $1); @} |
| 999 ; |
| 1000 |
| 1001 declarator : ID @{ printf ("\"%s\" ", $1); @} |
| 1002 | '(' declarator ')' |
| 1003 ; |
| 1004 @end example |
| 1005 |
| 1006 @noindent |
| 1007 This models a problematic part of the C++ grammar---the ambiguity between |
| 1008 certain declarations and statements. For example, |
| 1009 |
| 1010 @example |
| 1011 T (x) = y+z; |
| 1012 @end example |
| 1013 |
| 1014 @noindent |
| 1015 parses as either an @code{expr} or a @code{stmt} |
| 1016 (assuming that @samp{T} is recognized as a @code{TYPENAME} and |
| 1017 @samp{x} as an @code{ID}). |
| 1018 Bison detects this as a reduce/reduce conflict between the rules |
| 1019 @code{expr : ID} and @code{declarator : ID}, which it cannot resolve at the |
| 1020 time it encounters @code{x} in the example above. Since this is a |
| 1021 @acronym{GLR} parser, it therefore splits the problem into two parses, one for |
| 1022 each choice of resolving the reduce/reduce conflict. |
| 1023 Unlike the example from the previous section (@pxref{Simple GLR Parsers}), |
| 1024 however, neither of these parses ``dies,'' because the grammar as it stands is |
| 1025 ambiguous. One of the parsers eventually reduces @code{stmt : expr ';'} and |
| 1026 the other reduces @code{stmt : decl}, after which both parsers are in an |
| 1027 identical state: they've seen @samp{prog stmt} and have the same unprocessed |
| 1028 input remaining. We say that these parses have @dfn{merged.} |
| 1029 |
| 1030 At this point, the @acronym{GLR} parser requires a specification in the |
| 1031 grammar of how to choose between the competing parses. |
| 1032 In the example above, the two @code{%dprec} |
| 1033 declarations specify that Bison is to give precedence |
| 1034 to the parse that interprets the example as a |
| 1035 @code{decl}, which implies that @code{x} is a declarator. |
| 1036 The parser therefore prints |
| 1037 |
| 1038 @example |
| 1039 "x" y z + T <init-declare> |
| 1040 @end example |
| 1041 |
| 1042 The @code{%dprec} declarations only come into play when more than one |
| 1043 parse survives. Consider a different input string for this parser: |
| 1044 |
| 1045 @example |
| 1046 T (x) + y; |
| 1047 @end example |
| 1048 |
| 1049 @noindent |
| 1050 This is another example of using @acronym{GLR} to parse an unambiguous |
| 1051 construct, as shown in the previous section (@pxref{Simple GLR Parsers}). |
| 1052 Here, there is no ambiguity (this cannot be parsed as a declaration). |
| 1053 However, at the time the Bison parser encounters @code{x}, it does not |
| 1054 have enough information to resolve the reduce/reduce conflict (again, |
| 1055 between @code{x} as an @code{expr} or a @code{declarator}). In this |
| 1056 case, no precedence declaration is used. Again, the parser splits |
| 1057 into two, one assuming that @code{x} is an @code{expr}, and the other |
| 1058 assuming @code{x} is a @code{declarator}. The second of these parsers |
| 1059 then vanishes when it sees @code{+}, and the parser prints |
| 1060 |
| 1061 @example |
| 1062 x T <cast> y + |
| 1063 @end example |
| 1064 |
| 1065 Suppose that instead of resolving the ambiguity, you wanted to see all |
| 1066 the possibilities. For this purpose, you must merge the semantic |
| 1067 actions of the two possible parsers, rather than choosing one over the |
| 1068 other. To do so, you could change the declaration of @code{stmt} as |
| 1069 follows: |
| 1070 |
| 1071 @example |
| 1072 stmt : expr ';' %merge <stmtMerge> |
| 1073 | decl %merge <stmtMerge> |
| 1074 ; |
| 1075 @end example |
| 1076 |
| 1077 @noindent |
| 1078 and define the @code{stmtMerge} function as: |
| 1079 |
| 1080 @example |
| 1081 static YYSTYPE |
| 1082 stmtMerge (YYSTYPE x0, YYSTYPE x1) |
| 1083 @{ |
| 1084 printf ("<OR> "); |
| 1085 return ""; |
| 1086 @} |
| 1087 @end example |
| 1088 |
| 1089 @noindent |
| 1090 with an accompanying forward declaration |
| 1091 in the C declarations at the beginning of the file: |
| 1092 |
| 1093 @example |
| 1094 %@{ |
| 1095 #define YYSTYPE char const * |
| 1096 static YYSTYPE stmtMerge (YYSTYPE x0, YYSTYPE x1); |
| 1097 %@} |
| 1098 @end example |
| 1099 |
| 1100 @noindent |
| 1101 With these declarations, the resulting parser parses the first example |
| 1102 as both an @code{expr} and a @code{decl}, and prints |
| 1103 |
| 1104 @example |
| 1105 "x" y z + T <init-declare> x T <cast> y z + = <OR> |
| 1106 @end example |
| 1107 |
| 1108 Bison requires that all of the |
| 1109 productions that participate in any particular merge have identical |
| 1110 @samp{%merge} clauses. Otherwise, the ambiguity would be unresolvable, |
| 1111 and the parser will report an error during any parse that results in |
| 1112 the offending merge. |
| 1113 |
| 1114 @node GLR Semantic Actions |
| 1115 @subsection GLR Semantic Actions |
| 1116 |
| 1117 @cindex deferred semantic actions |
| 1118 By definition, a deferred semantic action is not performed at the same time as |
| 1119 the associated reduction. |
| 1120 This raises caveats for several Bison features you might use in a semantic |
| 1121 action in a @acronym{GLR} parser. |
| 1122 |
| 1123 @vindex yychar |
| 1124 @cindex @acronym{GLR} parsers and @code{yychar} |
| 1125 @vindex yylval |
| 1126 @cindex @acronym{GLR} parsers and @code{yylval} |
| 1127 @vindex yylloc |
| 1128 @cindex @acronym{GLR} parsers and @code{yylloc} |
| 1129 In any semantic action, you can examine @code{yychar} to determine the type of |
| 1130 the lookahead token present at the time of the associated reduction. |
| 1131 After checking that @code{yychar} is not set to @code{YYEMPTY} or @code{YYEOF}, |
| 1132 you can then examine @code{yylval} and @code{yylloc} to determine the |
| 1133 lookahead token's semantic value and location, if any. |
| 1134 In a nondeferred semantic action, you can also modify any of these variables to |
| 1135 influence syntax analysis. |
| 1136 @xref{Lookahead, ,Lookahead Tokens}. |
| 1137 |
| 1138 @findex yyclearin |
| 1139 @cindex @acronym{GLR} parsers and @code{yyclearin} |
| 1140 In a deferred semantic action, it's too late to influence syntax analysis. |
| 1141 In this case, @code{yychar}, @code{yylval}, and @code{yylloc} are set to |
| 1142 shallow copies of the values they had at the time of the associated reduction. |
| 1143 For this reason alone, modifying them is dangerous. |
| 1144 Moreover, the result of modifying them is undefined and subject to change with |
| 1145 future versions of Bison. |
| 1146 For example, if a semantic action might be deferred, you should never write it |
| 1147 to invoke @code{yyclearin} (@pxref{Action Features}) or to attempt to free |
| 1148 memory referenced by @code{yylval}. |
| 1149 |
| 1150 @findex YYERROR |
| 1151 @cindex @acronym{GLR} parsers and @code{YYERROR} |
| 1152 Another Bison feature requiring special consideration is @code{YYERROR} |
| 1153 (@pxref{Action Features}), which you can invoke in a semantic action to |
| 1154 initiate error recovery. |
| 1155 During deterministic @acronym{GLR} operation, the effect of @code{YYERROR} is |
| 1156 the same as its effect in an @acronym{LALR}(1) parser. |
| 1157 In a deferred semantic action, its effect is undefined. |
| 1158 @c The effect is probably a syntax error at the split point. |
| 1159 |
| 1160 Also, see @ref{Location Default Action, ,Default Action for Locations}, which |
| 1161 describes a special usage of @code{YYLLOC_DEFAULT} in @acronym{GLR} parsers. |
| 1162 |
| 1163 @node Compiler Requirements |
| 1164 @subsection Considerations when Compiling @acronym{GLR} Parsers |
| 1165 @cindex @code{inline} |
| 1166 @cindex @acronym{GLR} parsers and @code{inline} |
| 1167 |
| 1168 The @acronym{GLR} parsers require a compiler for @acronym{ISO} C89 or |
| 1169 later. In addition, they use the @code{inline} keyword, which is not |
| 1170 C89, but is C99 and is a common extension in pre-C99 compilers. It is |
| 1171 up to the user of these parsers to handle |
| 1172 portability issues. For instance, if using Autoconf and the Autoconf |
| 1173 macro @code{AC_C_INLINE}, a mere |
| 1174 |
| 1175 @example |
| 1176 %@{ |
| 1177 #include <config.h> |
| 1178 %@} |
| 1179 @end example |
| 1180 |
| 1181 @noindent |
| 1182 will suffice. Otherwise, we suggest |
| 1183 |
| 1184 @example |
| 1185 %@{ |
| 1186 #if __STDC_VERSION__ < 199901 && ! defined __GNUC__ && ! defined inline |
| 1187 #define inline |
| 1188 #endif |
| 1189 %@} |
| 1190 @end example |
| 1191 |
| 1192 @node Locations Overview |
| 1193 @section Locations |
| 1194 @cindex location |
| 1195 @cindex textual location |
| 1196 @cindex location, textual |
| 1197 |
| 1198 Many applications, like interpreters or compilers, have to produce verbose |
| 1199 and useful error messages. To achieve this, one must be able to keep track of |
| 1200 the @dfn{textual location}, or @dfn{location}, of each syntactic construct. |
| 1201 Bison provides a mechanism for handling these locations. |
| 1202 |
| 1203 Each token has a semantic value. In a similar fashion, each token has an |
| 1204 associated location, but the type of locations is the same for all tokens and |
| 1205 groupings. Moreover, the output parser is equipped with a default data |
| 1206 structure for storing locations (@pxref{Locations}, for more details). |
| 1207 |
| 1208 Like semantic values, locations can be reached in actions using a dedicated |
| 1209 set of constructs. In the example above, the location of the whole grouping |
| 1210 is @code{@@$}, while the locations of the subexpressions are @code{@@1} and |
| 1211 @code{@@3}. |
| 1212 |
| 1213 When a rule is matched, a default action is used to compute the semantic value |
| 1214 of its left hand side (@pxref{Actions}). In the same way, another default |
| 1215 action is used for locations. However, the action for locations is general |
| 1216 enough for most cases, meaning there is usually no need to describe for each |
| 1217 rule how @code{@@$} should be formed. When building a new location for a given |
| 1218 grouping, the default behavior of the output parser is to take the beginning |
| 1219 of the first symbol, and the end of the last symbol. |
| 1220 |
| 1221 @node Bison Parser |
| 1222 @section Bison Output: the Parser File |
| 1223 @cindex Bison parser |
| 1224 @cindex Bison utility |
| 1225 @cindex lexical analyzer, purpose |
| 1226 @cindex parser |
| 1227 |
| 1228 When you run Bison, you give it a Bison grammar file as input. The output |
| 1229 is a C source file that parses the language described by the grammar. |
| 1230 This file is called a @dfn{Bison parser}. Keep in mind that the Bison |
| 1231 utility and the Bison parser are two distinct programs: the Bison utility |
| 1232 is a program whose output is the Bison parser that becomes part of your |
| 1233 program. |
| 1234 |
| 1235 The job of the Bison parser is to group tokens into groupings according to |
| 1236 the grammar rules---for example, to build identifiers and operators into |
| 1237 expressions. As it does this, it runs the actions for the grammar rules it |
| 1238 uses. |
| 1239 |
| 1240 The tokens come from a function called the @dfn{lexical analyzer} that |
| 1241 you must supply in some fashion (such as by writing it in C). The Bison |
| 1242 parser calls the lexical analyzer each time it wants a new token. It |
| 1243 doesn't know what is ``inside'' the tokens (though their semantic values |
| 1244 may reflect this). Typically the lexical analyzer makes the tokens by |
| 1245 parsing characters of text, but Bison does not depend on this. |
| 1246 @xref{Lexical, ,The Lexical Analyzer Function @code{yylex}}. |
| 1247 |
| 1248 The Bison parser file is C code which defines a function named |
| 1249 @code{yyparse} which implements that grammar. This function does not make |
| 1250 a complete C program: you must supply some additional functions. One is |
| 1251 the lexical analyzer. Another is an error-reporting function which the |
| 1252 parser calls to report an error. In addition, a complete C program must |
| 1253 start with a function called @code{main}; you have to provide this, and |
| 1254 arrange for it to call @code{yyparse} or the parser will never run. |
| 1255 @xref{Interface, ,Parser C-Language Interface}. |
| 1256 |
| 1257 Aside from the token type names and the symbols in the actions you |
| 1258 write, all symbols defined in the Bison parser file itself |
| 1259 begin with @samp{yy} or @samp{YY}. This includes interface functions |
| 1260 such as the lexical analyzer function @code{yylex}, the error reporting |
| 1261 function @code{yyerror} and the parser function @code{yyparse} itself. |
| 1262 This also includes numerous identifiers used for internal purposes. |
| 1263 Therefore, you should avoid using C identifiers starting with @samp{yy} |
| 1264 or @samp{YY} in the Bison grammar file except for the ones defined in |
| 1265 this manual. Also, you should avoid using the C identifiers |
| 1266 @samp{malloc} and @samp{free} for anything other than their usual |
| 1267 meanings. |
| 1268 |
| 1269 In some cases the Bison parser file includes system headers, and in |
| 1270 those cases your code should respect the identifiers reserved by those |
| 1271 headers. On some non-@acronym{GNU} hosts, @code{<alloca.h>}, @code{<malloc.h>}, |
| 1272 @code{<stddef.h>}, and @code{<stdlib.h>} are included as needed to |
| 1273 declare memory allocators and related types. @code{<libintl.h>} is |
| 1274 included if message translation is in use |
| 1275 (@pxref{Internationalization}). Other system headers may |
| 1276 be included if you define @code{YYDEBUG} to a nonzero value |
| 1277 (@pxref{Tracing, ,Tracing Your Parser}). |
| 1278 |
| 1279 @node Stages |
| 1280 @section Stages in Using Bison |
| 1281 @cindex stages in using Bison |
| 1282 @cindex using Bison |
| 1283 |
| 1284 The actual language-design process using Bison, from grammar specification |
| 1285 to a working compiler or interpreter, has these parts: |
| 1286 |
| 1287 @enumerate |
| 1288 @item |
| 1289 Formally specify the grammar in a form recognized by Bison |
| 1290 (@pxref{Grammar File, ,Bison Grammar Files}). For each grammatical rule |
| 1291 in the language, describe the action that is to be taken when an |
| 1292 instance of that rule is recognized. The action is described by a |
| 1293 sequence of C statements. |
| 1294 |
| 1295 @item |
| 1296 Write a lexical analyzer to process input and pass tokens to the parser. |
| 1297 The lexical analyzer may be written by hand in C (@pxref{Lexical, ,The |
| 1298 Lexical Analyzer Function @code{yylex}}). It could also be produced |
| 1299 using Lex, but the use of Lex is not discussed in this manual. |
| 1300 |
| 1301 @item |
| 1302 Write a controlling function that calls the Bison-produced parser. |
| 1303 |
| 1304 @item |
| 1305 Write error-reporting routines. |
| 1306 @end enumerate |
| 1307 |
| 1308 To turn this source code as written into a runnable program, you |
| 1309 must follow these steps: |
| 1310 |
| 1311 @enumerate |
| 1312 @item |
| 1313 Run Bison on the grammar to produce the parser. |
| 1314 |
| 1315 @item |
| 1316 Compile the code output by Bison, as well as any other source files. |
| 1317 |
| 1318 @item |
| 1319 Link the object files to produce the finished product. |
| 1320 @end enumerate |
| 1321 |
| 1322 @node Grammar Layout |
| 1323 @section The Overall Layout of a Bison Grammar |
| 1324 @cindex grammar file |
| 1325 @cindex file format |
| 1326 @cindex format of grammar file |
| 1327 @cindex layout of Bison grammar |
| 1328 |
| 1329 The input file for the Bison utility is a @dfn{Bison grammar file}. The |
| 1330 general form of a Bison grammar file is as follows: |
| 1331 |
| 1332 @example |
| 1333 %@{ |
| 1334 @var{Prologue} |
| 1335 %@} |
| 1336 |
| 1337 @var{Bison declarations} |
| 1338 |
| 1339 %% |
| 1340 @var{Grammar rules} |
| 1341 %% |
| 1342 @var{Epilogue} |
| 1343 @end example |
| 1344 |
| 1345 @noindent |
| 1346 The @samp{%%}, @samp{%@{} and @samp{%@}} are punctuation that appears |
| 1347 in every Bison grammar file to separate the sections. |
| 1348 |
| 1349 The prologue may define types and variables used in the actions. You can |
| 1350 also use preprocessor commands to define macros used there, and use |
| 1351 @code{#include} to include header files that do any of these things. |
| 1352 You need to declare the lexical analyzer @code{yylex} and the error |
| 1353 printer @code{yyerror} here, along with any other global identifiers |
| 1354 used by the actions in the grammar rules. |
| 1355 |
| 1356 The Bison declarations declare the names of the terminal and nonterminal |
| 1357 symbols, and may also describe operator precedence and the data types of |
| 1358 semantic values of various symbols. |
| 1359 |
| 1360 The grammar rules define how to construct each nonterminal symbol from its |
| 1361 parts. |
| 1362 |
| 1363 The epilogue can contain any code you want to use. Often the |
| 1364 definitions of functions declared in the prologue go here. In a |
| 1365 simple program, all the rest of the program can go here. |
| 1366 |
| 1367 @node Examples |
| 1368 @chapter Examples |
| 1369 @cindex simple examples |
| 1370 @cindex examples, simple |
| 1371 |
| 1372 Now we show and explain three sample programs written using Bison: a |
| 1373 reverse polish notation calculator, an algebraic (infix) notation |
| 1374 calculator, and a multi-function calculator. All three have been tested |
| 1375 under BSD Unix 4.3; each produces a usable, though limited, interactive |
| 1376 desk-top calculator. |
| 1377 |
| 1378 These examples are simple, but Bison grammars for real programming |
| 1379 languages are written the same way. You can copy these examples into a |
| 1380 source file to try them. |
| 1381 |
| 1382 @menu |
| 1383 * RPN Calc:: Reverse polish notation calculator; |
| 1384 a first example with no operator precedence. |
| 1385 * Infix Calc:: Infix (algebraic) notation calculator. |
| 1386 Operator precedence is introduced. |
| 1387 * Simple Error Recovery:: Continuing after syntax errors. |
| 1388 * Location Tracking Calc:: Demonstrating the use of @@@var{n} and @@$. |
| 1389 * Multi-function Calc:: Calculator with memory and trig functions. |
| 1390 It uses multiple data-types for semantic values. |
| 1391 * Exercises:: Ideas for improving the multi-function calculator. |
| 1392 @end menu |
| 1393 |
| 1394 @node RPN Calc |
| 1395 @section Reverse Polish Notation Calculator |
| 1396 @cindex reverse polish notation |
| 1397 @cindex polish notation calculator |
| 1398 @cindex @code{rpcalc} |
| 1399 @cindex calculator, simple |
| 1400 |
| 1401 The first example is that of a simple double-precision @dfn{reverse polish |
| 1402 notation} calculator (a calculator using postfix operators). This example |
| 1403 provides a good starting point, since operator precedence is not an issue. |
| 1404 The second example will illustrate how operator precedence is handled. |
| 1405 |
| 1406 The source code for this calculator is named @file{rpcalc.y}. The |
| 1407 @samp{.y} extension is a convention used for Bison input files. |
| 1408 |
| 1409 @menu |
| 1410 * Rpcalc Declarations:: Prologue (declarations) for rpcalc. |
| 1411 * Rpcalc Rules:: Grammar Rules for rpcalc, with explanation. |
| 1412 * Rpcalc Lexer:: The lexical analyzer. |
| 1413 * Rpcalc Main:: The controlling function. |
| 1414 * Rpcalc Error:: The error reporting function. |
| 1415 * Rpcalc Generate:: Running Bison on the grammar file. |
| 1416 * Rpcalc Compile:: Run the C compiler on the output code. |
| 1417 @end menu |
| 1418 |
| 1419 @node Rpcalc Declarations |
| 1420 @subsection Declarations for @code{rpcalc} |
| 1421 |
| 1422 Here are the C and Bison declarations for the reverse polish notation |
| 1423 calculator. As in C, comments are placed between @samp{/*@dots{}*/}. |
| 1424 |
| 1425 @example |
| 1426 /* Reverse polish notation calculator. */ |
| 1427 |
| 1428 %@{ |
| 1429 #define YYSTYPE double |
| 1430 #include <math.h> |
| 1431 int yylex (void); |
| 1432 void yyerror (char const *); |
| 1433 %@} |
| 1434 |
| 1435 %token NUM |
| 1436 |
| 1437 %% /* Grammar rules and actions follow. */ |
| 1438 @end example |
| 1439 |
| 1440 The declarations section (@pxref{Prologue, , The prologue}) contains two |
| 1441 preprocessor directives and two forward declarations. |
| 1442 |
| 1443 The @code{#define} directive defines the macro @code{YYSTYPE}, thus |
| 1444 specifying the C data type for semantic values of both tokens and |
| 1445 groupings (@pxref{Value Type, ,Data Types of Semantic Values}). The |
| 1446 Bison parser will use whatever type @code{YYSTYPE} is defined as; if you |
| 1447 don't define it, @code{int} is the default. Because we specify |
| 1448 @code{double}, each token and each expression has an associated value, |
| 1449 which is a floating point number. |
| 1450 |
| 1451 The @code{#include} directive is used to declare the exponentiation |
| 1452 function @code{pow}. |
| 1453 |
| 1454 The forward declarations for @code{yylex} and @code{yyerror} are |
| 1455 needed because the C language requires that functions be declared |
| 1456 before they are used. These functions will be defined in the |
| 1457 epilogue, but the parser calls them so they must be declared in the |
| 1458 prologue. |
| 1459 |
| 1460 The second section, Bison declarations, provides information to Bison |
| 1461 about the token types (@pxref{Bison Declarations, ,The Bison |
| 1462 Declarations Section}). Each terminal symbol that is not a |
| 1463 single-character literal must be declared here. (Single-character |
| 1464 literals normally don't need to be declared.) In this example, all the |
| 1465 arithmetic operators are designated by single-character literals, so the |
| 1466 only terminal symbol that needs to be declared is @code{NUM}, the token |
| 1467 type for numeric constants. |
| 1468 |
| 1469 @node Rpcalc Rules |
| 1470 @subsection Grammar Rules for @code{rpcalc} |
| 1471 |
| 1472 Here are the grammar rules for the reverse polish notation calculator. |
| 1473 |
| 1474 @example |
| 1475 input: /* empty */ |
| 1476 | input line |
| 1477 ; |
| 1478 |
| 1479 line: '\n' |
| 1480 | exp '\n' @{ printf ("\t%.10g\n", $1); @} |
| 1481 ; |
| 1482 |
| 1483 exp: NUM @{ $$ = $1; @} |
| 1484 | exp exp '+' @{ $$ = $1 + $2; @} |
| 1485 | exp exp '-' @{ $$ = $1 - $2; @} |
| 1486 | exp exp '*' @{ $$ = $1 * $2; @} |
| 1487 | exp exp '/' @{ $$ = $1 / $2; @} |
| 1488 /* Exponentiation */ |
| 1489 | exp exp '^' @{ $$ = pow ($1, $2); @} |
| 1490 /* Unary minus */ |
| 1491 | exp 'n' @{ $$ = -$1; @} |
| 1492 ; |
| 1493 %% |
| 1494 @end example |
| 1495 |
| 1496 The groupings of the rpcalc ``language'' defined here are the expression |
| 1497 (given the name @code{exp}), the line of input (@code{line}), and the |
| 1498 complete input transcript (@code{input}). Each of these nonterminal |
| 1499 symbols has several alternate rules, joined by the vertical bar @samp{|} |
| 1500 which is read as ``or''. The following sections explain what these rules |
| 1501 mean. |
| 1502 |
| 1503 The semantics of the language is determined by the actions taken when a |
| 1504 grouping is recognized. The actions are the C code that appears inside |
| 1505 braces. @xref{Actions}. |
| 1506 |
| 1507 You must specify these actions in C, but Bison provides the means for |
| 1508 passing semantic values between the rules. In each action, the |
| 1509 pseudo-variable @code{$$} stands for the semantic value for the grouping |
| 1510 that the rule is going to construct. Assigning a value to @code{$$} is the |
| 1511 main job of most actions. The semantic values of the components of the |
| 1512 rule are referred to as @code{$1}, @code{$2}, and so on. |
| 1513 |
| 1514 @menu |
| 1515 * Rpcalc Input:: |
| 1516 * Rpcalc Line:: |
| 1517 * Rpcalc Expr:: |
| 1518 @end menu |
| 1519 |
| 1520 @node Rpcalc Input |
| 1521 @subsubsection Explanation of @code{input} |
| 1522 |
| 1523 Consider the definition of @code{input}: |
| 1524 |
| 1525 @example |
| 1526 input: /* empty */ |
| 1527 | input line |
| 1528 ; |
| 1529 @end example |
| 1530 |
| 1531 This definition reads as follows: ``A complete input is either an empty |
| 1532 string, or a complete input followed by an input line''. Notice that |
| 1533 ``complete input'' is defined in terms of itself. This definition is said |
| 1534 to be @dfn{left recursive} since @code{input} appears always as the |
| 1535 leftmost symbol in the sequence. @xref{Recursion, ,Recursive Rules}. |
| 1536 |
| 1537 The first alternative is empty because there are no symbols between the |
| 1538 colon and the first @samp{|}; this means that @code{input} can match an |
| 1539 empty string of input (no tokens). We write the rules this way because it |
| 1540 is legitimate to type @kbd{Ctrl-d} right after you start the calculator. |
| 1541 It's conventional to put an empty alternative first and write the comment |
| 1542 @samp{/* empty */} in it. |
| 1543 |
| 1544 The second alternate rule (@code{input line}) handles all nontrivial input. |
| 1545 It means, ``After reading any number of lines, read one more line if |
| 1546 possible.'' The left recursion makes this rule into a loop. Since the |
| 1547 first alternative matches empty input, the loop can be executed zero or |
| 1548 more times. |
| 1549 |
| 1550 The parser function @code{yyparse} continues to process input until a |
| 1551 grammatical error is seen or the lexical analyzer says there are no more |
| 1552 input tokens; we will arrange for the latter to happen at end-of-input. |
| 1553 |
| 1554 @node Rpcalc Line |
| 1555 @subsubsection Explanation of @code{line} |
| 1556 |
| 1557 Now consider the definition of @code{line}: |
| 1558 |
| 1559 @example |
| 1560 line: '\n' |
| 1561 | exp '\n' @{ printf ("\t%.10g\n", $1); @} |
| 1562 ; |
| 1563 @end example |
| 1564 |
| 1565 The first alternative is a token which is a newline character; this means |
| 1566 that rpcalc accepts a blank line (and ignores it, since there is no |
| 1567 action). The second alternative is an expression followed by a newline. |
| 1568 This is the alternative that makes rpcalc useful. The semantic value of |
| 1569 the @code{exp} grouping is the value of @code{$1} because the @code{exp} in |
| 1570 question is the first symbol in the alternative. The action prints this |
| 1571 value, which is the result of the computation the user asked for. |
| 1572 |
| 1573 This action is unusual because it does not assign a value to @code{$$}. As |
| 1574 a consequence, the semantic value associated with the @code{line} is |
| 1575 uninitialized (its value will be unpredictable). This would be a bug if |
| 1576 that value were ever used, but we don't use it: once rpcalc has printed the |
| 1577 value of the user's input line, that value is no longer needed. |
| 1578 |
| 1579 @node Rpcalc Expr |
| 1580 @subsubsection Explanation of @code{expr} |
| 1581 |
| 1582 The @code{exp} grouping has several rules, one for each kind of expression. |
| 1583 The first rule handles the simplest expressions: those that are just numbers. |
| 1584 The second handles an addition-expression, which looks like two expressions |
| 1585 followed by a plus-sign. The third handles subtraction, and so on. |
| 1586 |
| 1587 @example |
| 1588 exp: NUM |
| 1589 | exp exp '+' @{ $$ = $1 + $2; @} |
| 1590 | exp exp '-' @{ $$ = $1 - $2; @} |
| 1591 @dots{} |
| 1592 ; |
| 1593 @end example |
| 1594 |
| 1595 We have used @samp{|} to join all the rules for @code{exp}, but we could |
| 1596 equally well have written them separately: |
| 1597 |
| 1598 @example |
| 1599 exp: NUM ; |
| 1600 exp: exp exp '+' @{ $$ = $1 + $2; @} ; |
| 1601 exp: exp exp '-' @{ $$ = $1 - $2; @} ; |
| 1602 @dots{} |
| 1603 @end example |
| 1604 |
| 1605 Most of the rules have actions that compute the value of the expression in |
| 1606 terms of the value of its parts. For example, in the rule for addition, |
| 1607 @code{$1} refers to the first component @code{exp} and @code{$2} refers to |
| 1608 the second one. The third component, @code{'+'}, has no meaningful |
| 1609 associated semantic value, but if it had one you could refer to it as |
| 1610 @code{$3}. When @code{yyparse} recognizes a sum expression using this |
| 1611 rule, the sum of the two subexpressions' values is produced as the value of |
| 1612 the entire expression. @xref{Actions}. |
| 1613 |
| 1614 You don't have to give an action for every rule. When a rule has no |
| 1615 action, Bison by default copies the value of @code{$1} into @code{$$}. |
| 1616 This is what happens in the first rule (the one that uses @code{NUM}). |
| 1617 |
| 1618 The formatting shown here is the recommended convention, but Bison does |
| 1619 not require it. You can add or change white space as much as you wish. |
| 1620 For example, this: |
| 1621 |
| 1622 @example |
| 1623 exp : NUM | exp exp '+' @{$$ = $1 + $2; @} | @dots{} ; |
| 1624 @end example |
| 1625 |
| 1626 @noindent |
| 1627 means the same thing as this: |
| 1628 |
| 1629 @example |
| 1630 exp: NUM |
| 1631 | exp exp '+' @{ $$ = $1 + $2; @} |
| 1632 | @dots{} |
| 1633 ; |
| 1634 @end example |
| 1635 |
| 1636 @noindent |
| 1637 The latter, however, is much more readable. |
| 1638 |
| 1639 @node Rpcalc Lexer |
| 1640 @subsection The @code{rpcalc} Lexical Analyzer |
| 1641 @cindex writing a lexical analyzer |
| 1642 @cindex lexical analyzer, writing |
| 1643 |
| 1644 The lexical analyzer's job is low-level parsing: converting characters |
| 1645 or sequences of characters into tokens. The Bison parser gets its |
| 1646 tokens by calling the lexical analyzer. @xref{Lexical, ,The Lexical |
| 1647 Analyzer Function @code{yylex}}. |
| 1648 |
| 1649 Only a simple lexical analyzer is needed for the @acronym{RPN} |
| 1650 calculator. This |
| 1651 lexical analyzer skips blanks and tabs, then reads in numbers as |
| 1652 @code{double} and returns them as @code{NUM} tokens. Any other character |
| 1653 that isn't part of a number is a separate token. Note that the token-code |
| 1654 for such a single-character token is the character itself. |
| 1655 |
| 1656 The return value of the lexical analyzer function is a numeric code which |
| 1657 represents a token type. The same text used in Bison rules to stand for |
| 1658 this token type is also a C expression for the numeric code for the type. |
| 1659 This works in two ways. If the token type is a character literal, then its |
| 1660 numeric code is that of the character; you can use the same |
| 1661 character literal in the lexical analyzer to express the number. If the |
| 1662 token type is an identifier, that identifier is defined by Bison as a C |
| 1663 macro whose definition is the appropriate number. In this example, |
| 1664 therefore, @code{NUM} becomes a macro for @code{yylex} to use. |
| 1665 |
| 1666 The semantic value of the token (if it has one) is stored into the |
| 1667 global variable @code{yylval}, which is where the Bison parser will look |
| 1668 for it. (The C data type of @code{yylval} is @code{YYSTYPE}, which was |
| 1669 defined at the beginning of the grammar; @pxref{Rpcalc Declarations, |
| 1670 ,Declarations for @code{rpcalc}}.) |
| 1671 |
| 1672 A token type code of zero is returned if the end-of-input is encountered. |
| 1673 (Bison recognizes any nonpositive value as indicating end-of-input.) |
| 1674 |
| 1675 Here is the code for the lexical analyzer: |
| 1676 |
| 1677 @example |
| 1678 @group |
| 1679 /* The lexical analyzer returns a double floating point |
| 1680 number on the stack and the token NUM, or the numeric code |
| 1681 of the character read if not a number. It skips all blanks |
| 1682 and tabs, and returns 0 for end-of-input. */ |
| 1683 |
| 1684 #include <ctype.h> |
| 1685 @end group |
| 1686 |
| 1687 @group |
| 1688 int |
| 1689 yylex (void) |
| 1690 @{ |
| 1691 int c; |
| 1692 |
| 1693 /* Skip white space. */ |
| 1694 while ((c = getchar ()) == ' ' || c == '\t') |
| 1695 ; |
| 1696 @end group |
| 1697 @group |
| 1698 /* Process numbers. */ |
| 1699 if (c == '.' || isdigit (c)) |
| 1700 @{ |
| 1701 ungetc (c, stdin); |
| 1702 scanf ("%lf", &yylval); |
| 1703 return NUM; |
| 1704 @} |
| 1705 @end group |
| 1706 @group |
| 1707 /* Return end-of-input. */ |
| 1708 if (c == EOF) |
| 1709 return 0; |
| 1710 /* Return a single char. */ |
| 1711 return c; |
| 1712 @} |
| 1713 @end group |
| 1714 @end example |
| 1715 |
| 1716 @node Rpcalc Main |
| 1717 @subsection The Controlling Function |
| 1718 @cindex controlling function |
| 1719 @cindex main function in simple example |
| 1720 |
| 1721 In keeping with the spirit of this example, the controlling function is |
| 1722 kept to the bare minimum. The only requirement is that it call |
| 1723 @code{yyparse} to start the process of parsing. |
| 1724 |
| 1725 @example |
| 1726 @group |
| 1727 int |
| 1728 main (void) |
| 1729 @{ |
| 1730 return yyparse (); |
| 1731 @} |
| 1732 @end group |
| 1733 @end example |
| 1734 |
| 1735 @node Rpcalc Error |
| 1736 @subsection The Error Reporting Routine |
| 1737 @cindex error reporting routine |
| 1738 |
| 1739 When @code{yyparse} detects a syntax error, it calls the error reporting |
| 1740 function @code{yyerror} to print an error message (usually but not |
| 1741 always @code{"syntax error"}). It is up to the programmer to supply |
| 1742 @code{yyerror} (@pxref{Interface, ,Parser C-Language Interface}), so |
| 1743 here is the definition we will use: |
| 1744 |
| 1745 @example |
| 1746 @group |
| 1747 #include <stdio.h> |
| 1748 |
| 1749 /* Called by yyparse on error. */ |
| 1750 void |
| 1751 yyerror (char const *s) |
| 1752 @{ |
| 1753 fprintf (stderr, "%s\n", s); |
| 1754 @} |
| 1755 @end group |
| 1756 @end example |
| 1757 |
| 1758 After @code{yyerror} returns, the Bison parser may recover from the error |
| 1759 and continue parsing if the grammar contains a suitable error rule |
| 1760 (@pxref{Error Recovery}). Otherwise, @code{yyparse} returns nonzero. We |
| 1761 have not written any error rules in this example, so any invalid input will |
| 1762 cause the calculator program to exit. This is not clean behavior for a |
| 1763 real calculator, but it is adequate for the first example. |
| 1764 |
| 1765 @node Rpcalc Generate |
| 1766 @subsection Running Bison to Make the Parser |
| 1767 @cindex running Bison (introduction) |
| 1768 |
| 1769 Before running Bison to produce a parser, we need to decide how to |
| 1770 arrange all the source code in one or more source files. For such a |
| 1771 simple example, the easiest thing is to put everything in one file. The |
| 1772 definitions of @code{yylex}, @code{yyerror} and @code{main} go at the |
| 1773 end, in the epilogue of the file |
| 1774 (@pxref{Grammar Layout, ,The Overall Layout of a Bison Grammar}). |
| 1775 |
| 1776 For a large project, you would probably have several source files, and use |
| 1777 @code{make} to arrange to recompile them. |
| 1778 |
| 1779 With all the source in a single file, you use the following command to |
| 1780 convert it into a parser file: |
| 1781 |
| 1782 @example |
| 1783 bison @var{file}.y |
| 1784 @end example |
| 1785 |
| 1786 @noindent |
| 1787 In this example the file was called @file{rpcalc.y} (for ``Reverse Polish |
| 1788 @sc{calc}ulator''). Bison produces a file named @file{@var{file}.tab.c}, |
| 1789 removing the @samp{.y} from the original file name. The file output by |
| 1790 Bison contains the source code for @code{yyparse}. The additional |
| 1791 functions in the input file (@code{yylex}, @code{yyerror} and @code{main}) |
| 1792 are copied verbatim to the output. |
| 1793 |
| 1794 @node Rpcalc Compile |
| 1795 @subsection Compiling the Parser File |
| 1796 @cindex compiling the parser |
| 1797 |
| 1798 Here is how to compile and run the parser file: |
| 1799 |
| 1800 @example |
| 1801 @group |
| 1802 # @r{List files in current directory.} |
| 1803 $ @kbd{ls} |
| 1804 rpcalc.tab.c rpcalc.y |
| 1805 @end group |
| 1806 |
| 1807 @group |
| 1808 # @r{Compile the Bison parser.} |
| 1809 # @r{@samp{-lm} tells compiler to search math library for @code{pow}.} |
| 1810 $ @kbd{cc -lm -o rpcalc rpcalc.tab.c} |
| 1811 @end group |
| 1812 |
| 1813 @group |
| 1814 # @r{List files again.} |
| 1815 $ @kbd{ls} |
| 1816 rpcalc rpcalc.tab.c rpcalc.y |
| 1817 @end group |
| 1818 @end example |
| 1819 |
| 1820 The file @file{rpcalc} now contains the executable code. Here is an |
| 1821 example session using @code{rpcalc}. |
| 1822 |
| 1823 @example |
| 1824 $ @kbd{rpcalc} |
| 1825 @kbd{4 9 +} |
| 1826 13 |
| 1827 @kbd{3 7 + 3 4 5 *+-} |
| 1828 -13 |
| 1829 @kbd{3 7 + 3 4 5 * + - n} @r{Note the unary minus, @samp{n}} |
| 1830 13 |
| 1831 @kbd{5 6 / 4 n +} |
| 1832 -3.166666667 |
| 1833 @kbd{3 4 ^} @r{Exponentiation} |
| 1834 81 |
| 1835 @kbd{^D} @r{End-of-file indicator} |
| 1836 $ |
| 1837 @end example |
| 1838 |
| 1839 @node Infix Calc |
| 1840 @section Infix Notation Calculator: @code{calc} |
| 1841 @cindex infix notation calculator |
| 1842 @cindex @code{calc} |
| 1843 @cindex calculator, infix notation |
| 1844 |
| 1845 We now modify rpcalc to handle infix operators instead of postfix. Infix |
| 1846 notation involves the concept of operator precedence and the need for |
| 1847 parentheses nested to arbitrary depth. Here is the Bison code for |
| 1848 @file{calc.y}, an infix desk-top calculator. |
| 1849 |
| 1850 @example |
| 1851 /* Infix notation calculator. */ |
| 1852 |
| 1853 %@{ |
| 1854 #define YYSTYPE double |
| 1855 #include <math.h> |
| 1856 #include <stdio.h> |
| 1857 int yylex (void); |
| 1858 void yyerror (char const *); |
| 1859 %@} |
| 1860 |
| 1861 /* Bison declarations. */ |
| 1862 %token NUM |
| 1863 %left '-' '+' |
| 1864 %left '*' '/' |
| 1865 %left NEG /* negation--unary minus */ |
| 1866 %right '^' /* exponentiation */ |
| 1867 |
| 1868 %% /* The grammar follows. */ |
| 1869 input: /* empty */ |
| 1870 | input line |
| 1871 ; |
| 1872 |
| 1873 line: '\n' |
| 1874 | exp '\n' @{ printf ("\t%.10g\n", $1); @} |
| 1875 ; |
| 1876 |
| 1877 exp: NUM @{ $$ = $1; @} |
| 1878 | exp '+' exp @{ $$ = $1 + $3; @} |
| 1879 | exp '-' exp @{ $$ = $1 - $3; @} |
| 1880 | exp '*' exp @{ $$ = $1 * $3; @} |
| 1881 | exp '/' exp @{ $$ = $1 / $3; @} |
| 1882 | '-' exp %prec NEG @{ $$ = -$2; @} |
| 1883 | exp '^' exp @{ $$ = pow ($1, $3); @} |
| 1884 | '(' exp ')' @{ $$ = $2; @} |
| 1885 ; |
| 1886 %% |
| 1887 @end example |
| 1888 |
| 1889 @noindent |
| 1890 The functions @code{yylex}, @code{yyerror} and @code{main} can be the |
| 1891 same as before. |
| 1892 |
| 1893 There are two important new features shown in this code. |
| 1894 |
| 1895 In the second section (Bison declarations), @code{%left} declares token |
| 1896 types and says they are left-associative operators. The declarations |
| 1897 @code{%left} and @code{%right} (right associativity) take the place of |
| 1898 @code{%token} which is used to declare a token type name without |
| 1899 associativity. (These tokens are single-character literals, which |
| 1900 ordinarily don't need to be declared. We declare them here to specify |
| 1901 the associativity.) |
| 1902 |
| 1903 Operator precedence is determined by the line ordering of the |
| 1904 declarations; the higher the line number of the declaration (lower on |
| 1905 the page or screen), the higher the precedence. Hence, exponentiation |
| 1906 has the highest precedence, unary minus (@code{NEG}) is next, followed |
| 1907 by @samp{*} and @samp{/}, and so on. @xref{Precedence, ,Operator |
| 1908 Precedence}. |
| 1909 |
| 1910 The other important new feature is the @code{%prec} in the grammar |
| 1911 section for the unary minus operator. The @code{%prec} simply instructs |
| 1912 Bison that the rule @samp{| '-' exp} has the same precedence as |
| 1913 @code{NEG}---in this case the next-to-highest. @xref{Contextual |
| 1914 Precedence, ,Context-Dependent Precedence}. |
| 1915 |
| 1916 Here is a sample run of @file{calc.y}: |
| 1917 |
| 1918 @need 500 |
| 1919 @example |
| 1920 $ @kbd{calc} |
| 1921 @kbd{4 + 4.5 - (34/(8*3+-3))} |
| 1922 6.880952381 |
| 1923 @kbd{-56 + 2} |
| 1924 -54 |
| 1925 @kbd{3 ^ 2} |
| 1926 9 |
| 1927 @end example |
| 1928 |
| 1929 @node Simple Error Recovery |
| 1930 @section Simple Error Recovery |
| 1931 @cindex error recovery, simple |
| 1932 |
| 1933 Up to this point, this manual has not addressed the issue of @dfn{error |
| 1934 recovery}---how to continue parsing after the parser detects a syntax |
| 1935 error. All we have handled is error reporting with @code{yyerror}. |
| 1936 Recall that by default @code{yyparse} returns after calling |
| 1937 @code{yyerror}. This means that an erroneous input line causes the |
| 1938 calculator program to exit. Now we show how to rectify this deficiency. |
| 1939 |
| 1940 The Bison language itself includes the reserved word @code{error}, which |
| 1941 may be included in the grammar rules. In the example below it has |
| 1942 been added to one of the alternatives for @code{line}: |
| 1943 |
| 1944 @example |
| 1945 @group |
| 1946 line: '\n' |
| 1947 | exp '\n' @{ printf ("\t%.10g\n", $1); @} |
| 1948 | error '\n' @{ yyerrok; @} |
| 1949 ; |
| 1950 @end group |
| 1951 @end example |
| 1952 |
| 1953 This addition to the grammar allows for simple error recovery in the |
| 1954 event of a syntax error. If an expression that cannot be evaluated is |
| 1955 read, the error will be recognized by the third rule for @code{line}, |
| 1956 and parsing will continue. (The @code{yyerror} function is still called |
| 1957 upon to print its message as well.) The action executes the statement |
| 1958 @code{yyerrok}, a macro defined automatically by Bison; its meaning is |
| 1959 that error recovery is complete (@pxref{Error Recovery}). Note the |
| 1960 difference between @code{yyerrok} and @code{yyerror}; neither one is a |
| 1961 misprint. |
| 1962 |
| 1963 This form of error recovery deals with syntax errors. There are other |
| 1964 kinds of errors; for example, division by zero, which raises an exception |
| 1965 signal that is normally fatal. A real calculator program must handle this |
| 1966 signal and use @code{longjmp} to return to @code{main} and resume parsing |
| 1967 input lines; it would also have to discard the rest of the current line of |
| 1968 input. We won't discuss this issue further because it is not specific to |
| 1969 Bison programs. |
| 1970 |
| 1971 @node Location Tracking Calc |
| 1972 @section Location Tracking Calculator: @code{ltcalc} |
| 1973 @cindex location tracking calculator |
| 1974 @cindex @code{ltcalc} |
| 1975 @cindex calculator, location tracking |
| 1976 |
| 1977 This example extends the infix notation calculator with location |
| 1978 tracking. This feature will be used to improve the error messages. For |
| 1979 the sake of clarity, this example is a simple integer calculator, since |
| 1980 most of the work needed to use locations will be done in the lexical |
| 1981 analyzer. |
| 1982 |
| 1983 @menu |
| 1984 * Ltcalc Declarations:: Bison and C declarations for ltcalc. |
| 1985 * Ltcalc Rules:: Grammar rules for ltcalc, with explanations. |
| 1986 * Ltcalc Lexer:: The lexical analyzer. |
| 1987 @end menu |
| 1988 |
| 1989 @node Ltcalc Declarations |
| 1990 @subsection Declarations for @code{ltcalc} |
| 1991 |
| 1992 The C and Bison declarations for the location tracking calculator are |
| 1993 the same as the declarations for the infix notation calculator. |
| 1994 |
| 1995 @example |
| 1996 /* Location tracking calculator. */ |
| 1997 |
| 1998 %@{ |
| 1999 #define YYSTYPE int |
| 2000 #include <math.h> |
| 2001 int yylex (void); |
| 2002 void yyerror (char const *); |
| 2003 %@} |
| 2004 |
| 2005 /* Bison declarations. */ |
| 2006 %token NUM |
| 2007 |
| 2008 %left '-' '+' |
| 2009 %left '*' '/' |
| 2010 %left NEG |
| 2011 %right '^' |
| 2012 |
| 2013 %% /* The grammar follows. */ |
| 2014 @end example |
| 2015 |
| 2016 @noindent |
| 2017 Note there are no declarations specific to locations. Defining a data |
| 2018 type for storing locations is not needed: we will use the type provided |
| 2019 by default (@pxref{Location Type, ,Data Types of Locations}), which is a |
| 2020 four member structure with the following integer fields: |
| 2021 @code{first_line}, @code{first_column}, @code{last_line} and |
| 2022 @code{last_column}. By conventions, and in accordance with the GNU |
| 2023 Coding Standards and common practice, the line and column count both |
| 2024 start at 1. |
| 2025 |
| 2026 @node Ltcalc Rules |
| 2027 @subsection Grammar Rules for @code{ltcalc} |
| 2028 |
| 2029 Whether handling locations or not has no effect on the syntax of your |
| 2030 language. Therefore, grammar rules for this example will be very close |
| 2031 to those of the previous example: we will only modify them to benefit |
| 2032 from the new information. |
| 2033 |
| 2034 Here, we will use locations to report divisions by zero, and locate the |
| 2035 wrong expressions or subexpressions. |
| 2036 |
| 2037 @example |
| 2038 @group |
| 2039 input : /* empty */ |
| 2040 | input line |
| 2041 ; |
| 2042 @end group |
| 2043 |
| 2044 @group |
| 2045 line : '\n' |
| 2046 | exp '\n' @{ printf ("%d\n", $1); @} |
| 2047 ; |
| 2048 @end group |
| 2049 |
| 2050 @group |
| 2051 exp : NUM @{ $$ = $1; @} |
| 2052 | exp '+' exp @{ $$ = $1 + $3; @} |
| 2053 | exp '-' exp @{ $$ = $1 - $3; @} |
| 2054 | exp '*' exp @{ $$ = $1 * $3; @} |
| 2055 @end group |
| 2056 @group |
| 2057 | exp '/' exp |
| 2058 @{ |
| 2059 if ($3) |
| 2060 $$ = $1 / $3; |
| 2061 else |
| 2062 @{ |
| 2063 $$ = 1; |
| 2064 fprintf (stderr, "%d.%d-%d.%d: division by zero", |
| 2065 @@3.first_line, @@3.first_column, |
| 2066 @@3.last_line, @@3.last_column); |
| 2067 @} |
| 2068 @} |
| 2069 @end group |
| 2070 @group |
| 2071 | '-' exp %prec NEG @{ $$ = -$2; @} |
| 2072 | exp '^' exp @{ $$ = pow ($1, $3); @} |
| 2073 | '(' exp ')' @{ $$ = $2; @} |
| 2074 @end group |
| 2075 @end example |
| 2076 |
| 2077 This code shows how to reach locations inside of semantic actions, by |
| 2078 using the pseudo-variables @code{@@@var{n}} for rule components, and the |
| 2079 pseudo-variable @code{@@$} for groupings. |
| 2080 |
| 2081 We don't need to assign a value to @code{@@$}: the output parser does it |
| 2082 automatically. By default, before executing the C code of each action, |
| 2083 @code{@@$} is set to range from the beginning of @code{@@1} to the end |
| 2084 of @code{@@@var{n}}, for a rule with @var{n} components. This behavior |
| 2085 can be redefined (@pxref{Location Default Action, , Default Action for |
| 2086 Locations}), and for very specific rules, @code{@@$} can be computed by |
| 2087 hand. |
| 2088 |
| 2089 @node Ltcalc Lexer |
| 2090 @subsection The @code{ltcalc} Lexical Analyzer. |
| 2091 |
| 2092 Until now, we relied on Bison's defaults to enable location |
| 2093 tracking. The next step is to rewrite the lexical analyzer, and make it |
| 2094 able to feed the parser with the token locations, as it already does for |
| 2095 semantic values. |
| 2096 |
| 2097 To this end, we must take into account every single character of the |
| 2098 input text, to avoid the computed locations of being fuzzy or wrong: |
| 2099 |
| 2100 @example |
| 2101 @group |
| 2102 int |
| 2103 yylex (void) |
| 2104 @{ |
| 2105 int c; |
| 2106 @end group |
| 2107 |
| 2108 @group |
| 2109 /* Skip white space. */ |
| 2110 while ((c = getchar ()) == ' ' || c == '\t') |
| 2111 ++yylloc.last_column; |
| 2112 @end group |
| 2113 |
| 2114 @group |
| 2115 /* Step. */ |
| 2116 yylloc.first_line = yylloc.last_line; |
| 2117 yylloc.first_column = yylloc.last_column; |
| 2118 @end group |
| 2119 |
| 2120 @group |
| 2121 /* Process numbers. */ |
| 2122 if (isdigit (c)) |
| 2123 @{ |
| 2124 yylval = c - '0'; |
| 2125 ++yylloc.last_column; |
| 2126 while (isdigit (c = getchar ())) |
| 2127 @{ |
| 2128 ++yylloc.last_column; |
| 2129 yylval = yylval * 10 + c - '0'; |
| 2130 @} |
| 2131 ungetc (c, stdin); |
| 2132 return NUM; |
| 2133 @} |
| 2134 @end group |
| 2135 |
| 2136 /* Return end-of-input. */ |
| 2137 if (c == EOF) |
| 2138 return 0; |
| 2139 |
| 2140 /* Return a single char, and update location. */ |
| 2141 if (c == '\n') |
| 2142 @{ |
| 2143 ++yylloc.last_line; |
| 2144 yylloc.last_column = 0; |
| 2145 @} |
| 2146 else |
| 2147 ++yylloc.last_column; |
| 2148 return c; |
| 2149 @} |
| 2150 @end example |
| 2151 |
| 2152 Basically, the lexical analyzer performs the same processing as before: |
| 2153 it skips blanks and tabs, and reads numbers or single-character tokens. |
| 2154 In addition, it updates @code{yylloc}, the global variable (of type |
| 2155 @code{YYLTYPE}) containing the token's location. |
| 2156 |
| 2157 Now, each time this function returns a token, the parser has its number |
| 2158 as well as its semantic value, and its location in the text. The last |
| 2159 needed change is to initialize @code{yylloc}, for example in the |
| 2160 controlling function: |
| 2161 |
| 2162 @example |
| 2163 @group |
| 2164 int |
| 2165 main (void) |
| 2166 @{ |
| 2167 yylloc.first_line = yylloc.last_line = 1; |
| 2168 yylloc.first_column = yylloc.last_column = 0; |
| 2169 return yyparse (); |
| 2170 @} |
| 2171 @end group |
| 2172 @end example |
| 2173 |
| 2174 Remember that computing locations is not a matter of syntax. Every |
| 2175 character must be associated to a location update, whether it is in |
| 2176 valid input, in comments, in literal strings, and so on. |
| 2177 |
| 2178 @node Multi-function Calc |
| 2179 @section Multi-Function Calculator: @code{mfcalc} |
| 2180 @cindex multi-function calculator |
| 2181 @cindex @code{mfcalc} |
| 2182 @cindex calculator, multi-function |
| 2183 |
| 2184 Now that the basics of Bison have been discussed, it is time to move on to |
| 2185 a more advanced problem. The above calculators provided only five |
| 2186 functions, @samp{+}, @samp{-}, @samp{*}, @samp{/} and @samp{^}. It would |
| 2187 be nice to have a calculator that provides other mathematical functions such |
| 2188 as @code{sin}, @code{cos}, etc. |
| 2189 |
| 2190 It is easy to add new operators to the infix calculator as long as they are |
| 2191 only single-character literals. The lexical analyzer @code{yylex} passes |
| 2192 back all nonnumeric characters as tokens, so new grammar rules suffice for |
| 2193 adding a new operator. But we want something more flexible: built-in |
| 2194 functions whose syntax has this form: |
| 2195 |
| 2196 @example |
| 2197 @var{function_name} (@var{argument}) |
| 2198 @end example |
| 2199 |
| 2200 @noindent |
| 2201 At the same time, we will add memory to the calculator, by allowing you |
| 2202 to create named variables, store values in them, and use them later. |
| 2203 Here is a sample session with the multi-function calculator: |
| 2204 |
| 2205 @example |
| 2206 $ @kbd{mfcalc} |
| 2207 @kbd{pi = 3.141592653589} |
| 2208 3.1415926536 |
| 2209 @kbd{sin(pi)} |
| 2210 0.0000000000 |
| 2211 @kbd{alpha = beta1 = 2.3} |
| 2212 2.3000000000 |
| 2213 @kbd{alpha} |
| 2214 2.3000000000 |
| 2215 @kbd{ln(alpha)} |
| 2216 0.8329091229 |
| 2217 @kbd{exp(ln(beta1))} |
| 2218 2.3000000000 |
| 2219 $ |
| 2220 @end example |
| 2221 |
| 2222 Note that multiple assignment and nested function calls are permitted. |
| 2223 |
| 2224 @menu |
| 2225 * Mfcalc Declarations:: Bison declarations for multi-function calculator. |
| 2226 * Mfcalc Rules:: Grammar rules for the calculator. |
| 2227 * Mfcalc Symbol Table:: Symbol table management subroutines. |
| 2228 @end menu |
| 2229 |
| 2230 @node Mfcalc Declarations |
| 2231 @subsection Declarations for @code{mfcalc} |
| 2232 |
| 2233 Here are the C and Bison declarations for the multi-function calculator. |
| 2234 |
| 2235 @smallexample |
| 2236 @group |
| 2237 %@{ |
| 2238 #include <math.h> /* For math functions, cos(), sin(), etc. */ |
| 2239 #include "calc.h" /* Contains definition of `symrec'. */ |
| 2240 int yylex (void); |
| 2241 void yyerror (char const *); |
| 2242 %@} |
| 2243 @end group |
| 2244 @group |
| 2245 %union @{ |
| 2246 double val; /* For returning numbers. */ |
| 2247 symrec *tptr; /* For returning symbol-table pointers. */ |
| 2248 @} |
| 2249 @end group |
| 2250 %token <val> NUM /* Simple double precision number. */ |
| 2251 %token <tptr> VAR FNCT /* Variable and Function. */ |
| 2252 %type <val> exp |
| 2253 |
| 2254 @group |
| 2255 %right '=' |
| 2256 %left '-' '+' |
| 2257 %left '*' '/' |
| 2258 %left NEG /* negation--unary minus */ |
| 2259 %right '^' /* exponentiation */ |
| 2260 @end group |
| 2261 %% /* The grammar follows. */ |
| 2262 @end smallexample |
| 2263 |
| 2264 The above grammar introduces only two new features of the Bison language. |
| 2265 These features allow semantic values to have various data types |
| 2266 (@pxref{Multiple Types, ,More Than One Value Type}). |
| 2267 |
| 2268 The @code{%union} declaration specifies the entire list of possible types; |
| 2269 this is instead of defining @code{YYSTYPE}. The allowable types are now |
| 2270 double-floats (for @code{exp} and @code{NUM}) and pointers to entries in |
| 2271 the symbol table. @xref{Union Decl, ,The Collection of Value Types}. |
| 2272 |
| 2273 Since values can now have various types, it is necessary to associate a |
| 2274 type with each grammar symbol whose semantic value is used. These symbols |
| 2275 are @code{NUM}, @code{VAR}, @code{FNCT}, and @code{exp}. Their |
| 2276 declarations are augmented with information about their data type (placed |
| 2277 between angle brackets). |
| 2278 |
| 2279 The Bison construct @code{%type} is used for declaring nonterminal |
| 2280 symbols, just as @code{%token} is used for declaring token types. We |
| 2281 have not used @code{%type} before because nonterminal symbols are |
| 2282 normally declared implicitly by the rules that define them. But |
| 2283 @code{exp} must be declared explicitly so we can specify its value type. |
| 2284 @xref{Type Decl, ,Nonterminal Symbols}. |
| 2285 |
| 2286 @node Mfcalc Rules |
| 2287 @subsection Grammar Rules for @code{mfcalc} |
| 2288 |
| 2289 Here are the grammar rules for the multi-function calculator. |
| 2290 Most of them are copied directly from @code{calc}; three rules, |
| 2291 those which mention @code{VAR} or @code{FNCT}, are new. |
| 2292 |
| 2293 @smallexample |
| 2294 @group |
| 2295 input: /* empty */ |
| 2296 | input line |
| 2297 ; |
| 2298 @end group |
| 2299 |
| 2300 @group |
| 2301 line: |
| 2302 '\n' |
| 2303 | exp '\n' @{ printf ("\t%.10g\n", $1); @} |
| 2304 | error '\n' @{ yyerrok; @} |
| 2305 ; |
| 2306 @end group |
| 2307 |
| 2308 @group |
| 2309 exp: NUM @{ $$ = $1; @} |
| 2310 | VAR @{ $$ = $1->value.var; @} |
| 2311 | VAR '=' exp @{ $$ = $3; $1->value.var = $3; @} |
| 2312 | FNCT '(' exp ')' @{ $$ = (*($1->value.fnctptr))($3); @} |
| 2313 | exp '+' exp @{ $$ = $1 + $3; @} |
| 2314 | exp '-' exp @{ $$ = $1 - $3; @} |
| 2315 | exp '*' exp @{ $$ = $1 * $3; @} |
| 2316 | exp '/' exp @{ $$ = $1 / $3; @} |
| 2317 | '-' exp %prec NEG @{ $$ = -$2; @} |
| 2318 | exp '^' exp @{ $$ = pow ($1, $3); @} |
| 2319 | '(' exp ')' @{ $$ = $2; @} |
| 2320 ; |
| 2321 @end group |
| 2322 /* End of grammar. */ |
| 2323 %% |
| 2324 @end smallexample |
| 2325 |
| 2326 @node Mfcalc Symbol Table |
| 2327 @subsection The @code{mfcalc} Symbol Table |
| 2328 @cindex symbol table example |
| 2329 |
| 2330 The multi-function calculator requires a symbol table to keep track of the |
| 2331 names and meanings of variables and functions. This doesn't affect the |
| 2332 grammar rules (except for the actions) or the Bison declarations, but it |
| 2333 requires some additional C functions for support. |
| 2334 |
| 2335 The symbol table itself consists of a linked list of records. Its |
| 2336 definition, which is kept in the header @file{calc.h}, is as follows. It |
| 2337 provides for either functions or variables to be placed in the table. |
| 2338 |
| 2339 @smallexample |
| 2340 @group |
| 2341 /* Function type. */ |
| 2342 typedef double (*func_t) (double); |
| 2343 @end group |
| 2344 |
| 2345 @group |
| 2346 /* Data type for links in the chain of symbols. */ |
| 2347 struct symrec |
| 2348 @{ |
| 2349 char *name; /* name of symbol */ |
| 2350 int type; /* type of symbol: either VAR or FNCT */ |
| 2351 union |
| 2352 @{ |
| 2353 double var; /* value of a VAR */ |
| 2354 func_t fnctptr; /* value of a FNCT */ |
| 2355 @} value; |
| 2356 struct symrec *next; /* link field */ |
| 2357 @}; |
| 2358 @end group |
| 2359 |
| 2360 @group |
| 2361 typedef struct symrec symrec; |
| 2362 |
| 2363 /* The symbol table: a chain of `struct symrec'. */ |
| 2364 extern symrec *sym_table; |
| 2365 |
| 2366 symrec *putsym (char const *, int); |
| 2367 symrec *getsym (char const *); |
| 2368 @end group |
| 2369 @end smallexample |
| 2370 |
| 2371 The new version of @code{main} includes a call to @code{init_table}, a |
| 2372 function that initializes the symbol table. Here it is, and |
| 2373 @code{init_table} as well: |
| 2374 |
| 2375 @smallexample |
| 2376 #include <stdio.h> |
| 2377 |
| 2378 @group |
| 2379 /* Called by yyparse on error. */ |
| 2380 void |
| 2381 yyerror (char const *s) |
| 2382 @{ |
| 2383 printf ("%s\n", s); |
| 2384 @} |
| 2385 @end group |
| 2386 |
| 2387 @group |
| 2388 struct init |
| 2389 @{ |
| 2390 char const *fname; |
| 2391 double (*fnct) (double); |
| 2392 @}; |
| 2393 @end group |
| 2394 |
| 2395 @group |
| 2396 struct init const arith_fncts[] = |
| 2397 @{ |
| 2398 "sin", sin, |
| 2399 "cos", cos, |
| 2400 "atan", atan, |
| 2401 "ln", log, |
| 2402 "exp", exp, |
| 2403 "sqrt", sqrt, |
| 2404 0, 0 |
| 2405 @}; |
| 2406 @end group |
| 2407 |
| 2408 @group |
| 2409 /* The symbol table: a chain of `struct symrec'. */ |
| 2410 symrec *sym_table; |
| 2411 @end group |
| 2412 |
| 2413 @group |
| 2414 /* Put arithmetic functions in table. */ |
| 2415 void |
| 2416 init_table (void) |
| 2417 @{ |
| 2418 int i; |
| 2419 symrec *ptr; |
| 2420 for (i = 0; arith_fncts[i].fname != 0; i++) |
| 2421 @{ |
| 2422 ptr = putsym (arith_fncts[i].fname, FNCT); |
| 2423 ptr->value.fnctptr = arith_fncts[i].fnct; |
| 2424 @} |
| 2425 @} |
| 2426 @end group |
| 2427 |
| 2428 @group |
| 2429 int |
| 2430 main (void) |
| 2431 @{ |
| 2432 init_table (); |
| 2433 return yyparse (); |
| 2434 @} |
| 2435 @end group |
| 2436 @end smallexample |
| 2437 |
| 2438 By simply editing the initialization list and adding the necessary include |
| 2439 files, you can add additional functions to the calculator. |
| 2440 |
| 2441 Two important functions allow look-up and installation of symbols in the |
| 2442 symbol table. The function @code{putsym} is passed a name and the type |
| 2443 (@code{VAR} or @code{FNCT}) of the object to be installed. The object is |
| 2444 linked to the front of the list, and a pointer to the object is returned. |
| 2445 The function @code{getsym} is passed the name of the symbol to look up. If |
| 2446 found, a pointer to that symbol is returned; otherwise zero is returned. |
| 2447 |
| 2448 @smallexample |
| 2449 symrec * |
| 2450 putsym (char const *sym_name, int sym_type) |
| 2451 @{ |
| 2452 symrec *ptr; |
| 2453 ptr = (symrec *) malloc (sizeof (symrec)); |
| 2454 ptr->name = (char *) malloc (strlen (sym_name) + 1); |
| 2455 strcpy (ptr->name,sym_name); |
| 2456 ptr->type = sym_type; |
| 2457 ptr->value.var = 0; /* Set value to 0 even if fctn. */ |
| 2458 ptr->next = (struct symrec *)sym_table; |
| 2459 sym_table = ptr; |
| 2460 return ptr; |
| 2461 @} |
| 2462 |
| 2463 symrec * |
| 2464 getsym (char const *sym_name) |
| 2465 @{ |
| 2466 symrec *ptr; |
| 2467 for (ptr = sym_table; ptr != (symrec *) 0; |
| 2468 ptr = (symrec *)ptr->next) |
| 2469 if (strcmp (ptr->name,sym_name) == 0) |
| 2470 return ptr; |
| 2471 return 0; |
| 2472 @} |
| 2473 @end smallexample |
| 2474 |
| 2475 The function @code{yylex} must now recognize variables, numeric values, and |
| 2476 the single-character arithmetic operators. Strings of alphanumeric |
| 2477 characters with a leading letter are recognized as either variables or |
| 2478 functions depending on what the symbol table says about them. |
| 2479 |
| 2480 The string is passed to @code{getsym} for look up in the symbol table. If |
| 2481 the name appears in the table, a pointer to its location and its type |
| 2482 (@code{VAR} or @code{FNCT}) is returned to @code{yyparse}. If it is not |
| 2483 already in the table, then it is installed as a @code{VAR} using |
| 2484 @code{putsym}. Again, a pointer and its type (which must be @code{VAR}) is |
| 2485 returned to @code{yyparse}. |
| 2486 |
| 2487 No change is needed in the handling of numeric values and arithmetic |
| 2488 operators in @code{yylex}. |
| 2489 |
| 2490 @smallexample |
| 2491 @group |
| 2492 #include <ctype.h> |
| 2493 @end group |
| 2494 |
| 2495 @group |
| 2496 int |
| 2497 yylex (void) |
| 2498 @{ |
| 2499 int c; |
| 2500 |
| 2501 /* Ignore white space, get first nonwhite character. */ |
| 2502 while ((c = getchar ()) == ' ' || c == '\t'); |
| 2503 |
| 2504 if (c == EOF) |
| 2505 return 0; |
| 2506 @end group |
| 2507 |
| 2508 @group |
| 2509 /* Char starts a number => parse the number. */ |
| 2510 if (c == '.' || isdigit (c)) |
| 2511 @{ |
| 2512 ungetc (c, stdin); |
| 2513 scanf ("%lf", &yylval.val); |
| 2514 return NUM; |
| 2515 @} |
| 2516 @end group |
| 2517 |
| 2518 @group |
| 2519 /* Char starts an identifier => read the name. */ |
| 2520 if (isalpha (c)) |
| 2521 @{ |
| 2522 symrec *s; |
| 2523 static char *symbuf = 0; |
| 2524 static int length = 0; |
| 2525 int i; |
| 2526 @end group |
| 2527 |
| 2528 @group |
| 2529 /* Initially make the buffer long enough |
| 2530 for a 40-character symbol name. */ |
| 2531 if (length == 0) |
| 2532 length = 40, symbuf = (char *)malloc (length + 1); |
| 2533 |
| 2534 i = 0; |
| 2535 do |
| 2536 @end group |
| 2537 @group |
| 2538 @{ |
| 2539 /* If buffer is full, make it bigger. */ |
| 2540 if (i == length) |
| 2541 @{ |
| 2542 length *= 2; |
| 2543 symbuf = (char *) realloc (symbuf, length + 1); |
| 2544 @} |
| 2545 /* Add this character to the buffer. */ |
| 2546 symbuf[i++] = c; |
| 2547 /* Get another character. */ |
| 2548 c = getchar (); |
| 2549 @} |
| 2550 @end group |
| 2551 @group |
| 2552 while (isalnum (c)); |
| 2553 |
| 2554 ungetc (c, stdin); |
| 2555 symbuf[i] = '\0'; |
| 2556 @end group |
| 2557 |
| 2558 @group |
| 2559 s = getsym (symbuf); |
| 2560 if (s == 0) |
| 2561 s = putsym (symbuf, VAR); |
| 2562 yylval.tptr = s; |
| 2563 return s->type; |
| 2564 @} |
| 2565 |
| 2566 /* Any other character is a token by itself. */ |
| 2567 return c; |
| 2568 @} |
| 2569 @end group |
| 2570 @end smallexample |
| 2571 |
| 2572 This program is both powerful and flexible. You may easily add new |
| 2573 functions, and it is a simple job to modify this code to install |
| 2574 predefined variables such as @code{pi} or @code{e} as well. |
| 2575 |
| 2576 @node Exercises |
| 2577 @section Exercises |
| 2578 @cindex exercises |
| 2579 |
| 2580 @enumerate |
| 2581 @item |
| 2582 Add some new functions from @file{math.h} to the initialization list. |
| 2583 |
| 2584 @item |
| 2585 Add another array that contains constants and their values. Then |
| 2586 modify @code{init_table} to add these constants to the symbol table. |
| 2587 It will be easiest to give the constants type @code{VAR}. |
| 2588 |
| 2589 @item |
| 2590 Make the program report an error if the user refers to an |
| 2591 uninitialized variable in any way except to store a value in it. |
| 2592 @end enumerate |
| 2593 |
| 2594 @node Grammar File |
| 2595 @chapter Bison Grammar Files |
| 2596 |
| 2597 Bison takes as input a context-free grammar specification and produces a |
| 2598 C-language function that recognizes correct instances of the grammar. |
| 2599 |
| 2600 The Bison grammar input file conventionally has a name ending in @samp{.y}. |
| 2601 @xref{Invocation, ,Invoking Bison}. |
| 2602 |
| 2603 @menu |
| 2604 * Grammar Outline:: Overall layout of the grammar file. |
| 2605 * Symbols:: Terminal and nonterminal symbols. |
| 2606 * Rules:: How to write grammar rules. |
| 2607 * Recursion:: Writing recursive rules. |
| 2608 * Semantics:: Semantic values and actions. |
| 2609 * Locations:: Locations and actions. |
| 2610 * Declarations:: All kinds of Bison declarations are described here. |
| 2611 * Multiple Parsers:: Putting more than one Bison parser in one program. |
| 2612 @end menu |
| 2613 |
| 2614 @node Grammar Outline |
| 2615 @section Outline of a Bison Grammar |
| 2616 |
| 2617 A Bison grammar file has four main sections, shown here with the |
| 2618 appropriate delimiters: |
| 2619 |
| 2620 @example |
| 2621 %@{ |
| 2622 @var{Prologue} |
| 2623 %@} |
| 2624 |
| 2625 @var{Bison declarations} |
| 2626 |
| 2627 %% |
| 2628 @var{Grammar rules} |
| 2629 %% |
| 2630 |
| 2631 @var{Epilogue} |
| 2632 @end example |
| 2633 |
| 2634 Comments enclosed in @samp{/* @dots{} */} may appear in any of the sections. |
| 2635 As a @acronym{GNU} extension, @samp{//} introduces a comment that |
| 2636 continues until end of line. |
| 2637 |
| 2638 @menu |
| 2639 * Prologue:: Syntax and usage of the prologue. |
| 2640 * Prologue Alternatives:: Syntax and usage of alternatives to the prologue. |
| 2641 * Bison Declarations:: Syntax and usage of the Bison declarations section. |
| 2642 * Grammar Rules:: Syntax and usage of the grammar rules section. |
| 2643 * Epilogue:: Syntax and usage of the epilogue. |
| 2644 @end menu |
| 2645 |
| 2646 @node Prologue |
| 2647 @subsection The prologue |
| 2648 @cindex declarations section |
| 2649 @cindex Prologue |
| 2650 @cindex declarations |
| 2651 |
| 2652 The @var{Prologue} section contains macro definitions and declarations |
| 2653 of functions and variables that are used in the actions in the grammar |
| 2654 rules. These are copied to the beginning of the parser file so that |
| 2655 they precede the definition of @code{yyparse}. You can use |
| 2656 @samp{#include} to get the declarations from a header file. If you |
| 2657 don't need any C declarations, you may omit the @samp{%@{} and |
| 2658 @samp{%@}} delimiters that bracket this section. |
| 2659 |
| 2660 The @var{Prologue} section is terminated by the first occurrence |
| 2661 of @samp{%@}} that is outside a comment, a string literal, or a |
| 2662 character constant. |
| 2663 |
| 2664 You may have more than one @var{Prologue} section, intermixed with the |
| 2665 @var{Bison declarations}. This allows you to have C and Bison |
| 2666 declarations that refer to each other. For example, the @code{%union} |
| 2667 declaration may use types defined in a header file, and you may wish to |
| 2668 prototype functions that take arguments of type @code{YYSTYPE}. This |
| 2669 can be done with two @var{Prologue} blocks, one before and one after the |
| 2670 @code{%union} declaration. |
| 2671 |
| 2672 @smallexample |
| 2673 %@{ |
| 2674 #define _GNU_SOURCE |
| 2675 #include <stdio.h> |
| 2676 #include "ptypes.h" |
| 2677 %@} |
| 2678 |
| 2679 %union @{ |
| 2680 long int n; |
| 2681 tree t; /* @r{@code{tree} is defined in @file{ptypes.h}.} */ |
| 2682 @} |
| 2683 |
| 2684 %@{ |
| 2685 static void print_token_value (FILE *, int, YYSTYPE); |
| 2686 #define YYPRINT(F, N, L) print_token_value (F, N, L) |
| 2687 %@} |
| 2688 |
| 2689 @dots{} |
| 2690 @end smallexample |
| 2691 |
| 2692 When in doubt, it is usually safer to put prologue code before all |
| 2693 Bison declarations, rather than after. For example, any definitions |
| 2694 of feature test macros like @code{_GNU_SOURCE} or |
| 2695 @code{_POSIX_C_SOURCE} should appear before all Bison declarations, as |
| 2696 feature test macros can affect the behavior of Bison-generated |
| 2697 @code{#include} directives. |
| 2698 |
| 2699 @node Prologue Alternatives |
| 2700 @subsection Prologue Alternatives |
| 2701 @cindex Prologue Alternatives |
| 2702 |
| 2703 @findex %code |
| 2704 @findex %code requires |
| 2705 @findex %code provides |
| 2706 @findex %code top |
| 2707 (The prologue alternatives described here are experimental. |
| 2708 More user feedback will help to determine whether they should become permanent |
| 2709 features.) |
| 2710 |
| 2711 The functionality of @var{Prologue} sections can often be subtle and |
| 2712 inflexible. |
| 2713 As an alternative, Bison provides a %code directive with an explicit qualifier |
| 2714 field, which identifies the purpose of the code and thus the location(s) where |
| 2715 Bison should generate it. |
| 2716 For C/C++, the qualifier can be omitted for the default location, or it can be |
| 2717 one of @code{requires}, @code{provides}, @code{top}. |
| 2718 @xref{Decl Summary,,%code}. |
| 2719 |
| 2720 Look again at the example of the previous section: |
| 2721 |
| 2722 @smallexample |
| 2723 %@{ |
| 2724 #define _GNU_SOURCE |
| 2725 #include <stdio.h> |
| 2726 #include "ptypes.h" |
| 2727 %@} |
| 2728 |
| 2729 %union @{ |
| 2730 long int n; |
| 2731 tree t; /* @r{@code{tree} is defined in @file{ptypes.h}.} */ |
| 2732 @} |
| 2733 |
| 2734 %@{ |
| 2735 static void print_token_value (FILE *, int, YYSTYPE); |
| 2736 #define YYPRINT(F, N, L) print_token_value (F, N, L) |
| 2737 %@} |
| 2738 |
| 2739 @dots{} |
| 2740 @end smallexample |
| 2741 |
| 2742 @noindent |
| 2743 Notice that there are two @var{Prologue} sections here, but there's a subtle |
| 2744 distinction between their functionality. |
| 2745 For example, if you decide to override Bison's default definition for |
| 2746 @code{YYLTYPE}, in which @var{Prologue} section should you write your new |
| 2747 definition? |
| 2748 You should write it in the first since Bison will insert that code into the |
| 2749 parser source code file @emph{before} the default @code{YYLTYPE} definition. |
| 2750 In which @var{Prologue} section should you prototype an internal function, |
| 2751 @code{trace_token}, that accepts @code{YYLTYPE} and @code{yytokentype} as |
| 2752 arguments? |
| 2753 You should prototype it in the second since Bison will insert that code |
| 2754 @emph{after} the @code{YYLTYPE} and @code{yytokentype} definitions. |
| 2755 |
| 2756 This distinction in functionality between the two @var{Prologue} sections is |
| 2757 established by the appearance of the @code{%union} between them. |
| 2758 This behavior raises a few questions. |
| 2759 First, why should the position of a @code{%union} affect definitions related to |
| 2760 @code{YYLTYPE} and @code{yytokentype}? |
| 2761 Second, what if there is no @code{%union}? |
| 2762 In that case, the second kind of @var{Prologue} section is not available. |
| 2763 This behavior is not intuitive. |
| 2764 |
| 2765 To avoid this subtle @code{%union} dependency, rewrite the example using a |
| 2766 @code{%code top} and an unqualified @code{%code}. |
| 2767 Let's go ahead and add the new @code{YYLTYPE} definition and the |
| 2768 @code{trace_token} prototype at the same time: |
| 2769 |
| 2770 @smallexample |
| 2771 %code top @{ |
| 2772 #define _GNU_SOURCE |
| 2773 #include <stdio.h> |
| 2774 |
| 2775 /* WARNING: The following code really belongs |
| 2776 * in a `%code requires'; see below. */ |
| 2777 |
| 2778 #include "ptypes.h" |
| 2779 #define YYLTYPE YYLTYPE |
| 2780 typedef struct YYLTYPE |
| 2781 @{ |
| 2782 int first_line; |
| 2783 int first_column; |
| 2784 int last_line; |
| 2785 int last_column; |
| 2786 char *filename; |
| 2787 @} YYLTYPE; |
| 2788 @} |
| 2789 |
| 2790 %union @{ |
| 2791 long int n; |
| 2792 tree t; /* @r{@code{tree} is defined in @file{ptypes.h}.} */ |
| 2793 @} |
| 2794 |
| 2795 %code @{ |
| 2796 static void print_token_value (FILE *, int, YYSTYPE); |
| 2797 #define YYPRINT(F, N, L) print_token_value (F, N, L) |
| 2798 static void trace_token (enum yytokentype token, YYLTYPE loc); |
| 2799 @} |
| 2800 |
| 2801 @dots{} |
| 2802 @end smallexample |
| 2803 |
| 2804 @noindent |
| 2805 In this way, @code{%code top} and the unqualified @code{%code} achieve the same |
| 2806 functionality as the two kinds of @var{Prologue} sections, but it's always |
| 2807 explicit which kind you intend. |
| 2808 Moreover, both kinds are always available even in the absence of @code{%union}. |
| 2809 |
| 2810 The @code{%code top} block above logically contains two parts. |
| 2811 The first two lines before the warning need to appear near the top of the |
| 2812 parser source code file. |
| 2813 The first line after the warning is required by @code{YYSTYPE} and thus also |
| 2814 needs to appear in the parser source code file. |
| 2815 However, if you've instructed Bison to generate a parser header file |
| 2816 (@pxref{Decl Summary, ,%defines}), you probably want that line to appear before |
| 2817 the @code{YYSTYPE} definition in that header file as well. |
| 2818 The @code{YYLTYPE} definition should also appear in the parser header file to |
| 2819 override the default @code{YYLTYPE} definition there. |
| 2820 |
| 2821 In other words, in the @code{%code top} block above, all but the first two |
| 2822 lines are dependency code required by the @code{YYSTYPE} and @code{YYLTYPE} |
| 2823 definitions. |
| 2824 Thus, they belong in one or more @code{%code requires}: |
| 2825 |
| 2826 @smallexample |
| 2827 %code top @{ |
| 2828 #define _GNU_SOURCE |
| 2829 #include <stdio.h> |
| 2830 @} |
| 2831 |
| 2832 %code requires @{ |
| 2833 #include "ptypes.h" |
| 2834 @} |
| 2835 %union @{ |
| 2836 long int n; |
| 2837 tree t; /* @r{@code{tree} is defined in @file{ptypes.h}.} */ |
| 2838 @} |
| 2839 |
| 2840 %code requires @{ |
| 2841 #define YYLTYPE YYLTYPE |
| 2842 typedef struct YYLTYPE |
| 2843 @{ |
| 2844 int first_line; |
| 2845 int first_column; |
| 2846 int last_line; |
| 2847 int last_column; |
| 2848 char *filename; |
| 2849 @} YYLTYPE; |
| 2850 @} |
| 2851 |
| 2852 %code @{ |
| 2853 static void print_token_value (FILE *, int, YYSTYPE); |
| 2854 #define YYPRINT(F, N, L) print_token_value (F, N, L) |
| 2855 static void trace_token (enum yytokentype token, YYLTYPE loc); |
| 2856 @} |
| 2857 |
| 2858 @dots{} |
| 2859 @end smallexample |
| 2860 |
| 2861 @noindent |
| 2862 Now Bison will insert @code{#include "ptypes.h"} and the new @code{YYLTYPE} |
| 2863 definition before the Bison-generated @code{YYSTYPE} and @code{YYLTYPE} |
| 2864 definitions in both the parser source code file and the parser header file. |
| 2865 (By the same reasoning, @code{%code requires} would also be the appropriate |
| 2866 place to write your own definition for @code{YYSTYPE}.) |
| 2867 |
| 2868 When you are writing dependency code for @code{YYSTYPE} and @code{YYLTYPE}, you |
| 2869 should prefer @code{%code requires} over @code{%code top} regardless of whether |
| 2870 you instruct Bison to generate a parser header file. |
| 2871 When you are writing code that you need Bison to insert only into the parser |
| 2872 source code file and that has no special need to appear at the top of that |
| 2873 file, you should prefer the unqualified @code{%code} over @code{%code top}. |
| 2874 These practices will make the purpose of each block of your code explicit to |
| 2875 Bison and to other developers reading your grammar file. |
| 2876 Following these practices, we expect the unqualified @code{%code} and |
| 2877 @code{%code requires} to be the most important of the four @var{Prologue} |
| 2878 alternatives. |
| 2879 |
| 2880 At some point while developing your parser, you might decide to provide |
| 2881 @code{trace_token} to modules that are external to your parser. |
| 2882 Thus, you might wish for Bison to insert the prototype into both the parser |
| 2883 header file and the parser source code file. |
| 2884 Since this function is not a dependency required by @code{YYSTYPE} or |
| 2885 @code{YYLTYPE}, it doesn't make sense to move its prototype to a |
| 2886 @code{%code requires}. |
| 2887 More importantly, since it depends upon @code{YYLTYPE} and @code{yytokentype}, |
| 2888 @code{%code requires} is not sufficient. |
| 2889 Instead, move its prototype from the unqualified @code{%code} to a |
| 2890 @code{%code provides}: |
| 2891 |
| 2892 @smallexample |
| 2893 %code top @{ |
| 2894 #define _GNU_SOURCE |
| 2895 #include <stdio.h> |
| 2896 @} |
| 2897 |
| 2898 %code requires @{ |
| 2899 #include "ptypes.h" |
| 2900 @} |
| 2901 %union @{ |
| 2902 long int n; |
| 2903 tree t; /* @r{@code{tree} is defined in @file{ptypes.h}.} */ |
| 2904 @} |
| 2905 |
| 2906 %code requires @{ |
| 2907 #define YYLTYPE YYLTYPE |
| 2908 typedef struct YYLTYPE |
| 2909 @{ |
| 2910 int first_line; |
| 2911 int first_column; |
| 2912 int last_line; |
| 2913 int last_column; |
| 2914 char *filename; |
| 2915 @} YYLTYPE; |
| 2916 @} |
| 2917 |
| 2918 %code provides @{ |
| 2919 void trace_token (enum yytokentype token, YYLTYPE loc); |
| 2920 @} |
| 2921 |
| 2922 %code @{ |
| 2923 static void print_token_value (FILE *, int, YYSTYPE); |
| 2924 #define YYPRINT(F, N, L) print_token_value (F, N, L) |
| 2925 @} |
| 2926 |
| 2927 @dots{} |
| 2928 @end smallexample |
| 2929 |
| 2930 @noindent |
| 2931 Bison will insert the @code{trace_token} prototype into both the parser header |
| 2932 file and the parser source code file after the definitions for |
| 2933 @code{yytokentype}, @code{YYLTYPE}, and @code{YYSTYPE}. |
| 2934 |
| 2935 The above examples are careful to write directives in an order that reflects |
| 2936 the layout of the generated parser source code and header files: |
| 2937 @code{%code top}, @code{%code requires}, @code{%code provides}, and then |
| 2938 @code{%code}. |
| 2939 While your grammar files may generally be easier to read if you also follow |
| 2940 this order, Bison does not require it. |
| 2941 Instead, Bison lets you choose an organization that makes sense to you. |
| 2942 |
| 2943 You may declare any of these directives multiple times in the grammar file. |
| 2944 In that case, Bison concatenates the contained code in declaration order. |
| 2945 This is the only way in which the position of one of these directives within |
| 2946 the grammar file affects its functionality. |
| 2947 |
| 2948 The result of the previous two properties is greater flexibility in how you may |
| 2949 organize your grammar file. |
| 2950 For example, you may organize semantic-type-related directives by semantic |
| 2951 type: |
| 2952 |
| 2953 @smallexample |
| 2954 %code requires @{ #include "type1.h" @} |
| 2955 %union @{ type1 field1; @} |
| 2956 %destructor @{ type1_free ($$); @} <field1> |
| 2957 %printer @{ type1_print ($$); @} <field1> |
| 2958 |
| 2959 %code requires @{ #include "type2.h" @} |
| 2960 %union @{ type2 field2; @} |
| 2961 %destructor @{ type2_free ($$); @} <field2> |
| 2962 %printer @{ type2_print ($$); @} <field2> |
| 2963 @end smallexample |
| 2964 |
| 2965 @noindent |
| 2966 You could even place each of the above directive groups in the rules section of |
| 2967 the grammar file next to the set of rules that uses the associated semantic |
| 2968 type. |
| 2969 (In the rules section, you must terminate each of those directives with a |
| 2970 semicolon.) |
| 2971 And you don't have to worry that some directive (like a @code{%union}) in the |
| 2972 definitions section is going to adversely affect their functionality in some |
| 2973 counter-intuitive manner just because it comes first. |
| 2974 Such an organization is not possible using @var{Prologue} sections. |
| 2975 |
| 2976 This section has been concerned with explaining the advantages of the four |
| 2977 @var{Prologue} alternatives over the original Yacc @var{Prologue}. |
| 2978 However, in most cases when using these directives, you shouldn't need to |
| 2979 think about all the low-level ordering issues discussed here. |
| 2980 Instead, you should simply use these directives to label each block of your |
| 2981 code according to its purpose and let Bison handle the ordering. |
| 2982 @code{%code} is the most generic label. |
| 2983 Move code to @code{%code requires}, @code{%code provides}, or @code{%code top} |
| 2984 as needed. |
| 2985 |
| 2986 @node Bison Declarations |
| 2987 @subsection The Bison Declarations Section |
| 2988 @cindex Bison declarations (introduction) |
| 2989 @cindex declarations, Bison (introduction) |
| 2990 |
| 2991 The @var{Bison declarations} section contains declarations that define |
| 2992 terminal and nonterminal symbols, specify precedence, and so on. |
| 2993 In some simple grammars you may not need any declarations. |
| 2994 @xref{Declarations, ,Bison Declarations}. |
| 2995 |
| 2996 @node Grammar Rules |
| 2997 @subsection The Grammar Rules Section |
| 2998 @cindex grammar rules section |
| 2999 @cindex rules section for grammar |
| 3000 |
| 3001 The @dfn{grammar rules} section contains one or more Bison grammar |
| 3002 rules, and nothing else. @xref{Rules, ,Syntax of Grammar Rules}. |
| 3003 |
| 3004 There must always be at least one grammar rule, and the first |
| 3005 @samp{%%} (which precedes the grammar rules) may never be omitted even |
| 3006 if it is the first thing in the file. |
| 3007 |
| 3008 @node Epilogue |
| 3009 @subsection The epilogue |
| 3010 @cindex additional C code section |
| 3011 @cindex epilogue |
| 3012 @cindex C code, section for additional |
| 3013 |
| 3014 The @var{Epilogue} is copied verbatim to the end of the parser file, just as |
| 3015 the @var{Prologue} is copied to the beginning. This is the most convenient |
| 3016 place to put anything that you want to have in the parser file but which need |
| 3017 not come before the definition of @code{yyparse}. For example, the |
| 3018 definitions of @code{yylex} and @code{yyerror} often go here. Because |
| 3019 C requires functions to be declared before being used, you often need |
| 3020 to declare functions like @code{yylex} and @code{yyerror} in the Prologue, |
| 3021 even if you define them in the Epilogue. |
| 3022 @xref{Interface, ,Parser C-Language Interface}. |
| 3023 |
| 3024 If the last section is empty, you may omit the @samp{%%} that separates it |
| 3025 from the grammar rules. |
| 3026 |
| 3027 The Bison parser itself contains many macros and identifiers whose names |
| 3028 start with @samp{yy} or @samp{YY}, so it is a good idea to avoid using |
| 3029 any such names (except those documented in this manual) in the epilogue |
| 3030 of the grammar file. |
| 3031 |
| 3032 @node Symbols |
| 3033 @section Symbols, Terminal and Nonterminal |
| 3034 @cindex nonterminal symbol |
| 3035 @cindex terminal symbol |
| 3036 @cindex token type |
| 3037 @cindex symbol |
| 3038 |
| 3039 @dfn{Symbols} in Bison grammars represent the grammatical classifications |
| 3040 of the language. |
| 3041 |
| 3042 A @dfn{terminal symbol} (also known as a @dfn{token type}) represents a |
| 3043 class of syntactically equivalent tokens. You use the symbol in grammar |
| 3044 rules to mean that a token in that class is allowed. The symbol is |
| 3045 represented in the Bison parser by a numeric code, and the @code{yylex} |
| 3046 function returns a token type code to indicate what kind of token has |
| 3047 been read. You don't need to know what the code value is; you can use |
| 3048 the symbol to stand for it. |
| 3049 |
| 3050 A @dfn{nonterminal symbol} stands for a class of syntactically |
| 3051 equivalent groupings. The symbol name is used in writing grammar rules. |
| 3052 By convention, it should be all lower case. |
| 3053 |
| 3054 Symbol names can contain letters, digits (not at the beginning), |
| 3055 underscores and periods. Periods make sense only in nonterminals. |
| 3056 |
| 3057 There are three ways of writing terminal symbols in the grammar: |
| 3058 |
| 3059 @itemize @bullet |
| 3060 @item |
| 3061 A @dfn{named token type} is written with an identifier, like an |
| 3062 identifier in C@. By convention, it should be all upper case. Each |
| 3063 such name must be defined with a Bison declaration such as |
| 3064 @code{%token}. @xref{Token Decl, ,Token Type Names}. |
| 3065 |
| 3066 @item |
| 3067 @cindex character token |
| 3068 @cindex literal token |
| 3069 @cindex single-character literal |
| 3070 A @dfn{character token type} (or @dfn{literal character token}) is |
| 3071 written in the grammar using the same syntax used in C for character |
| 3072 constants; for example, @code{'+'} is a character token type. A |
| 3073 character token type doesn't need to be declared unless you need to |
| 3074 specify its semantic value data type (@pxref{Value Type, ,Data Types of |
| 3075 Semantic Values}), associativity, or precedence (@pxref{Precedence, |
| 3076 ,Operator Precedence}). |
| 3077 |
| 3078 By convention, a character token type is used only to represent a |
| 3079 token that consists of that particular character. Thus, the token |
| 3080 type @code{'+'} is used to represent the character @samp{+} as a |
| 3081 token. Nothing enforces this convention, but if you depart from it, |
| 3082 your program will confuse other readers. |
| 3083 |
| 3084 All the usual escape sequences used in character literals in C can be |
| 3085 used in Bison as well, but you must not use the null character as a |
| 3086 character literal because its numeric code, zero, signifies |
| 3087 end-of-input (@pxref{Calling Convention, ,Calling Convention |
| 3088 for @code{yylex}}). Also, unlike standard C, trigraphs have no |
| 3089 special meaning in Bison character literals, nor is backslash-newline |
| 3090 allowed. |
| 3091 |
| 3092 @item |
| 3093 @cindex string token |
| 3094 @cindex literal string token |
| 3095 @cindex multicharacter literal |
| 3096 A @dfn{literal string token} is written like a C string constant; for |
| 3097 example, @code{"<="} is a literal string token. A literal string token |
| 3098 doesn't need to be declared unless you need to specify its semantic |
| 3099 value data type (@pxref{Value Type}), associativity, or precedence |
| 3100 (@pxref{Precedence}). |
| 3101 |
| 3102 You can associate the literal string token with a symbolic name as an |
| 3103 alias, using the @code{%token} declaration (@pxref{Token Decl, ,Token |
| 3104 Declarations}). If you don't do that, the lexical analyzer has to |
| 3105 retrieve the token number for the literal string token from the |
| 3106 @code{yytname} table (@pxref{Calling Convention}). |
| 3107 |
| 3108 @strong{Warning}: literal string tokens do not work in Yacc. |
| 3109 |
| 3110 By convention, a literal string token is used only to represent a token |
| 3111 that consists of that particular string. Thus, you should use the token |
| 3112 type @code{"<="} to represent the string @samp{<=} as a token. Bison |
| 3113 does not enforce this convention, but if you depart from it, people who |
| 3114 read your program will be confused. |
| 3115 |
| 3116 All the escape sequences used in string literals in C can be used in |
| 3117 Bison as well, except that you must not use a null character within a |
| 3118 string literal. Also, unlike Standard C, trigraphs have no special |
| 3119 meaning in Bison string literals, nor is backslash-newline allowed. A |
| 3120 literal string token must contain two or more characters; for a token |
| 3121 containing just one character, use a character token (see above). |
| 3122 @end itemize |
| 3123 |
| 3124 How you choose to write a terminal symbol has no effect on its |
| 3125 grammatical meaning. That depends only on where it appears in rules and |
| 3126 on when the parser function returns that symbol. |
| 3127 |
| 3128 The value returned by @code{yylex} is always one of the terminal |
| 3129 symbols, except that a zero or negative value signifies end-of-input. |
| 3130 Whichever way you write the token type in the grammar rules, you write |
| 3131 it the same way in the definition of @code{yylex}. The numeric code |
| 3132 for a character token type is simply the positive numeric code of the |
| 3133 character, so @code{yylex} can use the identical value to generate the |
| 3134 requisite code, though you may need to convert it to @code{unsigned |
| 3135 char} to avoid sign-extension on hosts where @code{char} is signed. |
| 3136 Each named token type becomes a C macro in |
| 3137 the parser file, so @code{yylex} can use the name to stand for the code. |
| 3138 (This is why periods don't make sense in terminal symbols.) |
| 3139 @xref{Calling Convention, ,Calling Convention for @code{yylex}}. |
| 3140 |
| 3141 If @code{yylex} is defined in a separate file, you need to arrange for the |
| 3142 token-type macro definitions to be available there. Use the @samp{-d} |
| 3143 option when you run Bison, so that it will write these macro definitions |
| 3144 into a separate header file @file{@var{name}.tab.h} which you can include |
| 3145 in the other source files that need it. @xref{Invocation, ,Invoking Bison}. |
| 3146 |
| 3147 If you want to write a grammar that is portable to any Standard C |
| 3148 host, you must use only nonnull character tokens taken from the basic |
| 3149 execution character set of Standard C@. This set consists of the ten |
| 3150 digits, the 52 lower- and upper-case English letters, and the |
| 3151 characters in the following C-language string: |
| 3152 |
| 3153 @example |
| 3154 "\a\b\t\n\v\f\r !\"#%&'()*+,-./:;<=>?[\\]^_@{|@}~" |
| 3155 @end example |
| 3156 |
| 3157 The @code{yylex} function and Bison must use a consistent character set |
| 3158 and encoding for character tokens. For example, if you run Bison in an |
| 3159 @acronym{ASCII} environment, but then compile and run the resulting |
| 3160 program in an environment that uses an incompatible character set like |
| 3161 @acronym{EBCDIC}, the resulting program may not work because the tables |
| 3162 generated by Bison will assume @acronym{ASCII} numeric values for |
| 3163 character tokens. It is standard practice for software distributions to |
| 3164 contain C source files that were generated by Bison in an |
| 3165 @acronym{ASCII} environment, so installers on platforms that are |
| 3166 incompatible with @acronym{ASCII} must rebuild those files before |
| 3167 compiling them. |
| 3168 |
| 3169 The symbol @code{error} is a terminal symbol reserved for error recovery |
| 3170 (@pxref{Error Recovery}); you shouldn't use it for any other purpose. |
| 3171 In particular, @code{yylex} should never return this value. The default |
| 3172 value of the error token is 256, unless you explicitly assigned 256 to |
| 3173 one of your tokens with a @code{%token} declaration. |
| 3174 |
| 3175 @node Rules |
| 3176 @section Syntax of Grammar Rules |
| 3177 @cindex rule syntax |
| 3178 @cindex grammar rule syntax |
| 3179 @cindex syntax of grammar rules |
| 3180 |
| 3181 A Bison grammar rule has the following general form: |
| 3182 |
| 3183 @example |
| 3184 @group |
| 3185 @var{result}: @var{components}@dots{} |
| 3186 ; |
| 3187 @end group |
| 3188 @end example |
| 3189 |
| 3190 @noindent |
| 3191 where @var{result} is the nonterminal symbol that this rule describes, |
| 3192 and @var{components} are various terminal and nonterminal symbols that |
| 3193 are put together by this rule (@pxref{Symbols}). |
| 3194 |
| 3195 For example, |
| 3196 |
| 3197 @example |
| 3198 @group |
| 3199 exp: exp '+' exp |
| 3200 ; |
| 3201 @end group |
| 3202 @end example |
| 3203 |
| 3204 @noindent |
| 3205 says that two groupings of type @code{exp}, with a @samp{+} token in between, |
| 3206 can be combined into a larger grouping of type @code{exp}. |
| 3207 |
| 3208 White space in rules is significant only to separate symbols. You can add |
| 3209 extra white space as you wish. |
| 3210 |
| 3211 Scattered among the components can be @var{actions} that determine |
| 3212 the semantics of the rule. An action looks like this: |
| 3213 |
| 3214 @example |
| 3215 @{@var{C statements}@} |
| 3216 @end example |
| 3217 |
| 3218 @noindent |
| 3219 @cindex braced code |
| 3220 This is an example of @dfn{braced code}, that is, C code surrounded by |
| 3221 braces, much like a compound statement in C@. Braced code can contain |
| 3222 any sequence of C tokens, so long as its braces are balanced. Bison |
| 3223 does not check the braced code for correctness directly; it merely |
| 3224 copies the code to the output file, where the C compiler can check it. |
| 3225 |
| 3226 Within braced code, the balanced-brace count is not affected by braces |
| 3227 within comments, string literals, or character constants, but it is |
| 3228 affected by the C digraphs @samp{<%} and @samp{%>} that represent |
| 3229 braces. At the top level braced code must be terminated by @samp{@}} |
| 3230 and not by a digraph. Bison does not look for trigraphs, so if braced |
| 3231 code uses trigraphs you should ensure that they do not affect the |
| 3232 nesting of braces or the boundaries of comments, string literals, or |
| 3233 character constants. |
| 3234 |
| 3235 Usually there is only one action and it follows the components. |
| 3236 @xref{Actions}. |
| 3237 |
| 3238 @findex | |
| 3239 Multiple rules for the same @var{result} can be written separately or can |
| 3240 be joined with the vertical-bar character @samp{|} as follows: |
| 3241 |
| 3242 @example |
| 3243 @group |
| 3244 @var{result}: @var{rule1-components}@dots{} |
| 3245 | @var{rule2-components}@dots{} |
| 3246 @dots{} |
| 3247 ; |
| 3248 @end group |
| 3249 @end example |
| 3250 |
| 3251 @noindent |
| 3252 They are still considered distinct rules even when joined in this way. |
| 3253 |
| 3254 If @var{components} in a rule is empty, it means that @var{result} can |
| 3255 match the empty string. For example, here is how to define a |
| 3256 comma-separated sequence of zero or more @code{exp} groupings: |
| 3257 |
| 3258 @example |
| 3259 @group |
| 3260 expseq: /* empty */ |
| 3261 | expseq1 |
| 3262 ; |
| 3263 @end group |
| 3264 |
| 3265 @group |
| 3266 expseq1: exp |
| 3267 | expseq1 ',' exp |
| 3268 ; |
| 3269 @end group |
| 3270 @end example |
| 3271 |
| 3272 @noindent |
| 3273 It is customary to write a comment @samp{/* empty */} in each rule |
| 3274 with no components. |
| 3275 |
| 3276 @node Recursion |
| 3277 @section Recursive Rules |
| 3278 @cindex recursive rule |
| 3279 |
| 3280 A rule is called @dfn{recursive} when its @var{result} nonterminal |
| 3281 appears also on its right hand side. Nearly all Bison grammars need to |
| 3282 use recursion, because that is the only way to define a sequence of any |
| 3283 number of a particular thing. Consider this recursive definition of a |
| 3284 comma-separated sequence of one or more expressions: |
| 3285 |
| 3286 @example |
| 3287 @group |
| 3288 expseq1: exp |
| 3289 | expseq1 ',' exp |
| 3290 ; |
| 3291 @end group |
| 3292 @end example |
| 3293 |
| 3294 @cindex left recursion |
| 3295 @cindex right recursion |
| 3296 @noindent |
| 3297 Since the recursive use of @code{expseq1} is the leftmost symbol in the |
| 3298 right hand side, we call this @dfn{left recursion}. By contrast, here |
| 3299 the same construct is defined using @dfn{right recursion}: |
| 3300 |
| 3301 @example |
| 3302 @group |
| 3303 expseq1: exp |
| 3304 | exp ',' expseq1 |
| 3305 ; |
| 3306 @end group |
| 3307 @end example |
| 3308 |
| 3309 @noindent |
| 3310 Any kind of sequence can be defined using either left recursion or right |
| 3311 recursion, but you should always use left recursion, because it can |
| 3312 parse a sequence of any number of elements with bounded stack space. |
| 3313 Right recursion uses up space on the Bison stack in proportion to the |
| 3314 number of elements in the sequence, because all the elements must be |
| 3315 shifted onto the stack before the rule can be applied even once. |
| 3316 @xref{Algorithm, ,The Bison Parser Algorithm}, for further explanation |
| 3317 of this. |
| 3318 |
| 3319 @cindex mutual recursion |
| 3320 @dfn{Indirect} or @dfn{mutual} recursion occurs when the result of the |
| 3321 rule does not appear directly on its right hand side, but does appear |
| 3322 in rules for other nonterminals which do appear on its right hand |
| 3323 side. |
| 3324 |
| 3325 For example: |
| 3326 |
| 3327 @example |
| 3328 @group |
| 3329 expr: primary |
| 3330 | primary '+' primary |
| 3331 ; |
| 3332 @end group |
| 3333 |
| 3334 @group |
| 3335 primary: constant |
| 3336 | '(' expr ')' |
| 3337 ; |
| 3338 @end group |
| 3339 @end example |
| 3340 |
| 3341 @noindent |
| 3342 defines two mutually-recursive nonterminals, since each refers to the |
| 3343 other. |
| 3344 |
| 3345 @node Semantics |
| 3346 @section Defining Language Semantics |
| 3347 @cindex defining language semantics |
| 3348 @cindex language semantics, defining |
| 3349 |
| 3350 The grammar rules for a language determine only the syntax. The semantics |
| 3351 are determined by the semantic values associated with various tokens and |
| 3352 groupings, and by the actions taken when various groupings are recognized. |
| 3353 |
| 3354 For example, the calculator calculates properly because the value |
| 3355 associated with each expression is the proper number; it adds properly |
| 3356 because the action for the grouping @w{@samp{@var{x} + @var{y}}} is to add |
| 3357 the numbers associated with @var{x} and @var{y}. |
| 3358 |
| 3359 @menu |
| 3360 * Value Type:: Specifying one data type for all semantic values. |
| 3361 * Multiple Types:: Specifying several alternative data types. |
| 3362 * Actions:: An action is the semantic definition of a grammar rule. |
| 3363 * Action Types:: Specifying data types for actions to operate on. |
| 3364 * Mid-Rule Actions:: Most actions go at the end of a rule. |
| 3365 This says when, why and how to use the exceptional |
| 3366 action in the middle of a rule. |
| 3367 @end menu |
| 3368 |
| 3369 @node Value Type |
| 3370 @subsection Data Types of Semantic Values |
| 3371 @cindex semantic value type |
| 3372 @cindex value type, semantic |
| 3373 @cindex data types of semantic values |
| 3374 @cindex default data type |
| 3375 |
| 3376 In a simple program it may be sufficient to use the same data type for |
| 3377 the semantic values of all language constructs. This was true in the |
| 3378 @acronym{RPN} and infix calculator examples (@pxref{RPN Calc, ,Reverse Polish |
| 3379 Notation Calculator}). |
| 3380 |
| 3381 Bison normally uses the type @code{int} for semantic values if your |
| 3382 program uses the same data type for all language constructs. To |
| 3383 specify some other type, define @code{YYSTYPE} as a macro, like this: |
| 3384 |
| 3385 @example |
| 3386 #define YYSTYPE double |
| 3387 @end example |
| 3388 |
| 3389 @noindent |
| 3390 @code{YYSTYPE}'s replacement list should be a type name |
| 3391 that does not contain parentheses or square brackets. |
| 3392 This macro definition must go in the prologue of the grammar file |
| 3393 (@pxref{Grammar Outline, ,Outline of a Bison Grammar}). |
| 3394 |
| 3395 @node Multiple Types |
| 3396 @subsection More Than One Value Type |
| 3397 |
| 3398 In most programs, you will need different data types for different kinds |
| 3399 of tokens and groupings. For example, a numeric constant may need type |
| 3400 @code{int} or @code{long int}, while a string constant needs type |
| 3401 @code{char *}, and an identifier might need a pointer to an entry in the |
| 3402 symbol table. |
| 3403 |
| 3404 To use more than one data type for semantic values in one parser, Bison |
| 3405 requires you to do two things: |
| 3406 |
| 3407 @itemize @bullet |
| 3408 @item |
| 3409 Specify the entire collection of possible data types, either by using the |
| 3410 @code{%union} Bison declaration (@pxref{Union Decl, ,The Collection of |
| 3411 Value Types}), or by using a @code{typedef} or a @code{#define} to |
| 3412 define @code{YYSTYPE} to be a union type whose member names are |
| 3413 the type tags. |
| 3414 |
| 3415 @item |
| 3416 Choose one of those types for each symbol (terminal or nonterminal) for |
| 3417 which semantic values are used. This is done for tokens with the |
| 3418 @code{%token} Bison declaration (@pxref{Token Decl, ,Token Type Names}) |
| 3419 and for groupings with the @code{%type} Bison declaration (@pxref{Type |
| 3420 Decl, ,Nonterminal Symbols}). |
| 3421 @end itemize |
| 3422 |
| 3423 @node Actions |
| 3424 @subsection Actions |
| 3425 @cindex action |
| 3426 @vindex $$ |
| 3427 @vindex $@var{n} |
| 3428 |
| 3429 An action accompanies a syntactic rule and contains C code to be executed |
| 3430 each time an instance of that rule is recognized. The task of most actions |
| 3431 is to compute a semantic value for the grouping built by the rule from the |
| 3432 semantic values associated with tokens or smaller groupings. |
| 3433 |
| 3434 An action consists of braced code containing C statements, and can be |
| 3435 placed at any position in the rule; |
| 3436 it is executed at that position. Most rules have just one action at the |
| 3437 end of the rule, following all the components. Actions in the middle of |
| 3438 a rule are tricky and used only for special purposes (@pxref{Mid-Rule |
| 3439 Actions, ,Actions in Mid-Rule}). |
| 3440 |
| 3441 The C code in an action can refer to the semantic values of the components |
| 3442 matched by the rule with the construct @code{$@var{n}}, which stands for |
| 3443 the value of the @var{n}th component. The semantic value for the grouping |
| 3444 being constructed is @code{$$}. Bison translates both of these |
| 3445 constructs into expressions of the appropriate type when it copies the |
| 3446 actions into the parser file. @code{$$} is translated to a modifiable |
| 3447 lvalue, so it can be assigned to. |
| 3448 |
| 3449 Here is a typical example: |
| 3450 |
| 3451 @example |
| 3452 @group |
| 3453 exp: @dots{} |
| 3454 | exp '+' exp |
| 3455 @{ $$ = $1 + $3; @} |
| 3456 @end group |
| 3457 @end example |
| 3458 |
| 3459 @noindent |
| 3460 This rule constructs an @code{exp} from two smaller @code{exp} groupings |
| 3461 connected by a plus-sign token. In the action, @code{$1} and @code{$3} |
| 3462 refer to the semantic values of the two component @code{exp} groupings, |
| 3463 which are the first and third symbols on the right hand side of the rule. |
| 3464 The sum is stored into @code{$$} so that it becomes the semantic value of |
| 3465 the addition-expression just recognized by the rule. If there were a |
| 3466 useful semantic value associated with the @samp{+} token, it could be |
| 3467 referred to as @code{$2}. |
| 3468 |
| 3469 Note that the vertical-bar character @samp{|} is really a rule |
| 3470 separator, and actions are attached to a single rule. This is a |
| 3471 difference with tools like Flex, for which @samp{|} stands for either |
| 3472 ``or'', or ``the same action as that of the next rule''. In the |
| 3473 following example, the action is triggered only when @samp{b} is found: |
| 3474 |
| 3475 @example |
| 3476 @group |
| 3477 a-or-b: 'a'|'b' @{ a_or_b_found = 1; @}; |
| 3478 @end group |
| 3479 @end example |
| 3480 |
| 3481 @cindex default action |
| 3482 If you don't specify an action for a rule, Bison supplies a default: |
| 3483 @w{@code{$$ = $1}.} Thus, the value of the first symbol in the rule |
| 3484 becomes the value of the whole rule. Of course, the default action is |
| 3485 valid only if the two data types match. There is no meaningful default |
| 3486 action for an empty rule; every empty rule must have an explicit action |
| 3487 unless the rule's value does not matter. |
| 3488 |
| 3489 @code{$@var{n}} with @var{n} zero or negative is allowed for reference |
| 3490 to tokens and groupings on the stack @emph{before} those that match the |
| 3491 current rule. This is a very risky practice, and to use it reliably |
| 3492 you must be certain of the context in which the rule is applied. Here |
| 3493 is a case in which you can use this reliably: |
| 3494 |
| 3495 @example |
| 3496 @group |
| 3497 foo: expr bar '+' expr @{ @dots{} @} |
| 3498 | expr bar '-' expr @{ @dots{} @} |
| 3499 ; |
| 3500 @end group |
| 3501 |
| 3502 @group |
| 3503 bar: /* empty */ |
| 3504 @{ previous_expr = $0; @} |
| 3505 ; |
| 3506 @end group |
| 3507 @end example |
| 3508 |
| 3509 As long as @code{bar} is used only in the fashion shown here, @code{$0} |
| 3510 always refers to the @code{expr} which precedes @code{bar} in the |
| 3511 definition of @code{foo}. |
| 3512 |
| 3513 @vindex yylval |
| 3514 It is also possible to access the semantic value of the lookahead token, if |
| 3515 any, from a semantic action. |
| 3516 This semantic value is stored in @code{yylval}. |
| 3517 @xref{Action Features, ,Special Features for Use in Actions}. |
| 3518 |
| 3519 @node Action Types |
| 3520 @subsection Data Types of Values in Actions |
| 3521 @cindex action data types |
| 3522 @cindex data types in actions |
| 3523 |
| 3524 If you have chosen a single data type for semantic values, the @code{$$} |
| 3525 and @code{$@var{n}} constructs always have that data type. |
| 3526 |
| 3527 If you have used @code{%union} to specify a variety of data types, then you |
| 3528 must declare a choice among these types for each terminal or nonterminal |
| 3529 symbol that can have a semantic value. Then each time you use @code{$$} or |
| 3530 @code{$@var{n}}, its data type is determined by which symbol it refers to |
| 3531 in the rule. In this example, |
| 3532 |
| 3533 @example |
| 3534 @group |
| 3535 exp: @dots{} |
| 3536 | exp '+' exp |
| 3537 @{ $$ = $1 + $3; @} |
| 3538 @end group |
| 3539 @end example |
| 3540 |
| 3541 @noindent |
| 3542 @code{$1} and @code{$3} refer to instances of @code{exp}, so they all |
| 3543 have the data type declared for the nonterminal symbol @code{exp}. If |
| 3544 @code{$2} were used, it would have the data type declared for the |
| 3545 terminal symbol @code{'+'}, whatever that might be. |
| 3546 |
| 3547 Alternatively, you can specify the data type when you refer to the value, |
| 3548 by inserting @samp{<@var{type}>} after the @samp{$} at the beginning of the |
| 3549 reference. For example, if you have defined types as shown here: |
| 3550 |
| 3551 @example |
| 3552 @group |
| 3553 %union @{ |
| 3554 int itype; |
| 3555 double dtype; |
| 3556 @} |
| 3557 @end group |
| 3558 @end example |
| 3559 |
| 3560 @noindent |
| 3561 then you can write @code{$<itype>1} to refer to the first subunit of the |
| 3562 rule as an integer, or @code{$<dtype>1} to refer to it as a double. |
| 3563 |
| 3564 @node Mid-Rule Actions |
| 3565 @subsection Actions in Mid-Rule |
| 3566 @cindex actions in mid-rule |
| 3567 @cindex mid-rule actions |
| 3568 |
| 3569 Occasionally it is useful to put an action in the middle of a rule. |
| 3570 These actions are written just like usual end-of-rule actions, but they |
| 3571 are executed before the parser even recognizes the following components. |
| 3572 |
| 3573 A mid-rule action may refer to the components preceding it using |
| 3574 @code{$@var{n}}, but it may not refer to subsequent components because |
| 3575 it is run before they are parsed. |
| 3576 |
| 3577 The mid-rule action itself counts as one of the components of the rule. |
| 3578 This makes a difference when there is another action later in the same rule |
| 3579 (and usually there is another at the end): you have to count the actions |
| 3580 along with the symbols when working out which number @var{n} to use in |
| 3581 @code{$@var{n}}. |
| 3582 |
| 3583 The mid-rule action can also have a semantic value. The action can set |
| 3584 its value with an assignment to @code{$$}, and actions later in the rule |
| 3585 can refer to the value using @code{$@var{n}}. Since there is no symbol |
| 3586 to name the action, there is no way to declare a data type for the value |
| 3587 in advance, so you must use the @samp{$<@dots{}>@var{n}} construct to |
| 3588 specify a data type each time you refer to this value. |
| 3589 |
| 3590 There is no way to set the value of the entire rule with a mid-rule |
| 3591 action, because assignments to @code{$$} do not have that effect. The |
| 3592 only way to set the value for the entire rule is with an ordinary action |
| 3593 at the end of the rule. |
| 3594 |
| 3595 Here is an example from a hypothetical compiler, handling a @code{let} |
| 3596 statement that looks like @samp{let (@var{variable}) @var{statement}} and |
| 3597 serves to create a variable named @var{variable} temporarily for the |
| 3598 duration of @var{statement}. To parse this construct, we must put |
| 3599 @var{variable} into the symbol table while @var{statement} is parsed, then |
| 3600 remove it afterward. Here is how it is done: |
| 3601 |
| 3602 @example |
| 3603 @group |
| 3604 stmt: LET '(' var ')' |
| 3605 @{ $<context>$ = push_context (); |
| 3606 declare_variable ($3); @} |
| 3607 stmt @{ $$ = $6; |
| 3608 pop_context ($<context>5); @} |
| 3609 @end group |
| 3610 @end example |
| 3611 |
| 3612 @noindent |
| 3613 As soon as @samp{let (@var{variable})} has been recognized, the first |
| 3614 action is run. It saves a copy of the current semantic context (the |
| 3615 list of accessible variables) as its semantic value, using alternative |
| 3616 @code{context} in the data-type union. Then it calls |
| 3617 @code{declare_variable} to add the new variable to that list. Once the |
| 3618 first action is finished, the embedded statement @code{stmt} can be |
| 3619 parsed. Note that the mid-rule action is component number 5, so the |
| 3620 @samp{stmt} is component number 6. |
| 3621 |
| 3622 After the embedded statement is parsed, its semantic value becomes the |
| 3623 value of the entire @code{let}-statement. Then the semantic value from the |
| 3624 earlier action is used to restore the prior list of variables. This |
| 3625 removes the temporary @code{let}-variable from the list so that it won't |
| 3626 appear to exist while the rest of the program is parsed. |
| 3627 |
| 3628 @findex %destructor |
| 3629 @cindex discarded symbols, mid-rule actions |
| 3630 @cindex error recovery, mid-rule actions |
| 3631 In the above example, if the parser initiates error recovery (@pxref{Error |
| 3632 Recovery}) while parsing the tokens in the embedded statement @code{stmt}, |
| 3633 it might discard the previous semantic context @code{$<context>5} without |
| 3634 restoring it. |
| 3635 Thus, @code{$<context>5} needs a destructor (@pxref{Destructor Decl, , Freeing |
| 3636 Discarded Symbols}). |
| 3637 However, Bison currently provides no means to declare a destructor specific to |
| 3638 a particular mid-rule action's semantic value. |
| 3639 |
| 3640 One solution is to bury the mid-rule action inside a nonterminal symbol and to |
| 3641 declare a destructor for that symbol: |
| 3642 |
| 3643 @example |
| 3644 @group |
| 3645 %type <context> let |
| 3646 %destructor @{ pop_context ($$); @} let |
| 3647 |
| 3648 %% |
| 3649 |
| 3650 stmt: let stmt |
| 3651 @{ $$ = $2; |
| 3652 pop_context ($1); @} |
| 3653 ; |
| 3654 |
| 3655 let: LET '(' var ')' |
| 3656 @{ $$ = push_context (); |
| 3657 declare_variable ($3); @} |
| 3658 ; |
| 3659 |
| 3660 @end group |
| 3661 @end example |
| 3662 |
| 3663 @noindent |
| 3664 Note that the action is now at the end of its rule. |
| 3665 Any mid-rule action can be converted to an end-of-rule action in this way, and |
| 3666 this is what Bison actually does to implement mid-rule actions. |
| 3667 |
| 3668 Taking action before a rule is completely recognized often leads to |
| 3669 conflicts since the parser must commit to a parse in order to execute the |
| 3670 action. For example, the following two rules, without mid-rule actions, |
| 3671 can coexist in a working parser because the parser can shift the open-brace |
| 3672 token and look at what follows before deciding whether there is a |
| 3673 declaration or not: |
| 3674 |
| 3675 @example |
| 3676 @group |
| 3677 compound: '@{' declarations statements '@}' |
| 3678 | '@{' statements '@}' |
| 3679 ; |
| 3680 @end group |
| 3681 @end example |
| 3682 |
| 3683 @noindent |
| 3684 But when we add a mid-rule action as follows, the rules become nonfunctional: |
| 3685 |
| 3686 @example |
| 3687 @group |
| 3688 compound: @{ prepare_for_local_variables (); @} |
| 3689 '@{' declarations statements '@}' |
| 3690 @end group |
| 3691 @group |
| 3692 | '@{' statements '@}' |
| 3693 ; |
| 3694 @end group |
| 3695 @end example |
| 3696 |
| 3697 @noindent |
| 3698 Now the parser is forced to decide whether to run the mid-rule action |
| 3699 when it has read no farther than the open-brace. In other words, it |
| 3700 must commit to using one rule or the other, without sufficient |
| 3701 information to do it correctly. (The open-brace token is what is called |
| 3702 the @dfn{lookahead} token at this time, since the parser is still |
| 3703 deciding what to do about it. @xref{Lookahead, ,Lookahead Tokens}.) |
| 3704 |
| 3705 You might think that you could correct the problem by putting identical |
| 3706 actions into the two rules, like this: |
| 3707 |
| 3708 @example |
| 3709 @group |
| 3710 compound: @{ prepare_for_local_variables (); @} |
| 3711 '@{' declarations statements '@}' |
| 3712 | @{ prepare_for_local_variables (); @} |
| 3713 '@{' statements '@}' |
| 3714 ; |
| 3715 @end group |
| 3716 @end example |
| 3717 |
| 3718 @noindent |
| 3719 But this does not help, because Bison does not realize that the two actions |
| 3720 are identical. (Bison never tries to understand the C code in an action.) |
| 3721 |
| 3722 If the grammar is such that a declaration can be distinguished from a |
| 3723 statement by the first token (which is true in C), then one solution which |
| 3724 does work is to put the action after the open-brace, like this: |
| 3725 |
| 3726 @example |
| 3727 @group |
| 3728 compound: '@{' @{ prepare_for_local_variables (); @} |
| 3729 declarations statements '@}' |
| 3730 | '@{' statements '@}' |
| 3731 ; |
| 3732 @end group |
| 3733 @end example |
| 3734 |
| 3735 @noindent |
| 3736 Now the first token of the following declaration or statement, |
| 3737 which would in any case tell Bison which rule to use, can still do so. |
| 3738 |
| 3739 Another solution is to bury the action inside a nonterminal symbol which |
| 3740 serves as a subroutine: |
| 3741 |
| 3742 @example |
| 3743 @group |
| 3744 subroutine: /* empty */ |
| 3745 @{ prepare_for_local_variables (); @} |
| 3746 ; |
| 3747 |
| 3748 @end group |
| 3749 |
| 3750 @group |
| 3751 compound: subroutine |
| 3752 '@{' declarations statements '@}' |
| 3753 | subroutine |
| 3754 '@{' statements '@}' |
| 3755 ; |
| 3756 @end group |
| 3757 @end example |
| 3758 |
| 3759 @noindent |
| 3760 Now Bison can execute the action in the rule for @code{subroutine} without |
| 3761 deciding which rule for @code{compound} it will eventually use. |
| 3762 |
| 3763 @node Locations |
| 3764 @section Tracking Locations |
| 3765 @cindex location |
| 3766 @cindex textual location |
| 3767 @cindex location, textual |
| 3768 |
| 3769 Though grammar rules and semantic actions are enough to write a fully |
| 3770 functional parser, it can be useful to process some additional information, |
| 3771 especially symbol locations. |
| 3772 |
| 3773 The way locations are handled is defined by providing a data type, and |
| 3774 actions to take when rules are matched. |
| 3775 |
| 3776 @menu |
| 3777 * Location Type:: Specifying a data type for locations. |
| 3778 * Actions and Locations:: Using locations in actions. |
| 3779 * Location Default Action:: Defining a general way to compute locations. |
| 3780 @end menu |
| 3781 |
| 3782 @node Location Type |
| 3783 @subsection Data Type of Locations |
| 3784 @cindex data type of locations |
| 3785 @cindex default location type |
| 3786 |
| 3787 Defining a data type for locations is much simpler than for semantic values, |
| 3788 since all tokens and groupings always use the same type. |
| 3789 |
| 3790 You can specify the type of locations by defining a macro called |
| 3791 @code{YYLTYPE}, just as you can specify the semantic value type by |
| 3792 defining a @code{YYSTYPE} macro (@pxref{Value Type}). |
| 3793 When @code{YYLTYPE} is not defined, Bison uses a default structure type with |
| 3794 four members: |
| 3795 |
| 3796 @example |
| 3797 typedef struct YYLTYPE |
| 3798 @{ |
| 3799 int first_line; |
| 3800 int first_column; |
| 3801 int last_line; |
| 3802 int last_column; |
| 3803 @} YYLTYPE; |
| 3804 @end example |
| 3805 |
| 3806 At the beginning of the parsing, Bison initializes all these fields to 1 |
| 3807 for @code{yylloc}. |
| 3808 |
| 3809 @node Actions and Locations |
| 3810 @subsection Actions and Locations |
| 3811 @cindex location actions |
| 3812 @cindex actions, location |
| 3813 @vindex @@$ |
| 3814 @vindex @@@var{n} |
| 3815 |
| 3816 Actions are not only useful for defining language semantics, but also for |
| 3817 describing the behavior of the output parser with locations. |
| 3818 |
| 3819 The most obvious way for building locations of syntactic groupings is very |
| 3820 similar to the way semantic values are computed. In a given rule, several |
| 3821 constructs can be used to access the locations of the elements being matched. |
| 3822 The location of the @var{n}th component of the right hand side is |
| 3823 @code{@@@var{n}}, while the location of the left hand side grouping is |
| 3824 @code{@@$}. |
| 3825 |
| 3826 Here is a basic example using the default data type for locations: |
| 3827 |
| 3828 @example |
| 3829 @group |
| 3830 exp: @dots{} |
| 3831 | exp '/' exp |
| 3832 @{ |
| 3833 @@$.first_column = @@1.first_column; |
| 3834 @@$.first_line = @@1.first_line; |
| 3835 @@$.last_column = @@3.last_column; |
| 3836 @@$.last_line = @@3.last_line; |
| 3837 if ($3) |
| 3838 $$ = $1 / $3; |
| 3839 else |
| 3840 @{ |
| 3841 $$ = 1; |
| 3842 fprintf (stderr, |
| 3843 "Division by zero, l%d,c%d-l%d,c%d", |
| 3844 @@3.first_line, @@3.first_column, |
| 3845 @@3.last_line, @@3.last_column); |
| 3846 @} |
| 3847 @} |
| 3848 @end group |
| 3849 @end example |
| 3850 |
| 3851 As for semantic values, there is a default action for locations that is |
| 3852 run each time a rule is matched. It sets the beginning of @code{@@$} to the |
| 3853 beginning of the first symbol, and the end of @code{@@$} to the end of the |
| 3854 last symbol. |
| 3855 |
| 3856 With this default action, the location tracking can be fully automatic. The |
| 3857 example above simply rewrites this way: |
| 3858 |
| 3859 @example |
| 3860 @group |
| 3861 exp: @dots{} |
| 3862 | exp '/' exp |
| 3863 @{ |
| 3864 if ($3) |
| 3865 $$ = $1 / $3; |
| 3866 else |
| 3867 @{ |
| 3868 $$ = 1; |
| 3869 fprintf (stderr, |
| 3870 "Division by zero, l%d,c%d-l%d,c%d", |
| 3871 @@3.first_line, @@3.first_column, |
| 3872 @@3.last_line, @@3.last_column); |
| 3873 @} |
| 3874 @} |
| 3875 @end group |
| 3876 @end example |
| 3877 |
| 3878 @vindex yylloc |
| 3879 It is also possible to access the location of the lookahead token, if any, |
| 3880 from a semantic action. |
| 3881 This location is stored in @code{yylloc}. |
| 3882 @xref{Action Features, ,Special Features for Use in Actions}. |
| 3883 |
| 3884 @node Location Default Action |
| 3885 @subsection Default Action for Locations |
| 3886 @vindex YYLLOC_DEFAULT |
| 3887 @cindex @acronym{GLR} parsers and @code{YYLLOC_DEFAULT} |
| 3888 |
| 3889 Actually, actions are not the best place to compute locations. Since |
| 3890 locations are much more general than semantic values, there is room in |
| 3891 the output parser to redefine the default action to take for each |
| 3892 rule. The @code{YYLLOC_DEFAULT} macro is invoked each time a rule is |
| 3893 matched, before the associated action is run. It is also invoked |
| 3894 while processing a syntax error, to compute the error's location. |
| 3895 Before reporting an unresolvable syntactic ambiguity, a @acronym{GLR} |
| 3896 parser invokes @code{YYLLOC_DEFAULT} recursively to compute the location |
| 3897 of that ambiguity. |
| 3898 |
| 3899 Most of the time, this macro is general enough to suppress location |
| 3900 dedicated code from semantic actions. |
| 3901 |
| 3902 The @code{YYLLOC_DEFAULT} macro takes three parameters. The first one is |
| 3903 the location of the grouping (the result of the computation). When a |
| 3904 rule is matched, the second parameter identifies locations of |
| 3905 all right hand side elements of the rule being matched, and the third |
| 3906 parameter is the size of the rule's right hand side. |
| 3907 When a @acronym{GLR} parser reports an ambiguity, which of multiple candidate |
| 3908 right hand sides it passes to @code{YYLLOC_DEFAULT} is undefined. |
| 3909 When processing a syntax error, the second parameter identifies locations |
| 3910 of the symbols that were discarded during error processing, and the third |
| 3911 parameter is the number of discarded symbols. |
| 3912 |
| 3913 By default, @code{YYLLOC_DEFAULT} is defined this way: |
| 3914 |
| 3915 @smallexample |
| 3916 @group |
| 3917 # define YYLLOC_DEFAULT(Current, Rhs, N) \ |
| 3918 do \ |
| 3919 if (N) \ |
| 3920 @{ \ |
| 3921 (Current).first_line = YYRHSLOC(Rhs, 1).first_line; \ |
| 3922 (Current).first_column = YYRHSLOC(Rhs, 1).first_column; \ |
| 3923 (Current).last_line = YYRHSLOC(Rhs, N).last_line; \ |
| 3924 (Current).last_column = YYRHSLOC(Rhs, N).last_column; \ |
| 3925 @} \ |
| 3926 else \ |
| 3927 @{ \ |
| 3928 (Current).first_line = (Current).last_line = \ |
| 3929 YYRHSLOC(Rhs, 0).last_line; \ |
| 3930 (Current).first_column = (Current).last_column = \ |
| 3931 YYRHSLOC(Rhs, 0).last_column; \ |
| 3932 @} \ |
| 3933 while (0) |
| 3934 @end group |
| 3935 @end smallexample |
| 3936 |
| 3937 where @code{YYRHSLOC (rhs, k)} is the location of the @var{k}th symbol |
| 3938 in @var{rhs} when @var{k} is positive, and the location of the symbol |
| 3939 just before the reduction when @var{k} and @var{n} are both zero. |
| 3940 |
| 3941 When defining @code{YYLLOC_DEFAULT}, you should consider that: |
| 3942 |
| 3943 @itemize @bullet |
| 3944 @item |
| 3945 All arguments are free of side-effects. However, only the first one (the |
| 3946 result) should be modified by @code{YYLLOC_DEFAULT}. |
| 3947 |
| 3948 @item |
| 3949 For consistency with semantic actions, valid indexes within the |
| 3950 right hand side range from 1 to @var{n}. When @var{n} is zero, only 0 is a |
| 3951 valid index, and it refers to the symbol just before the reduction. |
| 3952 During error processing @var{n} is always positive. |
| 3953 |
| 3954 @item |
| 3955 Your macro should parenthesize its arguments, if need be, since the |
| 3956 actual arguments may not be surrounded by parentheses. Also, your |
| 3957 macro should expand to something that can be used as a single |
| 3958 statement when it is followed by a semicolon. |
| 3959 @end itemize |
| 3960 |
| 3961 @node Declarations |
| 3962 @section Bison Declarations |
| 3963 @cindex declarations, Bison |
| 3964 @cindex Bison declarations |
| 3965 |
| 3966 The @dfn{Bison declarations} section of a Bison grammar defines the symbols |
| 3967 used in formulating the grammar and the data types of semantic values. |
| 3968 @xref{Symbols}. |
| 3969 |
| 3970 All token type names (but not single-character literal tokens such as |
| 3971 @code{'+'} and @code{'*'}) must be declared. Nonterminal symbols must be |
| 3972 declared if you need to specify which data type to use for the semantic |
| 3973 value (@pxref{Multiple Types, ,More Than One Value Type}). |
| 3974 |
| 3975 The first rule in the file also specifies the start symbol, by default. |
| 3976 If you want some other symbol to be the start symbol, you must declare |
| 3977 it explicitly (@pxref{Language and Grammar, ,Languages and Context-Free |
| 3978 Grammars}). |
| 3979 |
| 3980 @menu |
| 3981 * Require Decl:: Requiring a Bison version. |
| 3982 * Token Decl:: Declaring terminal symbols. |
| 3983 * Precedence Decl:: Declaring terminals with precedence and associativity. |
| 3984 * Union Decl:: Declaring the set of all semantic value types. |
| 3985 * Type Decl:: Declaring the choice of type for a nonterminal symbol. |
| 3986 * Initial Action Decl:: Code run before parsing starts. |
| 3987 * Destructor Decl:: Declaring how symbols are freed. |
| 3988 * Expect Decl:: Suppressing warnings about parsing conflicts. |
| 3989 * Start Decl:: Specifying the start symbol. |
| 3990 * Pure Decl:: Requesting a reentrant parser. |
| 3991 * Push Decl:: Requesting a push parser. |
| 3992 * Decl Summary:: Table of all Bison declarations. |
| 3993 @end menu |
| 3994 |
| 3995 @node Require Decl |
| 3996 @subsection Require a Version of Bison |
| 3997 @cindex version requirement |
| 3998 @cindex requiring a version of Bison |
| 3999 @findex %require |
| 4000 |
| 4001 You may require the minimum version of Bison to process the grammar. If |
| 4002 the requirement is not met, @command{bison} exits with an error (exit |
| 4003 status 63). |
| 4004 |
| 4005 @example |
| 4006 %require "@var{version}" |
| 4007 @end example |
| 4008 |
| 4009 @node Token Decl |
| 4010 @subsection Token Type Names |
| 4011 @cindex declaring token type names |
| 4012 @cindex token type names, declaring |
| 4013 @cindex declaring literal string tokens |
| 4014 @findex %token |
| 4015 |
| 4016 The basic way to declare a token type name (terminal symbol) is as follows: |
| 4017 |
| 4018 @example |
| 4019 %token @var{name} |
| 4020 @end example |
| 4021 |
| 4022 Bison will convert this into a @code{#define} directive in |
| 4023 the parser, so that the function @code{yylex} (if it is in this file) |
| 4024 can use the name @var{name} to stand for this token type's code. |
| 4025 |
| 4026 Alternatively, you can use @code{%left}, @code{%right}, or |
| 4027 @code{%nonassoc} instead of @code{%token}, if you wish to specify |
| 4028 associativity and precedence. @xref{Precedence Decl, ,Operator |
| 4029 Precedence}. |
| 4030 |
| 4031 You can explicitly specify the numeric code for a token type by appending |
| 4032 a nonnegative decimal or hexadecimal integer value in the field immediately |
| 4033 following the token name: |
| 4034 |
| 4035 @example |
| 4036 %token NUM 300 |
| 4037 %token XNUM 0x12d // a GNU extension |
| 4038 @end example |
| 4039 |
| 4040 @noindent |
| 4041 It is generally best, however, to let Bison choose the numeric codes for |
| 4042 all token types. Bison will automatically select codes that don't conflict |
| 4043 with each other or with normal characters. |
| 4044 |
| 4045 In the event that the stack type is a union, you must augment the |
| 4046 @code{%token} or other token declaration to include the data type |
| 4047 alternative delimited by angle-brackets (@pxref{Multiple Types, ,More |
| 4048 Than One Value Type}). |
| 4049 |
| 4050 For example: |
| 4051 |
| 4052 @example |
| 4053 @group |
| 4054 %union @{ /* define stack type */ |
| 4055 double val; |
| 4056 symrec *tptr; |
| 4057 @} |
| 4058 %token <val> NUM /* define token NUM and its type */ |
| 4059 @end group |
| 4060 @end example |
| 4061 |
| 4062 You can associate a literal string token with a token type name by |
| 4063 writing the literal string at the end of a @code{%token} |
| 4064 declaration which declares the name. For example: |
| 4065 |
| 4066 @example |
| 4067 %token arrow "=>" |
| 4068 @end example |
| 4069 |
| 4070 @noindent |
| 4071 For example, a grammar for the C language might specify these names with |
| 4072 equivalent literal string tokens: |
| 4073 |
| 4074 @example |
| 4075 %token <operator> OR "||" |
| 4076 %token <operator> LE 134 "<=" |
| 4077 %left OR "<=" |
| 4078 @end example |
| 4079 |
| 4080 @noindent |
| 4081 Once you equate the literal string and the token name, you can use them |
| 4082 interchangeably in further declarations or the grammar rules. The |
| 4083 @code{yylex} function can use the token name or the literal string to |
| 4084 obtain the token type code number (@pxref{Calling Convention}). |
| 4085 Syntax error messages passed to @code{yyerror} from the parser will reference |
| 4086 the literal string instead of the token name. |
| 4087 |
| 4088 The token numbered as 0 corresponds to end of file; the following line |
| 4089 allows for nicer error messages referring to ``end of file'' instead |
| 4090 of ``$end'': |
| 4091 |
| 4092 @example |
| 4093 %token END 0 "end of file" |
| 4094 @end example |
| 4095 |
| 4096 @node Precedence Decl |
| 4097 @subsection Operator Precedence |
| 4098 @cindex precedence declarations |
| 4099 @cindex declaring operator precedence |
| 4100 @cindex operator precedence, declaring |
| 4101 |
| 4102 Use the @code{%left}, @code{%right} or @code{%nonassoc} declaration to |
| 4103 declare a token and specify its precedence and associativity, all at |
| 4104 once. These are called @dfn{precedence declarations}. |
| 4105 @xref{Precedence, ,Operator Precedence}, for general information on |
| 4106 operator precedence. |
| 4107 |
| 4108 The syntax of a precedence declaration is nearly the same as that of |
| 4109 @code{%token}: either |
| 4110 |
| 4111 @example |
| 4112 %left @var{symbols}@dots{} |
| 4113 @end example |
| 4114 |
| 4115 @noindent |
| 4116 or |
| 4117 |
| 4118 @example |
| 4119 %left <@var{type}> @var{symbols}@dots{} |
| 4120 @end example |
| 4121 |
| 4122 And indeed any of these declarations serves the purposes of @code{%token}. |
| 4123 But in addition, they specify the associativity and relative precedence for |
| 4124 all the @var{symbols}: |
| 4125 |
| 4126 @itemize @bullet |
| 4127 @item |
| 4128 The associativity of an operator @var{op} determines how repeated uses |
| 4129 of the operator nest: whether @samp{@var{x} @var{op} @var{y} @var{op} |
| 4130 @var{z}} is parsed by grouping @var{x} with @var{y} first or by |
| 4131 grouping @var{y} with @var{z} first. @code{%left} specifies |
| 4132 left-associativity (grouping @var{x} with @var{y} first) and |
| 4133 @code{%right} specifies right-associativity (grouping @var{y} with |
| 4134 @var{z} first). @code{%nonassoc} specifies no associativity, which |
| 4135 means that @samp{@var{x} @var{op} @var{y} @var{op} @var{z}} is |
| 4136 considered a syntax error. |
| 4137 |
| 4138 @item |
| 4139 The precedence of an operator determines how it nests with other operators. |
| 4140 All the tokens declared in a single precedence declaration have equal |
| 4141 precedence and nest together according to their associativity. |
| 4142 When two tokens declared in different precedence declarations associate, |
| 4143 the one declared later has the higher precedence and is grouped first. |
| 4144 @end itemize |
| 4145 |
| 4146 For backward compatibility, there is a confusing difference between the |
| 4147 argument lists of @code{%token} and precedence declarations. |
| 4148 Only a @code{%token} can associate a literal string with a token type name. |
| 4149 A precedence declaration always interprets a literal string as a reference to a |
| 4150 separate token. |
| 4151 For example: |
| 4152 |
| 4153 @example |
| 4154 %left OR "<=" // Does not declare an alias. |
| 4155 %left OR 134 "<=" 135 // Declares 134 for OR and 135 for "<=". |
| 4156 @end example |
| 4157 |
| 4158 @node Union Decl |
| 4159 @subsection The Collection of Value Types |
| 4160 @cindex declaring value types |
| 4161 @cindex value types, declaring |
| 4162 @findex %union |
| 4163 |
| 4164 The @code{%union} declaration specifies the entire collection of |
| 4165 possible data types for semantic values. The keyword @code{%union} is |
| 4166 followed by braced code containing the same thing that goes inside a |
| 4167 @code{union} in C@. |
| 4168 |
| 4169 For example: |
| 4170 |
| 4171 @example |
| 4172 @group |
| 4173 %union @{ |
| 4174 double val; |
| 4175 symrec *tptr; |
| 4176 @} |
| 4177 @end group |
| 4178 @end example |
| 4179 |
| 4180 @noindent |
| 4181 This says that the two alternative types are @code{double} and @code{symrec |
| 4182 *}. They are given names @code{val} and @code{tptr}; these names are used |
| 4183 in the @code{%token} and @code{%type} declarations to pick one of the types |
| 4184 for a terminal or nonterminal symbol (@pxref{Type Decl, ,Nonterminal Symbols}). |
| 4185 |
| 4186 As an extension to @acronym{POSIX}, a tag is allowed after the |
| 4187 @code{union}. For example: |
| 4188 |
| 4189 @example |
| 4190 @group |
| 4191 %union value @{ |
| 4192 double val; |
| 4193 symrec *tptr; |
| 4194 @} |
| 4195 @end group |
| 4196 @end example |
| 4197 |
| 4198 @noindent |
| 4199 specifies the union tag @code{value}, so the corresponding C type is |
| 4200 @code{union value}. If you do not specify a tag, it defaults to |
| 4201 @code{YYSTYPE}. |
| 4202 |
| 4203 As another extension to @acronym{POSIX}, you may specify multiple |
| 4204 @code{%union} declarations; their contents are concatenated. However, |
| 4205 only the first @code{%union} declaration can specify a tag. |
| 4206 |
| 4207 Note that, unlike making a @code{union} declaration in C, you need not write |
| 4208 a semicolon after the closing brace. |
| 4209 |
| 4210 Instead of @code{%union}, you can define and use your own union type |
| 4211 @code{YYSTYPE} if your grammar contains at least one |
| 4212 @samp{<@var{type}>} tag. For example, you can put the following into |
| 4213 a header file @file{parser.h}: |
| 4214 |
| 4215 @example |
| 4216 @group |
| 4217 union YYSTYPE @{ |
| 4218 double val; |
| 4219 symrec *tptr; |
| 4220 @}; |
| 4221 typedef union YYSTYPE YYSTYPE; |
| 4222 @end group |
| 4223 @end example |
| 4224 |
| 4225 @noindent |
| 4226 and then your grammar can use the following |
| 4227 instead of @code{%union}: |
| 4228 |
| 4229 @example |
| 4230 @group |
| 4231 %@{ |
| 4232 #include "parser.h" |
| 4233 %@} |
| 4234 %type <val> expr |
| 4235 %token <tptr> ID |
| 4236 @end group |
| 4237 @end example |
| 4238 |
| 4239 @node Type Decl |
| 4240 @subsection Nonterminal Symbols |
| 4241 @cindex declaring value types, nonterminals |
| 4242 @cindex value types, nonterminals, declaring |
| 4243 @findex %type |
| 4244 |
| 4245 @noindent |
| 4246 When you use @code{%union} to specify multiple value types, you must |
| 4247 declare the value type of each nonterminal symbol for which values are |
| 4248 used. This is done with a @code{%type} declaration, like this: |
| 4249 |
| 4250 @example |
| 4251 %type <@var{type}> @var{nonterminal}@dots{} |
| 4252 @end example |
| 4253 |
| 4254 @noindent |
| 4255 Here @var{nonterminal} is the name of a nonterminal symbol, and |
| 4256 @var{type} is the name given in the @code{%union} to the alternative |
| 4257 that you want (@pxref{Union Decl, ,The Collection of Value Types}). You |
| 4258 can give any number of nonterminal symbols in the same @code{%type} |
| 4259 declaration, if they have the same value type. Use spaces to separate |
| 4260 the symbol names. |
| 4261 |
| 4262 You can also declare the value type of a terminal symbol. To do this, |
| 4263 use the same @code{<@var{type}>} construction in a declaration for the |
| 4264 terminal symbol. All kinds of token declarations allow |
| 4265 @code{<@var{type}>}. |
| 4266 |
| 4267 @node Initial Action Decl |
| 4268 @subsection Performing Actions before Parsing |
| 4269 @findex %initial-action |
| 4270 |
| 4271 Sometimes your parser needs to perform some initializations before |
| 4272 parsing. The @code{%initial-action} directive allows for such arbitrary |
| 4273 code. |
| 4274 |
| 4275 @deffn {Directive} %initial-action @{ @var{code} @} |
| 4276 @findex %initial-action |
| 4277 Declare that the braced @var{code} must be invoked before parsing each time |
| 4278 @code{yyparse} is called. The @var{code} may use @code{$$} and |
| 4279 @code{@@$} --- initial value and location of the lookahead --- and the |
| 4280 @code{%parse-param}. |
| 4281 @end deffn |
| 4282 |
| 4283 For instance, if your locations use a file name, you may use |
| 4284 |
| 4285 @example |
| 4286 %parse-param @{ char const *file_name @}; |
| 4287 %initial-action |
| 4288 @{ |
| 4289 @@$.initialize (file_name); |
| 4290 @}; |
| 4291 @end example |
| 4292 |
| 4293 |
| 4294 @node Destructor Decl |
| 4295 @subsection Freeing Discarded Symbols |
| 4296 @cindex freeing discarded symbols |
| 4297 @findex %destructor |
| 4298 @findex <*> |
| 4299 @findex <> |
| 4300 During error recovery (@pxref{Error Recovery}), symbols already pushed |
| 4301 on the stack and tokens coming from the rest of the file are discarded |
| 4302 until the parser falls on its feet. If the parser runs out of memory, |
| 4303 or if it returns via @code{YYABORT} or @code{YYACCEPT}, all the |
| 4304 symbols on the stack must be discarded. Even if the parser succeeds, it |
| 4305 must discard the start symbol. |
| 4306 |
| 4307 When discarded symbols convey heap based information, this memory is |
| 4308 lost. While this behavior can be tolerable for batch parsers, such as |
| 4309 in traditional compilers, it is unacceptable for programs like shells or |
| 4310 protocol implementations that may parse and execute indefinitely. |
| 4311 |
| 4312 The @code{%destructor} directive defines code that is called when a |
| 4313 symbol is automatically discarded. |
| 4314 |
| 4315 @deffn {Directive} %destructor @{ @var{code} @} @var{symbols} |
| 4316 @findex %destructor |
| 4317 Invoke the braced @var{code} whenever the parser discards one of the |
| 4318 @var{symbols}. |
| 4319 Within @var{code}, @code{$$} designates the semantic value associated |
| 4320 with the discarded symbol, and @code{@@$} designates its location. |
| 4321 The additional parser parameters are also available (@pxref{Parser Function, , |
| 4322 The Parser Function @code{yyparse}}). |
| 4323 |
| 4324 When a symbol is listed among @var{symbols}, its @code{%destructor} is called a |
| 4325 per-symbol @code{%destructor}. |
| 4326 You may also define a per-type @code{%destructor} by listing a semantic type |
| 4327 tag among @var{symbols}. |
| 4328 In that case, the parser will invoke this @var{code} whenever it discards any |
| 4329 grammar symbol that has that semantic type tag unless that symbol has its own |
| 4330 per-symbol @code{%destructor}. |
| 4331 |
| 4332 Finally, you can define two different kinds of default @code{%destructor}s. |
| 4333 (These default forms are experimental. |
| 4334 More user feedback will help to determine whether they should become permanent |
| 4335 features.) |
| 4336 You can place each of @code{<*>} and @code{<>} in the @var{symbols} list of |
| 4337 exactly one @code{%destructor} declaration in your grammar file. |
| 4338 The parser will invoke the @var{code} associated with one of these whenever it |
| 4339 discards any user-defined grammar symbol that has no per-symbol and no per-type |
| 4340 @code{%destructor}. |
| 4341 The parser uses the @var{code} for @code{<*>} in the case of such a grammar |
| 4342 symbol for which you have formally declared a semantic type tag (@code{%type} |
| 4343 counts as such a declaration, but @code{$<tag>$} does not). |
| 4344 The parser uses the @var{code} for @code{<>} in the case of such a grammar |
| 4345 symbol that has no declared semantic type tag. |
| 4346 @end deffn |
| 4347 |
| 4348 @noindent |
| 4349 For example: |
| 4350 |
| 4351 @smallexample |
| 4352 %union @{ char *string; @} |
| 4353 %token <string> STRING1 |
| 4354 %token <string> STRING2 |
| 4355 %type <string> string1 |
| 4356 %type <string> string2 |
| 4357 %union @{ char character; @} |
| 4358 %token <character> CHR |
| 4359 %type <character> chr |
| 4360 %token TAGLESS |
| 4361 |
| 4362 %destructor @{ @} <character> |
| 4363 %destructor @{ free ($$); @} <*> |
| 4364 %destructor @{ free ($$); printf ("%d", @@$.first_line); @} STRING1 string1 |
| 4365 %destructor @{ printf ("Discarding tagless symbol.\n"); @} <> |
| 4366 @end smallexample |
| 4367 |
| 4368 @noindent |
| 4369 guarantees that, when the parser discards any user-defined symbol that has a |
| 4370 semantic type tag other than @code{<character>}, it passes its semantic value |
| 4371 to @code{free} by default. |
| 4372 However, when the parser discards a @code{STRING1} or a @code{string1}, it also |
| 4373 prints its line number to @code{stdout}. |
| 4374 It performs only the second @code{%destructor} in this case, so it invokes |
| 4375 @code{free} only once. |
| 4376 Finally, the parser merely prints a message whenever it discards any symbol, |
| 4377 such as @code{TAGLESS}, that has no semantic type tag. |
| 4378 |
| 4379 A Bison-generated parser invokes the default @code{%destructor}s only for |
| 4380 user-defined as opposed to Bison-defined symbols. |
| 4381 For example, the parser will not invoke either kind of default |
| 4382 @code{%destructor} for the special Bison-defined symbols @code{$accept}, |
| 4383 @code{$undefined}, or @code{$end} (@pxref{Table of Symbols, ,Bison Symbols}), |
| 4384 none of which you can reference in your grammar. |
| 4385 It also will not invoke either for the @code{error} token (@pxref{Table of |
| 4386 Symbols, ,error}), which is always defined by Bison regardless of whether you |
| 4387 reference it in your grammar. |
| 4388 However, it may invoke one of them for the end token (token 0) if you |
| 4389 redefine it from @code{$end} to, for example, @code{END}: |
| 4390 |
| 4391 @smallexample |
| 4392 %token END 0 |
| 4393 @end smallexample |
| 4394 |
| 4395 @cindex actions in mid-rule |
| 4396 @cindex mid-rule actions |
| 4397 Finally, Bison will never invoke a @code{%destructor} for an unreferenced |
| 4398 mid-rule semantic value (@pxref{Mid-Rule Actions,,Actions in Mid-Rule}). |
| 4399 That is, Bison does not consider a mid-rule to have a semantic value if you do |
| 4400 not reference @code{$$} in the mid-rule's action or @code{$@var{n}} (where |
| 4401 @var{n} is the RHS symbol position of the mid-rule) in any later action in that |
| 4402 rule. |
| 4403 However, if you do reference either, the Bison-generated parser will invoke the |
| 4404 @code{<>} @code{%destructor} whenever it discards the mid-rule symbol. |
| 4405 |
| 4406 @ignore |
| 4407 @noindent |
| 4408 In the future, it may be possible to redefine the @code{error} token as a |
| 4409 nonterminal that captures the discarded symbols. |
| 4410 In that case, the parser will invoke the default destructor for it as well. |
| 4411 @end ignore |
| 4412 |
| 4413 @sp 1 |
| 4414 |
| 4415 @cindex discarded symbols |
| 4416 @dfn{Discarded symbols} are the following: |
| 4417 |
| 4418 @itemize |
| 4419 @item |
| 4420 stacked symbols popped during the first phase of error recovery, |
| 4421 @item |
| 4422 incoming terminals during the second phase of error recovery, |
| 4423 @item |
| 4424 the current lookahead and the entire stack (except the current |
| 4425 right-hand side symbols) when the parser returns immediately, and |
| 4426 @item |
| 4427 the start symbol, when the parser succeeds. |
| 4428 @end itemize |
| 4429 |
| 4430 The parser can @dfn{return immediately} because of an explicit call to |
| 4431 @code{YYABORT} or @code{YYACCEPT}, or failed error recovery, or memory |
| 4432 exhaustion. |
| 4433 |
| 4434 Right-hand side symbols of a rule that explicitly triggers a syntax |
| 4435 error via @code{YYERROR} are not discarded automatically. As a rule |
| 4436 of thumb, destructors are invoked only when user actions cannot manage |
| 4437 the memory. |
| 4438 |
| 4439 @node Expect Decl |
| 4440 @subsection Suppressing Conflict Warnings |
| 4441 @cindex suppressing conflict warnings |
| 4442 @cindex preventing warnings about conflicts |
| 4443 @cindex warnings, preventing |
| 4444 @cindex conflicts, suppressing warnings of |
| 4445 @findex %expect |
| 4446 @findex %expect-rr |
| 4447 |
| 4448 Bison normally warns if there are any conflicts in the grammar |
| 4449 (@pxref{Shift/Reduce, ,Shift/Reduce Conflicts}), but most real grammars |
| 4450 have harmless shift/reduce conflicts which are resolved in a predictable |
| 4451 way and would be difficult to eliminate. It is desirable to suppress |
| 4452 the warning about these conflicts unless the number of conflicts |
| 4453 changes. You can do this with the @code{%expect} declaration. |
| 4454 |
| 4455 The declaration looks like this: |
| 4456 |
| 4457 @example |
| 4458 %expect @var{n} |
| 4459 @end example |
| 4460 |
| 4461 Here @var{n} is a decimal integer. The declaration says there should |
| 4462 be @var{n} shift/reduce conflicts and no reduce/reduce conflicts. |
| 4463 Bison reports an error if the number of shift/reduce conflicts differs |
| 4464 from @var{n}, or if there are any reduce/reduce conflicts. |
| 4465 |
| 4466 For normal @acronym{LALR}(1) parsers, reduce/reduce conflicts are more |
| 4467 serious, and should be eliminated entirely. Bison will always report |
| 4468 reduce/reduce conflicts for these parsers. With @acronym{GLR} |
| 4469 parsers, however, both kinds of conflicts are routine; otherwise, |
| 4470 there would be no need to use @acronym{GLR} parsing. Therefore, it is |
| 4471 also possible to specify an expected number of reduce/reduce conflicts |
| 4472 in @acronym{GLR} parsers, using the declaration: |
| 4473 |
| 4474 @example |
| 4475 %expect-rr @var{n} |
| 4476 @end example |
| 4477 |
| 4478 In general, using @code{%expect} involves these steps: |
| 4479 |
| 4480 @itemize @bullet |
| 4481 @item |
| 4482 Compile your grammar without @code{%expect}. Use the @samp{-v} option |
| 4483 to get a verbose list of where the conflicts occur. Bison will also |
| 4484 print the number of conflicts. |
| 4485 |
| 4486 @item |
| 4487 Check each of the conflicts to make sure that Bison's default |
| 4488 resolution is what you really want. If not, rewrite the grammar and |
| 4489 go back to the beginning. |
| 4490 |
| 4491 @item |
| 4492 Add an @code{%expect} declaration, copying the number @var{n} from the |
| 4493 number which Bison printed. With @acronym{GLR} parsers, add an |
| 4494 @code{%expect-rr} declaration as well. |
| 4495 @end itemize |
| 4496 |
| 4497 Now Bison will warn you if you introduce an unexpected conflict, but |
| 4498 will keep silent otherwise. |
| 4499 |
| 4500 @node Start Decl |
| 4501 @subsection The Start-Symbol |
| 4502 @cindex declaring the start symbol |
| 4503 @cindex start symbol, declaring |
| 4504 @cindex default start symbol |
| 4505 @findex %start |
| 4506 |
| 4507 Bison assumes by default that the start symbol for the grammar is the first |
| 4508 nonterminal specified in the grammar specification section. The programmer |
| 4509 may override this restriction with the @code{%start} declaration as follows: |
| 4510 |
| 4511 @example |
| 4512 %start @var{symbol} |
| 4513 @end example |
| 4514 |
| 4515 @node Pure Decl |
| 4516 @subsection A Pure (Reentrant) Parser |
| 4517 @cindex reentrant parser |
| 4518 @cindex pure parser |
| 4519 @findex %define api.pure |
| 4520 |
| 4521 A @dfn{reentrant} program is one which does not alter in the course of |
| 4522 execution; in other words, it consists entirely of @dfn{pure} (read-only) |
| 4523 code. Reentrancy is important whenever asynchronous execution is possible; |
| 4524 for example, a nonreentrant program may not be safe to call from a signal |
| 4525 handler. In systems with multiple threads of control, a nonreentrant |
| 4526 program must be called only within interlocks. |
| 4527 |
| 4528 Normally, Bison generates a parser which is not reentrant. This is |
| 4529 suitable for most uses, and it permits compatibility with Yacc. (The |
| 4530 standard Yacc interfaces are inherently nonreentrant, because they use |
| 4531 statically allocated variables for communication with @code{yylex}, |
| 4532 including @code{yylval} and @code{yylloc}.) |
| 4533 |
| 4534 Alternatively, you can generate a pure, reentrant parser. The Bison |
| 4535 declaration @code{%define api.pure} says that you want the parser to be |
| 4536 reentrant. It looks like this: |
| 4537 |
| 4538 @example |
| 4539 %define api.pure |
| 4540 @end example |
| 4541 |
| 4542 The result is that the communication variables @code{yylval} and |
| 4543 @code{yylloc} become local variables in @code{yyparse}, and a different |
| 4544 calling convention is used for the lexical analyzer function |
| 4545 @code{yylex}. @xref{Pure Calling, ,Calling Conventions for Pure |
| 4546 Parsers}, for the details of this. The variable @code{yynerrs} |
| 4547 becomes local in @code{yyparse} in pull mode but it becomes a member |
| 4548 of yypstate in push mode. (@pxref{Error Reporting, ,The Error |
| 4549 Reporting Function @code{yyerror}}). The convention for calling |
| 4550 @code{yyparse} itself is unchanged. |
| 4551 |
| 4552 Whether the parser is pure has nothing to do with the grammar rules. |
| 4553 You can generate either a pure parser or a nonreentrant parser from any |
| 4554 valid grammar. |
| 4555 |
| 4556 @node Push Decl |
| 4557 @subsection A Push Parser |
| 4558 @cindex push parser |
| 4559 @cindex push parser |
| 4560 @findex %define api.push_pull |
| 4561 |
| 4562 (The current push parsing interface is experimental and may evolve. |
| 4563 More user feedback will help to stabilize it.) |
| 4564 |
| 4565 A pull parser is called once and it takes control until all its input |
| 4566 is completely parsed. A push parser, on the other hand, is called |
| 4567 each time a new token is made available. |
| 4568 |
| 4569 A push parser is typically useful when the parser is part of a |
| 4570 main event loop in the client's application. This is typically |
| 4571 a requirement of a GUI, when the main event loop needs to be triggered |
| 4572 within a certain time period. |
| 4573 |
| 4574 Normally, Bison generates a pull parser. |
| 4575 The following Bison declaration says that you want the parser to be a push |
| 4576 parser (@pxref{Decl Summary,,%define api.push_pull}): |
| 4577 |
| 4578 @example |
| 4579 %define api.push_pull "push" |
| 4580 @end example |
| 4581 |
| 4582 In almost all cases, you want to ensure that your push parser is also |
| 4583 a pure parser (@pxref{Pure Decl, ,A Pure (Reentrant) Parser}). The only |
| 4584 time you should create an impure push parser is to have backwards |
| 4585 compatibility with the impure Yacc pull mode interface. Unless you know |
| 4586 what you are doing, your declarations should look like this: |
| 4587 |
| 4588 @example |
| 4589 %define api.pure |
| 4590 %define api.push_pull "push" |
| 4591 @end example |
| 4592 |
| 4593 There is a major notable functional difference between the pure push parser |
| 4594 and the impure push parser. It is acceptable for a pure push parser to have |
| 4595 many parser instances, of the same type of parser, in memory at the same time. |
| 4596 An impure push parser should only use one parser at a time. |
| 4597 |
| 4598 When a push parser is selected, Bison will generate some new symbols in |
| 4599 the generated parser. @code{yypstate} is a structure that the generated |
| 4600 parser uses to store the parser's state. @code{yypstate_new} is the |
| 4601 function that will create a new parser instance. @code{yypstate_delete} |
| 4602 will free the resources associated with the corresponding parser instance. |
| 4603 Finally, @code{yypush_parse} is the function that should be called whenever a |
| 4604 token is available to provide the parser. A trivial example |
| 4605 of using a pure push parser would look like this: |
| 4606 |
| 4607 @example |
| 4608 int status; |
| 4609 yypstate *ps = yypstate_new (); |
| 4610 do @{ |
| 4611 status = yypush_parse (ps, yylex (), NULL); |
| 4612 @} while (status == YYPUSH_MORE); |
| 4613 yypstate_delete (ps); |
| 4614 @end example |
| 4615 |
| 4616 If the user decided to use an impure push parser, a few things about |
| 4617 the generated parser will change. The @code{yychar} variable becomes |
| 4618 a global variable instead of a variable in the @code{yypush_parse} function. |
| 4619 For this reason, the signature of the @code{yypush_parse} function is |
| 4620 changed to remove the token as a parameter. A nonreentrant push parser |
| 4621 example would thus look like this: |
| 4622 |
| 4623 @example |
| 4624 extern int yychar; |
| 4625 int status; |
| 4626 yypstate *ps = yypstate_new (); |
| 4627 do @{ |
| 4628 yychar = yylex (); |
| 4629 status = yypush_parse (ps); |
| 4630 @} while (status == YYPUSH_MORE); |
| 4631 yypstate_delete (ps); |
| 4632 @end example |
| 4633 |
| 4634 That's it. Notice the next token is put into the global variable @code{yychar} |
| 4635 for use by the next invocation of the @code{yypush_parse} function. |
| 4636 |
| 4637 Bison also supports both the push parser interface along with the pull parser |
| 4638 interface in the same generated parser. In order to get this functionality, |
| 4639 you should replace the @code{%define api.push_pull "push"} declaration with the |
| 4640 @code{%define api.push_pull "both"} declaration. Doing this will create all of |
| 4641 the symbols mentioned earlier along with the two extra symbols, @code{yyparse} |
| 4642 and @code{yypull_parse}. @code{yyparse} can be used exactly as it normally |
| 4643 would be used. However, the user should note that it is implemented in the |
| 4644 generated parser by calling @code{yypull_parse}. |
| 4645 This makes the @code{yyparse} function that is generated with the |
| 4646 @code{%define api.push_pull "both"} declaration slower than the normal |
| 4647 @code{yyparse} function. If the user |
| 4648 calls the @code{yypull_parse} function it will parse the rest of the input |
| 4649 stream. It is possible to @code{yypush_parse} tokens to select a subgrammar |
| 4650 and then @code{yypull_parse} the rest of the input stream. If you would like |
| 4651 to switch back and forth between between parsing styles, you would have to |
| 4652 write your own @code{yypull_parse} function that knows when to quit looking |
| 4653 for input. An example of using the @code{yypull_parse} function would look |
| 4654 like this: |
| 4655 |
| 4656 @example |
| 4657 yypstate *ps = yypstate_new (); |
| 4658 yypull_parse (ps); /* Will call the lexer */ |
| 4659 yypstate_delete (ps); |
| 4660 @end example |
| 4661 |
| 4662 Adding the @code{%define api.pure} declaration does exactly the same thing to |
| 4663 the generated parser with @code{%define api.push_pull "both"} as it did for |
| 4664 @code{%define api.push_pull "push"}. |
| 4665 |
| 4666 @node Decl Summary |
| 4667 @subsection Bison Declaration Summary |
| 4668 @cindex Bison declaration summary |
| 4669 @cindex declaration summary |
| 4670 @cindex summary, Bison declaration |
| 4671 |
| 4672 Here is a summary of the declarations used to define a grammar: |
| 4673 |
| 4674 @deffn {Directive} %union |
| 4675 Declare the collection of data types that semantic values may have |
| 4676 (@pxref{Union Decl, ,The Collection of Value Types}). |
| 4677 @end deffn |
| 4678 |
| 4679 @deffn {Directive} %token |
| 4680 Declare a terminal symbol (token type name) with no precedence |
| 4681 or associativity specified (@pxref{Token Decl, ,Token Type Names}). |
| 4682 @end deffn |
| 4683 |
| 4684 @deffn {Directive} %right |
| 4685 Declare a terminal symbol (token type name) that is right-associative |
| 4686 (@pxref{Precedence Decl, ,Operator Precedence}). |
| 4687 @end deffn |
| 4688 |
| 4689 @deffn {Directive} %left |
| 4690 Declare a terminal symbol (token type name) that is left-associative |
| 4691 (@pxref{Precedence Decl, ,Operator Precedence}). |
| 4692 @end deffn |
| 4693 |
| 4694 @deffn {Directive} %nonassoc |
| 4695 Declare a terminal symbol (token type name) that is nonassociative |
| 4696 (@pxref{Precedence Decl, ,Operator Precedence}). |
| 4697 Using it in a way that would be associative is a syntax error. |
| 4698 @end deffn |
| 4699 |
| 4700 @ifset defaultprec |
| 4701 @deffn {Directive} %default-prec |
| 4702 Assign a precedence to rules lacking an explicit @code{%prec} modifier |
| 4703 (@pxref{Contextual Precedence, ,Context-Dependent Precedence}). |
| 4704 @end deffn |
| 4705 @end ifset |
| 4706 |
| 4707 @deffn {Directive} %type |
| 4708 Declare the type of semantic values for a nonterminal symbol |
| 4709 (@pxref{Type Decl, ,Nonterminal Symbols}). |
| 4710 @end deffn |
| 4711 |
| 4712 @deffn {Directive} %start |
| 4713 Specify the grammar's start symbol (@pxref{Start Decl, ,The |
| 4714 Start-Symbol}). |
| 4715 @end deffn |
| 4716 |
| 4717 @deffn {Directive} %expect |
| 4718 Declare the expected number of shift-reduce conflicts |
| 4719 (@pxref{Expect Decl, ,Suppressing Conflict Warnings}). |
| 4720 @end deffn |
| 4721 |
| 4722 |
| 4723 @sp 1 |
| 4724 @noindent |
| 4725 In order to change the behavior of @command{bison}, use the following |
| 4726 directives: |
| 4727 |
| 4728 @deffn {Directive} %code @{@var{code}@} |
| 4729 @findex %code |
| 4730 This is the unqualified form of the @code{%code} directive. |
| 4731 It inserts @var{code} verbatim at a language-dependent default location in the |
| 4732 output@footnote{The default location is actually skeleton-dependent; |
| 4733 writers of non-standard skeletons however should choose the default location |
| 4734 consistently with the behavior of the standard Bison skeletons.}. |
| 4735 |
| 4736 @cindex Prologue |
| 4737 For C/C++, the default location is the parser source code |
| 4738 file after the usual contents of the parser header file. |
| 4739 Thus, @code{%code} replaces the traditional Yacc prologue, |
| 4740 @code{%@{@var{code}%@}}, for most purposes. |
| 4741 For a detailed discussion, see @ref{Prologue Alternatives}. |
| 4742 |
| 4743 For Java, the default location is inside the parser class. |
| 4744 |
| 4745 (Like all the Yacc prologue alternatives, this directive is experimental. |
| 4746 More user feedback will help to determine whether it should become a permanent |
| 4747 feature.) |
| 4748 @end deffn |
| 4749 |
| 4750 @deffn {Directive} %code @var{qualifier} @{@var{code}@} |
| 4751 This is the qualified form of the @code{%code} directive. |
| 4752 If you need to specify location-sensitive verbatim @var{code} that does not |
| 4753 belong at the default location selected by the unqualified @code{%code} form, |
| 4754 use this form instead. |
| 4755 |
| 4756 @var{qualifier} identifies the purpose of @var{code} and thus the location(s) |
| 4757 where Bison should generate it. |
| 4758 Not all values of @var{qualifier} are available for all target languages: |
| 4759 |
| 4760 @itemize @bullet |
| 4761 @item requires |
| 4762 @findex %code requires |
| 4763 |
| 4764 @itemize @bullet |
| 4765 @item Language(s): C, C++ |
| 4766 |
| 4767 @item Purpose: This is the best place to write dependency code required for |
| 4768 @code{YYSTYPE} and @code{YYLTYPE}. |
| 4769 In other words, it's the best place to define types referenced in @code{%union} |
| 4770 directives, and it's the best place to override Bison's default @code{YYSTYPE} |
| 4771 and @code{YYLTYPE} definitions. |
| 4772 |
| 4773 @item Location(s): The parser header file and the parser source code file |
| 4774 before the Bison-generated @code{YYSTYPE} and @code{YYLTYPE} definitions. |
| 4775 @end itemize |
| 4776 |
| 4777 @item provides |
| 4778 @findex %code provides |
| 4779 |
| 4780 @itemize @bullet |
| 4781 @item Language(s): C, C++ |
| 4782 |
| 4783 @item Purpose: This is the best place to write additional definitions and |
| 4784 declarations that should be provided to other modules. |
| 4785 |
| 4786 @item Location(s): The parser header file and the parser source code file after |
| 4787 the Bison-generated @code{YYSTYPE}, @code{YYLTYPE}, and token definitions. |
| 4788 @end itemize |
| 4789 |
| 4790 @item top |
| 4791 @findex %code top |
| 4792 |
| 4793 @itemize @bullet |
| 4794 @item Language(s): C, C++ |
| 4795 |
| 4796 @item Purpose: The unqualified @code{%code} or @code{%code requires} should |
| 4797 usually be more appropriate than @code{%code top}. |
| 4798 However, occasionally it is necessary to insert code much nearer the top of the |
| 4799 parser source code file. |
| 4800 For example: |
| 4801 |
| 4802 @smallexample |
| 4803 %code top @{ |
| 4804 #define _GNU_SOURCE |
| 4805 #include <stdio.h> |
| 4806 @} |
| 4807 @end smallexample |
| 4808 |
| 4809 @item Location(s): Near the top of the parser source code file. |
| 4810 @end itemize |
| 4811 |
| 4812 @item imports |
| 4813 @findex %code imports |
| 4814 |
| 4815 @itemize @bullet |
| 4816 @item Language(s): Java |
| 4817 |
| 4818 @item Purpose: This is the best place to write Java import directives. |
| 4819 |
| 4820 @item Location(s): The parser Java file after any Java package directive and |
| 4821 before any class definitions. |
| 4822 @end itemize |
| 4823 @end itemize |
| 4824 |
| 4825 (Like all the Yacc prologue alternatives, this directive is experimental. |
| 4826 More user feedback will help to determine whether it should become a permanent |
| 4827 feature.) |
| 4828 |
| 4829 @cindex Prologue |
| 4830 For a detailed discussion of how to use @code{%code} in place of the |
| 4831 traditional Yacc prologue for C/C++, see @ref{Prologue Alternatives}. |
| 4832 @end deffn |
| 4833 |
| 4834 @deffn {Directive} %debug |
| 4835 In the parser file, define the macro @code{YYDEBUG} to 1 if it is not |
| 4836 already defined, so that the debugging facilities are compiled. |
| 4837 @end deffn |
| 4838 @xref{Tracing, ,Tracing Your Parser}. |
| 4839 |
| 4840 @deffn {Directive} %define @var{variable} |
| 4841 @deffnx {Directive} %define @var{variable} "@var{value}" |
| 4842 Define a variable to adjust Bison's behavior. |
| 4843 The possible choices for @var{variable}, as well as their meanings, depend on |
| 4844 the selected target language and/or the parser skeleton (@pxref{Decl |
| 4845 Summary,,%language}, @pxref{Decl Summary,,%skeleton}). |
| 4846 |
| 4847 Bison will warn if a @var{variable} is defined multiple times. |
| 4848 |
| 4849 Omitting @code{"@var{value}"} is always equivalent to specifying it as |
| 4850 @code{""}. |
| 4851 |
| 4852 Some @var{variable}s may be used as Booleans. |
| 4853 In this case, Bison will complain if the variable definition does not meet one |
| 4854 of the following four conditions: |
| 4855 |
| 4856 @enumerate |
| 4857 @item @code{"@var{value}"} is @code{"true"} |
| 4858 |
| 4859 @item @code{"@var{value}"} is omitted (or is @code{""}). |
| 4860 This is equivalent to @code{"true"}. |
| 4861 |
| 4862 @item @code{"@var{value}"} is @code{"false"}. |
| 4863 |
| 4864 @item @var{variable} is never defined. |
| 4865 In this case, Bison selects a default value, which may depend on the selected |
| 4866 target language and/or parser skeleton. |
| 4867 @end enumerate |
| 4868 |
| 4869 Some of the accepted @var{variable}s are: |
| 4870 |
| 4871 @itemize @bullet |
| 4872 @item api.pure |
| 4873 @findex %define api.pure |
| 4874 |
| 4875 @itemize @bullet |
| 4876 @item Language(s): C |
| 4877 |
| 4878 @item Purpose: Request a pure (reentrant) parser program. |
| 4879 @xref{Pure Decl, ,A Pure (Reentrant) Parser}. |
| 4880 |
| 4881 @item Accepted Values: Boolean |
| 4882 |
| 4883 @item Default Value: @code{"false"} |
| 4884 @end itemize |
| 4885 |
| 4886 @item api.push_pull |
| 4887 @findex %define api.push_pull |
| 4888 |
| 4889 @itemize @bullet |
| 4890 @item Language(s): C (LALR(1) only) |
| 4891 |
| 4892 @item Purpose: Requests a pull parser, a push parser, or both. |
| 4893 @xref{Push Decl, ,A Push Parser}. |
| 4894 (The current push parsing interface is experimental and may evolve. |
| 4895 More user feedback will help to stabilize it.) |
| 4896 |
| 4897 @item Accepted Values: @code{"pull"}, @code{"push"}, @code{"both"} |
| 4898 |
| 4899 @item Default Value: @code{"pull"} |
| 4900 @end itemize |
| 4901 |
| 4902 @item lr.keep_unreachable_states |
| 4903 @findex %define lr.keep_unreachable_states |
| 4904 |
| 4905 @itemize @bullet |
| 4906 @item Language(s): all |
| 4907 |
| 4908 @item Purpose: Requests that Bison allow unreachable parser states to remain in |
| 4909 the parser tables. |
| 4910 Bison considers a state to be unreachable if there exists no sequence of |
| 4911 transitions from the start state to that state. |
| 4912 A state can become unreachable during conflict resolution if Bison disables a |
| 4913 shift action leading to it from a predecessor state. |
| 4914 Keeping unreachable states is sometimes useful for analysis purposes, but they |
| 4915 are useless in the generated parser. |
| 4916 |
| 4917 @item Accepted Values: Boolean |
| 4918 |
| 4919 @item Default Value: @code{"false"} |
| 4920 |
| 4921 @item Caveats: |
| 4922 |
| 4923 @itemize @bullet |
| 4924 |
| 4925 @item Unreachable states may contain conflicts and may use rules not used in |
| 4926 any other state. |
| 4927 Thus, keeping unreachable states may induce warnings that are irrelevant to |
| 4928 your parser's behavior, and it may eliminate warnings that are relevant. |
| 4929 Of course, the change in warnings may actually be relevant to a parser table |
| 4930 analysis that wants to keep unreachable states, so this behavior will likely |
| 4931 remain in future Bison releases. |
| 4932 |
| 4933 @item While Bison is able to remove unreachable states, it is not guaranteed to |
| 4934 remove other kinds of useless states. |
| 4935 Specifically, when Bison disables reduce actions during conflict resolution, |
| 4936 some goto actions may become useless, and thus some additional states may |
| 4937 become useless. |
| 4938 If Bison were to compute which goto actions were useless and then disable those |
| 4939 actions, it could identify such states as unreachable and then remove those |
| 4940 states. |
| 4941 However, Bison does not compute which goto actions are useless. |
| 4942 @end itemize |
| 4943 @end itemize |
| 4944 |
| 4945 @item namespace |
| 4946 @findex %define namespace |
| 4947 |
| 4948 @itemize |
| 4949 @item Languages(s): C++ |
| 4950 |
| 4951 @item Purpose: Specifies the namespace for the parser class. |
| 4952 For example, if you specify: |
| 4953 |
| 4954 @smallexample |
| 4955 %define namespace "foo::bar" |
| 4956 @end smallexample |
| 4957 |
| 4958 Bison uses @code{foo::bar} verbatim in references such as: |
| 4959 |
| 4960 @smallexample |
| 4961 foo::bar::parser::semantic_type |
| 4962 @end smallexample |
| 4963 |
| 4964 However, to open a namespace, Bison removes any leading @code{::} and then |
| 4965 splits on any remaining occurrences: |
| 4966 |
| 4967 @smallexample |
| 4968 namespace foo @{ namespace bar @{ |
| 4969 class position; |
| 4970 class location; |
| 4971 @} @} |
| 4972 @end smallexample |
| 4973 |
| 4974 @item Accepted Values: Any absolute or relative C++ namespace reference without |
| 4975 a trailing @code{"::"}. |
| 4976 For example, @code{"foo"} or @code{"::foo::bar"}. |
| 4977 |
| 4978 @item Default Value: The value specified by @code{%name-prefix}, which defaults |
| 4979 to @code{yy}. |
| 4980 This usage of @code{%name-prefix} is for backward compatibility and can be |
| 4981 confusing since @code{%name-prefix} also specifies the textual prefix for the |
| 4982 lexical analyzer function. |
| 4983 Thus, if you specify @code{%name-prefix}, it is best to also specify |
| 4984 @code{%define namespace} so that @code{%name-prefix} @emph{only} affects the |
| 4985 lexical analyzer function. |
| 4986 For example, if you specify: |
| 4987 |
| 4988 @smallexample |
| 4989 %define namespace "foo" |
| 4990 %name-prefix "bar::" |
| 4991 @end smallexample |
| 4992 |
| 4993 The parser namespace is @code{foo} and @code{yylex} is referenced as |
| 4994 @code{bar::lex}. |
| 4995 @end itemize |
| 4996 @end itemize |
| 4997 |
| 4998 @end deffn |
| 4999 |
| 5000 @deffn {Directive} %defines |
| 5001 Write a header file containing macro definitions for the token type |
| 5002 names defined in the grammar as well as a few other declarations. |
| 5003 If the parser output file is named @file{@var{name}.c} then this file |
| 5004 is named @file{@var{name}.h}. |
| 5005 |
| 5006 For C parsers, the output header declares @code{YYSTYPE} unless |
| 5007 @code{YYSTYPE} is already defined as a macro or you have used a |
| 5008 @code{<@var{type}>} tag without using @code{%union}. |
| 5009 Therefore, if you are using a @code{%union} |
| 5010 (@pxref{Multiple Types, ,More Than One Value Type}) with components that |
| 5011 require other definitions, or if you have defined a @code{YYSTYPE} macro |
| 5012 or type definition |
| 5013 (@pxref{Value Type, ,Data Types of Semantic Values}), you need to |
| 5014 arrange for these definitions to be propagated to all modules, e.g., by |
| 5015 putting them in a prerequisite header that is included both by your |
| 5016 parser and by any other module that needs @code{YYSTYPE}. |
| 5017 |
| 5018 Unless your parser is pure, the output header declares @code{yylval} |
| 5019 as an external variable. @xref{Pure Decl, ,A Pure (Reentrant) |
| 5020 Parser}. |
| 5021 |
| 5022 If you have also used locations, the output header declares |
| 5023 @code{YYLTYPE} and @code{yylloc} using a protocol similar to that of |
| 5024 the @code{YYSTYPE} macro and @code{yylval}. @xref{Locations, ,Tracking |
| 5025 Locations}. |
| 5026 |
| 5027 This output file is normally essential if you wish to put the definition |
| 5028 of @code{yylex} in a separate source file, because @code{yylex} |
| 5029 typically needs to be able to refer to the above-mentioned declarations |
| 5030 and to the token type codes. @xref{Token Values, ,Semantic Values of |
| 5031 Tokens}. |
| 5032 |
| 5033 @findex %code requires |
| 5034 @findex %code provides |
| 5035 If you have declared @code{%code requires} or @code{%code provides}, the output |
| 5036 header also contains their code. |
| 5037 @xref{Decl Summary, ,%code}. |
| 5038 @end deffn |
| 5039 |
| 5040 @deffn {Directive} %defines @var{defines-file} |
| 5041 Same as above, but save in the file @var{defines-file}. |
| 5042 @end deffn |
| 5043 |
| 5044 @deffn {Directive} %destructor |
| 5045 Specify how the parser should reclaim the memory associated to |
| 5046 discarded symbols. @xref{Destructor Decl, , Freeing Discarded Symbols}. |
| 5047 @end deffn |
| 5048 |
| 5049 @deffn {Directive} %file-prefix "@var{prefix}" |
| 5050 Specify a prefix to use for all Bison output file names. The names are |
| 5051 chosen as if the input file were named @file{@var{prefix}.y}. |
| 5052 @end deffn |
| 5053 |
| 5054 @deffn {Directive} %language "@var{language}" |
| 5055 Specify the programming language for the generated parser. Currently |
| 5056 supported languages include C, C++, and Java. |
| 5057 @var{language} is case-insensitive. |
| 5058 |
| 5059 This directive is experimental and its effect may be modified in future |
| 5060 releases. |
| 5061 @end deffn |
| 5062 |
| 5063 @deffn {Directive} %locations |
| 5064 Generate the code processing the locations (@pxref{Action Features, |
| 5065 ,Special Features for Use in Actions}). This mode is enabled as soon as |
| 5066 the grammar uses the special @samp{@@@var{n}} tokens, but if your |
| 5067 grammar does not use it, using @samp{%locations} allows for more |
| 5068 accurate syntax error messages. |
| 5069 @end deffn |
| 5070 |
| 5071 @deffn {Directive} %name-prefix "@var{prefix}" |
| 5072 Rename the external symbols used in the parser so that they start with |
| 5073 @var{prefix} instead of @samp{yy}. The precise list of symbols renamed |
| 5074 in C parsers |
| 5075 is @code{yyparse}, @code{yylex}, @code{yyerror}, @code{yynerrs}, |
| 5076 @code{yylval}, @code{yychar}, @code{yydebug}, and |
| 5077 (if locations are used) @code{yylloc}. If you use a push parser, |
| 5078 @code{yypush_parse}, @code{yypull_parse}, @code{yypstate}, |
| 5079 @code{yypstate_new} and @code{yypstate_delete} will |
| 5080 also be renamed. For example, if you use @samp{%name-prefix "c_"}, the |
| 5081 names become @code{c_parse}, @code{c_lex}, and so on. |
| 5082 For C++ parsers, see the @code{%define namespace} documentation in this |
| 5083 section. |
| 5084 @xref{Multiple Parsers, ,Multiple Parsers in the Same Program}. |
| 5085 @end deffn |
| 5086 |
| 5087 @ifset defaultprec |
| 5088 @deffn {Directive} %no-default-prec |
| 5089 Do not assign a precedence to rules lacking an explicit @code{%prec} |
| 5090 modifier (@pxref{Contextual Precedence, ,Context-Dependent |
| 5091 Precedence}). |
| 5092 @end deffn |
| 5093 @end ifset |
| 5094 |
| 5095 @deffn {Directive} %no-lines |
| 5096 Don't generate any @code{#line} preprocessor commands in the parser |
| 5097 file. Ordinarily Bison writes these commands in the parser file so that |
| 5098 the C compiler and debuggers will associate errors and object code with |
| 5099 your source file (the grammar file). This directive causes them to |
| 5100 associate errors with the parser file, treating it an independent source |
| 5101 file in its own right. |
| 5102 @end deffn |
| 5103 |
| 5104 @deffn {Directive} %output "@var{file}" |
| 5105 Specify @var{file} for the parser file. |
| 5106 @end deffn |
| 5107 |
| 5108 @deffn {Directive} %pure-parser |
| 5109 Deprecated version of @code{%define api.pure} (@pxref{Decl Summary, ,%define}), |
| 5110 for which Bison is more careful to warn about unreasonable usage. |
| 5111 @end deffn |
| 5112 |
| 5113 @deffn {Directive} %require "@var{version}" |
| 5114 Require version @var{version} or higher of Bison. @xref{Require Decl, , |
| 5115 Require a Version of Bison}. |
| 5116 @end deffn |
| 5117 |
| 5118 @deffn {Directive} %skeleton "@var{file}" |
| 5119 Specify the skeleton to use. |
| 5120 |
| 5121 @c You probably don't need this option unless you are developing Bison. |
| 5122 @c You should use @code{%language} if you want to specify the skeleton for a |
| 5123 @c different language, because it is clearer and because it will always choose t
he |
| 5124 @c correct skeleton for non-deterministic or push parsers. |
| 5125 |
| 5126 If @var{file} does not contain a @code{/}, @var{file} is the name of a skeleton |
| 5127 file in the Bison installation directory. |
| 5128 If it does, @var{file} is an absolute file name or a file name relative to the |
| 5129 directory of the grammar file. |
| 5130 This is similar to how most shells resolve commands. |
| 5131 @end deffn |
| 5132 |
| 5133 @deffn {Directive} %token-table |
| 5134 Generate an array of token names in the parser file. The name of the |
| 5135 array is @code{yytname}; @code{yytname[@var{i}]} is the name of the |
| 5136 token whose internal Bison token code number is @var{i}. The first |
| 5137 three elements of @code{yytname} correspond to the predefined tokens |
| 5138 @code{"$end"}, |
| 5139 @code{"error"}, and @code{"$undefined"}; after these come the symbols |
| 5140 defined in the grammar file. |
| 5141 |
| 5142 The name in the table includes all the characters needed to represent |
| 5143 the token in Bison. For single-character literals and literal |
| 5144 strings, this includes the surrounding quoting characters and any |
| 5145 escape sequences. For example, the Bison single-character literal |
| 5146 @code{'+'} corresponds to a three-character name, represented in C as |
| 5147 @code{"'+'"}; and the Bison two-character literal string @code{"\\/"} |
| 5148 corresponds to a five-character name, represented in C as |
| 5149 @code{"\"\\\\/\""}. |
| 5150 |
| 5151 When you specify @code{%token-table}, Bison also generates macro |
| 5152 definitions for macros @code{YYNTOKENS}, @code{YYNNTS}, and |
| 5153 @code{YYNRULES}, and @code{YYNSTATES}: |
| 5154 |
| 5155 @table @code |
| 5156 @item YYNTOKENS |
| 5157 The highest token number, plus one. |
| 5158 @item YYNNTS |
| 5159 The number of nonterminal symbols. |
| 5160 @item YYNRULES |
| 5161 The number of grammar rules, |
| 5162 @item YYNSTATES |
| 5163 The number of parser states (@pxref{Parser States}). |
| 5164 @end table |
| 5165 @end deffn |
| 5166 |
| 5167 @deffn {Directive} %verbose |
| 5168 Write an extra output file containing verbose descriptions of the |
| 5169 parser states and what is done for each type of lookahead token in |
| 5170 that state. @xref{Understanding, , Understanding Your Parser}, for more |
| 5171 information. |
| 5172 @end deffn |
| 5173 |
| 5174 @deffn {Directive} %yacc |
| 5175 Pretend the option @option{--yacc} was given, i.e., imitate Yacc, |
| 5176 including its naming conventions. @xref{Bison Options}, for more. |
| 5177 @end deffn |
| 5178 |
| 5179 |
| 5180 @node Multiple Parsers |
| 5181 @section Multiple Parsers in the Same Program |
| 5182 |
| 5183 Most programs that use Bison parse only one language and therefore contain |
| 5184 only one Bison parser. But what if you want to parse more than one |
| 5185 language with the same program? Then you need to avoid a name conflict |
| 5186 between different definitions of @code{yyparse}, @code{yylval}, and so on. |
| 5187 |
| 5188 The easy way to do this is to use the option @samp{-p @var{prefix}} |
| 5189 (@pxref{Invocation, ,Invoking Bison}). This renames the interface |
| 5190 functions and variables of the Bison parser to start with @var{prefix} |
| 5191 instead of @samp{yy}. You can use this to give each parser distinct |
| 5192 names that do not conflict. |
| 5193 |
| 5194 The precise list of symbols renamed is @code{yyparse}, @code{yylex}, |
| 5195 @code{yyerror}, @code{yynerrs}, @code{yylval}, @code{yylloc}, |
| 5196 @code{yychar} and @code{yydebug}. If you use a push parser, |
| 5197 @code{yypush_parse}, @code{yypull_parse}, @code{yypstate}, |
| 5198 @code{yypstate_new} and @code{yypstate_delete} will also be renamed. |
| 5199 For example, if you use @samp{-p c}, the names become @code{cparse}, |
| 5200 @code{clex}, and so on. |
| 5201 |
| 5202 @strong{All the other variables and macros associated with Bison are not |
| 5203 renamed.} These others are not global; there is no conflict if the same |
| 5204 name is used in different parsers. For example, @code{YYSTYPE} is not |
| 5205 renamed, but defining this in different ways in different parsers causes |
| 5206 no trouble (@pxref{Value Type, ,Data Types of Semantic Values}). |
| 5207 |
| 5208 The @samp{-p} option works by adding macro definitions to the beginning |
| 5209 of the parser source file, defining @code{yyparse} as |
| 5210 @code{@var{prefix}parse}, and so on. This effectively substitutes one |
| 5211 name for the other in the entire parser file. |
| 5212 |
| 5213 @node Interface |
| 5214 @chapter Parser C-Language Interface |
| 5215 @cindex C-language interface |
| 5216 @cindex interface |
| 5217 |
| 5218 The Bison parser is actually a C function named @code{yyparse}. Here we |
| 5219 describe the interface conventions of @code{yyparse} and the other |
| 5220 functions that it needs to use. |
| 5221 |
| 5222 Keep in mind that the parser uses many C identifiers starting with |
| 5223 @samp{yy} and @samp{YY} for internal purposes. If you use such an |
| 5224 identifier (aside from those in this manual) in an action or in epilogue |
| 5225 in the grammar file, you are likely to run into trouble. |
| 5226 |
| 5227 @menu |
| 5228 * Parser Function:: How to call @code{yyparse} and what it returns. |
| 5229 * Push Parser Function:: How to call @code{yypush_parse} and what it returns. |
| 5230 * Pull Parser Function:: How to call @code{yypull_parse} and what it returns. |
| 5231 * Parser Create Function:: How to call @code{yypstate_new} and what it returns. |
| 5232 * Parser Delete Function:: How to call @code{yypstate_delete} and what it retur
ns. |
| 5233 * Lexical:: You must supply a function @code{yylex} |
| 5234 which reads tokens. |
| 5235 * Error Reporting:: You must supply a function @code{yyerror}. |
| 5236 * Action Features:: Special features for use in actions. |
| 5237 * Internationalization:: How to let the parser speak in the user's |
| 5238 native language. |
| 5239 @end menu |
| 5240 |
| 5241 @node Parser Function |
| 5242 @section The Parser Function @code{yyparse} |
| 5243 @findex yyparse |
| 5244 |
| 5245 You call the function @code{yyparse} to cause parsing to occur. This |
| 5246 function reads tokens, executes actions, and ultimately returns when it |
| 5247 encounters end-of-input or an unrecoverable syntax error. You can also |
| 5248 write an action which directs @code{yyparse} to return immediately |
| 5249 without reading further. |
| 5250 |
| 5251 |
| 5252 @deftypefun int yyparse (void) |
| 5253 The value returned by @code{yyparse} is 0 if parsing was successful (return |
| 5254 is due to end-of-input). |
| 5255 |
| 5256 The value is 1 if parsing failed because of invalid input, i.e., input |
| 5257 that contains a syntax error or that causes @code{YYABORT} to be |
| 5258 invoked. |
| 5259 |
| 5260 The value is 2 if parsing failed due to memory exhaustion. |
| 5261 @end deftypefun |
| 5262 |
| 5263 In an action, you can cause immediate return from @code{yyparse} by using |
| 5264 these macros: |
| 5265 |
| 5266 @defmac YYACCEPT |
| 5267 @findex YYACCEPT |
| 5268 Return immediately with value 0 (to report success). |
| 5269 @end defmac |
| 5270 |
| 5271 @defmac YYABORT |
| 5272 @findex YYABORT |
| 5273 Return immediately with value 1 (to report failure). |
| 5274 @end defmac |
| 5275 |
| 5276 If you use a reentrant parser, you can optionally pass additional |
| 5277 parameter information to it in a reentrant way. To do so, use the |
| 5278 declaration @code{%parse-param}: |
| 5279 |
| 5280 @deffn {Directive} %parse-param @{@var{argument-declaration}@} |
| 5281 @findex %parse-param |
| 5282 Declare that an argument declared by the braced-code |
| 5283 @var{argument-declaration} is an additional @code{yyparse} argument. |
| 5284 The @var{argument-declaration} is used when declaring |
| 5285 functions or prototypes. The last identifier in |
| 5286 @var{argument-declaration} must be the argument name. |
| 5287 @end deffn |
| 5288 |
| 5289 Here's an example. Write this in the parser: |
| 5290 |
| 5291 @example |
| 5292 %parse-param @{int *nastiness@} |
| 5293 %parse-param @{int *randomness@} |
| 5294 @end example |
| 5295 |
| 5296 @noindent |
| 5297 Then call the parser like this: |
| 5298 |
| 5299 @example |
| 5300 @{ |
| 5301 int nastiness, randomness; |
| 5302 @dots{} /* @r{Store proper data in @code{nastiness} and @code{randomness}.}
*/ |
| 5303 value = yyparse (&nastiness, &randomness); |
| 5304 @dots{} |
| 5305 @} |
| 5306 @end example |
| 5307 |
| 5308 @noindent |
| 5309 In the grammar actions, use expressions like this to refer to the data: |
| 5310 |
| 5311 @example |
| 5312 exp: @dots{} @{ @dots{}; *randomness += 1; @dots{} @} |
| 5313 @end example |
| 5314 |
| 5315 @node Push Parser Function |
| 5316 @section The Push Parser Function @code{yypush_parse} |
| 5317 @findex yypush_parse |
| 5318 |
| 5319 (The current push parsing interface is experimental and may evolve. |
| 5320 More user feedback will help to stabilize it.) |
| 5321 |
| 5322 You call the function @code{yypush_parse} to parse a single token. This |
| 5323 function is available if either the @code{%define api.push_pull "push"} or |
| 5324 @code{%define api.push_pull "both"} declaration is used. |
| 5325 @xref{Push Decl, ,A Push Parser}. |
| 5326 |
| 5327 @deftypefun int yypush_parse (yypstate *yyps) |
| 5328 The value returned by @code{yypush_parse} is the same as for yyparse with the |
| 5329 following exception. @code{yypush_parse} will return YYPUSH_MORE if more input |
| 5330 is required to finish parsing the grammar. |
| 5331 @end deftypefun |
| 5332 |
| 5333 @node Pull Parser Function |
| 5334 @section The Pull Parser Function @code{yypull_parse} |
| 5335 @findex yypull_parse |
| 5336 |
| 5337 (The current push parsing interface is experimental and may evolve. |
| 5338 More user feedback will help to stabilize it.) |
| 5339 |
| 5340 You call the function @code{yypull_parse} to parse the rest of the input |
| 5341 stream. This function is available if the @code{%define api.push_pull "both"} |
| 5342 declaration is used. |
| 5343 @xref{Push Decl, ,A Push Parser}. |
| 5344 |
| 5345 @deftypefun int yypull_parse (yypstate *yyps) |
| 5346 The value returned by @code{yypull_parse} is the same as for @code{yyparse}. |
| 5347 @end deftypefun |
| 5348 |
| 5349 @node Parser Create Function |
| 5350 @section The Parser Create Function @code{yystate_new} |
| 5351 @findex yypstate_new |
| 5352 |
| 5353 (The current push parsing interface is experimental and may evolve. |
| 5354 More user feedback will help to stabilize it.) |
| 5355 |
| 5356 You call the function @code{yypstate_new} to create a new parser instance. |
| 5357 This function is available if either the @code{%define api.push_pull "push"} or |
| 5358 @code{%define api.push_pull "both"} declaration is used. |
| 5359 @xref{Push Decl, ,A Push Parser}. |
| 5360 |
| 5361 @deftypefun yypstate *yypstate_new (void) |
| 5362 The fuction will return a valid parser instance if there was memory available |
| 5363 or 0 if no memory was available. |
| 5364 In impure mode, it will also return 0 if a parser instance is currently |
| 5365 allocated. |
| 5366 @end deftypefun |
| 5367 |
| 5368 @node Parser Delete Function |
| 5369 @section The Parser Delete Function @code{yystate_delete} |
| 5370 @findex yypstate_delete |
| 5371 |
| 5372 (The current push parsing interface is experimental and may evolve. |
| 5373 More user feedback will help to stabilize it.) |
| 5374 |
| 5375 You call the function @code{yypstate_delete} to delete a parser instance. |
| 5376 function is available if either the @code{%define api.push_pull "push"} or |
| 5377 @code{%define api.push_pull "both"} declaration is used. |
| 5378 @xref{Push Decl, ,A Push Parser}. |
| 5379 |
| 5380 @deftypefun void yypstate_delete (yypstate *yyps) |
| 5381 This function will reclaim the memory associated with a parser instance. |
| 5382 After this call, you should no longer attempt to use the parser instance. |
| 5383 @end deftypefun |
| 5384 |
| 5385 @node Lexical |
| 5386 @section The Lexical Analyzer Function @code{yylex} |
| 5387 @findex yylex |
| 5388 @cindex lexical analyzer |
| 5389 |
| 5390 The @dfn{lexical analyzer} function, @code{yylex}, recognizes tokens from |
| 5391 the input stream and returns them to the parser. Bison does not create |
| 5392 this function automatically; you must write it so that @code{yyparse} can |
| 5393 call it. The function is sometimes referred to as a lexical scanner. |
| 5394 |
| 5395 In simple programs, @code{yylex} is often defined at the end of the Bison |
| 5396 grammar file. If @code{yylex} is defined in a separate source file, you |
| 5397 need to arrange for the token-type macro definitions to be available there. |
| 5398 To do this, use the @samp{-d} option when you run Bison, so that it will |
| 5399 write these macro definitions into a separate header file |
| 5400 @file{@var{name}.tab.h} which you can include in the other source files |
| 5401 that need it. @xref{Invocation, ,Invoking Bison}. |
| 5402 |
| 5403 @menu |
| 5404 * Calling Convention:: How @code{yyparse} calls @code{yylex}. |
| 5405 * Token Values:: How @code{yylex} must return the semantic value |
| 5406 of the token it has read. |
| 5407 * Token Locations:: How @code{yylex} must return the text location |
| 5408 (line number, etc.) of the token, if the |
| 5409 actions want that. |
| 5410 * Pure Calling:: How the calling convention differs in a pure parser |
| 5411 (@pxref{Pure Decl, ,A Pure (Reentrant) Parser}). |
| 5412 @end menu |
| 5413 |
| 5414 @node Calling Convention |
| 5415 @subsection Calling Convention for @code{yylex} |
| 5416 |
| 5417 The value that @code{yylex} returns must be the positive numeric code |
| 5418 for the type of token it has just found; a zero or negative value |
| 5419 signifies end-of-input. |
| 5420 |
| 5421 When a token is referred to in the grammar rules by a name, that name |
| 5422 in the parser file becomes a C macro whose definition is the proper |
| 5423 numeric code for that token type. So @code{yylex} can use the name |
| 5424 to indicate that type. @xref{Symbols}. |
| 5425 |
| 5426 When a token is referred to in the grammar rules by a character literal, |
| 5427 the numeric code for that character is also the code for the token type. |
| 5428 So @code{yylex} can simply return that character code, possibly converted |
| 5429 to @code{unsigned char} to avoid sign-extension. The null character |
| 5430 must not be used this way, because its code is zero and that |
| 5431 signifies end-of-input. |
| 5432 |
| 5433 Here is an example showing these things: |
| 5434 |
| 5435 @example |
| 5436 int |
| 5437 yylex (void) |
| 5438 @{ |
| 5439 @dots{} |
| 5440 if (c == EOF) /* Detect end-of-input. */ |
| 5441 return 0; |
| 5442 @dots{} |
| 5443 if (c == '+' || c == '-') |
| 5444 return c; /* Assume token type for `+' is '+'. */ |
| 5445 @dots{} |
| 5446 return INT; /* Return the type of the token. */ |
| 5447 @dots{} |
| 5448 @} |
| 5449 @end example |
| 5450 |
| 5451 @noindent |
| 5452 This interface has been designed so that the output from the @code{lex} |
| 5453 utility can be used without change as the definition of @code{yylex}. |
| 5454 |
| 5455 If the grammar uses literal string tokens, there are two ways that |
| 5456 @code{yylex} can determine the token type codes for them: |
| 5457 |
| 5458 @itemize @bullet |
| 5459 @item |
| 5460 If the grammar defines symbolic token names as aliases for the |
| 5461 literal string tokens, @code{yylex} can use these symbolic names like |
| 5462 all others. In this case, the use of the literal string tokens in |
| 5463 the grammar file has no effect on @code{yylex}. |
| 5464 |
| 5465 @item |
| 5466 @code{yylex} can find the multicharacter token in the @code{yytname} |
| 5467 table. The index of the token in the table is the token type's code. |
| 5468 The name of a multicharacter token is recorded in @code{yytname} with a |
| 5469 double-quote, the token's characters, and another double-quote. The |
| 5470 token's characters are escaped as necessary to be suitable as input |
| 5471 to Bison. |
| 5472 |
| 5473 Here's code for looking up a multicharacter token in @code{yytname}, |
| 5474 assuming that the characters of the token are stored in |
| 5475 @code{token_buffer}, and assuming that the token does not contain any |
| 5476 characters like @samp{"} that require escaping. |
| 5477 |
| 5478 @smallexample |
| 5479 for (i = 0; i < YYNTOKENS; i++) |
| 5480 @{ |
| 5481 if (yytname[i] != 0 |
| 5482 && yytname[i][0] == '"' |
| 5483 && ! strncmp (yytname[i] + 1, token_buffer, |
| 5484 strlen (token_buffer)) |
| 5485 && yytname[i][strlen (token_buffer) + 1] == '"' |
| 5486 && yytname[i][strlen (token_buffer) + 2] == 0) |
| 5487 break; |
| 5488 @} |
| 5489 @end smallexample |
| 5490 |
| 5491 The @code{yytname} table is generated only if you use the |
| 5492 @code{%token-table} declaration. @xref{Decl Summary}. |
| 5493 @end itemize |
| 5494 |
| 5495 @node Token Values |
| 5496 @subsection Semantic Values of Tokens |
| 5497 |
| 5498 @vindex yylval |
| 5499 In an ordinary (nonreentrant) parser, the semantic value of the token must |
| 5500 be stored into the global variable @code{yylval}. When you are using |
| 5501 just one data type for semantic values, @code{yylval} has that type. |
| 5502 Thus, if the type is @code{int} (the default), you might write this in |
| 5503 @code{yylex}: |
| 5504 |
| 5505 @example |
| 5506 @group |
| 5507 @dots{} |
| 5508 yylval = value; /* Put value onto Bison stack. */ |
| 5509 return INT; /* Return the type of the token. */ |
| 5510 @dots{} |
| 5511 @end group |
| 5512 @end example |
| 5513 |
| 5514 When you are using multiple data types, @code{yylval}'s type is a union |
| 5515 made from the @code{%union} declaration (@pxref{Union Decl, ,The |
| 5516 Collection of Value Types}). So when you store a token's value, you |
| 5517 must use the proper member of the union. If the @code{%union} |
| 5518 declaration looks like this: |
| 5519 |
| 5520 @example |
| 5521 @group |
| 5522 %union @{ |
| 5523 int intval; |
| 5524 double val; |
| 5525 symrec *tptr; |
| 5526 @} |
| 5527 @end group |
| 5528 @end example |
| 5529 |
| 5530 @noindent |
| 5531 then the code in @code{yylex} might look like this: |
| 5532 |
| 5533 @example |
| 5534 @group |
| 5535 @dots{} |
| 5536 yylval.intval = value; /* Put value onto Bison stack. */ |
| 5537 return INT; /* Return the type of the token. */ |
| 5538 @dots{} |
| 5539 @end group |
| 5540 @end example |
| 5541 |
| 5542 @node Token Locations |
| 5543 @subsection Textual Locations of Tokens |
| 5544 |
| 5545 @vindex yylloc |
| 5546 If you are using the @samp{@@@var{n}}-feature (@pxref{Locations, , |
| 5547 Tracking Locations}) in actions to keep track of the textual locations |
| 5548 of tokens and groupings, then you must provide this information in |
| 5549 @code{yylex}. The function @code{yyparse} expects to find the textual |
| 5550 location of a token just parsed in the global variable @code{yylloc}. |
| 5551 So @code{yylex} must store the proper data in that variable. |
| 5552 |
| 5553 By default, the value of @code{yylloc} is a structure and you need only |
| 5554 initialize the members that are going to be used by the actions. The |
| 5555 four members are called @code{first_line}, @code{first_column}, |
| 5556 @code{last_line} and @code{last_column}. Note that the use of this |
| 5557 feature makes the parser noticeably slower. |
| 5558 |
| 5559 @tindex YYLTYPE |
| 5560 The data type of @code{yylloc} has the name @code{YYLTYPE}. |
| 5561 |
| 5562 @node Pure Calling |
| 5563 @subsection Calling Conventions for Pure Parsers |
| 5564 |
| 5565 When you use the Bison declaration @code{%define api.pure} to request a |
| 5566 pure, reentrant parser, the global communication variables @code{yylval} |
| 5567 and @code{yylloc} cannot be used. (@xref{Pure Decl, ,A Pure (Reentrant) |
| 5568 Parser}.) In such parsers the two global variables are replaced by |
| 5569 pointers passed as arguments to @code{yylex}. You must declare them as |
| 5570 shown here, and pass the information back by storing it through those |
| 5571 pointers. |
| 5572 |
| 5573 @example |
| 5574 int |
| 5575 yylex (YYSTYPE *lvalp, YYLTYPE *llocp) |
| 5576 @{ |
| 5577 @dots{} |
| 5578 *lvalp = value; /* Put value onto Bison stack. */ |
| 5579 return INT; /* Return the type of the token. */ |
| 5580 @dots{} |
| 5581 @} |
| 5582 @end example |
| 5583 |
| 5584 If the grammar file does not use the @samp{@@} constructs to refer to |
| 5585 textual locations, then the type @code{YYLTYPE} will not be defined. In |
| 5586 this case, omit the second argument; @code{yylex} will be called with |
| 5587 only one argument. |
| 5588 |
| 5589 |
| 5590 If you wish to pass the additional parameter data to @code{yylex}, use |
| 5591 @code{%lex-param} just like @code{%parse-param} (@pxref{Parser |
| 5592 Function}). |
| 5593 |
| 5594 @deffn {Directive} lex-param @{@var{argument-declaration}@} |
| 5595 @findex %lex-param |
| 5596 Declare that the braced-code @var{argument-declaration} is an |
| 5597 additional @code{yylex} argument declaration. |
| 5598 @end deffn |
| 5599 |
| 5600 For instance: |
| 5601 |
| 5602 @example |
| 5603 %parse-param @{int *nastiness@} |
| 5604 %lex-param @{int *nastiness@} |
| 5605 %parse-param @{int *randomness@} |
| 5606 @end example |
| 5607 |
| 5608 @noindent |
| 5609 results in the following signature: |
| 5610 |
| 5611 @example |
| 5612 int yylex (int *nastiness); |
| 5613 int yyparse (int *nastiness, int *randomness); |
| 5614 @end example |
| 5615 |
| 5616 If @code{%define api.pure} is added: |
| 5617 |
| 5618 @example |
| 5619 int yylex (YYSTYPE *lvalp, int *nastiness); |
| 5620 int yyparse (int *nastiness, int *randomness); |
| 5621 @end example |
| 5622 |
| 5623 @noindent |
| 5624 and finally, if both @code{%define api.pure} and @code{%locations} are used: |
| 5625 |
| 5626 @example |
| 5627 int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness); |
| 5628 int yyparse (int *nastiness, int *randomness); |
| 5629 @end example |
| 5630 |
| 5631 @node Error Reporting |
| 5632 @section The Error Reporting Function @code{yyerror} |
| 5633 @cindex error reporting function |
| 5634 @findex yyerror |
| 5635 @cindex parse error |
| 5636 @cindex syntax error |
| 5637 |
| 5638 The Bison parser detects a @dfn{syntax error} or @dfn{parse error} |
| 5639 whenever it reads a token which cannot satisfy any syntax rule. An |
| 5640 action in the grammar can also explicitly proclaim an error, using the |
| 5641 macro @code{YYERROR} (@pxref{Action Features, ,Special Features for Use |
| 5642 in Actions}). |
| 5643 |
| 5644 The Bison parser expects to report the error by calling an error |
| 5645 reporting function named @code{yyerror}, which you must supply. It is |
| 5646 called by @code{yyparse} whenever a syntax error is found, and it |
| 5647 receives one argument. For a syntax error, the string is normally |
| 5648 @w{@code{"syntax error"}}. |
| 5649 |
| 5650 @findex %error-verbose |
| 5651 If you invoke the directive @code{%error-verbose} in the Bison |
| 5652 declarations section (@pxref{Bison Declarations, ,The Bison Declarations |
| 5653 Section}), then Bison provides a more verbose and specific error message |
| 5654 string instead of just plain @w{@code{"syntax error"}}. |
| 5655 |
| 5656 The parser can detect one other kind of error: memory exhaustion. This |
| 5657 can happen when the input contains constructions that are very deeply |
| 5658 nested. It isn't likely you will encounter this, since the Bison |
| 5659 parser normally extends its stack automatically up to a very large limit. But |
| 5660 if memory is exhausted, @code{yyparse} calls @code{yyerror} in the usual |
| 5661 fashion, except that the argument string is @w{@code{"memory exhausted"}}. |
| 5662 |
| 5663 In some cases diagnostics like @w{@code{"syntax error"}} are |
| 5664 translated automatically from English to some other language before |
| 5665 they are passed to @code{yyerror}. @xref{Internationalization}. |
| 5666 |
| 5667 The following definition suffices in simple programs: |
| 5668 |
| 5669 @example |
| 5670 @group |
| 5671 void |
| 5672 yyerror (char const *s) |
| 5673 @{ |
| 5674 @end group |
| 5675 @group |
| 5676 fprintf (stderr, "%s\n", s); |
| 5677 @} |
| 5678 @end group |
| 5679 @end example |
| 5680 |
| 5681 After @code{yyerror} returns to @code{yyparse}, the latter will attempt |
| 5682 error recovery if you have written suitable error recovery grammar rules |
| 5683 (@pxref{Error Recovery}). If recovery is impossible, @code{yyparse} will |
| 5684 immediately return 1. |
| 5685 |
| 5686 Obviously, in location tracking pure parsers, @code{yyerror} should have |
| 5687 an access to the current location. |
| 5688 This is indeed the case for the @acronym{GLR} |
| 5689 parsers, but not for the Yacc parser, for historical reasons. I.e., if |
| 5690 @samp{%locations %define api.pure} is passed then the prototypes for |
| 5691 @code{yyerror} are: |
| 5692 |
| 5693 @example |
| 5694 void yyerror (char const *msg); /* Yacc parsers. */ |
| 5695 void yyerror (YYLTYPE *locp, char const *msg); /* GLR parsers. */ |
| 5696 @end example |
| 5697 |
| 5698 If @samp{%parse-param @{int *nastiness@}} is used, then: |
| 5699 |
| 5700 @example |
| 5701 void yyerror (int *nastiness, char const *msg); /* Yacc parsers. */ |
| 5702 void yyerror (int *nastiness, char const *msg); /* GLR parsers. */ |
| 5703 @end example |
| 5704 |
| 5705 Finally, @acronym{GLR} and Yacc parsers share the same @code{yyerror} calling |
| 5706 convention for absolutely pure parsers, i.e., when the calling |
| 5707 convention of @code{yylex} @emph{and} the calling convention of |
| 5708 @code{%define api.pure} are pure. |
| 5709 I.e.: |
| 5710 |
| 5711 @example |
| 5712 /* Location tracking. */ |
| 5713 %locations |
| 5714 /* Pure yylex. */ |
| 5715 %define api.pure |
| 5716 %lex-param @{int *nastiness@} |
| 5717 /* Pure yyparse. */ |
| 5718 %parse-param @{int *nastiness@} |
| 5719 %parse-param @{int *randomness@} |
| 5720 @end example |
| 5721 |
| 5722 @noindent |
| 5723 results in the following signatures for all the parser kinds: |
| 5724 |
| 5725 @example |
| 5726 int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness); |
| 5727 int yyparse (int *nastiness, int *randomness); |
| 5728 void yyerror (YYLTYPE *locp, |
| 5729 int *nastiness, int *randomness, |
| 5730 char const *msg); |
| 5731 @end example |
| 5732 |
| 5733 @noindent |
| 5734 The prototypes are only indications of how the code produced by Bison |
| 5735 uses @code{yyerror}. Bison-generated code always ignores the returned |
| 5736 value, so @code{yyerror} can return any type, including @code{void}. |
| 5737 Also, @code{yyerror} can be a variadic function; that is why the |
| 5738 message is always passed last. |
| 5739 |
| 5740 Traditionally @code{yyerror} returns an @code{int} that is always |
| 5741 ignored, but this is purely for historical reasons, and @code{void} is |
| 5742 preferable since it more accurately describes the return type for |
| 5743 @code{yyerror}. |
| 5744 |
| 5745 @vindex yynerrs |
| 5746 The variable @code{yynerrs} contains the number of syntax errors |
| 5747 reported so far. Normally this variable is global; but if you |
| 5748 request a pure parser (@pxref{Pure Decl, ,A Pure (Reentrant) Parser}) |
| 5749 then it is a local variable which only the actions can access. |
| 5750 |
| 5751 @node Action Features |
| 5752 @section Special Features for Use in Actions |
| 5753 @cindex summary, action features |
| 5754 @cindex action features summary |
| 5755 |
| 5756 Here is a table of Bison constructs, variables and macros that |
| 5757 are useful in actions. |
| 5758 |
| 5759 @deffn {Variable} $$ |
| 5760 Acts like a variable that contains the semantic value for the |
| 5761 grouping made by the current rule. @xref{Actions}. |
| 5762 @end deffn |
| 5763 |
| 5764 @deffn {Variable} $@var{n} |
| 5765 Acts like a variable that contains the semantic value for the |
| 5766 @var{n}th component of the current rule. @xref{Actions}. |
| 5767 @end deffn |
| 5768 |
| 5769 @deffn {Variable} $<@var{typealt}>$ |
| 5770 Like @code{$$} but specifies alternative @var{typealt} in the union |
| 5771 specified by the @code{%union} declaration. @xref{Action Types, ,Data |
| 5772 Types of Values in Actions}. |
| 5773 @end deffn |
| 5774 |
| 5775 @deffn {Variable} $<@var{typealt}>@var{n} |
| 5776 Like @code{$@var{n}} but specifies alternative @var{typealt} in the |
| 5777 union specified by the @code{%union} declaration. |
| 5778 @xref{Action Types, ,Data Types of Values in Actions}. |
| 5779 @end deffn |
| 5780 |
| 5781 @deffn {Macro} YYABORT; |
| 5782 Return immediately from @code{yyparse}, indicating failure. |
| 5783 @xref{Parser Function, ,The Parser Function @code{yyparse}}. |
| 5784 @end deffn |
| 5785 |
| 5786 @deffn {Macro} YYACCEPT; |
| 5787 Return immediately from @code{yyparse}, indicating success. |
| 5788 @xref{Parser Function, ,The Parser Function @code{yyparse}}. |
| 5789 @end deffn |
| 5790 |
| 5791 @deffn {Macro} YYBACKUP (@var{token}, @var{value}); |
| 5792 @findex YYBACKUP |
| 5793 Unshift a token. This macro is allowed only for rules that reduce |
| 5794 a single value, and only when there is no lookahead token. |
| 5795 It is also disallowed in @acronym{GLR} parsers. |
| 5796 It installs a lookahead token with token type @var{token} and |
| 5797 semantic value @var{value}; then it discards the value that was |
| 5798 going to be reduced by this rule. |
| 5799 |
| 5800 If the macro is used when it is not valid, such as when there is |
| 5801 a lookahead token already, then it reports a syntax error with |
| 5802 a message @samp{cannot back up} and performs ordinary error |
| 5803 recovery. |
| 5804 |
| 5805 In either case, the rest of the action is not executed. |
| 5806 @end deffn |
| 5807 |
| 5808 @deffn {Macro} YYEMPTY |
| 5809 @vindex YYEMPTY |
| 5810 Value stored in @code{yychar} when there is no lookahead token. |
| 5811 @end deffn |
| 5812 |
| 5813 @deffn {Macro} YYEOF |
| 5814 @vindex YYEOF |
| 5815 Value stored in @code{yychar} when the lookahead is the end of the input |
| 5816 stream. |
| 5817 @end deffn |
| 5818 |
| 5819 @deffn {Macro} YYERROR; |
| 5820 @findex YYERROR |
| 5821 Cause an immediate syntax error. This statement initiates error |
| 5822 recovery just as if the parser itself had detected an error; however, it |
| 5823 does not call @code{yyerror}, and does not print any message. If you |
| 5824 want to print an error message, call @code{yyerror} explicitly before |
| 5825 the @samp{YYERROR;} statement. @xref{Error Recovery}. |
| 5826 @end deffn |
| 5827 |
| 5828 @deffn {Macro} YYRECOVERING |
| 5829 @findex YYRECOVERING |
| 5830 The expression @code{YYRECOVERING ()} yields 1 when the parser |
| 5831 is recovering from a syntax error, and 0 otherwise. |
| 5832 @xref{Error Recovery}. |
| 5833 @end deffn |
| 5834 |
| 5835 @deffn {Variable} yychar |
| 5836 Variable containing either the lookahead token, or @code{YYEOF} when the |
| 5837 lookahead is the end of the input stream, or @code{YYEMPTY} when no lookahead |
| 5838 has been performed so the next token is not yet known. |
| 5839 Do not modify @code{yychar} in a deferred semantic action (@pxref{GLR Semantic |
| 5840 Actions}). |
| 5841 @xref{Lookahead, ,Lookahead Tokens}. |
| 5842 @end deffn |
| 5843 |
| 5844 @deffn {Macro} yyclearin; |
| 5845 Discard the current lookahead token. This is useful primarily in |
| 5846 error rules. |
| 5847 Do not invoke @code{yyclearin} in a deferred semantic action (@pxref{GLR |
| 5848 Semantic Actions}). |
| 5849 @xref{Error Recovery}. |
| 5850 @end deffn |
| 5851 |
| 5852 @deffn {Macro} yyerrok; |
| 5853 Resume generating error messages immediately for subsequent syntax |
| 5854 errors. This is useful primarily in error rules. |
| 5855 @xref{Error Recovery}. |
| 5856 @end deffn |
| 5857 |
| 5858 @deffn {Variable} yylloc |
| 5859 Variable containing the lookahead token location when @code{yychar} is not set |
| 5860 to @code{YYEMPTY} or @code{YYEOF}. |
| 5861 Do not modify @code{yylloc} in a deferred semantic action (@pxref{GLR Semantic |
| 5862 Actions}). |
| 5863 @xref{Actions and Locations, ,Actions and Locations}. |
| 5864 @end deffn |
| 5865 |
| 5866 @deffn {Variable} yylval |
| 5867 Variable containing the lookahead token semantic value when @code{yychar} is |
| 5868 not set to @code{YYEMPTY} or @code{YYEOF}. |
| 5869 Do not modify @code{yylval} in a deferred semantic action (@pxref{GLR Semantic |
| 5870 Actions}). |
| 5871 @xref{Actions, ,Actions}. |
| 5872 @end deffn |
| 5873 |
| 5874 @deffn {Value} @@$ |
| 5875 @findex @@$ |
| 5876 Acts like a structure variable containing information on the textual location |
| 5877 of the grouping made by the current rule. @xref{Locations, , |
| 5878 Tracking Locations}. |
| 5879 |
| 5880 @c Check if those paragraphs are still useful or not. |
| 5881 |
| 5882 @c @example |
| 5883 @c struct @{ |
| 5884 @c int first_line, last_line; |
| 5885 @c int first_column, last_column; |
| 5886 @c @}; |
| 5887 @c @end example |
| 5888 |
| 5889 @c Thus, to get the starting line number of the third component, you would |
| 5890 @c use @samp{@@3.first_line}. |
| 5891 |
| 5892 @c In order for the members of this structure to contain valid information, |
| 5893 @c you must make @code{yylex} supply this information about each token. |
| 5894 @c If you need only certain members, then @code{yylex} need only fill in |
| 5895 @c those members. |
| 5896 |
| 5897 @c The use of this feature makes the parser noticeably slower. |
| 5898 @end deffn |
| 5899 |
| 5900 @deffn {Value} @@@var{n} |
| 5901 @findex @@@var{n} |
| 5902 Acts like a structure variable containing information on the textual location |
| 5903 of the @var{n}th component of the current rule. @xref{Locations, , |
| 5904 Tracking Locations}. |
| 5905 @end deffn |
| 5906 |
| 5907 @node Internationalization |
| 5908 @section Parser Internationalization |
| 5909 @cindex internationalization |
| 5910 @cindex i18n |
| 5911 @cindex NLS |
| 5912 @cindex gettext |
| 5913 @cindex bison-po |
| 5914 |
| 5915 A Bison-generated parser can print diagnostics, including error and |
| 5916 tracing messages. By default, they appear in English. However, Bison |
| 5917 also supports outputting diagnostics in the user's native language. To |
| 5918 make this work, the user should set the usual environment variables. |
| 5919 @xref{Users, , The User's View, gettext, GNU @code{gettext} utilities}. |
| 5920 For example, the shell command @samp{export LC_ALL=fr_CA.UTF-8} might |
| 5921 set the user's locale to French Canadian using the @acronym{UTF}-8 |
| 5922 encoding. The exact set of available locales depends on the user's |
| 5923 installation. |
| 5924 |
| 5925 The maintainer of a package that uses a Bison-generated parser enables |
| 5926 the internationalization of the parser's output through the following |
| 5927 steps. Here we assume a package that uses @acronym{GNU} Autoconf and |
| 5928 @acronym{GNU} Automake. |
| 5929 |
| 5930 @enumerate |
| 5931 @item |
| 5932 @cindex bison-i18n.m4 |
| 5933 Into the directory containing the @acronym{GNU} Autoconf macros used |
| 5934 by the package---often called @file{m4}---copy the |
| 5935 @file{bison-i18n.m4} file installed by Bison under |
| 5936 @samp{share/aclocal/bison-i18n.m4} in Bison's installation directory. |
| 5937 For example: |
| 5938 |
| 5939 @example |
| 5940 cp /usr/local/share/aclocal/bison-i18n.m4 m4/bison-i18n.m4 |
| 5941 @end example |
| 5942 |
| 5943 @item |
| 5944 @findex BISON_I18N |
| 5945 @vindex BISON_LOCALEDIR |
| 5946 @vindex YYENABLE_NLS |
| 5947 In the top-level @file{configure.ac}, after the @code{AM_GNU_GETTEXT} |
| 5948 invocation, add an invocation of @code{BISON_I18N}. This macro is |
| 5949 defined in the file @file{bison-i18n.m4} that you copied earlier. It |
| 5950 causes @samp{configure} to find the value of the |
| 5951 @code{BISON_LOCALEDIR} variable, and it defines the source-language |
| 5952 symbol @code{YYENABLE_NLS} to enable translations in the |
| 5953 Bison-generated parser. |
| 5954 |
| 5955 @item |
| 5956 In the @code{main} function of your program, designate the directory |
| 5957 containing Bison's runtime message catalog, through a call to |
| 5958 @samp{bindtextdomain} with domain name @samp{bison-runtime}. |
| 5959 For example: |
| 5960 |
| 5961 @example |
| 5962 bindtextdomain ("bison-runtime", BISON_LOCALEDIR); |
| 5963 @end example |
| 5964 |
| 5965 Typically this appears after any other call @code{bindtextdomain |
| 5966 (PACKAGE, LOCALEDIR)} that your package already has. Here we rely on |
| 5967 @samp{BISON_LOCALEDIR} to be defined as a string through the |
| 5968 @file{Makefile}. |
| 5969 |
| 5970 @item |
| 5971 In the @file{Makefile.am} that controls the compilation of the @code{main} |
| 5972 function, make @samp{BISON_LOCALEDIR} available as a C preprocessor macro, |
| 5973 either in @samp{DEFS} or in @samp{AM_CPPFLAGS}. For example: |
| 5974 |
| 5975 @example |
| 5976 DEFS = @@DEFS@@ -DBISON_LOCALEDIR='"$(BISON_LOCALEDIR)"' |
| 5977 @end example |
| 5978 |
| 5979 or: |
| 5980 |
| 5981 @example |
| 5982 AM_CPPFLAGS = -DBISON_LOCALEDIR='"$(BISON_LOCALEDIR)"' |
| 5983 @end example |
| 5984 |
| 5985 @item |
| 5986 Finally, invoke the command @command{autoreconf} to generate the build |
| 5987 infrastructure. |
| 5988 @end enumerate |
| 5989 |
| 5990 |
| 5991 @node Algorithm |
| 5992 @chapter The Bison Parser Algorithm |
| 5993 @cindex Bison parser algorithm |
| 5994 @cindex algorithm of parser |
| 5995 @cindex shifting |
| 5996 @cindex reduction |
| 5997 @cindex parser stack |
| 5998 @cindex stack, parser |
| 5999 |
| 6000 As Bison reads tokens, it pushes them onto a stack along with their |
| 6001 semantic values. The stack is called the @dfn{parser stack}. Pushing a |
| 6002 token is traditionally called @dfn{shifting}. |
| 6003 |
| 6004 For example, suppose the infix calculator has read @samp{1 + 5 *}, with a |
| 6005 @samp{3} to come. The stack will have four elements, one for each token |
| 6006 that was shifted. |
| 6007 |
| 6008 But the stack does not always have an element for each token read. When |
| 6009 the last @var{n} tokens and groupings shifted match the components of a |
| 6010 grammar rule, they can be combined according to that rule. This is called |
| 6011 @dfn{reduction}. Those tokens and groupings are replaced on the stack by a |
| 6012 single grouping whose symbol is the result (left hand side) of that rule. |
| 6013 Running the rule's action is part of the process of reduction, because this |
| 6014 is what computes the semantic value of the resulting grouping. |
| 6015 |
| 6016 For example, if the infix calculator's parser stack contains this: |
| 6017 |
| 6018 @example |
| 6019 1 + 5 * 3 |
| 6020 @end example |
| 6021 |
| 6022 @noindent |
| 6023 and the next input token is a newline character, then the last three |
| 6024 elements can be reduced to 15 via the rule: |
| 6025 |
| 6026 @example |
| 6027 expr: expr '*' expr; |
| 6028 @end example |
| 6029 |
| 6030 @noindent |
| 6031 Then the stack contains just these three elements: |
| 6032 |
| 6033 @example |
| 6034 1 + 15 |
| 6035 @end example |
| 6036 |
| 6037 @noindent |
| 6038 At this point, another reduction can be made, resulting in the single value |
| 6039 16. Then the newline token can be shifted. |
| 6040 |
| 6041 The parser tries, by shifts and reductions, to reduce the entire input down |
| 6042 to a single grouping whose symbol is the grammar's start-symbol |
| 6043 (@pxref{Language and Grammar, ,Languages and Context-Free Grammars}). |
| 6044 |
| 6045 This kind of parser is known in the literature as a bottom-up parser. |
| 6046 |
| 6047 @menu |
| 6048 * Lookahead:: Parser looks one token ahead when deciding what to do. |
| 6049 * Shift/Reduce:: Conflicts: when either shifting or reduction is valid. |
| 6050 * Precedence:: Operator precedence works by resolving conflicts. |
| 6051 * Contextual Precedence:: When an operator's precedence depends on context. |
| 6052 * Parser States:: The parser is a finite-state-machine with stack. |
| 6053 * Reduce/Reduce:: When two rules are applicable in the same situation. |
| 6054 * Mystery Conflicts:: Reduce/reduce conflicts that look unjustified. |
| 6055 * Generalized LR Parsing:: Parsing arbitrary context-free grammars. |
| 6056 * Memory Management:: What happens when memory is exhausted. How to avoid it. |
| 6057 @end menu |
| 6058 |
| 6059 @node Lookahead |
| 6060 @section Lookahead Tokens |
| 6061 @cindex lookahead token |
| 6062 |
| 6063 The Bison parser does @emph{not} always reduce immediately as soon as the |
| 6064 last @var{n} tokens and groupings match a rule. This is because such a |
| 6065 simple strategy is inadequate to handle most languages. Instead, when a |
| 6066 reduction is possible, the parser sometimes ``looks ahead'' at the next |
| 6067 token in order to decide what to do. |
| 6068 |
| 6069 When a token is read, it is not immediately shifted; first it becomes the |
| 6070 @dfn{lookahead token}, which is not on the stack. Now the parser can |
| 6071 perform one or more reductions of tokens and groupings on the stack, while |
| 6072 the lookahead token remains off to the side. When no more reductions |
| 6073 should take place, the lookahead token is shifted onto the stack. This |
| 6074 does not mean that all possible reductions have been done; depending on the |
| 6075 token type of the lookahead token, some rules may choose to delay their |
| 6076 application. |
| 6077 |
| 6078 Here is a simple case where lookahead is needed. These three rules define |
| 6079 expressions which contain binary addition operators and postfix unary |
| 6080 factorial operators (@samp{!}), and allow parentheses for grouping. |
| 6081 |
| 6082 @example |
| 6083 @group |
| 6084 expr: term '+' expr |
| 6085 | term |
| 6086 ; |
| 6087 @end group |
| 6088 |
| 6089 @group |
| 6090 term: '(' expr ')' |
| 6091 | term '!' |
| 6092 | NUMBER |
| 6093 ; |
| 6094 @end group |
| 6095 @end example |
| 6096 |
| 6097 Suppose that the tokens @w{@samp{1 + 2}} have been read and shifted; what |
| 6098 should be done? If the following token is @samp{)}, then the first three |
| 6099 tokens must be reduced to form an @code{expr}. This is the only valid |
| 6100 course, because shifting the @samp{)} would produce a sequence of symbols |
| 6101 @w{@code{term ')'}}, and no rule allows this. |
| 6102 |
| 6103 If the following token is @samp{!}, then it must be shifted immediately so |
| 6104 that @w{@samp{2 !}} can be reduced to make a @code{term}. If instead the |
| 6105 parser were to reduce before shifting, @w{@samp{1 + 2}} would become an |
| 6106 @code{expr}. It would then be impossible to shift the @samp{!} because |
| 6107 doing so would produce on the stack the sequence of symbols @code{expr |
| 6108 '!'}. No rule allows that sequence. |
| 6109 |
| 6110 @vindex yychar |
| 6111 @vindex yylval |
| 6112 @vindex yylloc |
| 6113 The lookahead token is stored in the variable @code{yychar}. |
| 6114 Its semantic value and location, if any, are stored in the variables |
| 6115 @code{yylval} and @code{yylloc}. |
| 6116 @xref{Action Features, ,Special Features for Use in Actions}. |
| 6117 |
| 6118 @node Shift/Reduce |
| 6119 @section Shift/Reduce Conflicts |
| 6120 @cindex conflicts |
| 6121 @cindex shift/reduce conflicts |
| 6122 @cindex dangling @code{else} |
| 6123 @cindex @code{else}, dangling |
| 6124 |
| 6125 Suppose we are parsing a language which has if-then and if-then-else |
| 6126 statements, with a pair of rules like this: |
| 6127 |
| 6128 @example |
| 6129 @group |
| 6130 if_stmt: |
| 6131 IF expr THEN stmt |
| 6132 | IF expr THEN stmt ELSE stmt |
| 6133 ; |
| 6134 @end group |
| 6135 @end example |
| 6136 |
| 6137 @noindent |
| 6138 Here we assume that @code{IF}, @code{THEN} and @code{ELSE} are |
| 6139 terminal symbols for specific keyword tokens. |
| 6140 |
| 6141 When the @code{ELSE} token is read and becomes the lookahead token, the |
| 6142 contents of the stack (assuming the input is valid) are just right for |
| 6143 reduction by the first rule. But it is also legitimate to shift the |
| 6144 @code{ELSE}, because that would lead to eventual reduction by the second |
| 6145 rule. |
| 6146 |
| 6147 This situation, where either a shift or a reduction would be valid, is |
| 6148 called a @dfn{shift/reduce conflict}. Bison is designed to resolve |
| 6149 these conflicts by choosing to shift, unless otherwise directed by |
| 6150 operator precedence declarations. To see the reason for this, let's |
| 6151 contrast it with the other alternative. |
| 6152 |
| 6153 Since the parser prefers to shift the @code{ELSE}, the result is to attach |
| 6154 the else-clause to the innermost if-statement, making these two inputs |
| 6155 equivalent: |
| 6156 |
| 6157 @example |
| 6158 if x then if y then win (); else lose; |
| 6159 |
| 6160 if x then do; if y then win (); else lose; end; |
| 6161 @end example |
| 6162 |
| 6163 But if the parser chose to reduce when possible rather than shift, the |
| 6164 result would be to attach the else-clause to the outermost if-statement, |
| 6165 making these two inputs equivalent: |
| 6166 |
| 6167 @example |
| 6168 if x then if y then win (); else lose; |
| 6169 |
| 6170 if x then do; if y then win (); end; else lose; |
| 6171 @end example |
| 6172 |
| 6173 The conflict exists because the grammar as written is ambiguous: either |
| 6174 parsing of the simple nested if-statement is legitimate. The established |
| 6175 convention is that these ambiguities are resolved by attaching the |
| 6176 else-clause to the innermost if-statement; this is what Bison accomplishes |
| 6177 by choosing to shift rather than reduce. (It would ideally be cleaner to |
| 6178 write an unambiguous grammar, but that is very hard to do in this case.) |
| 6179 This particular ambiguity was first encountered in the specifications of |
| 6180 Algol 60 and is called the ``dangling @code{else}'' ambiguity. |
| 6181 |
| 6182 To avoid warnings from Bison about predictable, legitimate shift/reduce |
| 6183 conflicts, use the @code{%expect @var{n}} declaration. There will be no |
| 6184 warning as long as the number of shift/reduce conflicts is exactly @var{n}. |
| 6185 @xref{Expect Decl, ,Suppressing Conflict Warnings}. |
| 6186 |
| 6187 The definition of @code{if_stmt} above is solely to blame for the |
| 6188 conflict, but the conflict does not actually appear without additional |
| 6189 rules. Here is a complete Bison input file that actually manifests the |
| 6190 conflict: |
| 6191 |
| 6192 @example |
| 6193 @group |
| 6194 %token IF THEN ELSE variable |
| 6195 %% |
| 6196 @end group |
| 6197 @group |
| 6198 stmt: expr |
| 6199 | if_stmt |
| 6200 ; |
| 6201 @end group |
| 6202 |
| 6203 @group |
| 6204 if_stmt: |
| 6205 IF expr THEN stmt |
| 6206 | IF expr THEN stmt ELSE stmt |
| 6207 ; |
| 6208 @end group |
| 6209 |
| 6210 expr: variable |
| 6211 ; |
| 6212 @end example |
| 6213 |
| 6214 @node Precedence |
| 6215 @section Operator Precedence |
| 6216 @cindex operator precedence |
| 6217 @cindex precedence of operators |
| 6218 |
| 6219 Another situation where shift/reduce conflicts appear is in arithmetic |
| 6220 expressions. Here shifting is not always the preferred resolution; the |
| 6221 Bison declarations for operator precedence allow you to specify when to |
| 6222 shift and when to reduce. |
| 6223 |
| 6224 @menu |
| 6225 * Why Precedence:: An example showing why precedence is needed. |
| 6226 * Using Precedence:: How to specify precedence in Bison grammars. |
| 6227 * Precedence Examples:: How these features are used in the previous example. |
| 6228 * How Precedence:: How they work. |
| 6229 @end menu |
| 6230 |
| 6231 @node Why Precedence |
| 6232 @subsection When Precedence is Needed |
| 6233 |
| 6234 Consider the following ambiguous grammar fragment (ambiguous because the |
| 6235 input @w{@samp{1 - 2 * 3}} can be parsed in two different ways): |
| 6236 |
| 6237 @example |
| 6238 @group |
| 6239 expr: expr '-' expr |
| 6240 | expr '*' expr |
| 6241 | expr '<' expr |
| 6242 | '(' expr ')' |
| 6243 @dots{} |
| 6244 ; |
| 6245 @end group |
| 6246 @end example |
| 6247 |
| 6248 @noindent |
| 6249 Suppose the parser has seen the tokens @samp{1}, @samp{-} and @samp{2}; |
| 6250 should it reduce them via the rule for the subtraction operator? It |
| 6251 depends on the next token. Of course, if the next token is @samp{)}, we |
| 6252 must reduce; shifting is invalid because no single rule can reduce the |
| 6253 token sequence @w{@samp{- 2 )}} or anything starting with that. But if |
| 6254 the next token is @samp{*} or @samp{<}, we have a choice: either |
| 6255 shifting or reduction would allow the parse to complete, but with |
| 6256 different results. |
| 6257 |
| 6258 To decide which one Bison should do, we must consider the results. If |
| 6259 the next operator token @var{op} is shifted, then it must be reduced |
| 6260 first in order to permit another opportunity to reduce the difference. |
| 6261 The result is (in effect) @w{@samp{1 - (2 @var{op} 3)}}. On the other |
| 6262 hand, if the subtraction is reduced before shifting @var{op}, the result |
| 6263 is @w{@samp{(1 - 2) @var{op} 3}}. Clearly, then, the choice of shift or |
| 6264 reduce should depend on the relative precedence of the operators |
| 6265 @samp{-} and @var{op}: @samp{*} should be shifted first, but not |
| 6266 @samp{<}. |
| 6267 |
| 6268 @cindex associativity |
| 6269 What about input such as @w{@samp{1 - 2 - 5}}; should this be |
| 6270 @w{@samp{(1 - 2) - 5}} or should it be @w{@samp{1 - (2 - 5)}}? For most |
| 6271 operators we prefer the former, which is called @dfn{left association}. |
| 6272 The latter alternative, @dfn{right association}, is desirable for |
| 6273 assignment operators. The choice of left or right association is a |
| 6274 matter of whether the parser chooses to shift or reduce when the stack |
| 6275 contains @w{@samp{1 - 2}} and the lookahead token is @samp{-}: shifting |
| 6276 makes right-associativity. |
| 6277 |
| 6278 @node Using Precedence |
| 6279 @subsection Specifying Operator Precedence |
| 6280 @findex %left |
| 6281 @findex %right |
| 6282 @findex %nonassoc |
| 6283 |
| 6284 Bison allows you to specify these choices with the operator precedence |
| 6285 declarations @code{%left} and @code{%right}. Each such declaration |
| 6286 contains a list of tokens, which are operators whose precedence and |
| 6287 associativity is being declared. The @code{%left} declaration makes all |
| 6288 those operators left-associative and the @code{%right} declaration makes |
| 6289 them right-associative. A third alternative is @code{%nonassoc}, which |
| 6290 declares that it is a syntax error to find the same operator twice ``in a |
| 6291 row''. |
| 6292 |
| 6293 The relative precedence of different operators is controlled by the |
| 6294 order in which they are declared. The first @code{%left} or |
| 6295 @code{%right} declaration in the file declares the operators whose |
| 6296 precedence is lowest, the next such declaration declares the operators |
| 6297 whose precedence is a little higher, and so on. |
| 6298 |
| 6299 @node Precedence Examples |
| 6300 @subsection Precedence Examples |
| 6301 |
| 6302 In our example, we would want the following declarations: |
| 6303 |
| 6304 @example |
| 6305 %left '<' |
| 6306 %left '-' |
| 6307 %left '*' |
| 6308 @end example |
| 6309 |
| 6310 In a more complete example, which supports other operators as well, we |
| 6311 would declare them in groups of equal precedence. For example, @code{'+'} is |
| 6312 declared with @code{'-'}: |
| 6313 |
| 6314 @example |
| 6315 %left '<' '>' '=' NE LE GE |
| 6316 %left '+' '-' |
| 6317 %left '*' '/' |
| 6318 @end example |
| 6319 |
| 6320 @noindent |
| 6321 (Here @code{NE} and so on stand for the operators for ``not equal'' |
| 6322 and so on. We assume that these tokens are more than one character long |
| 6323 and therefore are represented by names, not character literals.) |
| 6324 |
| 6325 @node How Precedence |
| 6326 @subsection How Precedence Works |
| 6327 |
| 6328 The first effect of the precedence declarations is to assign precedence |
| 6329 levels to the terminal symbols declared. The second effect is to assign |
| 6330 precedence levels to certain rules: each rule gets its precedence from |
| 6331 the last terminal symbol mentioned in the components. (You can also |
| 6332 specify explicitly the precedence of a rule. @xref{Contextual |
| 6333 Precedence, ,Context-Dependent Precedence}.) |
| 6334 |
| 6335 Finally, the resolution of conflicts works by comparing the precedence |
| 6336 of the rule being considered with that of the lookahead token. If the |
| 6337 token's precedence is higher, the choice is to shift. If the rule's |
| 6338 precedence is higher, the choice is to reduce. If they have equal |
| 6339 precedence, the choice is made based on the associativity of that |
| 6340 precedence level. The verbose output file made by @samp{-v} |
| 6341 (@pxref{Invocation, ,Invoking Bison}) says how each conflict was |
| 6342 resolved. |
| 6343 |
| 6344 Not all rules and not all tokens have precedence. If either the rule or |
| 6345 the lookahead token has no precedence, then the default is to shift. |
| 6346 |
| 6347 @node Contextual Precedence |
| 6348 @section Context-Dependent Precedence |
| 6349 @cindex context-dependent precedence |
| 6350 @cindex unary operator precedence |
| 6351 @cindex precedence, context-dependent |
| 6352 @cindex precedence, unary operator |
| 6353 @findex %prec |
| 6354 |
| 6355 Often the precedence of an operator depends on the context. This sounds |
| 6356 outlandish at first, but it is really very common. For example, a minus |
| 6357 sign typically has a very high precedence as a unary operator, and a |
| 6358 somewhat lower precedence (lower than multiplication) as a binary operator. |
| 6359 |
| 6360 The Bison precedence declarations, @code{%left}, @code{%right} and |
| 6361 @code{%nonassoc}, can only be used once for a given token; so a token has |
| 6362 only one precedence declared in this way. For context-dependent |
| 6363 precedence, you need to use an additional mechanism: the @code{%prec} |
| 6364 modifier for rules. |
| 6365 |
| 6366 The @code{%prec} modifier declares the precedence of a particular rule by |
| 6367 specifying a terminal symbol whose precedence should be used for that rule. |
| 6368 It's not necessary for that symbol to appear otherwise in the rule. The |
| 6369 modifier's syntax is: |
| 6370 |
| 6371 @example |
| 6372 %prec @var{terminal-symbol} |
| 6373 @end example |
| 6374 |
| 6375 @noindent |
| 6376 and it is written after the components of the rule. Its effect is to |
| 6377 assign the rule the precedence of @var{terminal-symbol}, overriding |
| 6378 the precedence that would be deduced for it in the ordinary way. The |
| 6379 altered rule precedence then affects how conflicts involving that rule |
| 6380 are resolved (@pxref{Precedence, ,Operator Precedence}). |
| 6381 |
| 6382 Here is how @code{%prec} solves the problem of unary minus. First, declare |
| 6383 a precedence for a fictitious terminal symbol named @code{UMINUS}. There |
| 6384 are no tokens of this type, but the symbol serves to stand for its |
| 6385 precedence: |
| 6386 |
| 6387 @example |
| 6388 @dots{} |
| 6389 %left '+' '-' |
| 6390 %left '*' |
| 6391 %left UMINUS |
| 6392 @end example |
| 6393 |
| 6394 Now the precedence of @code{UMINUS} can be used in specific rules: |
| 6395 |
| 6396 @example |
| 6397 @group |
| 6398 exp: @dots{} |
| 6399 | exp '-' exp |
| 6400 @dots{} |
| 6401 | '-' exp %prec UMINUS |
| 6402 @end group |
| 6403 @end example |
| 6404 |
| 6405 @ifset defaultprec |
| 6406 If you forget to append @code{%prec UMINUS} to the rule for unary |
| 6407 minus, Bison silently assumes that minus has its usual precedence. |
| 6408 This kind of problem can be tricky to debug, since one typically |
| 6409 discovers the mistake only by testing the code. |
| 6410 |
| 6411 The @code{%no-default-prec;} declaration makes it easier to discover |
| 6412 this kind of problem systematically. It causes rules that lack a |
| 6413 @code{%prec} modifier to have no precedence, even if the last terminal |
| 6414 symbol mentioned in their components has a declared precedence. |
| 6415 |
| 6416 If @code{%no-default-prec;} is in effect, you must specify @code{%prec} |
| 6417 for all rules that participate in precedence conflict resolution. |
| 6418 Then you will see any shift/reduce conflict until you tell Bison how |
| 6419 to resolve it, either by changing your grammar or by adding an |
| 6420 explicit precedence. This will probably add declarations to the |
| 6421 grammar, but it helps to protect against incorrect rule precedences. |
| 6422 |
| 6423 The effect of @code{%no-default-prec;} can be reversed by giving |
| 6424 @code{%default-prec;}, which is the default. |
| 6425 @end ifset |
| 6426 |
| 6427 @node Parser States |
| 6428 @section Parser States |
| 6429 @cindex finite-state machine |
| 6430 @cindex parser state |
| 6431 @cindex state (of parser) |
| 6432 |
| 6433 The function @code{yyparse} is implemented using a finite-state machine. |
| 6434 The values pushed on the parser stack are not simply token type codes; they |
| 6435 represent the entire sequence of terminal and nonterminal symbols at or |
| 6436 near the top of the stack. The current state collects all the information |
| 6437 about previous input which is relevant to deciding what to do next. |
| 6438 |
| 6439 Each time a lookahead token is read, the current parser state together |
| 6440 with the type of lookahead token are looked up in a table. This table |
| 6441 entry can say, ``Shift the lookahead token.'' In this case, it also |
| 6442 specifies the new parser state, which is pushed onto the top of the |
| 6443 parser stack. Or it can say, ``Reduce using rule number @var{n}.'' |
| 6444 This means that a certain number of tokens or groupings are taken off |
| 6445 the top of the stack, and replaced by one grouping. In other words, |
| 6446 that number of states are popped from the stack, and one new state is |
| 6447 pushed. |
| 6448 |
| 6449 There is one other alternative: the table can say that the lookahead token |
| 6450 is erroneous in the current state. This causes error processing to begin |
| 6451 (@pxref{Error Recovery}). |
| 6452 |
| 6453 @node Reduce/Reduce |
| 6454 @section Reduce/Reduce Conflicts |
| 6455 @cindex reduce/reduce conflict |
| 6456 @cindex conflicts, reduce/reduce |
| 6457 |
| 6458 A reduce/reduce conflict occurs if there are two or more rules that apply |
| 6459 to the same sequence of input. This usually indicates a serious error |
| 6460 in the grammar. |
| 6461 |
| 6462 For example, here is an erroneous attempt to define a sequence |
| 6463 of zero or more @code{word} groupings. |
| 6464 |
| 6465 @example |
| 6466 sequence: /* empty */ |
| 6467 @{ printf ("empty sequence\n"); @} |
| 6468 | maybeword |
| 6469 | sequence word |
| 6470 @{ printf ("added word %s\n", $2); @} |
| 6471 ; |
| 6472 |
| 6473 maybeword: /* empty */ |
| 6474 @{ printf ("empty maybeword\n"); @} |
| 6475 | word |
| 6476 @{ printf ("single word %s\n", $1); @} |
| 6477 ; |
| 6478 @end example |
| 6479 |
| 6480 @noindent |
| 6481 The error is an ambiguity: there is more than one way to parse a single |
| 6482 @code{word} into a @code{sequence}. It could be reduced to a |
| 6483 @code{maybeword} and then into a @code{sequence} via the second rule. |
| 6484 Alternatively, nothing-at-all could be reduced into a @code{sequence} |
| 6485 via the first rule, and this could be combined with the @code{word} |
| 6486 using the third rule for @code{sequence}. |
| 6487 |
| 6488 There is also more than one way to reduce nothing-at-all into a |
| 6489 @code{sequence}. This can be done directly via the first rule, |
| 6490 or indirectly via @code{maybeword} and then the second rule. |
| 6491 |
| 6492 You might think that this is a distinction without a difference, because it |
| 6493 does not change whether any particular input is valid or not. But it does |
| 6494 affect which actions are run. One parsing order runs the second rule's |
| 6495 action; the other runs the first rule's action and the third rule's action. |
| 6496 In this example, the output of the program changes. |
| 6497 |
| 6498 Bison resolves a reduce/reduce conflict by choosing to use the rule that |
| 6499 appears first in the grammar, but it is very risky to rely on this. Every |
| 6500 reduce/reduce conflict must be studied and usually eliminated. Here is the |
| 6501 proper way to define @code{sequence}: |
| 6502 |
| 6503 @example |
| 6504 sequence: /* empty */ |
| 6505 @{ printf ("empty sequence\n"); @} |
| 6506 | sequence word |
| 6507 @{ printf ("added word %s\n", $2); @} |
| 6508 ; |
| 6509 @end example |
| 6510 |
| 6511 Here is another common error that yields a reduce/reduce conflict: |
| 6512 |
| 6513 @example |
| 6514 sequence: /* empty */ |
| 6515 | sequence words |
| 6516 | sequence redirects |
| 6517 ; |
| 6518 |
| 6519 words: /* empty */ |
| 6520 | words word |
| 6521 ; |
| 6522 |
| 6523 redirects:/* empty */ |
| 6524 | redirects redirect |
| 6525 ; |
| 6526 @end example |
| 6527 |
| 6528 @noindent |
| 6529 The intention here is to define a sequence which can contain either |
| 6530 @code{word} or @code{redirect} groupings. The individual definitions of |
| 6531 @code{sequence}, @code{words} and @code{redirects} are error-free, but the |
| 6532 three together make a subtle ambiguity: even an empty input can be parsed |
| 6533 in infinitely many ways! |
| 6534 |
| 6535 Consider: nothing-at-all could be a @code{words}. Or it could be two |
| 6536 @code{words} in a row, or three, or any number. It could equally well be a |
| 6537 @code{redirects}, or two, or any number. Or it could be a @code{words} |
| 6538 followed by three @code{redirects} and another @code{words}. And so on. |
| 6539 |
| 6540 Here are two ways to correct these rules. First, to make it a single level |
| 6541 of sequence: |
| 6542 |
| 6543 @example |
| 6544 sequence: /* empty */ |
| 6545 | sequence word |
| 6546 | sequence redirect |
| 6547 ; |
| 6548 @end example |
| 6549 |
| 6550 Second, to prevent either a @code{words} or a @code{redirects} |
| 6551 from being empty: |
| 6552 |
| 6553 @example |
| 6554 sequence: /* empty */ |
| 6555 | sequence words |
| 6556 | sequence redirects |
| 6557 ; |
| 6558 |
| 6559 words: word |
| 6560 | words word |
| 6561 ; |
| 6562 |
| 6563 redirects:redirect |
| 6564 | redirects redirect |
| 6565 ; |
| 6566 @end example |
| 6567 |
| 6568 @node Mystery Conflicts |
| 6569 @section Mysterious Reduce/Reduce Conflicts |
| 6570 |
| 6571 Sometimes reduce/reduce conflicts can occur that don't look warranted. |
| 6572 Here is an example: |
| 6573 |
| 6574 @example |
| 6575 @group |
| 6576 %token ID |
| 6577 |
| 6578 %% |
| 6579 def: param_spec return_spec ',' |
| 6580 ; |
| 6581 param_spec: |
| 6582 type |
| 6583 | name_list ':' type |
| 6584 ; |
| 6585 @end group |
| 6586 @group |
| 6587 return_spec: |
| 6588 type |
| 6589 | name ':' type |
| 6590 ; |
| 6591 @end group |
| 6592 @group |
| 6593 type: ID |
| 6594 ; |
| 6595 @end group |
| 6596 @group |
| 6597 name: ID |
| 6598 ; |
| 6599 name_list: |
| 6600 name |
| 6601 | name ',' name_list |
| 6602 ; |
| 6603 @end group |
| 6604 @end example |
| 6605 |
| 6606 It would seem that this grammar can be parsed with only a single token |
| 6607 of lookahead: when a @code{param_spec} is being read, an @code{ID} is |
| 6608 a @code{name} if a comma or colon follows, or a @code{type} if another |
| 6609 @code{ID} follows. In other words, this grammar is @acronym{LR}(1). |
| 6610 |
| 6611 @cindex @acronym{LR}(1) |
| 6612 @cindex @acronym{LALR}(1) |
| 6613 However, Bison, like most parser generators, cannot actually handle all |
| 6614 @acronym{LR}(1) grammars. In this grammar, two contexts, that after |
| 6615 an @code{ID} |
| 6616 at the beginning of a @code{param_spec} and likewise at the beginning of |
| 6617 a @code{return_spec}, are similar enough that Bison assumes they are the |
| 6618 same. They appear similar because the same set of rules would be |
| 6619 active---the rule for reducing to a @code{name} and that for reducing to |
| 6620 a @code{type}. Bison is unable to determine at that stage of processing |
| 6621 that the rules would require different lookahead tokens in the two |
| 6622 contexts, so it makes a single parser state for them both. Combining |
| 6623 the two contexts causes a conflict later. In parser terminology, this |
| 6624 occurrence means that the grammar is not @acronym{LALR}(1). |
| 6625 |
| 6626 In general, it is better to fix deficiencies than to document them. But |
| 6627 this particular deficiency is intrinsically hard to fix; parser |
| 6628 generators that can handle @acronym{LR}(1) grammars are hard to write |
| 6629 and tend to |
| 6630 produce parsers that are very large. In practice, Bison is more useful |
| 6631 as it is now. |
| 6632 |
| 6633 When the problem arises, you can often fix it by identifying the two |
| 6634 parser states that are being confused, and adding something to make them |
| 6635 look distinct. In the above example, adding one rule to |
| 6636 @code{return_spec} as follows makes the problem go away: |
| 6637 |
| 6638 @example |
| 6639 @group |
| 6640 %token BOGUS |
| 6641 @dots{} |
| 6642 %% |
| 6643 @dots{} |
| 6644 return_spec: |
| 6645 type |
| 6646 | name ':' type |
| 6647 /* This rule is never used. */ |
| 6648 | ID BOGUS |
| 6649 ; |
| 6650 @end group |
| 6651 @end example |
| 6652 |
| 6653 This corrects the problem because it introduces the possibility of an |
| 6654 additional active rule in the context after the @code{ID} at the beginning of |
| 6655 @code{return_spec}. This rule is not active in the corresponding context |
| 6656 in a @code{param_spec}, so the two contexts receive distinct parser states. |
| 6657 As long as the token @code{BOGUS} is never generated by @code{yylex}, |
| 6658 the added rule cannot alter the way actual input is parsed. |
| 6659 |
| 6660 In this particular example, there is another way to solve the problem: |
| 6661 rewrite the rule for @code{return_spec} to use @code{ID} directly |
| 6662 instead of via @code{name}. This also causes the two confusing |
| 6663 contexts to have different sets of active rules, because the one for |
| 6664 @code{return_spec} activates the altered rule for @code{return_spec} |
| 6665 rather than the one for @code{name}. |
| 6666 |
| 6667 @example |
| 6668 param_spec: |
| 6669 type |
| 6670 | name_list ':' type |
| 6671 ; |
| 6672 return_spec: |
| 6673 type |
| 6674 | ID ':' type |
| 6675 ; |
| 6676 @end example |
| 6677 |
| 6678 For a more detailed exposition of @acronym{LALR}(1) parsers and parser |
| 6679 generators, please see: |
| 6680 Frank DeRemer and Thomas Pennello, Efficient Computation of |
| 6681 @acronym{LALR}(1) Look-Ahead Sets, @cite{@acronym{ACM} Transactions on |
| 6682 Programming Languages and Systems}, Vol.@: 4, No.@: 4 (October 1982), |
| 6683 pp.@: 615--649 @uref{http://doi.acm.org/10.1145/69622.357187}. |
| 6684 |
| 6685 @node Generalized LR Parsing |
| 6686 @section Generalized @acronym{LR} (@acronym{GLR}) Parsing |
| 6687 @cindex @acronym{GLR} parsing |
| 6688 @cindex generalized @acronym{LR} (@acronym{GLR}) parsing |
| 6689 @cindex ambiguous grammars |
| 6690 @cindex nondeterministic parsing |
| 6691 |
| 6692 Bison produces @emph{deterministic} parsers that choose uniquely |
| 6693 when to reduce and which reduction to apply |
| 6694 based on a summary of the preceding input and on one extra token of lookahead. |
| 6695 As a result, normal Bison handles a proper subset of the family of |
| 6696 context-free languages. |
| 6697 Ambiguous grammars, since they have strings with more than one possible |
| 6698 sequence of reductions cannot have deterministic parsers in this sense. |
| 6699 The same is true of languages that require more than one symbol of |
| 6700 lookahead, since the parser lacks the information necessary to make a |
| 6701 decision at the point it must be made in a shift-reduce parser. |
| 6702 Finally, as previously mentioned (@pxref{Mystery Conflicts}), |
| 6703 there are languages where Bison's particular choice of how to |
| 6704 summarize the input seen so far loses necessary information. |
| 6705 |
| 6706 When you use the @samp{%glr-parser} declaration in your grammar file, |
| 6707 Bison generates a parser that uses a different algorithm, called |
| 6708 Generalized @acronym{LR} (or @acronym{GLR}). A Bison @acronym{GLR} |
| 6709 parser uses the same basic |
| 6710 algorithm for parsing as an ordinary Bison parser, but behaves |
| 6711 differently in cases where there is a shift-reduce conflict that has not |
| 6712 been resolved by precedence rules (@pxref{Precedence}) or a |
| 6713 reduce-reduce conflict. When a @acronym{GLR} parser encounters such a |
| 6714 situation, it |
| 6715 effectively @emph{splits} into a several parsers, one for each possible |
| 6716 shift or reduction. These parsers then proceed as usual, consuming |
| 6717 tokens in lock-step. Some of the stacks may encounter other conflicts |
| 6718 and split further, with the result that instead of a sequence of states, |
| 6719 a Bison @acronym{GLR} parsing stack is what is in effect a tree of states. |
| 6720 |
| 6721 In effect, each stack represents a guess as to what the proper parse |
| 6722 is. Additional input may indicate that a guess was wrong, in which case |
| 6723 the appropriate stack silently disappears. Otherwise, the semantics |
| 6724 actions generated in each stack are saved, rather than being executed |
| 6725 immediately. When a stack disappears, its saved semantic actions never |
| 6726 get executed. When a reduction causes two stacks to become equivalent, |
| 6727 their sets of semantic actions are both saved with the state that |
| 6728 results from the reduction. We say that two stacks are equivalent |
| 6729 when they both represent the same sequence of states, |
| 6730 and each pair of corresponding states represents a |
| 6731 grammar symbol that produces the same segment of the input token |
| 6732 stream. |
| 6733 |
| 6734 Whenever the parser makes a transition from having multiple |
| 6735 states to having one, it reverts to the normal @acronym{LALR}(1) parsing |
| 6736 algorithm, after resolving and executing the saved-up actions. |
| 6737 At this transition, some of the states on the stack will have semantic |
| 6738 values that are sets (actually multisets) of possible actions. The |
| 6739 parser tries to pick one of the actions by first finding one whose rule |
| 6740 has the highest dynamic precedence, as set by the @samp{%dprec} |
| 6741 declaration. Otherwise, if the alternative actions are not ordered by |
| 6742 precedence, but there the same merging function is declared for both |
| 6743 rules by the @samp{%merge} declaration, |
| 6744 Bison resolves and evaluates both and then calls the merge function on |
| 6745 the result. Otherwise, it reports an ambiguity. |
| 6746 |
| 6747 It is possible to use a data structure for the @acronym{GLR} parsing tree that |
| 6748 permits the processing of any @acronym{LALR}(1) grammar in linear time (in the |
| 6749 size of the input), any unambiguous (not necessarily |
| 6750 @acronym{LALR}(1)) grammar in |
| 6751 quadratic worst-case time, and any general (possibly ambiguous) |
| 6752 context-free grammar in cubic worst-case time. However, Bison currently |
| 6753 uses a simpler data structure that requires time proportional to the |
| 6754 length of the input times the maximum number of stacks required for any |
| 6755 prefix of the input. Thus, really ambiguous or nondeterministic |
| 6756 grammars can require exponential time and space to process. Such badly |
| 6757 behaving examples, however, are not generally of practical interest. |
| 6758 Usually, nondeterminism in a grammar is local---the parser is ``in |
| 6759 doubt'' only for a few tokens at a time. Therefore, the current data |
| 6760 structure should generally be adequate. On @acronym{LALR}(1) portions of a |
| 6761 grammar, in particular, it is only slightly slower than with the default |
| 6762 Bison parser. |
| 6763 |
| 6764 For a more detailed exposition of @acronym{GLR} parsers, please see: Elizabeth |
| 6765 Scott, Adrian Johnstone and Shamsa Sadaf Hussain, Tomita-Style |
| 6766 Generalised @acronym{LR} Parsers, Royal Holloway, University of |
| 6767 London, Department of Computer Science, TR-00-12, |
| 6768 @uref{http://www.cs.rhul.ac.uk/research/languages/publications/tomita_style_1.ps
}, |
| 6769 (2000-12-24). |
| 6770 |
| 6771 @node Memory Management |
| 6772 @section Memory Management, and How to Avoid Memory Exhaustion |
| 6773 @cindex memory exhaustion |
| 6774 @cindex memory management |
| 6775 @cindex stack overflow |
| 6776 @cindex parser stack overflow |
| 6777 @cindex overflow of parser stack |
| 6778 |
| 6779 The Bison parser stack can run out of memory if too many tokens are shifted and |
| 6780 not reduced. When this happens, the parser function @code{yyparse} |
| 6781 calls @code{yyerror} and then returns 2. |
| 6782 |
| 6783 Because Bison parsers have growing stacks, hitting the upper limit |
| 6784 usually results from using a right recursion instead of a left |
| 6785 recursion, @xref{Recursion, ,Recursive Rules}. |
| 6786 |
| 6787 @vindex YYMAXDEPTH |
| 6788 By defining the macro @code{YYMAXDEPTH}, you can control how deep the |
| 6789 parser stack can become before memory is exhausted. Define the |
| 6790 macro with a value that is an integer. This value is the maximum number |
| 6791 of tokens that can be shifted (and not reduced) before overflow. |
| 6792 |
| 6793 The stack space allowed is not necessarily allocated. If you specify a |
| 6794 large value for @code{YYMAXDEPTH}, the parser normally allocates a small |
| 6795 stack at first, and then makes it bigger by stages as needed. This |
| 6796 increasing allocation happens automatically and silently. Therefore, |
| 6797 you do not need to make @code{YYMAXDEPTH} painfully small merely to save |
| 6798 space for ordinary inputs that do not need much stack. |
| 6799 |
| 6800 However, do not allow @code{YYMAXDEPTH} to be a value so large that |
| 6801 arithmetic overflow could occur when calculating the size of the stack |
| 6802 space. Also, do not allow @code{YYMAXDEPTH} to be less than |
| 6803 @code{YYINITDEPTH}. |
| 6804 |
| 6805 @cindex default stack limit |
| 6806 The default value of @code{YYMAXDEPTH}, if you do not define it, is |
| 6807 10000. |
| 6808 |
| 6809 @vindex YYINITDEPTH |
| 6810 You can control how much stack is allocated initially by defining the |
| 6811 macro @code{YYINITDEPTH} to a positive integer. For the C |
| 6812 @acronym{LALR}(1) parser, this value must be a compile-time constant |
| 6813 unless you are assuming C99 or some other target language or compiler |
| 6814 that allows variable-length arrays. The default is 200. |
| 6815 |
| 6816 Do not allow @code{YYINITDEPTH} to be greater than @code{YYMAXDEPTH}. |
| 6817 |
| 6818 @c FIXME: C++ output. |
| 6819 Because of semantical differences between C and C++, the |
| 6820 @acronym{LALR}(1) parsers in C produced by Bison cannot grow when compiled |
| 6821 by C++ compilers. In this precise case (compiling a C parser as C++) you are |
| 6822 suggested to grow @code{YYINITDEPTH}. The Bison maintainers hope to fix |
| 6823 this deficiency in a future release. |
| 6824 |
| 6825 @node Error Recovery |
| 6826 @chapter Error Recovery |
| 6827 @cindex error recovery |
| 6828 @cindex recovery from errors |
| 6829 |
| 6830 It is not usually acceptable to have a program terminate on a syntax |
| 6831 error. For example, a compiler should recover sufficiently to parse the |
| 6832 rest of the input file and check it for errors; a calculator should accept |
| 6833 another expression. |
| 6834 |
| 6835 In a simple interactive command parser where each input is one line, it may |
| 6836 be sufficient to allow @code{yyparse} to return 1 on error and have the |
| 6837 caller ignore the rest of the input line when that happens (and then call |
| 6838 @code{yyparse} again). But this is inadequate for a compiler, because it |
| 6839 forgets all the syntactic context leading up to the error. A syntax error |
| 6840 deep within a function in the compiler input should not cause the compiler |
| 6841 to treat the following line like the beginning of a source file. |
| 6842 |
| 6843 @findex error |
| 6844 You can define how to recover from a syntax error by writing rules to |
| 6845 recognize the special token @code{error}. This is a terminal symbol that |
| 6846 is always defined (you need not declare it) and reserved for error |
| 6847 handling. The Bison parser generates an @code{error} token whenever a |
| 6848 syntax error happens; if you have provided a rule to recognize this token |
| 6849 in the current context, the parse can continue. |
| 6850 |
| 6851 For example: |
| 6852 |
| 6853 @example |
| 6854 stmnts: /* empty string */ |
| 6855 | stmnts '\n' |
| 6856 | stmnts exp '\n' |
| 6857 | stmnts error '\n' |
| 6858 @end example |
| 6859 |
| 6860 The fourth rule in this example says that an error followed by a newline |
| 6861 makes a valid addition to any @code{stmnts}. |
| 6862 |
| 6863 What happens if a syntax error occurs in the middle of an @code{exp}? The |
| 6864 error recovery rule, interpreted strictly, applies to the precise sequence |
| 6865 of a @code{stmnts}, an @code{error} and a newline. If an error occurs in |
| 6866 the middle of an @code{exp}, there will probably be some additional tokens |
| 6867 and subexpressions on the stack after the last @code{stmnts}, and there |
| 6868 will be tokens to read before the next newline. So the rule is not |
| 6869 applicable in the ordinary way. |
| 6870 |
| 6871 But Bison can force the situation to fit the rule, by discarding part of |
| 6872 the semantic context and part of the input. First it discards states |
| 6873 and objects from the stack until it gets back to a state in which the |
| 6874 @code{error} token is acceptable. (This means that the subexpressions |
| 6875 already parsed are discarded, back to the last complete @code{stmnts}.) |
| 6876 At this point the @code{error} token can be shifted. Then, if the old |
| 6877 lookahead token is not acceptable to be shifted next, the parser reads |
| 6878 tokens and discards them until it finds a token which is acceptable. In |
| 6879 this example, Bison reads and discards input until the next newline so |
| 6880 that the fourth rule can apply. Note that discarded symbols are |
| 6881 possible sources of memory leaks, see @ref{Destructor Decl, , Freeing |
| 6882 Discarded Symbols}, for a means to reclaim this memory. |
| 6883 |
| 6884 The choice of error rules in the grammar is a choice of strategies for |
| 6885 error recovery. A simple and useful strategy is simply to skip the rest of |
| 6886 the current input line or current statement if an error is detected: |
| 6887 |
| 6888 @example |
| 6889 stmnt: error ';' /* On error, skip until ';' is read. */ |
| 6890 @end example |
| 6891 |
| 6892 It is also useful to recover to the matching close-delimiter of an |
| 6893 opening-delimiter that has already been parsed. Otherwise the |
| 6894 close-delimiter will probably appear to be unmatched, and generate another, |
| 6895 spurious error message: |
| 6896 |
| 6897 @example |
| 6898 primary: '(' expr ')' |
| 6899 | '(' error ')' |
| 6900 @dots{} |
| 6901 ; |
| 6902 @end example |
| 6903 |
| 6904 Error recovery strategies are necessarily guesses. When they guess wrong, |
| 6905 one syntax error often leads to another. In the above example, the error |
| 6906 recovery rule guesses that an error is due to bad input within one |
| 6907 @code{stmnt}. Suppose that instead a spurious semicolon is inserted in the |
| 6908 middle of a valid @code{stmnt}. After the error recovery rule recovers |
| 6909 from the first error, another syntax error will be found straightaway, |
| 6910 since the text following the spurious semicolon is also an invalid |
| 6911 @code{stmnt}. |
| 6912 |
| 6913 To prevent an outpouring of error messages, the parser will output no error |
| 6914 message for another syntax error that happens shortly after the first; only |
| 6915 after three consecutive input tokens have been successfully shifted will |
| 6916 error messages resume. |
| 6917 |
| 6918 Note that rules which accept the @code{error} token may have actions, just |
| 6919 as any other rules can. |
| 6920 |
| 6921 @findex yyerrok |
| 6922 You can make error messages resume immediately by using the macro |
| 6923 @code{yyerrok} in an action. If you do this in the error rule's action, no |
| 6924 error messages will be suppressed. This macro requires no arguments; |
| 6925 @samp{yyerrok;} is a valid C statement. |
| 6926 |
| 6927 @findex yyclearin |
| 6928 The previous lookahead token is reanalyzed immediately after an error. If |
| 6929 this is unacceptable, then the macro @code{yyclearin} may be used to clear |
| 6930 this token. Write the statement @samp{yyclearin;} in the error rule's |
| 6931 action. |
| 6932 @xref{Action Features, ,Special Features for Use in Actions}. |
| 6933 |
| 6934 For example, suppose that on a syntax error, an error handling routine is |
| 6935 called that advances the input stream to some point where parsing should |
| 6936 once again commence. The next symbol returned by the lexical scanner is |
| 6937 probably correct. The previous lookahead token ought to be discarded |
| 6938 with @samp{yyclearin;}. |
| 6939 |
| 6940 @vindex YYRECOVERING |
| 6941 The expression @code{YYRECOVERING ()} yields 1 when the parser |
| 6942 is recovering from a syntax error, and 0 otherwise. |
| 6943 Syntax error diagnostics are suppressed while recovering from a syntax |
| 6944 error. |
| 6945 |
| 6946 @node Context Dependency |
| 6947 @chapter Handling Context Dependencies |
| 6948 |
| 6949 The Bison paradigm is to parse tokens first, then group them into larger |
| 6950 syntactic units. In many languages, the meaning of a token is affected by |
| 6951 its context. Although this violates the Bison paradigm, certain techniques |
| 6952 (known as @dfn{kludges}) may enable you to write Bison parsers for such |
| 6953 languages. |
| 6954 |
| 6955 @menu |
| 6956 * Semantic Tokens:: Token parsing can depend on the semantic context. |
| 6957 * Lexical Tie-ins:: Token parsing can depend on the syntactic context. |
| 6958 * Tie-in Recovery:: Lexical tie-ins have implications for how |
| 6959 error recovery rules must be written. |
| 6960 @end menu |
| 6961 |
| 6962 (Actually, ``kludge'' means any technique that gets its job done but is |
| 6963 neither clean nor robust.) |
| 6964 |
| 6965 @node Semantic Tokens |
| 6966 @section Semantic Info in Token Types |
| 6967 |
| 6968 The C language has a context dependency: the way an identifier is used |
| 6969 depends on what its current meaning is. For example, consider this: |
| 6970 |
| 6971 @example |
| 6972 foo (x); |
| 6973 @end example |
| 6974 |
| 6975 This looks like a function call statement, but if @code{foo} is a typedef |
| 6976 name, then this is actually a declaration of @code{x}. How can a Bison |
| 6977 parser for C decide how to parse this input? |
| 6978 |
| 6979 The method used in @acronym{GNU} C is to have two different token types, |
| 6980 @code{IDENTIFIER} and @code{TYPENAME}. When @code{yylex} finds an |
| 6981 identifier, it looks up the current declaration of the identifier in order |
| 6982 to decide which token type to return: @code{TYPENAME} if the identifier is |
| 6983 declared as a typedef, @code{IDENTIFIER} otherwise. |
| 6984 |
| 6985 The grammar rules can then express the context dependency by the choice of |
| 6986 token type to recognize. @code{IDENTIFIER} is accepted as an expression, |
| 6987 but @code{TYPENAME} is not. @code{TYPENAME} can start a declaration, but |
| 6988 @code{IDENTIFIER} cannot. In contexts where the meaning of the identifier |
| 6989 is @emph{not} significant, such as in declarations that can shadow a |
| 6990 typedef name, either @code{TYPENAME} or @code{IDENTIFIER} is |
| 6991 accepted---there is one rule for each of the two token types. |
| 6992 |
| 6993 This technique is simple to use if the decision of which kinds of |
| 6994 identifiers to allow is made at a place close to where the identifier is |
| 6995 parsed. But in C this is not always so: C allows a declaration to |
| 6996 redeclare a typedef name provided an explicit type has been specified |
| 6997 earlier: |
| 6998 |
| 6999 @example |
| 7000 typedef int foo, bar; |
| 7001 int baz (void) |
| 7002 @{ |
| 7003 static bar (bar); /* @r{redeclare @code{bar} as static variable} */ |
| 7004 extern foo foo (foo); /* @r{redeclare @code{foo} as function} */ |
| 7005 return foo (bar); |
| 7006 @} |
| 7007 @end example |
| 7008 |
| 7009 Unfortunately, the name being declared is separated from the declaration |
| 7010 construct itself by a complicated syntactic structure---the ``declarator''. |
| 7011 |
| 7012 As a result, part of the Bison parser for C needs to be duplicated, with |
| 7013 all the nonterminal names changed: once for parsing a declaration in |
| 7014 which a typedef name can be redefined, and once for parsing a |
| 7015 declaration in which that can't be done. Here is a part of the |
| 7016 duplication, with actions omitted for brevity: |
| 7017 |
| 7018 @example |
| 7019 initdcl: |
| 7020 declarator maybeasm '=' |
| 7021 init |
| 7022 | declarator maybeasm |
| 7023 ; |
| 7024 |
| 7025 notype_initdcl: |
| 7026 notype_declarator maybeasm '=' |
| 7027 init |
| 7028 | notype_declarator maybeasm |
| 7029 ; |
| 7030 @end example |
| 7031 |
| 7032 @noindent |
| 7033 Here @code{initdcl} can redeclare a typedef name, but @code{notype_initdcl} |
| 7034 cannot. The distinction between @code{declarator} and |
| 7035 @code{notype_declarator} is the same sort of thing. |
| 7036 |
| 7037 There is some similarity between this technique and a lexical tie-in |
| 7038 (described next), in that information which alters the lexical analysis is |
| 7039 changed during parsing by other parts of the program. The difference is |
| 7040 here the information is global, and is used for other purposes in the |
| 7041 program. A true lexical tie-in has a special-purpose flag controlled by |
| 7042 the syntactic context. |
| 7043 |
| 7044 @node Lexical Tie-ins |
| 7045 @section Lexical Tie-ins |
| 7046 @cindex lexical tie-in |
| 7047 |
| 7048 One way to handle context-dependency is the @dfn{lexical tie-in}: a flag |
| 7049 which is set by Bison actions, whose purpose is to alter the way tokens are |
| 7050 parsed. |
| 7051 |
| 7052 For example, suppose we have a language vaguely like C, but with a special |
| 7053 construct @samp{hex (@var{hex-expr})}. After the keyword @code{hex} comes |
| 7054 an expression in parentheses in which all integers are hexadecimal. In |
| 7055 particular, the token @samp{a1b} must be treated as an integer rather than |
| 7056 as an identifier if it appears in that context. Here is how you can do it: |
| 7057 |
| 7058 @example |
| 7059 @group |
| 7060 %@{ |
| 7061 int hexflag; |
| 7062 int yylex (void); |
| 7063 void yyerror (char const *); |
| 7064 %@} |
| 7065 %% |
| 7066 @dots{} |
| 7067 @end group |
| 7068 @group |
| 7069 expr: IDENTIFIER |
| 7070 | constant |
| 7071 | HEX '(' |
| 7072 @{ hexflag = 1; @} |
| 7073 expr ')' |
| 7074 @{ hexflag = 0; |
| 7075 $$ = $4; @} |
| 7076 | expr '+' expr |
| 7077 @{ $$ = make_sum ($1, $3); @} |
| 7078 @dots{} |
| 7079 ; |
| 7080 @end group |
| 7081 |
| 7082 @group |
| 7083 constant: |
| 7084 INTEGER |
| 7085 | STRING |
| 7086 ; |
| 7087 @end group |
| 7088 @end example |
| 7089 |
| 7090 @noindent |
| 7091 Here we assume that @code{yylex} looks at the value of @code{hexflag}; when |
| 7092 it is nonzero, all integers are parsed in hexadecimal, and tokens starting |
| 7093 with letters are parsed as integers if possible. |
| 7094 |
| 7095 The declaration of @code{hexflag} shown in the prologue of the parser file |
| 7096 is needed to make it accessible to the actions (@pxref{Prologue, ,The Prologue})
. |
| 7097 You must also write the code in @code{yylex} to obey the flag. |
| 7098 |
| 7099 @node Tie-in Recovery |
| 7100 @section Lexical Tie-ins and Error Recovery |
| 7101 |
| 7102 Lexical tie-ins make strict demands on any error recovery rules you have. |
| 7103 @xref{Error Recovery}. |
| 7104 |
| 7105 The reason for this is that the purpose of an error recovery rule is to |
| 7106 abort the parsing of one construct and resume in some larger construct. |
| 7107 For example, in C-like languages, a typical error recovery rule is to skip |
| 7108 tokens until the next semicolon, and then start a new statement, like this: |
| 7109 |
| 7110 @example |
| 7111 stmt: expr ';' |
| 7112 | IF '(' expr ')' stmt @{ @dots{} @} |
| 7113 @dots{} |
| 7114 error ';' |
| 7115 @{ hexflag = 0; @} |
| 7116 ; |
| 7117 @end example |
| 7118 |
| 7119 If there is a syntax error in the middle of a @samp{hex (@var{expr})} |
| 7120 construct, this error rule will apply, and then the action for the |
| 7121 completed @samp{hex (@var{expr})} will never run. So @code{hexflag} would |
| 7122 remain set for the entire rest of the input, or until the next @code{hex} |
| 7123 keyword, causing identifiers to be misinterpreted as integers. |
| 7124 |
| 7125 To avoid this problem the error recovery rule itself clears @code{hexflag}. |
| 7126 |
| 7127 There may also be an error recovery rule that works within expressions. |
| 7128 For example, there could be a rule which applies within parentheses |
| 7129 and skips to the close-parenthesis: |
| 7130 |
| 7131 @example |
| 7132 @group |
| 7133 expr: @dots{} |
| 7134 | '(' expr ')' |
| 7135 @{ $$ = $2; @} |
| 7136 | '(' error ')' |
| 7137 @dots{} |
| 7138 @end group |
| 7139 @end example |
| 7140 |
| 7141 If this rule acts within the @code{hex} construct, it is not going to abort |
| 7142 that construct (since it applies to an inner level of parentheses within |
| 7143 the construct). Therefore, it should not clear the flag: the rest of |
| 7144 the @code{hex} construct should be parsed with the flag still in effect. |
| 7145 |
| 7146 What if there is an error recovery rule which might abort out of the |
| 7147 @code{hex} construct or might not, depending on circumstances? There is no |
| 7148 way you can write the action to determine whether a @code{hex} construct is |
| 7149 being aborted or not. So if you are using a lexical tie-in, you had better |
| 7150 make sure your error recovery rules are not of this kind. Each rule must |
| 7151 be such that you can be sure that it always will, or always won't, have to |
| 7152 clear the flag. |
| 7153 |
| 7154 @c ================================================== Debugging Your Parser |
| 7155 |
| 7156 @node Debugging |
| 7157 @chapter Debugging Your Parser |
| 7158 |
| 7159 Developing a parser can be a challenge, especially if you don't |
| 7160 understand the algorithm (@pxref{Algorithm, ,The Bison Parser |
| 7161 Algorithm}). Even so, sometimes a detailed description of the automaton |
| 7162 can help (@pxref{Understanding, , Understanding Your Parser}), or |
| 7163 tracing the execution of the parser can give some insight on why it |
| 7164 behaves improperly (@pxref{Tracing, , Tracing Your Parser}). |
| 7165 |
| 7166 @menu |
| 7167 * Understanding:: Understanding the structure of your parser. |
| 7168 * Tracing:: Tracing the execution of your parser. |
| 7169 @end menu |
| 7170 |
| 7171 @node Understanding |
| 7172 @section Understanding Your Parser |
| 7173 |
| 7174 As documented elsewhere (@pxref{Algorithm, ,The Bison Parser Algorithm}) |
| 7175 Bison parsers are @dfn{shift/reduce automata}. In some cases (much more |
| 7176 frequent than one would hope), looking at this automaton is required to |
| 7177 tune or simply fix a parser. Bison provides two different |
| 7178 representation of it, either textually or graphically (as a DOT file). |
| 7179 |
| 7180 The textual file is generated when the options @option{--report} or |
| 7181 @option{--verbose} are specified, see @xref{Invocation, , Invoking |
| 7182 Bison}. Its name is made by removing @samp{.tab.c} or @samp{.c} from |
| 7183 the parser output file name, and adding @samp{.output} instead. |
| 7184 Therefore, if the input file is @file{foo.y}, then the parser file is |
| 7185 called @file{foo.tab.c} by default. As a consequence, the verbose |
| 7186 output file is called @file{foo.output}. |
| 7187 |
| 7188 The following grammar file, @file{calc.y}, will be used in the sequel: |
| 7189 |
| 7190 @example |
| 7191 %token NUM STR |
| 7192 %left '+' '-' |
| 7193 %left '*' |
| 7194 %% |
| 7195 exp: exp '+' exp |
| 7196 | exp '-' exp |
| 7197 | exp '*' exp |
| 7198 | exp '/' exp |
| 7199 | NUM |
| 7200 ; |
| 7201 useless: STR; |
| 7202 %% |
| 7203 @end example |
| 7204 |
| 7205 @command{bison} reports: |
| 7206 |
| 7207 @example |
| 7208 calc.y: warning: 1 nonterminal and 1 rule useless in grammar |
| 7209 calc.y:11.1-7: warning: nonterminal useless in grammar: useless |
| 7210 calc.y:11.10-12: warning: rule useless in grammar: useless: STR |
| 7211 calc.y: conflicts: 7 shift/reduce |
| 7212 @end example |
| 7213 |
| 7214 When given @option{--report=state}, in addition to @file{calc.tab.c}, it |
| 7215 creates a file @file{calc.output} with contents detailed below. The |
| 7216 order of the output and the exact presentation might vary, but the |
| 7217 interpretation is the same. |
| 7218 |
| 7219 The first section includes details on conflicts that were solved thanks |
| 7220 to precedence and/or associativity: |
| 7221 |
| 7222 @example |
| 7223 Conflict in state 8 between rule 2 and token '+' resolved as reduce. |
| 7224 Conflict in state 8 between rule 2 and token '-' resolved as reduce. |
| 7225 Conflict in state 8 between rule 2 and token '*' resolved as shift. |
| 7226 @exdent @dots{} |
| 7227 @end example |
| 7228 |
| 7229 @noindent |
| 7230 The next section lists states that still have conflicts. |
| 7231 |
| 7232 @example |
| 7233 State 8 conflicts: 1 shift/reduce |
| 7234 State 9 conflicts: 1 shift/reduce |
| 7235 State 10 conflicts: 1 shift/reduce |
| 7236 State 11 conflicts: 4 shift/reduce |
| 7237 @end example |
| 7238 |
| 7239 @noindent |
| 7240 @cindex token, useless |
| 7241 @cindex useless token |
| 7242 @cindex nonterminal, useless |
| 7243 @cindex useless nonterminal |
| 7244 @cindex rule, useless |
| 7245 @cindex useless rule |
| 7246 The next section reports useless tokens, nonterminal and rules. Useless |
| 7247 nonterminals and rules are removed in order to produce a smaller parser, |
| 7248 but useless tokens are preserved, since they might be used by the |
| 7249 scanner (note the difference between ``useless'' and ``unused'' |
| 7250 below): |
| 7251 |
| 7252 @example |
| 7253 Nonterminals useless in grammar: |
| 7254 useless |
| 7255 |
| 7256 Terminals unused in grammar: |
| 7257 STR |
| 7258 |
| 7259 Rules useless in grammar: |
| 7260 #6 useless: STR; |
| 7261 @end example |
| 7262 |
| 7263 @noindent |
| 7264 The next section reproduces the exact grammar that Bison used: |
| 7265 |
| 7266 @example |
| 7267 Grammar |
| 7268 |
| 7269 Number, Line, Rule |
| 7270 0 5 $accept -> exp $end |
| 7271 1 5 exp -> exp '+' exp |
| 7272 2 6 exp -> exp '-' exp |
| 7273 3 7 exp -> exp '*' exp |
| 7274 4 8 exp -> exp '/' exp |
| 7275 5 9 exp -> NUM |
| 7276 @end example |
| 7277 |
| 7278 @noindent |
| 7279 and reports the uses of the symbols: |
| 7280 |
| 7281 @example |
| 7282 Terminals, with rules where they appear |
| 7283 |
| 7284 $end (0) 0 |
| 7285 '*' (42) 3 |
| 7286 '+' (43) 1 |
| 7287 '-' (45) 2 |
| 7288 '/' (47) 4 |
| 7289 error (256) |
| 7290 NUM (258) 5 |
| 7291 |
| 7292 Nonterminals, with rules where they appear |
| 7293 |
| 7294 $accept (8) |
| 7295 on left: 0 |
| 7296 exp (9) |
| 7297 on left: 1 2 3 4 5, on right: 0 1 2 3 4 |
| 7298 @end example |
| 7299 |
| 7300 @noindent |
| 7301 @cindex item |
| 7302 @cindex pointed rule |
| 7303 @cindex rule, pointed |
| 7304 Bison then proceeds onto the automaton itself, describing each state |
| 7305 with it set of @dfn{items}, also known as @dfn{pointed rules}. Each |
| 7306 item is a production rule together with a point (marked by @samp{.}) |
| 7307 that the input cursor. |
| 7308 |
| 7309 @example |
| 7310 state 0 |
| 7311 |
| 7312 $accept -> . exp $ (rule 0) |
| 7313 |
| 7314 NUM shift, and go to state 1 |
| 7315 |
| 7316 exp go to state 2 |
| 7317 @end example |
| 7318 |
| 7319 This reads as follows: ``state 0 corresponds to being at the very |
| 7320 beginning of the parsing, in the initial rule, right before the start |
| 7321 symbol (here, @code{exp}). When the parser returns to this state right |
| 7322 after having reduced a rule that produced an @code{exp}, the control |
| 7323 flow jumps to state 2. If there is no such transition on a nonterminal |
| 7324 symbol, and the lookahead is a @code{NUM}, then this token is shifted on |
| 7325 the parse stack, and the control flow jumps to state 1. Any other |
| 7326 lookahead triggers a syntax error.'' |
| 7327 |
| 7328 @cindex core, item set |
| 7329 @cindex item set core |
| 7330 @cindex kernel, item set |
| 7331 @cindex item set core |
| 7332 Even though the only active rule in state 0 seems to be rule 0, the |
| 7333 report lists @code{NUM} as a lookahead token because @code{NUM} can be |
| 7334 at the beginning of any rule deriving an @code{exp}. By default Bison |
| 7335 reports the so-called @dfn{core} or @dfn{kernel} of the item set, but if |
| 7336 you want to see more detail you can invoke @command{bison} with |
| 7337 @option{--report=itemset} to list all the items, include those that can |
| 7338 be derived: |
| 7339 |
| 7340 @example |
| 7341 state 0 |
| 7342 |
| 7343 $accept -> . exp $ (rule 0) |
| 7344 exp -> . exp '+' exp (rule 1) |
| 7345 exp -> . exp '-' exp (rule 2) |
| 7346 exp -> . exp '*' exp (rule 3) |
| 7347 exp -> . exp '/' exp (rule 4) |
| 7348 exp -> . NUM (rule 5) |
| 7349 |
| 7350 NUM shift, and go to state 1 |
| 7351 |
| 7352 exp go to state 2 |
| 7353 @end example |
| 7354 |
| 7355 @noindent |
| 7356 In the state 1... |
| 7357 |
| 7358 @example |
| 7359 state 1 |
| 7360 |
| 7361 exp -> NUM . (rule 5) |
| 7362 |
| 7363 $default reduce using rule 5 (exp) |
| 7364 @end example |
| 7365 |
| 7366 @noindent |
| 7367 the rule 5, @samp{exp: NUM;}, is completed. Whatever the lookahead token |
| 7368 (@samp{$default}), the parser will reduce it. If it was coming from |
| 7369 state 0, then, after this reduction it will return to state 0, and will |
| 7370 jump to state 2 (@samp{exp: go to state 2}). |
| 7371 |
| 7372 @example |
| 7373 state 2 |
| 7374 |
| 7375 $accept -> exp . $ (rule 0) |
| 7376 exp -> exp . '+' exp (rule 1) |
| 7377 exp -> exp . '-' exp (rule 2) |
| 7378 exp -> exp . '*' exp (rule 3) |
| 7379 exp -> exp . '/' exp (rule 4) |
| 7380 |
| 7381 $ shift, and go to state 3 |
| 7382 '+' shift, and go to state 4 |
| 7383 '-' shift, and go to state 5 |
| 7384 '*' shift, and go to state 6 |
| 7385 '/' shift, and go to state 7 |
| 7386 @end example |
| 7387 |
| 7388 @noindent |
| 7389 In state 2, the automaton can only shift a symbol. For instance, |
| 7390 because of the item @samp{exp -> exp . '+' exp}, if the lookahead if |
| 7391 @samp{+}, it will be shifted on the parse stack, and the automaton |
| 7392 control will jump to state 4, corresponding to the item @samp{exp -> exp |
| 7393 '+' . exp}. Since there is no default action, any other token than |
| 7394 those listed above will trigger a syntax error. |
| 7395 |
| 7396 The state 3 is named the @dfn{final state}, or the @dfn{accepting |
| 7397 state}: |
| 7398 |
| 7399 @example |
| 7400 state 3 |
| 7401 |
| 7402 $accept -> exp $ . (rule 0) |
| 7403 |
| 7404 $default accept |
| 7405 @end example |
| 7406 |
| 7407 @noindent |
| 7408 the initial rule is completed (the start symbol and the end |
| 7409 of input were read), the parsing exits successfully. |
| 7410 |
| 7411 The interpretation of states 4 to 7 is straightforward, and is left to |
| 7412 the reader. |
| 7413 |
| 7414 @example |
| 7415 state 4 |
| 7416 |
| 7417 exp -> exp '+' . exp (rule 1) |
| 7418 |
| 7419 NUM shift, and go to state 1 |
| 7420 |
| 7421 exp go to state 8 |
| 7422 |
| 7423 state 5 |
| 7424 |
| 7425 exp -> exp '-' . exp (rule 2) |
| 7426 |
| 7427 NUM shift, and go to state 1 |
| 7428 |
| 7429 exp go to state 9 |
| 7430 |
| 7431 state 6 |
| 7432 |
| 7433 exp -> exp '*' . exp (rule 3) |
| 7434 |
| 7435 NUM shift, and go to state 1 |
| 7436 |
| 7437 exp go to state 10 |
| 7438 |
| 7439 state 7 |
| 7440 |
| 7441 exp -> exp '/' . exp (rule 4) |
| 7442 |
| 7443 NUM shift, and go to state 1 |
| 7444 |
| 7445 exp go to state 11 |
| 7446 @end example |
| 7447 |
| 7448 As was announced in beginning of the report, @samp{State 8 conflicts: |
| 7449 1 shift/reduce}: |
| 7450 |
| 7451 @example |
| 7452 state 8 |
| 7453 |
| 7454 exp -> exp . '+' exp (rule 1) |
| 7455 exp -> exp '+' exp . (rule 1) |
| 7456 exp -> exp . '-' exp (rule 2) |
| 7457 exp -> exp . '*' exp (rule 3) |
| 7458 exp -> exp . '/' exp (rule 4) |
| 7459 |
| 7460 '*' shift, and go to state 6 |
| 7461 '/' shift, and go to state 7 |
| 7462 |
| 7463 '/' [reduce using rule 1 (exp)] |
| 7464 $default reduce using rule 1 (exp) |
| 7465 @end example |
| 7466 |
| 7467 Indeed, there are two actions associated to the lookahead @samp{/}: |
| 7468 either shifting (and going to state 7), or reducing rule 1. The |
| 7469 conflict means that either the grammar is ambiguous, or the parser lacks |
| 7470 information to make the right decision. Indeed the grammar is |
| 7471 ambiguous, as, since we did not specify the precedence of @samp{/}, the |
| 7472 sentence @samp{NUM + NUM / NUM} can be parsed as @samp{NUM + (NUM / |
| 7473 NUM)}, which corresponds to shifting @samp{/}, or as @samp{(NUM + NUM) / |
| 7474 NUM}, which corresponds to reducing rule 1. |
| 7475 |
| 7476 Because in @acronym{LALR}(1) parsing a single decision can be made, Bison |
| 7477 arbitrarily chose to disable the reduction, see @ref{Shift/Reduce, , |
| 7478 Shift/Reduce Conflicts}. Discarded actions are reported in between |
| 7479 square brackets. |
| 7480 |
| 7481 Note that all the previous states had a single possible action: either |
| 7482 shifting the next token and going to the corresponding state, or |
| 7483 reducing a single rule. In the other cases, i.e., when shifting |
| 7484 @emph{and} reducing is possible or when @emph{several} reductions are |
| 7485 possible, the lookahead is required to select the action. State 8 is |
| 7486 one such state: if the lookahead is @samp{*} or @samp{/} then the action |
| 7487 is shifting, otherwise the action is reducing rule 1. In other words, |
| 7488 the first two items, corresponding to rule 1, are not eligible when the |
| 7489 lookahead token is @samp{*}, since we specified that @samp{*} has higher |
| 7490 precedence than @samp{+}. More generally, some items are eligible only |
| 7491 with some set of possible lookahead tokens. When run with |
| 7492 @option{--report=lookahead}, Bison specifies these lookahead tokens: |
| 7493 |
| 7494 @example |
| 7495 state 8 |
| 7496 |
| 7497 exp -> exp . '+' exp (rule 1) |
| 7498 exp -> exp '+' exp . [$, '+', '-', '/'] (rule 1) |
| 7499 exp -> exp . '-' exp (rule 2) |
| 7500 exp -> exp . '*' exp (rule 3) |
| 7501 exp -> exp . '/' exp (rule 4) |
| 7502 |
| 7503 '*' shift, and go to state 6 |
| 7504 '/' shift, and go to state 7 |
| 7505 |
| 7506 '/' [reduce using rule 1 (exp)] |
| 7507 $default reduce using rule 1 (exp) |
| 7508 @end example |
| 7509 |
| 7510 The remaining states are similar: |
| 7511 |
| 7512 @example |
| 7513 state 9 |
| 7514 |
| 7515 exp -> exp . '+' exp (rule 1) |
| 7516 exp -> exp . '-' exp (rule 2) |
| 7517 exp -> exp '-' exp . (rule 2) |
| 7518 exp -> exp . '*' exp (rule 3) |
| 7519 exp -> exp . '/' exp (rule 4) |
| 7520 |
| 7521 '*' shift, and go to state 6 |
| 7522 '/' shift, and go to state 7 |
| 7523 |
| 7524 '/' [reduce using rule 2 (exp)] |
| 7525 $default reduce using rule 2 (exp) |
| 7526 |
| 7527 state 10 |
| 7528 |
| 7529 exp -> exp . '+' exp (rule 1) |
| 7530 exp -> exp . '-' exp (rule 2) |
| 7531 exp -> exp . '*' exp (rule 3) |
| 7532 exp -> exp '*' exp . (rule 3) |
| 7533 exp -> exp . '/' exp (rule 4) |
| 7534 |
| 7535 '/' shift, and go to state 7 |
| 7536 |
| 7537 '/' [reduce using rule 3 (exp)] |
| 7538 $default reduce using rule 3 (exp) |
| 7539 |
| 7540 state 11 |
| 7541 |
| 7542 exp -> exp . '+' exp (rule 1) |
| 7543 exp -> exp . '-' exp (rule 2) |
| 7544 exp -> exp . '*' exp (rule 3) |
| 7545 exp -> exp . '/' exp (rule 4) |
| 7546 exp -> exp '/' exp . (rule 4) |
| 7547 |
| 7548 '+' shift, and go to state 4 |
| 7549 '-' shift, and go to state 5 |
| 7550 '*' shift, and go to state 6 |
| 7551 '/' shift, and go to state 7 |
| 7552 |
| 7553 '+' [reduce using rule 4 (exp)] |
| 7554 '-' [reduce using rule 4 (exp)] |
| 7555 '*' [reduce using rule 4 (exp)] |
| 7556 '/' [reduce using rule 4 (exp)] |
| 7557 $default reduce using rule 4 (exp) |
| 7558 @end example |
| 7559 |
| 7560 @noindent |
| 7561 Observe that state 11 contains conflicts not only due to the lack of |
| 7562 precedence of @samp{/} with respect to @samp{+}, @samp{-}, and |
| 7563 @samp{*}, but also because the |
| 7564 associativity of @samp{/} is not specified. |
| 7565 |
| 7566 |
| 7567 @node Tracing |
| 7568 @section Tracing Your Parser |
| 7569 @findex yydebug |
| 7570 @cindex debugging |
| 7571 @cindex tracing the parser |
| 7572 |
| 7573 If a Bison grammar compiles properly but doesn't do what you want when it |
| 7574 runs, the @code{yydebug} parser-trace feature can help you figure out why. |
| 7575 |
| 7576 There are several means to enable compilation of trace facilities: |
| 7577 |
| 7578 @table @asis |
| 7579 @item the macro @code{YYDEBUG} |
| 7580 @findex YYDEBUG |
| 7581 Define the macro @code{YYDEBUG} to a nonzero value when you compile the |
| 7582 parser. This is compliant with @acronym{POSIX} Yacc. You could use |
| 7583 @samp{-DYYDEBUG=1} as a compiler option or you could put @samp{#define |
| 7584 YYDEBUG 1} in the prologue of the grammar file (@pxref{Prologue, , The |
| 7585 Prologue}). |
| 7586 |
| 7587 @item the option @option{-t}, @option{--debug} |
| 7588 Use the @samp{-t} option when you run Bison (@pxref{Invocation, |
| 7589 ,Invoking Bison}). This is @acronym{POSIX} compliant too. |
| 7590 |
| 7591 @item the directive @samp{%debug} |
| 7592 @findex %debug |
| 7593 Add the @code{%debug} directive (@pxref{Decl Summary, ,Bison |
| 7594 Declaration Summary}). This is a Bison extension, which will prove |
| 7595 useful when Bison will output parsers for languages that don't use a |
| 7596 preprocessor. Unless @acronym{POSIX} and Yacc portability matter to |
| 7597 you, this is |
| 7598 the preferred solution. |
| 7599 @end table |
| 7600 |
| 7601 We suggest that you always enable the debug option so that debugging is |
| 7602 always possible. |
| 7603 |
| 7604 The trace facility outputs messages with macro calls of the form |
| 7605 @code{YYFPRINTF (stderr, @var{format}, @var{args})} where |
| 7606 @var{format} and @var{args} are the usual @code{printf} format and variadic |
| 7607 arguments. If you define @code{YYDEBUG} to a nonzero value but do not |
| 7608 define @code{YYFPRINTF}, @code{<stdio.h>} is automatically included |
| 7609 and @code{YYFPRINTF} is defined to @code{fprintf}. |
| 7610 |
| 7611 Once you have compiled the program with trace facilities, the way to |
| 7612 request a trace is to store a nonzero value in the variable @code{yydebug}. |
| 7613 You can do this by making the C code do it (in @code{main}, perhaps), or |
| 7614 you can alter the value with a C debugger. |
| 7615 |
| 7616 Each step taken by the parser when @code{yydebug} is nonzero produces a |
| 7617 line or two of trace information, written on @code{stderr}. The trace |
| 7618 messages tell you these things: |
| 7619 |
| 7620 @itemize @bullet |
| 7621 @item |
| 7622 Each time the parser calls @code{yylex}, what kind of token was read. |
| 7623 |
| 7624 @item |
| 7625 Each time a token is shifted, the depth and complete contents of the |
| 7626 state stack (@pxref{Parser States}). |
| 7627 |
| 7628 @item |
| 7629 Each time a rule is reduced, which rule it is, and the complete contents |
| 7630 of the state stack afterward. |
| 7631 @end itemize |
| 7632 |
| 7633 To make sense of this information, it helps to refer to the listing file |
| 7634 produced by the Bison @samp{-v} option (@pxref{Invocation, ,Invoking |
| 7635 Bison}). This file shows the meaning of each state in terms of |
| 7636 positions in various rules, and also what each state will do with each |
| 7637 possible input token. As you read the successive trace messages, you |
| 7638 can see that the parser is functioning according to its specification in |
| 7639 the listing file. Eventually you will arrive at the place where |
| 7640 something undesirable happens, and you will see which parts of the |
| 7641 grammar are to blame. |
| 7642 |
| 7643 The parser file is a C program and you can use C debuggers on it, but it's |
| 7644 not easy to interpret what it is doing. The parser function is a |
| 7645 finite-state machine interpreter, and aside from the actions it executes |
| 7646 the same code over and over. Only the values of variables show where in |
| 7647 the grammar it is working. |
| 7648 |
| 7649 @findex YYPRINT |
| 7650 The debugging information normally gives the token type of each token |
| 7651 read, but not its semantic value. You can optionally define a macro |
| 7652 named @code{YYPRINT} to provide a way to print the value. If you define |
| 7653 @code{YYPRINT}, it should take three arguments. The parser will pass a |
| 7654 standard I/O stream, the numeric code for the token type, and the token |
| 7655 value (from @code{yylval}). |
| 7656 |
| 7657 Here is an example of @code{YYPRINT} suitable for the multi-function |
| 7658 calculator (@pxref{Mfcalc Declarations, ,Declarations for @code{mfcalc}}): |
| 7659 |
| 7660 @smallexample |
| 7661 %@{ |
| 7662 static void print_token_value (FILE *, int, YYSTYPE); |
| 7663 #define YYPRINT(file, type, value) print_token_value (file, type, value) |
| 7664 %@} |
| 7665 |
| 7666 @dots{} %% @dots{} %% @dots{} |
| 7667 |
| 7668 static void |
| 7669 print_token_value (FILE *file, int type, YYSTYPE value) |
| 7670 @{ |
| 7671 if (type == VAR) |
| 7672 fprintf (file, "%s", value.tptr->name); |
| 7673 else if (type == NUM) |
| 7674 fprintf (file, "%d", value.val); |
| 7675 @} |
| 7676 @end smallexample |
| 7677 |
| 7678 @c ================================================= Invoking Bison |
| 7679 |
| 7680 @node Invocation |
| 7681 @chapter Invoking Bison |
| 7682 @cindex invoking Bison |
| 7683 @cindex Bison invocation |
| 7684 @cindex options for invoking Bison |
| 7685 |
| 7686 The usual way to invoke Bison is as follows: |
| 7687 |
| 7688 @example |
| 7689 bison @var{infile} |
| 7690 @end example |
| 7691 |
| 7692 Here @var{infile} is the grammar file name, which usually ends in |
| 7693 @samp{.y}. The parser file's name is made by replacing the @samp{.y} |
| 7694 with @samp{.tab.c} and removing any leading directory. Thus, the |
| 7695 @samp{bison foo.y} file name yields |
| 7696 @file{foo.tab.c}, and the @samp{bison hack/foo.y} file name yields |
| 7697 @file{foo.tab.c}. It's also possible, in case you are writing |
| 7698 C++ code instead of C in your grammar file, to name it @file{foo.ypp} |
| 7699 or @file{foo.y++}. Then, the output files will take an extension like |
| 7700 the given one as input (respectively @file{foo.tab.cpp} and |
| 7701 @file{foo.tab.c++}). |
| 7702 This feature takes effect with all options that manipulate file names like |
| 7703 @samp{-o} or @samp{-d}. |
| 7704 |
| 7705 For example : |
| 7706 |
| 7707 @example |
| 7708 bison -d @var{infile.yxx} |
| 7709 @end example |
| 7710 @noindent |
| 7711 will produce @file{infile.tab.cxx} and @file{infile.tab.hxx}, and |
| 7712 |
| 7713 @example |
| 7714 bison -d -o @var{output.c++} @var{infile.y} |
| 7715 @end example |
| 7716 @noindent |
| 7717 will produce @file{output.c++} and @file{outfile.h++}. |
| 7718 |
| 7719 For compatibility with @acronym{POSIX}, the standard Bison |
| 7720 distribution also contains a shell script called @command{yacc} that |
| 7721 invokes Bison with the @option{-y} option. |
| 7722 |
| 7723 @menu |
| 7724 * Bison Options:: All the options described in detail, |
| 7725 in alphabetical order by short options. |
| 7726 * Option Cross Key:: Alphabetical list of long options. |
| 7727 * Yacc Library:: Yacc-compatible @code{yylex} and @code{main}. |
| 7728 @end menu |
| 7729 |
| 7730 @node Bison Options |
| 7731 @section Bison Options |
| 7732 |
| 7733 Bison supports both traditional single-letter options and mnemonic long |
| 7734 option names. Long option names are indicated with @samp{--} instead of |
| 7735 @samp{-}. Abbreviations for option names are allowed as long as they |
| 7736 are unique. When a long option takes an argument, like |
| 7737 @samp{--file-prefix}, connect the option name and the argument with |
| 7738 @samp{=}. |
| 7739 |
| 7740 Here is a list of options that can be used with Bison, alphabetized by |
| 7741 short option. It is followed by a cross key alphabetized by long |
| 7742 option. |
| 7743 |
| 7744 @c Please, keep this ordered as in `bison --help'. |
| 7745 @noindent |
| 7746 Operations modes: |
| 7747 @table @option |
| 7748 @item -h |
| 7749 @itemx --help |
| 7750 Print a summary of the command-line options to Bison and exit. |
| 7751 |
| 7752 @item -V |
| 7753 @itemx --version |
| 7754 Print the version number of Bison and exit. |
| 7755 |
| 7756 @item --print-localedir |
| 7757 Print the name of the directory containing locale-dependent data. |
| 7758 |
| 7759 @item --print-datadir |
| 7760 Print the name of the directory containing skeletons and XSLT. |
| 7761 |
| 7762 @item -y |
| 7763 @itemx --yacc |
| 7764 Act more like the traditional Yacc command. This can cause |
| 7765 different diagnostics to be generated, and may change behavior in |
| 7766 other minor ways. Most importantly, imitate Yacc's output |
| 7767 file name conventions, so that the parser output file is called |
| 7768 @file{y.tab.c}, and the other outputs are called @file{y.output} and |
| 7769 @file{y.tab.h}. |
| 7770 Also, if generating an @acronym{LALR}(1) parser in C, generate @code{#define} |
| 7771 statements in addition to an @code{enum} to associate token numbers with token |
| 7772 names. |
| 7773 Thus, the following shell script can substitute for Yacc, and the Bison |
| 7774 distribution contains such a script for compatibility with @acronym{POSIX}: |
| 7775 |
| 7776 @example |
| 7777 #! /bin/sh |
| 7778 bison -y "$@@" |
| 7779 @end example |
| 7780 |
| 7781 The @option{-y}/@option{--yacc} option is intended for use with |
| 7782 traditional Yacc grammars. If your grammar uses a Bison extension |
| 7783 like @samp{%glr-parser}, Bison might not be Yacc-compatible even if |
| 7784 this option is specified. |
| 7785 |
| 7786 @item -W |
| 7787 @itemx --warnings |
| 7788 Output warnings falling in @var{category}. @var{category} can be one |
| 7789 of: |
| 7790 @table @code |
| 7791 @item midrule-values |
| 7792 Warn about mid-rule values that are set but not used within any of the actions |
| 7793 of the parent rule. |
| 7794 For example, warn about unused @code{$2} in: |
| 7795 |
| 7796 @example |
| 7797 exp: '1' @{ $$ = 1; @} '+' exp @{ $$ = $1 + $4; @}; |
| 7798 @end example |
| 7799 |
| 7800 Also warn about mid-rule values that are used but not set. |
| 7801 For example, warn about unset @code{$$} in the mid-rule action in: |
| 7802 |
| 7803 @example |
| 7804 exp: '1' @{ $1 = 1; @} '+' exp @{ $$ = $2 + $4; @}; |
| 7805 @end example |
| 7806 |
| 7807 These warnings are not enabled by default since they sometimes prove to |
| 7808 be false alarms in existing grammars employing the Yacc constructs |
| 7809 @code{$0} or @code{$-@var{n}} (where @var{n} is some positive integer). |
| 7810 |
| 7811 |
| 7812 @item yacc |
| 7813 Incompatibilities with @acronym{POSIX} Yacc. |
| 7814 |
| 7815 @item all |
| 7816 All the warnings. |
| 7817 @item none |
| 7818 Turn off all the warnings. |
| 7819 @item error |
| 7820 Treat warnings as errors. |
| 7821 @end table |
| 7822 |
| 7823 A category can be turned off by prefixing its name with @samp{no-}. For |
| 7824 instance, @option{-Wno-syntax} will hide the warnings about unused |
| 7825 variables. |
| 7826 @end table |
| 7827 |
| 7828 @noindent |
| 7829 Tuning the parser: |
| 7830 |
| 7831 @table @option |
| 7832 @item -t |
| 7833 @itemx --debug |
| 7834 In the parser file, define the macro @code{YYDEBUG} to 1 if it is not |
| 7835 already defined, so that the debugging facilities are compiled. |
| 7836 @xref{Tracing, ,Tracing Your Parser}. |
| 7837 |
| 7838 @item -L @var{language} |
| 7839 @itemx --language=@var{language} |
| 7840 Specify the programming language for the generated parser, as if |
| 7841 @code{%language} was specified (@pxref{Decl Summary, , Bison Declaration |
| 7842 Summary}). Currently supported languages include C, C++, and Java. |
| 7843 @var{language} is case-insensitive. |
| 7844 |
| 7845 This option is experimental and its effect may be modified in future |
| 7846 releases. |
| 7847 |
| 7848 @item --locations |
| 7849 Pretend that @code{%locations} was specified. @xref{Decl Summary}. |
| 7850 |
| 7851 @item -p @var{prefix} |
| 7852 @itemx --name-prefix=@var{prefix} |
| 7853 Pretend that @code{%name-prefix "@var{prefix}"} was specified. |
| 7854 @xref{Decl Summary}. |
| 7855 |
| 7856 @item -l |
| 7857 @itemx --no-lines |
| 7858 Don't put any @code{#line} preprocessor commands in the parser file. |
| 7859 Ordinarily Bison puts them in the parser file so that the C compiler |
| 7860 and debuggers will associate errors with your source file, the |
| 7861 grammar file. This option causes them to associate errors with the |
| 7862 parser file, treating it as an independent source file in its own right. |
| 7863 |
| 7864 @item -S @var{file} |
| 7865 @itemx --skeleton=@var{file} |
| 7866 Specify the skeleton to use, similar to @code{%skeleton} |
| 7867 (@pxref{Decl Summary, , Bison Declaration Summary}). |
| 7868 |
| 7869 @c You probably don't need this option unless you are developing Bison. |
| 7870 @c You should use @option{--language} if you want to specify the skeleton for a |
| 7871 @c different language, because it is clearer and because it will always |
| 7872 @c choose the correct skeleton for non-deterministic or push parsers. |
| 7873 |
| 7874 If @var{file} does not contain a @code{/}, @var{file} is the name of a skeleton |
| 7875 file in the Bison installation directory. |
| 7876 If it does, @var{file} is an absolute file name or a file name relative to the |
| 7877 current working directory. |
| 7878 This is similar to how most shells resolve commands. |
| 7879 |
| 7880 @item -k |
| 7881 @itemx --token-table |
| 7882 Pretend that @code{%token-table} was specified. @xref{Decl Summary}. |
| 7883 @end table |
| 7884 |
| 7885 @noindent |
| 7886 Adjust the output: |
| 7887 |
| 7888 @table @option |
| 7889 @item --defines[=@var{file}] |
| 7890 Pretend that @code{%defines} was specified, i.e., write an extra output |
| 7891 file containing macro definitions for the token type names defined in |
| 7892 the grammar, as well as a few other declarations. @xref{Decl Summary}. |
| 7893 |
| 7894 @item -d |
| 7895 This is the same as @code{--defines} except @code{-d} does not accept a |
| 7896 @var{file} argument since POSIX Yacc requires that @code{-d} can be bundled |
| 7897 with other short options. |
| 7898 |
| 7899 @item -b @var{file-prefix} |
| 7900 @itemx --file-prefix=@var{prefix} |
| 7901 Pretend that @code{%file-prefix} was specified, i.e., specify prefix to use |
| 7902 for all Bison output file names. @xref{Decl Summary}. |
| 7903 |
| 7904 @item -r @var{things} |
| 7905 @itemx --report=@var{things} |
| 7906 Write an extra output file containing verbose description of the comma |
| 7907 separated list of @var{things} among: |
| 7908 |
| 7909 @table @code |
| 7910 @item state |
| 7911 Description of the grammar, conflicts (resolved and unresolved), and |
| 7912 @acronym{LALR} automaton. |
| 7913 |
| 7914 @item lookahead |
| 7915 Implies @code{state} and augments the description of the automaton with |
| 7916 each rule's lookahead set. |
| 7917 |
| 7918 @item itemset |
| 7919 Implies @code{state} and augments the description of the automaton with |
| 7920 the full set of items for each state, instead of its core only. |
| 7921 @end table |
| 7922 |
| 7923 @item --report-file=@var{file} |
| 7924 Specify the @var{file} for the verbose description. |
| 7925 |
| 7926 @item -v |
| 7927 @itemx --verbose |
| 7928 Pretend that @code{%verbose} was specified, i.e., write an extra output |
| 7929 file containing verbose descriptions of the grammar and |
| 7930 parser. @xref{Decl Summary}. |
| 7931 |
| 7932 @item -o @var{file} |
| 7933 @itemx --output=@var{file} |
| 7934 Specify the @var{file} for the parser file. |
| 7935 |
| 7936 The other output files' names are constructed from @var{file} as |
| 7937 described under the @samp{-v} and @samp{-d} options. |
| 7938 |
| 7939 @item -g[@var{file}] |
| 7940 @itemx --graph[=@var{file}] |
| 7941 Output a graphical representation of the @acronym{LALR}(1) grammar |
| 7942 automaton computed by Bison, in @uref{http://www.graphviz.org/, Graphviz} |
| 7943 @uref{http://www.graphviz.org/doc/info/lang.html, @acronym{DOT}} format. |
| 7944 @code{@var{file}} is optional. |
| 7945 If omitted and the grammar file is @file{foo.y}, the output file will be |
| 7946 @file{foo.dot}. |
| 7947 |
| 7948 @item -x[@var{file}] |
| 7949 @itemx --xml[=@var{file}] |
| 7950 Output an XML report of the @acronym{LALR}(1) automaton computed by Bison. |
| 7951 @code{@var{file}} is optional. |
| 7952 If omitted and the grammar file is @file{foo.y}, the output file will be |
| 7953 @file{foo.xml}. |
| 7954 (The current XML schema is experimental and may evolve. |
| 7955 More user feedback will help to stabilize it.) |
| 7956 @end table |
| 7957 |
| 7958 @node Option Cross Key |
| 7959 @section Option Cross Key |
| 7960 |
| 7961 @c FIXME: How about putting the directives too? |
| 7962 Here is a list of options, alphabetized by long option, to help you find |
| 7963 the corresponding short option. |
| 7964 |
| 7965 @multitable {@option{--defines=@var{defines-file}}} {@option{-b @var{file-prefix
}XXX}} |
| 7966 @headitem Long Option @tab Short Option |
| 7967 @include cross-options.texi |
| 7968 @end multitable |
| 7969 |
| 7970 @node Yacc Library |
| 7971 @section Yacc Library |
| 7972 |
| 7973 The Yacc library contains default implementations of the |
| 7974 @code{yyerror} and @code{main} functions. These default |
| 7975 implementations are normally not useful, but @acronym{POSIX} requires |
| 7976 them. To use the Yacc library, link your program with the |
| 7977 @option{-ly} option. Note that Bison's implementation of the Yacc |
| 7978 library is distributed under the terms of the @acronym{GNU} General |
| 7979 Public License (@pxref{Copying}). |
| 7980 |
| 7981 If you use the Yacc library's @code{yyerror} function, you should |
| 7982 declare @code{yyerror} as follows: |
| 7983 |
| 7984 @example |
| 7985 int yyerror (char const *); |
| 7986 @end example |
| 7987 |
| 7988 Bison ignores the @code{int} value returned by this @code{yyerror}. |
| 7989 If you use the Yacc library's @code{main} function, your |
| 7990 @code{yyparse} function should have the following type signature: |
| 7991 |
| 7992 @example |
| 7993 int yyparse (void); |
| 7994 @end example |
| 7995 |
| 7996 @c ================================================= C++ Bison |
| 7997 |
| 7998 @node Other Languages |
| 7999 @chapter Parsers Written In Other Languages |
| 8000 |
| 8001 @menu |
| 8002 * C++ Parsers:: The interface to generate C++ parser classes |
| 8003 * Java Parsers:: The interface to generate Java parser classes |
| 8004 @end menu |
| 8005 |
| 8006 @node C++ Parsers |
| 8007 @section C++ Parsers |
| 8008 |
| 8009 @menu |
| 8010 * C++ Bison Interface:: Asking for C++ parser generation |
| 8011 * C++ Semantic Values:: %union vs. C++ |
| 8012 * C++ Location Values:: The position and location classes |
| 8013 * C++ Parser Interface:: Instantiating and running the parser |
| 8014 * C++ Scanner Interface:: Exchanges between yylex and parse |
| 8015 * A Complete C++ Example:: Demonstrating their use |
| 8016 @end menu |
| 8017 |
| 8018 @node C++ Bison Interface |
| 8019 @subsection C++ Bison Interface |
| 8020 @c - %skeleton "lalr1.cc" |
| 8021 @c - Always pure |
| 8022 @c - initial action |
| 8023 |
| 8024 The C++ @acronym{LALR}(1) parser is selected using the skeleton directive, |
| 8025 @samp{%skeleton "lalr1.c"}, or the synonymous command-line option |
| 8026 @option{--skeleton=lalr1.c}. |
| 8027 @xref{Decl Summary}. |
| 8028 |
| 8029 When run, @command{bison} will create several entities in the @samp{yy} |
| 8030 namespace. |
| 8031 @findex %define namespace |
| 8032 Use the @samp{%define namespace} directive to change the namespace name, see |
| 8033 @ref{Decl Summary}. |
| 8034 The various classes are generated in the following files: |
| 8035 |
| 8036 @table @file |
| 8037 @item position.hh |
| 8038 @itemx location.hh |
| 8039 The definition of the classes @code{position} and @code{location}, |
| 8040 used for location tracking. @xref{C++ Location Values}. |
| 8041 |
| 8042 @item stack.hh |
| 8043 An auxiliary class @code{stack} used by the parser. |
| 8044 |
| 8045 @item @var{file}.hh |
| 8046 @itemx @var{file}.cc |
| 8047 (Assuming the extension of the input file was @samp{.yy}.) The |
| 8048 declaration and implementation of the C++ parser class. The basename |
| 8049 and extension of these two files follow the same rules as with regular C |
| 8050 parsers (@pxref{Invocation}). |
| 8051 |
| 8052 The header is @emph{mandatory}; you must either pass |
| 8053 @option{-d}/@option{--defines} to @command{bison}, or use the |
| 8054 @samp{%defines} directive. |
| 8055 @end table |
| 8056 |
| 8057 All these files are documented using Doxygen; run @command{doxygen} |
| 8058 for a complete and accurate documentation. |
| 8059 |
| 8060 @node C++ Semantic Values |
| 8061 @subsection C++ Semantic Values |
| 8062 @c - No objects in unions |
| 8063 @c - YYSTYPE |
| 8064 @c - Printer and destructor |
| 8065 |
| 8066 The @code{%union} directive works as for C, see @ref{Union Decl, ,The |
| 8067 Collection of Value Types}. In particular it produces a genuine |
| 8068 @code{union}@footnote{In the future techniques to allow complex types |
| 8069 within pseudo-unions (similar to Boost variants) might be implemented to |
| 8070 alleviate these issues.}, which have a few specific features in C++. |
| 8071 @itemize @minus |
| 8072 @item |
| 8073 The type @code{YYSTYPE} is defined but its use is discouraged: rather |
| 8074 you should refer to the parser's encapsulated type |
| 8075 @code{yy::parser::semantic_type}. |
| 8076 @item |
| 8077 Non POD (Plain Old Data) types cannot be used. C++ forbids any |
| 8078 instance of classes with constructors in unions: only @emph{pointers} |
| 8079 to such objects are allowed. |
| 8080 @end itemize |
| 8081 |
| 8082 Because objects have to be stored via pointers, memory is not |
| 8083 reclaimed automatically: using the @code{%destructor} directive is the |
| 8084 only means to avoid leaks. @xref{Destructor Decl, , Freeing Discarded |
| 8085 Symbols}. |
| 8086 |
| 8087 |
| 8088 @node C++ Location Values |
| 8089 @subsection C++ Location Values |
| 8090 @c - %locations |
| 8091 @c - class Position |
| 8092 @c - class Location |
| 8093 @c - %define filename_type "const symbol::Symbol" |
| 8094 |
| 8095 When the directive @code{%locations} is used, the C++ parser supports |
| 8096 location tracking, see @ref{Locations, , Locations Overview}. Two |
| 8097 auxiliary classes define a @code{position}, a single point in a file, |
| 8098 and a @code{location}, a range composed of a pair of |
| 8099 @code{position}s (possibly spanning several files). |
| 8100 |
| 8101 @deftypemethod {position} {std::string*} file |
| 8102 The name of the file. It will always be handled as a pointer, the |
| 8103 parser will never duplicate nor deallocate it. As an experimental |
| 8104 feature you may change it to @samp{@var{type}*} using @samp{%define |
| 8105 filename_type "@var{type}"}. |
| 8106 @end deftypemethod |
| 8107 |
| 8108 @deftypemethod {position} {unsigned int} line |
| 8109 The line, starting at 1. |
| 8110 @end deftypemethod |
| 8111 |
| 8112 @deftypemethod {position} {unsigned int} lines (int @var{height} = 1) |
| 8113 Advance by @var{height} lines, resetting the column number. |
| 8114 @end deftypemethod |
| 8115 |
| 8116 @deftypemethod {position} {unsigned int} column |
| 8117 The column, starting at 0. |
| 8118 @end deftypemethod |
| 8119 |
| 8120 @deftypemethod {position} {unsigned int} columns (int @var{width} = 1) |
| 8121 Advance by @var{width} columns, without changing the line number. |
| 8122 @end deftypemethod |
| 8123 |
| 8124 @deftypemethod {position} {position&} operator+= (position& @var{pos}, int @var{
width}) |
| 8125 @deftypemethodx {position} {position} operator+ (const position& @var{pos}, int
@var{width}) |
| 8126 @deftypemethodx {position} {position&} operator-= (const position& @var{pos}, in
t @var{width}) |
| 8127 @deftypemethodx {position} {position} operator- (position& @var{pos}, int @var{w
idth}) |
| 8128 Various forms of syntactic sugar for @code{columns}. |
| 8129 @end deftypemethod |
| 8130 |
| 8131 @deftypemethod {position} {position} operator<< (std::ostream @var{o}, const pos
ition& @var{p}) |
| 8132 Report @var{p} on @var{o} like this: |
| 8133 @samp{@var{file}:@var{line}.@var{column}}, or |
| 8134 @samp{@var{line}.@var{column}} if @var{file} is null. |
| 8135 @end deftypemethod |
| 8136 |
| 8137 @deftypemethod {location} {position} begin |
| 8138 @deftypemethodx {location} {position} end |
| 8139 The first, inclusive, position of the range, and the first beyond. |
| 8140 @end deftypemethod |
| 8141 |
| 8142 @deftypemethod {location} {unsigned int} columns (int @var{width} = 1) |
| 8143 @deftypemethodx {location} {unsigned int} lines (int @var{height} = 1) |
| 8144 Advance the @code{end} position. |
| 8145 @end deftypemethod |
| 8146 |
| 8147 @deftypemethod {location} {location} operator+ (const location& @var{begin}, con
st location& @var{end}) |
| 8148 @deftypemethodx {location} {location} operator+ (const location& @var{begin}, in
t @var{width}) |
| 8149 @deftypemethodx {location} {location} operator+= (const location& @var{loc}, int
@var{width}) |
| 8150 Various forms of syntactic sugar. |
| 8151 @end deftypemethod |
| 8152 |
| 8153 @deftypemethod {location} {void} step () |
| 8154 Move @code{begin} onto @code{end}. |
| 8155 @end deftypemethod |
| 8156 |
| 8157 |
| 8158 @node C++ Parser Interface |
| 8159 @subsection C++ Parser Interface |
| 8160 @c - define parser_class_name |
| 8161 @c - Ctor |
| 8162 @c - parse, error, set_debug_level, debug_level, set_debug_stream, |
| 8163 @c debug_stream. |
| 8164 @c - Reporting errors |
| 8165 |
| 8166 The output files @file{@var{output}.hh} and @file{@var{output}.cc} |
| 8167 declare and define the parser class in the namespace @code{yy}. The |
| 8168 class name defaults to @code{parser}, but may be changed using |
| 8169 @samp{%define parser_class_name "@var{name}"}. The interface of |
| 8170 this class is detailed below. It can be extended using the |
| 8171 @code{%parse-param} feature: its semantics is slightly changed since |
| 8172 it describes an additional member of the parser class, and an |
| 8173 additional argument for its constructor. |
| 8174 |
| 8175 @defcv {Type} {parser} {semantic_value_type} |
| 8176 @defcvx {Type} {parser} {location_value_type} |
| 8177 The types for semantics value and locations. |
| 8178 @end defcv |
| 8179 |
| 8180 @deftypemethod {parser} {} parser (@var{type1} @var{arg1}, ...) |
| 8181 Build a new parser object. There are no arguments by default, unless |
| 8182 @samp{%parse-param @{@var{type1} @var{arg1}@}} was used. |
| 8183 @end deftypemethod |
| 8184 |
| 8185 @deftypemethod {parser} {int} parse () |
| 8186 Run the syntactic analysis, and return 0 on success, 1 otherwise. |
| 8187 @end deftypemethod |
| 8188 |
| 8189 @deftypemethod {parser} {std::ostream&} debug_stream () |
| 8190 @deftypemethodx {parser} {void} set_debug_stream (std::ostream& @var{o}) |
| 8191 Get or set the stream used for tracing the parsing. It defaults to |
| 8192 @code{std::cerr}. |
| 8193 @end deftypemethod |
| 8194 |
| 8195 @deftypemethod {parser} {debug_level_type} debug_level () |
| 8196 @deftypemethodx {parser} {void} set_debug_level (debug_level @var{l}) |
| 8197 Get or set the tracing level. Currently its value is either 0, no trace, |
| 8198 or nonzero, full tracing. |
| 8199 @end deftypemethod |
| 8200 |
| 8201 @deftypemethod {parser} {void} error (const location_type& @var{l}, const std::s
tring& @var{m}) |
| 8202 The definition for this member function must be supplied by the user: |
| 8203 the parser uses it to report a parser error occurring at @var{l}, |
| 8204 described by @var{m}. |
| 8205 @end deftypemethod |
| 8206 |
| 8207 |
| 8208 @node C++ Scanner Interface |
| 8209 @subsection C++ Scanner Interface |
| 8210 @c - prefix for yylex. |
| 8211 @c - Pure interface to yylex |
| 8212 @c - %lex-param |
| 8213 |
| 8214 The parser invokes the scanner by calling @code{yylex}. Contrary to C |
| 8215 parsers, C++ parsers are always pure: there is no point in using the |
| 8216 @code{%define api.pure} directive. Therefore the interface is as follows. |
| 8217 |
| 8218 @deftypemethod {parser} {int} yylex (semantic_value_type& @var{yylval}, location
_type& @var{yylloc}, @var{type1} @var{arg1}, ...) |
| 8219 Return the next token. Its type is the return value, its semantic |
| 8220 value and location being @var{yylval} and @var{yylloc}. Invocations of |
| 8221 @samp{%lex-param @{@var{type1} @var{arg1}@}} yield additional arguments. |
| 8222 @end deftypemethod |
| 8223 |
| 8224 |
| 8225 @node A Complete C++ Example |
| 8226 @subsection A Complete C++ Example |
| 8227 |
| 8228 This section demonstrates the use of a C++ parser with a simple but |
| 8229 complete example. This example should be available on your system, |
| 8230 ready to compile, in the directory @dfn{../bison/examples/calc++}. It |
| 8231 focuses on the use of Bison, therefore the design of the various C++ |
| 8232 classes is very naive: no accessors, no encapsulation of members etc. |
| 8233 We will use a Lex scanner, and more precisely, a Flex scanner, to |
| 8234 demonstrate the various interaction. A hand written scanner is |
| 8235 actually easier to interface with. |
| 8236 |
| 8237 @menu |
| 8238 * Calc++ --- C++ Calculator:: The specifications |
| 8239 * Calc++ Parsing Driver:: An active parsing context |
| 8240 * Calc++ Parser:: A parser class |
| 8241 * Calc++ Scanner:: A pure C++ Flex scanner |
| 8242 * Calc++ Top Level:: Conducting the band |
| 8243 @end menu |
| 8244 |
| 8245 @node Calc++ --- C++ Calculator |
| 8246 @subsubsection Calc++ --- C++ Calculator |
| 8247 |
| 8248 Of course the grammar is dedicated to arithmetics, a single |
| 8249 expression, possibly preceded by variable assignments. An |
| 8250 environment containing possibly predefined variables such as |
| 8251 @code{one} and @code{two}, is exchanged with the parser. An example |
| 8252 of valid input follows. |
| 8253 |
| 8254 @example |
| 8255 three := 3 |
| 8256 seven := one + two * three |
| 8257 seven * seven |
| 8258 @end example |
| 8259 |
| 8260 @node Calc++ Parsing Driver |
| 8261 @subsubsection Calc++ Parsing Driver |
| 8262 @c - An env |
| 8263 @c - A place to store error messages |
| 8264 @c - A place for the result |
| 8265 |
| 8266 To support a pure interface with the parser (and the scanner) the |
| 8267 technique of the ``parsing context'' is convenient: a structure |
| 8268 containing all the data to exchange. Since, in addition to simply |
| 8269 launch the parsing, there are several auxiliary tasks to execute (open |
| 8270 the file for parsing, instantiate the parser etc.), we recommend |
| 8271 transforming the simple parsing context structure into a fully blown |
| 8272 @dfn{parsing driver} class. |
| 8273 |
| 8274 The declaration of this driver class, @file{calc++-driver.hh}, is as |
| 8275 follows. The first part includes the CPP guard and imports the |
| 8276 required standard library components, and the declaration of the parser |
| 8277 class. |
| 8278 |
| 8279 @comment file: calc++-driver.hh |
| 8280 @example |
| 8281 #ifndef CALCXX_DRIVER_HH |
| 8282 # define CALCXX_DRIVER_HH |
| 8283 # include <string> |
| 8284 # include <map> |
| 8285 # include "calc++-parser.hh" |
| 8286 @end example |
| 8287 |
| 8288 |
| 8289 @noindent |
| 8290 Then comes the declaration of the scanning function. Flex expects |
| 8291 the signature of @code{yylex} to be defined in the macro |
| 8292 @code{YY_DECL}, and the C++ parser expects it to be declared. We can |
| 8293 factor both as follows. |
| 8294 |
| 8295 @comment file: calc++-driver.hh |
| 8296 @example |
| 8297 // Tell Flex the lexer's prototype ... |
| 8298 # define YY_DECL \ |
| 8299 yy::calcxx_parser::token_type \ |
| 8300 yylex (yy::calcxx_parser::semantic_type* yylval, \ |
| 8301 yy::calcxx_parser::location_type* yylloc, \ |
| 8302 calcxx_driver& driver) |
| 8303 // ... and declare it for the parser's sake. |
| 8304 YY_DECL; |
| 8305 @end example |
| 8306 |
| 8307 @noindent |
| 8308 The @code{calcxx_driver} class is then declared with its most obvious |
| 8309 members. |
| 8310 |
| 8311 @comment file: calc++-driver.hh |
| 8312 @example |
| 8313 // Conducting the whole scanning and parsing of Calc++. |
| 8314 class calcxx_driver |
| 8315 @{ |
| 8316 public: |
| 8317 calcxx_driver (); |
| 8318 virtual ~calcxx_driver (); |
| 8319 |
| 8320 std::map<std::string, int> variables; |
| 8321 |
| 8322 int result; |
| 8323 @end example |
| 8324 |
| 8325 @noindent |
| 8326 To encapsulate the coordination with the Flex scanner, it is useful to |
| 8327 have two members function to open and close the scanning phase. |
| 8328 |
| 8329 @comment file: calc++-driver.hh |
| 8330 @example |
| 8331 // Handling the scanner. |
| 8332 void scan_begin (); |
| 8333 void scan_end (); |
| 8334 bool trace_scanning; |
| 8335 @end example |
| 8336 |
| 8337 @noindent |
| 8338 Similarly for the parser itself. |
| 8339 |
| 8340 @comment file: calc++-driver.hh |
| 8341 @example |
| 8342 // Run the parser. Return 0 on success. |
| 8343 int parse (const std::string& f); |
| 8344 std::string file; |
| 8345 bool trace_parsing; |
| 8346 @end example |
| 8347 |
| 8348 @noindent |
| 8349 To demonstrate pure handling of parse errors, instead of simply |
| 8350 dumping them on the standard error output, we will pass them to the |
| 8351 compiler driver using the following two member functions. Finally, we |
| 8352 close the class declaration and CPP guard. |
| 8353 |
| 8354 @comment file: calc++-driver.hh |
| 8355 @example |
| 8356 // Error handling. |
| 8357 void error (const yy::location& l, const std::string& m); |
| 8358 void error (const std::string& m); |
| 8359 @}; |
| 8360 #endif // ! CALCXX_DRIVER_HH |
| 8361 @end example |
| 8362 |
| 8363 The implementation of the driver is straightforward. The @code{parse} |
| 8364 member function deserves some attention. The @code{error} functions |
| 8365 are simple stubs, they should actually register the located error |
| 8366 messages and set error state. |
| 8367 |
| 8368 @comment file: calc++-driver.cc |
| 8369 @example |
| 8370 #include "calc++-driver.hh" |
| 8371 #include "calc++-parser.hh" |
| 8372 |
| 8373 calcxx_driver::calcxx_driver () |
| 8374 : trace_scanning (false), trace_parsing (false) |
| 8375 @{ |
| 8376 variables["one"] = 1; |
| 8377 variables["two"] = 2; |
| 8378 @} |
| 8379 |
| 8380 calcxx_driver::~calcxx_driver () |
| 8381 @{ |
| 8382 @} |
| 8383 |
| 8384 int |
| 8385 calcxx_driver::parse (const std::string &f) |
| 8386 @{ |
| 8387 file = f; |
| 8388 scan_begin (); |
| 8389 yy::calcxx_parser parser (*this); |
| 8390 parser.set_debug_level (trace_parsing); |
| 8391 int res = parser.parse (); |
| 8392 scan_end (); |
| 8393 return res; |
| 8394 @} |
| 8395 |
| 8396 void |
| 8397 calcxx_driver::error (const yy::location& l, const std::string& m) |
| 8398 @{ |
| 8399 std::cerr << l << ": " << m << std::endl; |
| 8400 @} |
| 8401 |
| 8402 void |
| 8403 calcxx_driver::error (const std::string& m) |
| 8404 @{ |
| 8405 std::cerr << m << std::endl; |
| 8406 @} |
| 8407 @end example |
| 8408 |
| 8409 @node Calc++ Parser |
| 8410 @subsubsection Calc++ Parser |
| 8411 |
| 8412 The parser definition file @file{calc++-parser.yy} starts by asking for |
| 8413 the C++ LALR(1) skeleton, the creation of the parser header file, and |
| 8414 specifies the name of the parser class. Because the C++ skeleton |
| 8415 changed several times, it is safer to require the version you designed |
| 8416 the grammar for. |
| 8417 |
| 8418 @comment file: calc++-parser.yy |
| 8419 @example |
| 8420 %skeleton "lalr1.cc" /* -*- C++ -*- */ |
| 8421 %require "@value{VERSION}" |
| 8422 %defines |
| 8423 %define parser_class_name "calcxx_parser" |
| 8424 @end example |
| 8425 |
| 8426 @noindent |
| 8427 @findex %code requires |
| 8428 Then come the declarations/inclusions needed to define the |
| 8429 @code{%union}. Because the parser uses the parsing driver and |
| 8430 reciprocally, both cannot include the header of the other. Because the |
| 8431 driver's header needs detailed knowledge about the parser class (in |
| 8432 particular its inner types), it is the parser's header which will simply |
| 8433 use a forward declaration of the driver. |
| 8434 @xref{Decl Summary, ,%code}. |
| 8435 |
| 8436 @comment file: calc++-parser.yy |
| 8437 @example |
| 8438 %code requires @{ |
| 8439 # include <string> |
| 8440 class calcxx_driver; |
| 8441 @} |
| 8442 @end example |
| 8443 |
| 8444 @noindent |
| 8445 The driver is passed by reference to the parser and to the scanner. |
| 8446 This provides a simple but effective pure interface, not relying on |
| 8447 global variables. |
| 8448 |
| 8449 @comment file: calc++-parser.yy |
| 8450 @example |
| 8451 // The parsing context. |
| 8452 %parse-param @{ calcxx_driver& driver @} |
| 8453 %lex-param @{ calcxx_driver& driver @} |
| 8454 @end example |
| 8455 |
| 8456 @noindent |
| 8457 Then we request the location tracking feature, and initialize the |
| 8458 first location's file name. Afterwards new locations are computed |
| 8459 relatively to the previous locations: the file name will be |
| 8460 automatically propagated. |
| 8461 |
| 8462 @comment file: calc++-parser.yy |
| 8463 @example |
| 8464 %locations |
| 8465 %initial-action |
| 8466 @{ |
| 8467 // Initialize the initial location. |
| 8468 @@$.begin.filename = @@$.end.filename = &driver.file; |
| 8469 @}; |
| 8470 @end example |
| 8471 |
| 8472 @noindent |
| 8473 Use the two following directives to enable parser tracing and verbose |
| 8474 error messages. |
| 8475 |
| 8476 @comment file: calc++-parser.yy |
| 8477 @example |
| 8478 %debug |
| 8479 %error-verbose |
| 8480 @end example |
| 8481 |
| 8482 @noindent |
| 8483 Semantic values cannot use ``real'' objects, but only pointers to |
| 8484 them. |
| 8485 |
| 8486 @comment file: calc++-parser.yy |
| 8487 @example |
| 8488 // Symbols. |
| 8489 %union |
| 8490 @{ |
| 8491 int ival; |
| 8492 std::string *sval; |
| 8493 @}; |
| 8494 @end example |
| 8495 |
| 8496 @noindent |
| 8497 @findex %code |
| 8498 The code between @samp{%code @{} and @samp{@}} is output in the |
| 8499 @file{*.cc} file; it needs detailed knowledge about the driver. |
| 8500 |
| 8501 @comment file: calc++-parser.yy |
| 8502 @example |
| 8503 %code @{ |
| 8504 # include "calc++-driver.hh" |
| 8505 @} |
| 8506 @end example |
| 8507 |
| 8508 |
| 8509 @noindent |
| 8510 The token numbered as 0 corresponds to end of file; the following line |
| 8511 allows for nicer error messages referring to ``end of file'' instead |
| 8512 of ``$end''. Similarly user friendly named are provided for each |
| 8513 symbol. Note that the tokens names are prefixed by @code{TOKEN_} to |
| 8514 avoid name clashes. |
| 8515 |
| 8516 @comment file: calc++-parser.yy |
| 8517 @example |
| 8518 %token END 0 "end of file" |
| 8519 %token ASSIGN ":=" |
| 8520 %token <sval> IDENTIFIER "identifier" |
| 8521 %token <ival> NUMBER "number" |
| 8522 %type <ival> exp |
| 8523 @end example |
| 8524 |
| 8525 @noindent |
| 8526 To enable memory deallocation during error recovery, use |
| 8527 @code{%destructor}. |
| 8528 |
| 8529 @c FIXME: Document %printer, and mention that it takes a braced-code operand. |
| 8530 @comment file: calc++-parser.yy |
| 8531 @example |
| 8532 %printer @{ debug_stream () << *$$; @} "identifier" |
| 8533 %destructor @{ delete $$; @} "identifier" |
| 8534 |
| 8535 %printer @{ debug_stream () << $$; @} <ival> |
| 8536 @end example |
| 8537 |
| 8538 @noindent |
| 8539 The grammar itself is straightforward. |
| 8540 |
| 8541 @comment file: calc++-parser.yy |
| 8542 @example |
| 8543 %% |
| 8544 %start unit; |
| 8545 unit: assignments exp @{ driver.result = $2; @}; |
| 8546 |
| 8547 assignments: assignments assignment @{@} |
| 8548 | /* Nothing. */ @{@}; |
| 8549 |
| 8550 assignment: |
| 8551 "identifier" ":=" exp |
| 8552 @{ driver.variables[*$1] = $3; delete $1; @}; |
| 8553 |
| 8554 %left '+' '-'; |
| 8555 %left '*' '/'; |
| 8556 exp: exp '+' exp @{ $$ = $1 + $3; @} |
| 8557 | exp '-' exp @{ $$ = $1 - $3; @} |
| 8558 | exp '*' exp @{ $$ = $1 * $3; @} |
| 8559 | exp '/' exp @{ $$ = $1 / $3; @} |
| 8560 | "identifier" @{ $$ = driver.variables[*$1]; delete $1; @} |
| 8561 | "number" @{ $$ = $1; @}; |
| 8562 %% |
| 8563 @end example |
| 8564 |
| 8565 @noindent |
| 8566 Finally the @code{error} member function registers the errors to the |
| 8567 driver. |
| 8568 |
| 8569 @comment file: calc++-parser.yy |
| 8570 @example |
| 8571 void |
| 8572 yy::calcxx_parser::error (const yy::calcxx_parser::location_type& l, |
| 8573 const std::string& m) |
| 8574 @{ |
| 8575 driver.error (l, m); |
| 8576 @} |
| 8577 @end example |
| 8578 |
| 8579 @node Calc++ Scanner |
| 8580 @subsubsection Calc++ Scanner |
| 8581 |
| 8582 The Flex scanner first includes the driver declaration, then the |
| 8583 parser's to get the set of defined tokens. |
| 8584 |
| 8585 @comment file: calc++-scanner.ll |
| 8586 @example |
| 8587 %@{ /* -*- C++ -*- */ |
| 8588 # include <cstdlib> |
| 8589 # include <errno.h> |
| 8590 # include <limits.h> |
| 8591 # include <string> |
| 8592 # include "calc++-driver.hh" |
| 8593 # include "calc++-parser.hh" |
| 8594 |
| 8595 /* Work around an incompatibility in flex (at least versions |
| 8596 2.5.31 through 2.5.33): it generates code that does |
| 8597 not conform to C89. See Debian bug 333231 |
| 8598 <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=333231>. */ |
| 8599 # undef yywrap |
| 8600 # define yywrap() 1 |
| 8601 |
| 8602 /* By default yylex returns int, we use token_type. |
| 8603 Unfortunately yyterminate by default returns 0, which is |
| 8604 not of token_type. */ |
| 8605 #define yyterminate() return token::END |
| 8606 %@} |
| 8607 @end example |
| 8608 |
| 8609 @noindent |
| 8610 Because there is no @code{#include}-like feature we don't need |
| 8611 @code{yywrap}, we don't need @code{unput} either, and we parse an |
| 8612 actual file, this is not an interactive session with the user. |
| 8613 Finally we enable the scanner tracing features. |
| 8614 |
| 8615 @comment file: calc++-scanner.ll |
| 8616 @example |
| 8617 %option noyywrap nounput batch debug |
| 8618 @end example |
| 8619 |
| 8620 @noindent |
| 8621 Abbreviations allow for more readable rules. |
| 8622 |
| 8623 @comment file: calc++-scanner.ll |
| 8624 @example |
| 8625 id [a-zA-Z][a-zA-Z_0-9]* |
| 8626 int [0-9]+ |
| 8627 blank [ \t] |
| 8628 @end example |
| 8629 |
| 8630 @noindent |
| 8631 The following paragraph suffices to track locations accurately. Each |
| 8632 time @code{yylex} is invoked, the begin position is moved onto the end |
| 8633 position. Then when a pattern is matched, the end position is |
| 8634 advanced of its width. In case it matched ends of lines, the end |
| 8635 cursor is adjusted, and each time blanks are matched, the begin cursor |
| 8636 is moved onto the end cursor to effectively ignore the blanks |
| 8637 preceding tokens. Comments would be treated equally. |
| 8638 |
| 8639 @comment file: calc++-scanner.ll |
| 8640 @example |
| 8641 %@{ |
| 8642 # define YY_USER_ACTION yylloc->columns (yyleng); |
| 8643 %@} |
| 8644 %% |
| 8645 %@{ |
| 8646 yylloc->step (); |
| 8647 %@} |
| 8648 @{blank@}+ yylloc->step (); |
| 8649 [\n]+ yylloc->lines (yyleng); yylloc->step (); |
| 8650 @end example |
| 8651 |
| 8652 @noindent |
| 8653 The rules are simple, just note the use of the driver to report errors. |
| 8654 It is convenient to use a typedef to shorten |
| 8655 @code{yy::calcxx_parser::token::identifier} into |
| 8656 @code{token::identifier} for instance. |
| 8657 |
| 8658 @comment file: calc++-scanner.ll |
| 8659 @example |
| 8660 %@{ |
| 8661 typedef yy::calcxx_parser::token token; |
| 8662 %@} |
| 8663 /* Convert ints to the actual type of tokens. */ |
| 8664 [-+*/] return yy::calcxx_parser::token_type (yytext[0]); |
| 8665 ":=" return token::ASSIGN; |
| 8666 @{int@} @{ |
| 8667 errno = 0; |
| 8668 long n = strtol (yytext, NULL, 10); |
| 8669 if (! (INT_MIN <= n && n <= INT_MAX && errno != ERANGE)) |
| 8670 driver.error (*yylloc, "integer is out of range"); |
| 8671 yylval->ival = n; |
| 8672 return token::NUMBER; |
| 8673 @} |
| 8674 @{id@} yylval->sval = new std::string (yytext); return token::IDENTIFIER; |
| 8675 . driver.error (*yylloc, "invalid character"); |
| 8676 %% |
| 8677 @end example |
| 8678 |
| 8679 @noindent |
| 8680 Finally, because the scanner related driver's member function depend |
| 8681 on the scanner's data, it is simpler to implement them in this file. |
| 8682 |
| 8683 @comment file: calc++-scanner.ll |
| 8684 @example |
| 8685 void |
| 8686 calcxx_driver::scan_begin () |
| 8687 @{ |
| 8688 yy_flex_debug = trace_scanning; |
| 8689 if (file == "-") |
| 8690 yyin = stdin; |
| 8691 else if (!(yyin = fopen (file.c_str (), "r"))) |
| 8692 @{ |
| 8693 error (std::string ("cannot open ") + file); |
| 8694 exit (1); |
| 8695 @} |
| 8696 @} |
| 8697 |
| 8698 void |
| 8699 calcxx_driver::scan_end () |
| 8700 @{ |
| 8701 fclose (yyin); |
| 8702 @} |
| 8703 @end example |
| 8704 |
| 8705 @node Calc++ Top Level |
| 8706 @subsubsection Calc++ Top Level |
| 8707 |
| 8708 The top level file, @file{calc++.cc}, poses no problem. |
| 8709 |
| 8710 @comment file: calc++.cc |
| 8711 @example |
| 8712 #include <iostream> |
| 8713 #include "calc++-driver.hh" |
| 8714 |
| 8715 int |
| 8716 main (int argc, char *argv[]) |
| 8717 @{ |
| 8718 calcxx_driver driver; |
| 8719 for (++argv; argv[0]; ++argv) |
| 8720 if (*argv == std::string ("-p")) |
| 8721 driver.trace_parsing = true; |
| 8722 else if (*argv == std::string ("-s")) |
| 8723 driver.trace_scanning = true; |
| 8724 else if (!driver.parse (*argv)) |
| 8725 std::cout << driver.result << std::endl; |
| 8726 @} |
| 8727 @end example |
| 8728 |
| 8729 @node Java Parsers |
| 8730 @section Java Parsers |
| 8731 |
| 8732 @menu |
| 8733 * Java Bison Interface:: Asking for Java parser generation |
| 8734 * Java Semantic Values:: %type and %token vs. Java |
| 8735 * Java Location Values:: The position and location classes |
| 8736 * Java Parser Interface:: Instantiating and running the parser |
| 8737 * Java Scanner Interface:: Specifying the scanner for the parser |
| 8738 * Java Action Features:: Special features for use in actions |
| 8739 * Java Differences:: Differences between C/C++ and Java Grammars |
| 8740 * Java Declarations Summary:: List of Bison declarations used with Java |
| 8741 @end menu |
| 8742 |
| 8743 @node Java Bison Interface |
| 8744 @subsection Java Bison Interface |
| 8745 @c - %language "Java" |
| 8746 |
| 8747 (The current Java interface is experimental and may evolve. |
| 8748 More user feedback will help to stabilize it.) |
| 8749 |
| 8750 The Java parser skeletons are selected using the @code{%language "Java"} |
| 8751 directive or the @option{-L java}/@option{--language=java} option. |
| 8752 |
| 8753 @c FIXME: Documented bug. |
| 8754 When generating a Java parser, @code{bison @var{basename}.y} will create |
| 8755 a single Java source file named @file{@var{basename}.java}. Using an |
| 8756 input file without a @file{.y} suffix is currently broken. The basename |
| 8757 of the output file can be changed by the @code{%file-prefix} directive |
| 8758 or the @option{-p}/@option{--name-prefix} option. The entire output file |
| 8759 name can be changed by the @code{%output} directive or the |
| 8760 @option{-o}/@option{--output} option. The output file contains a single |
| 8761 class for the parser. |
| 8762 |
| 8763 You can create documentation for generated parsers using Javadoc. |
| 8764 |
| 8765 Contrary to C parsers, Java parsers do not use global variables; the |
| 8766 state of the parser is always local to an instance of the parser class. |
| 8767 Therefore, all Java parsers are ``pure'', and the @code{%pure-parser} |
| 8768 and @code{%define api.pure} directives does not do anything when used in |
| 8769 Java. |
| 8770 |
| 8771 Push parsers are currently unsupported in Java and @code{%define |
| 8772 api.push_pull} have no effect. |
| 8773 |
| 8774 @acronym{GLR} parsers are currently unsupported in Java. Do not use the |
| 8775 @code{glr-parser} directive. |
| 8776 |
| 8777 No header file can be generated for Java parsers. Do not use the |
| 8778 @code{%defines} directive or the @option{-d}/@option{--defines} options. |
| 8779 |
| 8780 @c FIXME: Possible code change. |
| 8781 Currently, support for debugging and verbose errors are always compiled |
| 8782 in. Thus the @code{%debug} and @code{%token-table} directives and the |
| 8783 @option{-t}/@option{--debug} and @option{-k}/@option{--token-table} |
| 8784 options have no effect. This may change in the future to eliminate |
| 8785 unused code in the generated parser, so use @code{%debug} and |
| 8786 @code{%verbose-error} explicitly if needed. Also, in the future the |
| 8787 @code{%token-table} directive might enable a public interface to |
| 8788 access the token names and codes. |
| 8789 |
| 8790 @node Java Semantic Values |
| 8791 @subsection Java Semantic Values |
| 8792 @c - No %union, specify type in %type/%token. |
| 8793 @c - YYSTYPE |
| 8794 @c - Printer and destructor |
| 8795 |
| 8796 There is no @code{%union} directive in Java parsers. Instead, the |
| 8797 semantic values' types (class names) should be specified in the |
| 8798 @code{%type} or @code{%token} directive: |
| 8799 |
| 8800 @example |
| 8801 %type <Expression> expr assignment_expr term factor |
| 8802 %type <Integer> number |
| 8803 @end example |
| 8804 |
| 8805 By default, the semantic stack is declared to have @code{Object} members, |
| 8806 which means that the class types you specify can be of any class. |
| 8807 To improve the type safety of the parser, you can declare the common |
| 8808 superclass of all the semantic values using the @code{%define stype} |
| 8809 directive. For example, after the following declaration: |
| 8810 |
| 8811 @example |
| 8812 %define stype "ASTNode" |
| 8813 @end example |
| 8814 |
| 8815 @noindent |
| 8816 any @code{%type} or @code{%token} specifying a semantic type which |
| 8817 is not a subclass of ASTNode, will cause a compile-time error. |
| 8818 |
| 8819 @c FIXME: Documented bug. |
| 8820 Types used in the directives may be qualified with a package name. |
| 8821 Primitive data types are accepted for Java version 1.5 or later. Note |
| 8822 that in this case the autoboxing feature of Java 1.5 will be used. |
| 8823 Generic types may not be used; this is due to a limitation in the |
| 8824 implementation of Bison, and may change in future releases. |
| 8825 |
| 8826 Java parsers do not support @code{%destructor}, since the language |
| 8827 adopts garbage collection. The parser will try to hold references |
| 8828 to semantic values for as little time as needed. |
| 8829 |
| 8830 Java parsers do not support @code{%printer}, as @code{toString()} |
| 8831 can be used to print the semantic values. This however may change |
| 8832 (in a backwards-compatible way) in future versions of Bison. |
| 8833 |
| 8834 |
| 8835 @node Java Location Values |
| 8836 @subsection Java Location Values |
| 8837 @c - %locations |
| 8838 @c - class Position |
| 8839 @c - class Location |
| 8840 |
| 8841 When the directive @code{%locations} is used, the Java parser |
| 8842 supports location tracking, see @ref{Locations, , Locations Overview}. |
| 8843 An auxiliary user-defined class defines a @dfn{position}, a single point |
| 8844 in a file; Bison itself defines a class representing a @dfn{location}, |
| 8845 a range composed of a pair of positions (possibly spanning several |
| 8846 files). The location class is an inner class of the parser; the name |
| 8847 is @code{Location} by default, and may also be renamed using |
| 8848 @code{%define location_type "@var{class-name}}. |
| 8849 |
| 8850 The location class treats the position as a completely opaque value. |
| 8851 By default, the class name is @code{Position}, but this can be changed |
| 8852 with @code{%define position_type "@var{class-name}"}. This class must |
| 8853 be supplied by the user. |
| 8854 |
| 8855 |
| 8856 @deftypeivar {Location} {Position} begin |
| 8857 @deftypeivarx {Location} {Position} end |
| 8858 The first, inclusive, position of the range, and the first beyond. |
| 8859 @end deftypeivar |
| 8860 |
| 8861 @deftypeop {Constructor} {Location} {} Location (Position @var{loc}) |
| 8862 Create a @code{Location} denoting an empty range located at a given point. |
| 8863 @end deftypeop |
| 8864 |
| 8865 @deftypeop {Constructor} {Location} {} Location (Position @var{begin}, Position
@var{end}) |
| 8866 Create a @code{Location} from the endpoints of the range. |
| 8867 @end deftypeop |
| 8868 |
| 8869 @deftypemethod {Location} {String} toString () |
| 8870 Prints the range represented by the location. For this to work |
| 8871 properly, the position class should override the @code{equals} and |
| 8872 @code{toString} methods appropriately. |
| 8873 @end deftypemethod |
| 8874 |
| 8875 |
| 8876 @node Java Parser Interface |
| 8877 @subsection Java Parser Interface |
| 8878 @c - define parser_class_name |
| 8879 @c - Ctor |
| 8880 @c - parse, error, set_debug_level, debug_level, set_debug_stream, |
| 8881 @c debug_stream. |
| 8882 @c - Reporting errors |
| 8883 |
| 8884 The name of the generated parser class defaults to @code{YYParser}. The |
| 8885 @code{YY} prefix may be changed using the @code{%name-prefix} directive |
| 8886 or the @option{-p}/@option{--name-prefix} option. Alternatively, use |
| 8887 @code{%define parser_class_name "@var{name}"} to give a custom name to |
| 8888 the class. The interface of this class is detailed below. |
| 8889 |
| 8890 By default, the parser class has package visibility. A declaration |
| 8891 @code{%define public} will change to public visibility. Remember that, |
| 8892 according to the Java language specification, the name of the @file{.java} |
| 8893 file should match the name of the class in this case. Similarly, you can |
| 8894 use @code{abstract}, @code{final} and @code{strictfp} with the |
| 8895 @code{%define} declaration to add other modifiers to the parser class. |
| 8896 |
| 8897 The Java package name of the parser class can be specified using the |
| 8898 @code{%define package} directive. The superclass and the implemented |
| 8899 interfaces of the parser class can be specified with the @code{%define |
| 8900 extends} and @code{%define implements} directives. |
| 8901 |
| 8902 The parser class defines an inner class, @code{Location}, that is used |
| 8903 for location tracking (see @ref{Java Location Values}), and a inner |
| 8904 interface, @code{Lexer} (see @ref{Java Scanner Interface}). Other than |
| 8905 these inner class/interface, and the members described in the interface |
| 8906 below, all the other members and fields are preceded with a @code{yy} or |
| 8907 @code{YY} prefix to avoid clashes with user code. |
| 8908 |
| 8909 @c FIXME: The following constants and variables are still undocumented: |
| 8910 @c @code{bisonVersion}, @code{bisonSkeleton} and @code{errorVerbose}. |
| 8911 |
| 8912 The parser class can be extended using the @code{%parse-param} |
| 8913 directive. Each occurrence of the directive will add a @code{protected |
| 8914 final} field to the parser class, and an argument to its constructor, |
| 8915 which initialize them automatically. |
| 8916 |
| 8917 Token names defined by @code{%token} and the predefined @code{EOF} token |
| 8918 name are added as constant fields to the parser class. |
| 8919 |
| 8920 @deftypeop {Constructor} {YYParser} {} YYParser (@var{lex_param}, @dots{}, @var{
parse_param}, @dots{}) |
| 8921 Build a new parser object with embedded @code{%code lexer}. There are |
| 8922 no parameters, unless @code{%parse-param}s and/or @code{%lex-param}s are |
| 8923 used. |
| 8924 @end deftypeop |
| 8925 |
| 8926 @deftypeop {Constructor} {YYParser} {} YYParser (Lexer @var{lexer}, @var{parse_p
aram}, @dots{}) |
| 8927 Build a new parser object using the specified scanner. There are no |
| 8928 additional parameters unless @code{%parse-param}s are used. |
| 8929 |
| 8930 If the scanner is defined by @code{%code lexer}, this constructor is |
| 8931 declared @code{protected} and is called automatically with a scanner |
| 8932 created with the correct @code{%lex-param}s. |
| 8933 @end deftypeop |
| 8934 |
| 8935 @deftypemethod {YYParser} {boolean} parse () |
| 8936 Run the syntactic analysis, and return @code{true} on success, |
| 8937 @code{false} otherwise. |
| 8938 @end deftypemethod |
| 8939 |
| 8940 @deftypemethod {YYParser} {boolean} recovering () |
| 8941 During the syntactic analysis, return @code{true} if recovering |
| 8942 from a syntax error. |
| 8943 @xref{Error Recovery}. |
| 8944 @end deftypemethod |
| 8945 |
| 8946 @deftypemethod {YYParser} {java.io.PrintStream} getDebugStream () |
| 8947 @deftypemethodx {YYParser} {void} setDebugStream (java.io.printStream @var{o}) |
| 8948 Get or set the stream used for tracing the parsing. It defaults to |
| 8949 @code{System.err}. |
| 8950 @end deftypemethod |
| 8951 |
| 8952 @deftypemethod {YYParser} {int} getDebugLevel () |
| 8953 @deftypemethodx {YYParser} {void} setDebugLevel (int @var{l}) |
| 8954 Get or set the tracing level. Currently its value is either 0, no trace, |
| 8955 or nonzero, full tracing. |
| 8956 @end deftypemethod |
| 8957 |
| 8958 |
| 8959 @node Java Scanner Interface |
| 8960 @subsection Java Scanner Interface |
| 8961 @c - %code lexer |
| 8962 @c - %lex-param |
| 8963 @c - Lexer interface |
| 8964 |
| 8965 There are two possible ways to interface a Bison-generated Java parser |
| 8966 with a scanner: the scanner may be defined by @code{%code lexer}, or |
| 8967 defined elsewhere. In either case, the scanner has to implement the |
| 8968 @code{Lexer} inner interface of the parser class. |
| 8969 |
| 8970 In the first case, the body of the scanner class is placed in |
| 8971 @code{%code lexer} blocks. If you want to pass parameters from the |
| 8972 parser constructor to the scanner constructor, specify them with |
| 8973 @code{%lex-param}; they are passed before @code{%parse-param}s to the |
| 8974 constructor. |
| 8975 |
| 8976 In the second case, the scanner has to implement the @code{Lexer} interface, |
| 8977 which is defined within the parser class (e.g., @code{YYParser.Lexer}). |
| 8978 The constructor of the parser object will then accept an object |
| 8979 implementing the interface; @code{%lex-param} is not used in this |
| 8980 case. |
| 8981 |
| 8982 In both cases, the scanner has to implement the following methods. |
| 8983 |
| 8984 @deftypemethod {Lexer} {void} yyerror (Location @var{loc}, String @var{msg}) |
| 8985 This method is defined by the user to emit an error message. The first |
| 8986 parameter is omitted if location tracking is not active. Its type can be |
| 8987 changed using @code{%define location_type "@var{class-name}".} |
| 8988 @end deftypemethod |
| 8989 |
| 8990 @deftypemethod {Lexer} {int} yylex () |
| 8991 Return the next token. Its type is the return value, its semantic |
| 8992 value and location are saved and returned by the ther methods in the |
| 8993 interface. |
| 8994 |
| 8995 Use @code{%define lex_throws} to specify any uncaught exceptions. |
| 8996 Default is @code{java.io.IOException}. |
| 8997 @end deftypemethod |
| 8998 |
| 8999 @deftypemethod {Lexer} {Position} getStartPos () |
| 9000 @deftypemethodx {Lexer} {Position} getEndPos () |
| 9001 Return respectively the first position of the last token that |
| 9002 @code{yylex} returned, and the first position beyond it. These |
| 9003 methods are not needed unless location tracking is active. |
| 9004 |
| 9005 The return type can be changed using @code{%define position_type |
| 9006 "@var{class-name}".} |
| 9007 @end deftypemethod |
| 9008 |
| 9009 @deftypemethod {Lexer} {Object} getLVal () |
| 9010 Return the semantical value of the last token that yylex returned. |
| 9011 |
| 9012 The return type can be changed using @code{%define stype |
| 9013 "@var{class-name}".} |
| 9014 @end deftypemethod |
| 9015 |
| 9016 |
| 9017 @node Java Action Features |
| 9018 @subsection Special Features for Use in Java Actions |
| 9019 |
| 9020 The following special constructs can be uses in Java actions. |
| 9021 Other analogous C action features are currently unavailable for Java. |
| 9022 |
| 9023 Use @code{%define throws} to specify any uncaught exceptions from parser |
| 9024 actions, and initial actions specified by @code{%initial-action}. |
| 9025 |
| 9026 @defvar $@var{n} |
| 9027 The semantic value for the @var{n}th component of the current rule. |
| 9028 This may not be assigned to. |
| 9029 @xref{Java Semantic Values}. |
| 9030 @end defvar |
| 9031 |
| 9032 @defvar $<@var{typealt}>@var{n} |
| 9033 Like @code{$@var{n}} but specifies a alternative type @var{typealt}. |
| 9034 @xref{Java Semantic Values}. |
| 9035 @end defvar |
| 9036 |
| 9037 @defvar $$ |
| 9038 The semantic value for the grouping made by the current rule. As a |
| 9039 value, this is in the base type (@code{Object} or as specified by |
| 9040 @code{%define stype}) as in not cast to the declared subtype because |
| 9041 casts are not allowed on the left-hand side of Java assignments. |
| 9042 Use an explicit Java cast if the correct subtype is needed. |
| 9043 @xref{Java Semantic Values}. |
| 9044 @end defvar |
| 9045 |
| 9046 @defvar $<@var{typealt}>$ |
| 9047 Same as @code{$$} since Java always allow assigning to the base type. |
| 9048 Perhaps we should use this and @code{$<>$} for the value and @code{$$} |
| 9049 for setting the value but there is currently no easy way to distinguish |
| 9050 these constructs. |
| 9051 @xref{Java Semantic Values}. |
| 9052 @end defvar |
| 9053 |
| 9054 @defvar @@@var{n} |
| 9055 The location information of the @var{n}th component of the current rule. |
| 9056 This may not be assigned to. |
| 9057 @xref{Java Location Values}. |
| 9058 @end defvar |
| 9059 |
| 9060 @defvar @@$ |
| 9061 The location information of the grouping made by the current rule. |
| 9062 @xref{Java Location Values}. |
| 9063 @end defvar |
| 9064 |
| 9065 @deffn {Statement} {return YYABORT;} |
| 9066 Return immediately from the parser, indicating failure. |
| 9067 @xref{Java Parser Interface}. |
| 9068 @end deffn |
| 9069 |
| 9070 @deffn {Statement} {return YYACCEPT;} |
| 9071 Return immediately from the parser, indicating success. |
| 9072 @xref{Java Parser Interface}. |
| 9073 @end deffn |
| 9074 |
| 9075 @deffn {Statement} {return YYERROR;} |
| 9076 Start error recovery without printing an error message. |
| 9077 @xref{Error Recovery}. |
| 9078 @end deffn |
| 9079 |
| 9080 @deffn {Statement} {return YYFAIL;} |
| 9081 Print an error message and start error recovery. |
| 9082 @xref{Error Recovery}. |
| 9083 @end deffn |
| 9084 |
| 9085 @deftypefn {Function} {boolean} recovering () |
| 9086 Return whether error recovery is being done. In this state, the parser |
| 9087 reads token until it reaches a known state, and then restarts normal |
| 9088 operation. |
| 9089 @xref{Error Recovery}. |
| 9090 @end deftypefn |
| 9091 |
| 9092 @deftypefn {Function} {protected void} yyerror (String msg) |
| 9093 @deftypefnx {Function} {protected void} yyerror (Position pos, String msg) |
| 9094 @deftypefnx {Function} {protected void} yyerror (Location loc, String msg) |
| 9095 Print an error message using the @code{yyerror} method of the scanner |
| 9096 instance in use. |
| 9097 @end deftypefn |
| 9098 |
| 9099 |
| 9100 @node Java Differences |
| 9101 @subsection Differences between C/C++ and Java Grammars |
| 9102 |
| 9103 The different structure of the Java language forces several differences |
| 9104 between C/C++ grammars, and grammars designed for Java parsers. This |
| 9105 section summarizes these differences. |
| 9106 |
| 9107 @itemize |
| 9108 @item |
| 9109 Java lacks a preprocessor, so the @code{YYERROR}, @code{YYACCEPT}, |
| 9110 @code{YYABORT} symbols (@pxref{Table of Symbols}) cannot obviously be |
| 9111 macros. Instead, they should be preceded by @code{return} when they |
| 9112 appear in an action. The actual definition of these symbols is |
| 9113 opaque to the Bison grammar, and it might change in the future. The |
| 9114 only meaningful operation that you can do, is to return them. |
| 9115 See @pxref{Java Action Features}. |
| 9116 |
| 9117 Note that of these three symbols, only @code{YYACCEPT} and |
| 9118 @code{YYABORT} will cause a return from the @code{yyparse} |
| 9119 method@footnote{Java parsers include the actions in a separate |
| 9120 method than @code{yyparse} in order to have an intuitive syntax that |
| 9121 corresponds to these C macros.}. |
| 9122 |
| 9123 @item |
| 9124 Java lacks unions, so @code{%union} has no effect. Instead, semantic |
| 9125 values have a common base type: @code{Object} or as specified by |
| 9126 @code{%define stype}. Angle backets on @code{%token}, @code{type}, |
| 9127 @code{$@var{n}} and @code{$$} specify subtypes rather than fields of |
| 9128 an union. The type of @code{$$}, even with angle brackets, is the base |
| 9129 type since Java casts are not allow on the left-hand side of assignments. |
| 9130 Also, @code{$@var{n}} and @code{@@@var{n}} are not allowed on the |
| 9131 left-hand side of assignments. See @pxref{Java Semantic Values} and |
| 9132 @pxref{Java Action Features}. |
| 9133 |
| 9134 @item |
| 9135 The prolog declarations have a different meaning than in C/C++ code. |
| 9136 @table @asis |
| 9137 @item @code{%code imports} |
| 9138 blocks are placed at the beginning of the Java source code. They may |
| 9139 include copyright notices. For a @code{package} declarations, it is |
| 9140 suggested to use @code{%define package} instead. |
| 9141 |
| 9142 @item unqualified @code{%code} |
| 9143 blocks are placed inside the parser class. |
| 9144 |
| 9145 @item @code{%code lexer} |
| 9146 blocks, if specified, should include the implementation of the |
| 9147 scanner. If there is no such block, the scanner can be any class |
| 9148 that implements the appropriate interface (see @pxref{Java Scanner |
| 9149 Interface}). |
| 9150 @end table |
| 9151 |
| 9152 Other @code{%code} blocks are not supported in Java parsers. |
| 9153 In particular, @code{%@{ @dots{} %@}} blocks should not be used |
| 9154 and may give an error in future versions of Bison. |
| 9155 |
| 9156 The epilogue has the same meaning as in C/C++ code and it can |
| 9157 be used to define other classes used by the parser @emph{outside} |
| 9158 the parser class. |
| 9159 @end itemize |
| 9160 |
| 9161 |
| 9162 @node Java Declarations Summary |
| 9163 @subsection Java Declarations Summary |
| 9164 |
| 9165 This summary only include declarations specific to Java or have special |
| 9166 meaning when used in a Java parser. |
| 9167 |
| 9168 @deffn {Directive} {%language "Java"} |
| 9169 Generate a Java class for the parser. |
| 9170 @end deffn |
| 9171 |
| 9172 @deffn {Directive} %lex-param @{@var{type} @var{name}@} |
| 9173 A parameter for the lexer class defined by @code{%code lexer} |
| 9174 @emph{only}, added as parameters to the lexer constructor and the parser |
| 9175 constructor that @emph{creates} a lexer. Default is none. |
| 9176 @xref{Java Scanner Interface}. |
| 9177 @end deffn |
| 9178 |
| 9179 @deffn {Directive} %name-prefix "@var{prefix}" |
| 9180 The prefix of the parser class name @code{@var{prefix}Parser} if |
| 9181 @code{%define parser_class_name} is not used. Default is @code{YY}. |
| 9182 @xref{Java Bison Interface}. |
| 9183 @end deffn |
| 9184 |
| 9185 @deffn {Directive} %parse-param @{@var{type} @var{name}@} |
| 9186 A parameter for the parser class added as parameters to constructor(s) |
| 9187 and as fields initialized by the constructor(s). Default is none. |
| 9188 @xref{Java Parser Interface}. |
| 9189 @end deffn |
| 9190 |
| 9191 @deffn {Directive} %token <@var{type}> @var{token} @dots{} |
| 9192 Declare tokens. Note that the angle brackets enclose a Java @emph{type}. |
| 9193 @xref{Java Semantic Values}. |
| 9194 @end deffn |
| 9195 |
| 9196 @deffn {Directive} %type <@var{type}> @var{nonterminal} @dots{} |
| 9197 Declare the type of nonterminals. Note that the angle brackets enclose |
| 9198 a Java @emph{type}. |
| 9199 @xref{Java Semantic Values}. |
| 9200 @end deffn |
| 9201 |
| 9202 @deffn {Directive} %code @{ @var{code} @dots{} @} |
| 9203 Code appended to the inside of the parser class. |
| 9204 @xref{Java Differences}. |
| 9205 @end deffn |
| 9206 |
| 9207 @deffn {Directive} {%code imports} @{ @var{code} @dots{} @} |
| 9208 Code inserted just after the @code{package} declaration. |
| 9209 @xref{Java Differences}. |
| 9210 @end deffn |
| 9211 |
| 9212 @deffn {Directive} {%code lexer} @{ @var{code} @dots{} @} |
| 9213 Code added to the body of a inner lexer class within the parser class. |
| 9214 @xref{Java Scanner Interface}. |
| 9215 @end deffn |
| 9216 |
| 9217 @deffn {Directive} %% @var{code} @dots{} |
| 9218 Code (after the second @code{%%}) appended to the end of the file, |
| 9219 @emph{outside} the parser class. |
| 9220 @xref{Java Differences}. |
| 9221 @end deffn |
| 9222 |
| 9223 @deffn {Directive} %@{ @var{code} @dots{} %@} |
| 9224 Not supported. Use @code{%code import} instead. |
| 9225 @xref{Java Differences}. |
| 9226 @end deffn |
| 9227 |
| 9228 @deffn {Directive} {%define abstract} |
| 9229 Whether the parser class is declared @code{abstract}. Default is false. |
| 9230 @xref{Java Bison Interface}. |
| 9231 @end deffn |
| 9232 |
| 9233 @deffn {Directive} {%define extends} "@var{superclass}" |
| 9234 The superclass of the parser class. Default is none. |
| 9235 @xref{Java Bison Interface}. |
| 9236 @end deffn |
| 9237 |
| 9238 @deffn {Directive} {%define final} |
| 9239 Whether the parser class is declared @code{final}. Default is false. |
| 9240 @xref{Java Bison Interface}. |
| 9241 @end deffn |
| 9242 |
| 9243 @deffn {Directive} {%define implements} "@var{interfaces}" |
| 9244 The implemented interfaces of the parser class, a comma-separated list. |
| 9245 Default is none. |
| 9246 @xref{Java Bison Interface}. |
| 9247 @end deffn |
| 9248 |
| 9249 @deffn {Directive} {%define lex_throws} "@var{exceptions}" |
| 9250 The exceptions thrown by the @code{yylex} method of the lexer, a |
| 9251 comma-separated list. Default is @code{java.io.IOException}. |
| 9252 @xref{Java Scanner Interface}. |
| 9253 @end deffn |
| 9254 |
| 9255 @deffn {Directive} {%define location_type} "@var{class}" |
| 9256 The name of the class used for locations (a range between two |
| 9257 positions). This class is generated as an inner class of the parser |
| 9258 class by @command{bison}. Default is @code{Location}. |
| 9259 @xref{Java Location Values}. |
| 9260 @end deffn |
| 9261 |
| 9262 @deffn {Directive} {%define package} "@var{package}" |
| 9263 The package to put the parser class in. Default is none. |
| 9264 @xref{Java Bison Interface}. |
| 9265 @end deffn |
| 9266 |
| 9267 @deffn {Directive} {%define parser_class_name} "@var{name}" |
| 9268 The name of the parser class. Default is @code{YYParser} or |
| 9269 @code{@var{name-prefix}Parser}. |
| 9270 @xref{Java Bison Interface}. |
| 9271 @end deffn |
| 9272 |
| 9273 @deffn {Directive} {%define position_type} "@var{class}" |
| 9274 The name of the class used for positions. This class must be supplied by |
| 9275 the user. Default is @code{Position}. |
| 9276 @xref{Java Location Values}. |
| 9277 @end deffn |
| 9278 |
| 9279 @deffn {Directive} {%define public} |
| 9280 Whether the parser class is declared @code{public}. Default is false. |
| 9281 @xref{Java Bison Interface}. |
| 9282 @end deffn |
| 9283 |
| 9284 @deffn {Directive} {%define stype} "@var{class}" |
| 9285 The base type of semantic values. Default is @code{Object}. |
| 9286 @xref{Java Semantic Values}. |
| 9287 @end deffn |
| 9288 |
| 9289 @deffn {Directive} {%define strictfp} |
| 9290 Whether the parser class is declared @code{strictfp}. Default is false. |
| 9291 @xref{Java Bison Interface}. |
| 9292 @end deffn |
| 9293 |
| 9294 @deffn {Directive} {%define throws} "@var{exceptions}" |
| 9295 The exceptions thrown by user-supplied parser actions and |
| 9296 @code{%initial-action}, a comma-separated list. Default is none. |
| 9297 @xref{Java Parser Interface}. |
| 9298 @end deffn |
| 9299 |
| 9300 |
| 9301 @c ================================================= FAQ |
| 9302 |
| 9303 @node FAQ |
| 9304 @chapter Frequently Asked Questions |
| 9305 @cindex frequently asked questions |
| 9306 @cindex questions |
| 9307 |
| 9308 Several questions about Bison come up occasionally. Here some of them |
| 9309 are addressed. |
| 9310 |
| 9311 @menu |
| 9312 * Memory Exhausted:: Breaking the Stack Limits |
| 9313 * How Can I Reset the Parser:: @code{yyparse} Keeps some State |
| 9314 * Strings are Destroyed:: @code{yylval} Loses Track of Strings |
| 9315 * Implementing Gotos/Loops:: Control Flow in the Calculator |
| 9316 * Multiple start-symbols:: Factoring closely related grammars |
| 9317 * Secure? Conform?:: Is Bison @acronym{POSIX} safe? |
| 9318 * I can't build Bison:: Troubleshooting |
| 9319 * Where can I find help?:: Troubleshouting |
| 9320 * Bug Reports:: Troublereporting |
| 9321 * More Languages:: Parsers in C++, Java, and so on |
| 9322 * Beta Testing:: Experimenting development versions |
| 9323 * Mailing Lists:: Meeting other Bison users |
| 9324 @end menu |
| 9325 |
| 9326 @node Memory Exhausted |
| 9327 @section Memory Exhausted |
| 9328 |
| 9329 @display |
| 9330 My parser returns with error with a @samp{memory exhausted} |
| 9331 message. What can I do? |
| 9332 @end display |
| 9333 |
| 9334 This question is already addressed elsewhere, @xref{Recursion, |
| 9335 ,Recursive Rules}. |
| 9336 |
| 9337 @node How Can I Reset the Parser |
| 9338 @section How Can I Reset the Parser |
| 9339 |
| 9340 The following phenomenon has several symptoms, resulting in the |
| 9341 following typical questions: |
| 9342 |
| 9343 @display |
| 9344 I invoke @code{yyparse} several times, and on correct input it works |
| 9345 properly; but when a parse error is found, all the other calls fail |
| 9346 too. How can I reset the error flag of @code{yyparse}? |
| 9347 @end display |
| 9348 |
| 9349 @noindent |
| 9350 or |
| 9351 |
| 9352 @display |
| 9353 My parser includes support for an @samp{#include}-like feature, in |
| 9354 which case I run @code{yyparse} from @code{yyparse}. This fails |
| 9355 although I did specify @code{%define api.pure}. |
| 9356 @end display |
| 9357 |
| 9358 These problems typically come not from Bison itself, but from |
| 9359 Lex-generated scanners. Because these scanners use large buffers for |
| 9360 speed, they might not notice a change of input file. As a |
| 9361 demonstration, consider the following source file, |
| 9362 @file{first-line.l}: |
| 9363 |
| 9364 @verbatim |
| 9365 %{ |
| 9366 #include <stdio.h> |
| 9367 #include <stdlib.h> |
| 9368 %} |
| 9369 %% |
| 9370 .*\n ECHO; return 1; |
| 9371 %% |
| 9372 int |
| 9373 yyparse (char const *file) |
| 9374 { |
| 9375 yyin = fopen (file, "r"); |
| 9376 if (!yyin) |
| 9377 exit (2); |
| 9378 /* One token only. */ |
| 9379 yylex (); |
| 9380 if (fclose (yyin) != 0) |
| 9381 exit (3); |
| 9382 return 0; |
| 9383 } |
| 9384 |
| 9385 int |
| 9386 main (void) |
| 9387 { |
| 9388 yyparse ("input"); |
| 9389 yyparse ("input"); |
| 9390 return 0; |
| 9391 } |
| 9392 @end verbatim |
| 9393 |
| 9394 @noindent |
| 9395 If the file @file{input} contains |
| 9396 |
| 9397 @verbatim |
| 9398 input:1: Hello, |
| 9399 input:2: World! |
| 9400 @end verbatim |
| 9401 |
| 9402 @noindent |
| 9403 then instead of getting the first line twice, you get: |
| 9404 |
| 9405 @example |
| 9406 $ @kbd{flex -ofirst-line.c first-line.l} |
| 9407 $ @kbd{gcc -ofirst-line first-line.c -ll} |
| 9408 $ @kbd{./first-line} |
| 9409 input:1: Hello, |
| 9410 input:2: World! |
| 9411 @end example |
| 9412 |
| 9413 Therefore, whenever you change @code{yyin}, you must tell the |
| 9414 Lex-generated scanner to discard its current buffer and switch to the |
| 9415 new one. This depends upon your implementation of Lex; see its |
| 9416 documentation for more. For Flex, it suffices to call |
| 9417 @samp{YY_FLUSH_BUFFER} after each change to @code{yyin}. If your |
| 9418 Flex-generated scanner needs to read from several input streams to |
| 9419 handle features like include files, you might consider using Flex |
| 9420 functions like @samp{yy_switch_to_buffer} that manipulate multiple |
| 9421 input buffers. |
| 9422 |
| 9423 If your Flex-generated scanner uses start conditions (@pxref{Start |
| 9424 conditions, , Start conditions, flex, The Flex Manual}), you might |
| 9425 also want to reset the scanner's state, i.e., go back to the initial |
| 9426 start condition, through a call to @samp{BEGIN (0)}. |
| 9427 |
| 9428 @node Strings are Destroyed |
| 9429 @section Strings are Destroyed |
| 9430 |
| 9431 @display |
| 9432 My parser seems to destroy old strings, or maybe it loses track of |
| 9433 them. Instead of reporting @samp{"foo", "bar"}, it reports |
| 9434 @samp{"bar", "bar"}, or even @samp{"foo\nbar", "bar"}. |
| 9435 @end display |
| 9436 |
| 9437 This error is probably the single most frequent ``bug report'' sent to |
| 9438 Bison lists, but is only concerned with a misunderstanding of the role |
| 9439 of the scanner. Consider the following Lex code: |
| 9440 |
| 9441 @verbatim |
| 9442 %{ |
| 9443 #include <stdio.h> |
| 9444 char *yylval = NULL; |
| 9445 %} |
| 9446 %% |
| 9447 .* yylval = yytext; return 1; |
| 9448 \n /* IGNORE */ |
| 9449 %% |
| 9450 int |
| 9451 main () |
| 9452 { |
| 9453 /* Similar to using $1, $2 in a Bison action. */ |
| 9454 char *fst = (yylex (), yylval); |
| 9455 char *snd = (yylex (), yylval); |
| 9456 printf ("\"%s\", \"%s\"\n", fst, snd); |
| 9457 return 0; |
| 9458 } |
| 9459 @end verbatim |
| 9460 |
| 9461 If you compile and run this code, you get: |
| 9462 |
| 9463 @example |
| 9464 $ @kbd{flex -osplit-lines.c split-lines.l} |
| 9465 $ @kbd{gcc -osplit-lines split-lines.c -ll} |
| 9466 $ @kbd{printf 'one\ntwo\n' | ./split-lines} |
| 9467 "one |
| 9468 two", "two" |
| 9469 @end example |
| 9470 |
| 9471 @noindent |
| 9472 this is because @code{yytext} is a buffer provided for @emph{reading} |
| 9473 in the action, but if you want to keep it, you have to duplicate it |
| 9474 (e.g., using @code{strdup}). Note that the output may depend on how |
| 9475 your implementation of Lex handles @code{yytext}. For instance, when |
| 9476 given the Lex compatibility option @option{-l} (which triggers the |
| 9477 option @samp{%array}) Flex generates a different behavior: |
| 9478 |
| 9479 @example |
| 9480 $ @kbd{flex -l -osplit-lines.c split-lines.l} |
| 9481 $ @kbd{gcc -osplit-lines split-lines.c -ll} |
| 9482 $ @kbd{printf 'one\ntwo\n' | ./split-lines} |
| 9483 "two", "two" |
| 9484 @end example |
| 9485 |
| 9486 |
| 9487 @node Implementing Gotos/Loops |
| 9488 @section Implementing Gotos/Loops |
| 9489 |
| 9490 @display |
| 9491 My simple calculator supports variables, assignments, and functions, |
| 9492 but how can I implement gotos, or loops? |
| 9493 @end display |
| 9494 |
| 9495 Although very pedagogical, the examples included in the document blur |
| 9496 the distinction to make between the parser---whose job is to recover |
| 9497 the structure of a text and to transmit it to subsequent modules of |
| 9498 the program---and the processing (such as the execution) of this |
| 9499 structure. This works well with so called straight line programs, |
| 9500 i.e., precisely those that have a straightforward execution model: |
| 9501 execute simple instructions one after the others. |
| 9502 |
| 9503 @cindex abstract syntax tree |
| 9504 @cindex @acronym{AST} |
| 9505 If you want a richer model, you will probably need to use the parser |
| 9506 to construct a tree that does represent the structure it has |
| 9507 recovered; this tree is usually called the @dfn{abstract syntax tree}, |
| 9508 or @dfn{@acronym{AST}} for short. Then, walking through this tree, |
| 9509 traversing it in various ways, will enable treatments such as its |
| 9510 execution or its translation, which will result in an interpreter or a |
| 9511 compiler. |
| 9512 |
| 9513 This topic is way beyond the scope of this manual, and the reader is |
| 9514 invited to consult the dedicated literature. |
| 9515 |
| 9516 |
| 9517 @node Multiple start-symbols |
| 9518 @section Multiple start-symbols |
| 9519 |
| 9520 @display |
| 9521 I have several closely related grammars, and I would like to share their |
| 9522 implementations. In fact, I could use a single grammar but with |
| 9523 multiple entry points. |
| 9524 @end display |
| 9525 |
| 9526 Bison does not support multiple start-symbols, but there is a very |
| 9527 simple means to simulate them. If @code{foo} and @code{bar} are the two |
| 9528 pseudo start-symbols, then introduce two new tokens, say |
| 9529 @code{START_FOO} and @code{START_BAR}, and use them as switches from the |
| 9530 real start-symbol: |
| 9531 |
| 9532 @example |
| 9533 %token START_FOO START_BAR; |
| 9534 %start start; |
| 9535 start: START_FOO foo |
| 9536 | START_BAR bar; |
| 9537 @end example |
| 9538 |
| 9539 These tokens prevents the introduction of new conflicts. As far as the |
| 9540 parser goes, that is all that is needed. |
| 9541 |
| 9542 Now the difficult part is ensuring that the scanner will send these |
| 9543 tokens first. If your scanner is hand-written, that should be |
| 9544 straightforward. If your scanner is generated by Lex, them there is |
| 9545 simple means to do it: recall that anything between @samp{%@{ ... %@}} |
| 9546 after the first @code{%%} is copied verbatim in the top of the generated |
| 9547 @code{yylex} function. Make sure a variable @code{start_token} is |
| 9548 available in the scanner (e.g., a global variable or using |
| 9549 @code{%lex-param} etc.), and use the following: |
| 9550 |
| 9551 @example |
| 9552 /* @r{Prologue.} */ |
| 9553 %% |
| 9554 %@{ |
| 9555 if (start_token) |
| 9556 @{ |
| 9557 int t = start_token; |
| 9558 start_token = 0; |
| 9559 return t; |
| 9560 @} |
| 9561 %@} |
| 9562 /* @r{The rules.} */ |
| 9563 @end example |
| 9564 |
| 9565 |
| 9566 @node Secure? Conform? |
| 9567 @section Secure? Conform? |
| 9568 |
| 9569 @display |
| 9570 Is Bison secure? Does it conform to POSIX? |
| 9571 @end display |
| 9572 |
| 9573 If you're looking for a guarantee or certification, we don't provide it. |
| 9574 However, Bison is intended to be a reliable program that conforms to the |
| 9575 @acronym{POSIX} specification for Yacc. If you run into problems, |
| 9576 please send us a bug report. |
| 9577 |
| 9578 @node I can't build Bison |
| 9579 @section I can't build Bison |
| 9580 |
| 9581 @display |
| 9582 I can't build Bison because @command{make} complains that |
| 9583 @code{msgfmt} is not found. |
| 9584 What should I do? |
| 9585 @end display |
| 9586 |
| 9587 Like most GNU packages with internationalization support, that feature |
| 9588 is turned on by default. If you have problems building in the @file{po} |
| 9589 subdirectory, it indicates that your system's internationalization |
| 9590 support is lacking. You can re-configure Bison with |
| 9591 @option{--disable-nls} to turn off this support, or you can install GNU |
| 9592 gettext from @url{ftp://ftp.gnu.org/gnu/gettext/} and re-configure |
| 9593 Bison. See the file @file{ABOUT-NLS} for more information. |
| 9594 |
| 9595 |
| 9596 @node Where can I find help? |
| 9597 @section Where can I find help? |
| 9598 |
| 9599 @display |
| 9600 I'm having trouble using Bison. Where can I find help? |
| 9601 @end display |
| 9602 |
| 9603 First, read this fine manual. Beyond that, you can send mail to |
| 9604 @email{help-bison@@gnu.org}. This mailing list is intended to be |
| 9605 populated with people who are willing to answer questions about using |
| 9606 and installing Bison. Please keep in mind that (most of) the people on |
| 9607 the list have aspects of their lives which are not related to Bison (!), |
| 9608 so you may not receive an answer to your question right away. This can |
| 9609 be frustrating, but please try not to honk them off; remember that any |
| 9610 help they provide is purely voluntary and out of the kindness of their |
| 9611 hearts. |
| 9612 |
| 9613 @node Bug Reports |
| 9614 @section Bug Reports |
| 9615 |
| 9616 @display |
| 9617 I found a bug. What should I include in the bug report? |
| 9618 @end display |
| 9619 |
| 9620 Before you send a bug report, make sure you are using the latest |
| 9621 version. Check @url{ftp://ftp.gnu.org/pub/gnu/bison/} or one of its |
| 9622 mirrors. Be sure to include the version number in your bug report. If |
| 9623 the bug is present in the latest version but not in a previous version, |
| 9624 try to determine the most recent version which did not contain the bug. |
| 9625 |
| 9626 If the bug is parser-related, you should include the smallest grammar |
| 9627 you can which demonstrates the bug. The grammar file should also be |
| 9628 complete (i.e., I should be able to run it through Bison without having |
| 9629 to edit or add anything). The smaller and simpler the grammar, the |
| 9630 easier it will be to fix the bug. |
| 9631 |
| 9632 Include information about your compilation environment, including your |
| 9633 operating system's name and version and your compiler's name and |
| 9634 version. If you have trouble compiling, you should also include a |
| 9635 transcript of the build session, starting with the invocation of |
| 9636 `configure'. Depending on the nature of the bug, you may be asked to |
| 9637 send additional files as well (such as `config.h' or `config.cache'). |
| 9638 |
| 9639 Patches are most welcome, but not required. That is, do not hesitate to |
| 9640 send a bug report just because you can not provide a fix. |
| 9641 |
| 9642 Send bug reports to @email{bug-bison@@gnu.org}. |
| 9643 |
| 9644 @node More Languages |
| 9645 @section More Languages |
| 9646 |
| 9647 @display |
| 9648 Will Bison ever have C++ and Java support? How about @var{insert your |
| 9649 favorite language here}? |
| 9650 @end display |
| 9651 |
| 9652 C++ and Java support is there now, and is documented. We'd love to add other |
| 9653 languages; contributions are welcome. |
| 9654 |
| 9655 @node Beta Testing |
| 9656 @section Beta Testing |
| 9657 |
| 9658 @display |
| 9659 What is involved in being a beta tester? |
| 9660 @end display |
| 9661 |
| 9662 It's not terribly involved. Basically, you would download a test |
| 9663 release, compile it, and use it to build and run a parser or two. After |
| 9664 that, you would submit either a bug report or a message saying that |
| 9665 everything is okay. It is important to report successes as well as |
| 9666 failures because test releases eventually become mainstream releases, |
| 9667 but only if they are adequately tested. If no one tests, development is |
| 9668 essentially halted. |
| 9669 |
| 9670 Beta testers are particularly needed for operating systems to which the |
| 9671 developers do not have easy access. They currently have easy access to |
| 9672 recent GNU/Linux and Solaris versions. Reports about other operating |
| 9673 systems are especially welcome. |
| 9674 |
| 9675 @node Mailing Lists |
| 9676 @section Mailing Lists |
| 9677 |
| 9678 @display |
| 9679 How do I join the help-bison and bug-bison mailing lists? |
| 9680 @end display |
| 9681 |
| 9682 See @url{http://lists.gnu.org/}. |
| 9683 |
| 9684 @c ================================================= Table of Symbols |
| 9685 |
| 9686 @node Table of Symbols |
| 9687 @appendix Bison Symbols |
| 9688 @cindex Bison symbols, table of |
| 9689 @cindex symbols in Bison, table of |
| 9690 |
| 9691 @deffn {Variable} @@$ |
| 9692 In an action, the location of the left-hand side of the rule. |
| 9693 @xref{Locations, , Locations Overview}. |
| 9694 @end deffn |
| 9695 |
| 9696 @deffn {Variable} @@@var{n} |
| 9697 In an action, the location of the @var{n}-th symbol of the right-hand |
| 9698 side of the rule. @xref{Locations, , Locations Overview}. |
| 9699 @end deffn |
| 9700 |
| 9701 @deffn {Variable} $$ |
| 9702 In an action, the semantic value of the left-hand side of the rule. |
| 9703 @xref{Actions}. |
| 9704 @end deffn |
| 9705 |
| 9706 @deffn {Variable} $@var{n} |
| 9707 In an action, the semantic value of the @var{n}-th symbol of the |
| 9708 right-hand side of the rule. @xref{Actions}. |
| 9709 @end deffn |
| 9710 |
| 9711 @deffn {Delimiter} %% |
| 9712 Delimiter used to separate the grammar rule section from the |
| 9713 Bison declarations section or the epilogue. |
| 9714 @xref{Grammar Layout, ,The Overall Layout of a Bison Grammar}. |
| 9715 @end deffn |
| 9716 |
| 9717 @c Don't insert spaces, or check the DVI output. |
| 9718 @deffn {Delimiter} %@{@var{code}%@} |
| 9719 All code listed between @samp{%@{} and @samp{%@}} is copied directly to |
| 9720 the output file uninterpreted. Such code forms the prologue of the input |
| 9721 file. @xref{Grammar Outline, ,Outline of a Bison |
| 9722 Grammar}. |
| 9723 @end deffn |
| 9724 |
| 9725 @deffn {Construct} /*@dots{}*/ |
| 9726 Comment delimiters, as in C. |
| 9727 @end deffn |
| 9728 |
| 9729 @deffn {Delimiter} : |
| 9730 Separates a rule's result from its components. @xref{Rules, ,Syntax of |
| 9731 Grammar Rules}. |
| 9732 @end deffn |
| 9733 |
| 9734 @deffn {Delimiter} ; |
| 9735 Terminates a rule. @xref{Rules, ,Syntax of Grammar Rules}. |
| 9736 @end deffn |
| 9737 |
| 9738 @deffn {Delimiter} | |
| 9739 Separates alternate rules for the same result nonterminal. |
| 9740 @xref{Rules, ,Syntax of Grammar Rules}. |
| 9741 @end deffn |
| 9742 |
| 9743 @deffn {Directive} <*> |
| 9744 Used to define a default tagged @code{%destructor} or default tagged |
| 9745 @code{%printer}. |
| 9746 |
| 9747 This feature is experimental. |
| 9748 More user feedback will help to determine whether it should become a permanent |
| 9749 feature. |
| 9750 |
| 9751 @xref{Destructor Decl, , Freeing Discarded Symbols}. |
| 9752 @end deffn |
| 9753 |
| 9754 @deffn {Directive} <> |
| 9755 Used to define a default tagless @code{%destructor} or default tagless |
| 9756 @code{%printer}. |
| 9757 |
| 9758 This feature is experimental. |
| 9759 More user feedback will help to determine whether it should become a permanent |
| 9760 feature. |
| 9761 |
| 9762 @xref{Destructor Decl, , Freeing Discarded Symbols}. |
| 9763 @end deffn |
| 9764 |
| 9765 @deffn {Symbol} $accept |
| 9766 The predefined nonterminal whose only rule is @samp{$accept: @var{start} |
| 9767 $end}, where @var{start} is the start symbol. @xref{Start Decl, , The |
| 9768 Start-Symbol}. It cannot be used in the grammar. |
| 9769 @end deffn |
| 9770 |
| 9771 @deffn {Directive} %code @{@var{code}@} |
| 9772 @deffnx {Directive} %code @var{qualifier} @{@var{code}@} |
| 9773 Insert @var{code} verbatim into output parser source. |
| 9774 @xref{Decl Summary,,%code}. |
| 9775 @end deffn |
| 9776 |
| 9777 @deffn {Directive} %debug |
| 9778 Equip the parser for debugging. @xref{Decl Summary}. |
| 9779 @end deffn |
| 9780 |
| 9781 @deffn {Directive} %debug |
| 9782 Equip the parser for debugging. @xref{Decl Summary}. |
| 9783 @end deffn |
| 9784 |
| 9785 @ifset defaultprec |
| 9786 @deffn {Directive} %default-prec |
| 9787 Assign a precedence to rules that lack an explicit @samp{%prec} |
| 9788 modifier. @xref{Contextual Precedence, ,Context-Dependent |
| 9789 Precedence}. |
| 9790 @end deffn |
| 9791 @end ifset |
| 9792 |
| 9793 @deffn {Directive} %define @var{define-variable} |
| 9794 @deffnx {Directive} %define @var{define-variable} @var{value} |
| 9795 Define a variable to adjust Bison's behavior. |
| 9796 @xref{Decl Summary,,%define}. |
| 9797 @end deffn |
| 9798 |
| 9799 @deffn {Directive} %defines |
| 9800 Bison declaration to create a header file meant for the scanner. |
| 9801 @xref{Decl Summary}. |
| 9802 @end deffn |
| 9803 |
| 9804 @deffn {Directive} %defines @var{defines-file} |
| 9805 Same as above, but save in the file @var{defines-file}. |
| 9806 @xref{Decl Summary}. |
| 9807 @end deffn |
| 9808 |
| 9809 @deffn {Directive} %destructor |
| 9810 Specify how the parser should reclaim the memory associated to |
| 9811 discarded symbols. @xref{Destructor Decl, , Freeing Discarded Symbols}. |
| 9812 @end deffn |
| 9813 |
| 9814 @deffn {Directive} %dprec |
| 9815 Bison declaration to assign a precedence to a rule that is used at parse |
| 9816 time to resolve reduce/reduce conflicts. @xref{GLR Parsers, ,Writing |
| 9817 @acronym{GLR} Parsers}. |
| 9818 @end deffn |
| 9819 |
| 9820 @deffn {Symbol} $end |
| 9821 The predefined token marking the end of the token stream. It cannot be |
| 9822 used in the grammar. |
| 9823 @end deffn |
| 9824 |
| 9825 @deffn {Symbol} error |
| 9826 A token name reserved for error recovery. This token may be used in |
| 9827 grammar rules so as to allow the Bison parser to recognize an error in |
| 9828 the grammar without halting the process. In effect, a sentence |
| 9829 containing an error may be recognized as valid. On a syntax error, the |
| 9830 token @code{error} becomes the current lookahead token. Actions |
| 9831 corresponding to @code{error} are then executed, and the lookahead |
| 9832 token is reset to the token that originally caused the violation. |
| 9833 @xref{Error Recovery}. |
| 9834 @end deffn |
| 9835 |
| 9836 @deffn {Directive} %error-verbose |
| 9837 Bison declaration to request verbose, specific error message strings |
| 9838 when @code{yyerror} is called. |
| 9839 @end deffn |
| 9840 |
| 9841 @deffn {Directive} %file-prefix "@var{prefix}" |
| 9842 Bison declaration to set the prefix of the output files. @xref{Decl |
| 9843 Summary}. |
| 9844 @end deffn |
| 9845 |
| 9846 @deffn {Directive} %glr-parser |
| 9847 Bison declaration to produce a @acronym{GLR} parser. @xref{GLR |
| 9848 Parsers, ,Writing @acronym{GLR} Parsers}. |
| 9849 @end deffn |
| 9850 |
| 9851 @deffn {Directive} %initial-action |
| 9852 Run user code before parsing. @xref{Initial Action Decl, , Performing Actions b
efore Parsing}. |
| 9853 @end deffn |
| 9854 |
| 9855 @deffn {Directive} %language |
| 9856 Specify the programming language for the generated parser. |
| 9857 @xref{Decl Summary}. |
| 9858 @end deffn |
| 9859 |
| 9860 @deffn {Directive} %left |
| 9861 Bison declaration to assign left associativity to token(s). |
| 9862 @xref{Precedence Decl, ,Operator Precedence}. |
| 9863 @end deffn |
| 9864 |
| 9865 @deffn {Directive} %lex-param @{@var{argument-declaration}@} |
| 9866 Bison declaration to specifying an additional parameter that |
| 9867 @code{yylex} should accept. @xref{Pure Calling,, Calling Conventions |
| 9868 for Pure Parsers}. |
| 9869 @end deffn |
| 9870 |
| 9871 @deffn {Directive} %merge |
| 9872 Bison declaration to assign a merging function to a rule. If there is a |
| 9873 reduce/reduce conflict with a rule having the same merging function, the |
| 9874 function is applied to the two semantic values to get a single result. |
| 9875 @xref{GLR Parsers, ,Writing @acronym{GLR} Parsers}. |
| 9876 @end deffn |
| 9877 |
| 9878 @deffn {Directive} %name-prefix "@var{prefix}" |
| 9879 Bison declaration to rename the external symbols. @xref{Decl Summary}. |
| 9880 @end deffn |
| 9881 |
| 9882 @ifset defaultprec |
| 9883 @deffn {Directive} %no-default-prec |
| 9884 Do not assign a precedence to rules that lack an explicit @samp{%prec} |
| 9885 modifier. @xref{Contextual Precedence, ,Context-Dependent |
| 9886 Precedence}. |
| 9887 @end deffn |
| 9888 @end ifset |
| 9889 |
| 9890 @deffn {Directive} %no-lines |
| 9891 Bison declaration to avoid generating @code{#line} directives in the |
| 9892 parser file. @xref{Decl Summary}. |
| 9893 @end deffn |
| 9894 |
| 9895 @deffn {Directive} %nonassoc |
| 9896 Bison declaration to assign nonassociativity to token(s). |
| 9897 @xref{Precedence Decl, ,Operator Precedence}. |
| 9898 @end deffn |
| 9899 |
| 9900 @deffn {Directive} %output "@var{file}" |
| 9901 Bison declaration to set the name of the parser file. @xref{Decl |
| 9902 Summary}. |
| 9903 @end deffn |
| 9904 |
| 9905 @deffn {Directive} %parse-param @{@var{argument-declaration}@} |
| 9906 Bison declaration to specifying an additional parameter that |
| 9907 @code{yyparse} should accept. @xref{Parser Function,, The Parser |
| 9908 Function @code{yyparse}}. |
| 9909 @end deffn |
| 9910 |
| 9911 @deffn {Directive} %prec |
| 9912 Bison declaration to assign a precedence to a specific rule. |
| 9913 @xref{Contextual Precedence, ,Context-Dependent Precedence}. |
| 9914 @end deffn |
| 9915 |
| 9916 @deffn {Directive} %pure-parser |
| 9917 Deprecated version of @code{%define api.pure} (@pxref{Decl Summary, ,%define}), |
| 9918 for which Bison is more careful to warn about unreasonable usage. |
| 9919 @end deffn |
| 9920 |
| 9921 @deffn {Directive} %require "@var{version}" |
| 9922 Require version @var{version} or higher of Bison. @xref{Require Decl, , |
| 9923 Require a Version of Bison}. |
| 9924 @end deffn |
| 9925 |
| 9926 @deffn {Directive} %right |
| 9927 Bison declaration to assign right associativity to token(s). |
| 9928 @xref{Precedence Decl, ,Operator Precedence}. |
| 9929 @end deffn |
| 9930 |
| 9931 @deffn {Directive} %skeleton |
| 9932 Specify the skeleton to use; usually for development. |
| 9933 @xref{Decl Summary}. |
| 9934 @end deffn |
| 9935 |
| 9936 @deffn {Directive} %start |
| 9937 Bison declaration to specify the start symbol. @xref{Start Decl, ,The |
| 9938 Start-Symbol}. |
| 9939 @end deffn |
| 9940 |
| 9941 @deffn {Directive} %token |
| 9942 Bison declaration to declare token(s) without specifying precedence. |
| 9943 @xref{Token Decl, ,Token Type Names}. |
| 9944 @end deffn |
| 9945 |
| 9946 @deffn {Directive} %token-table |
| 9947 Bison declaration to include a token name table in the parser file. |
| 9948 @xref{Decl Summary}. |
| 9949 @end deffn |
| 9950 |
| 9951 @deffn {Directive} %type |
| 9952 Bison declaration to declare nonterminals. @xref{Type Decl, |
| 9953 ,Nonterminal Symbols}. |
| 9954 @end deffn |
| 9955 |
| 9956 @deffn {Symbol} $undefined |
| 9957 The predefined token onto which all undefined values returned by |
| 9958 @code{yylex} are mapped. It cannot be used in the grammar, rather, use |
| 9959 @code{error}. |
| 9960 @end deffn |
| 9961 |
| 9962 @deffn {Directive} %union |
| 9963 Bison declaration to specify several possible data types for semantic |
| 9964 values. @xref{Union Decl, ,The Collection of Value Types}. |
| 9965 @end deffn |
| 9966 |
| 9967 @deffn {Macro} YYABORT |
| 9968 Macro to pretend that an unrecoverable syntax error has occurred, by |
| 9969 making @code{yyparse} return 1 immediately. The error reporting |
| 9970 function @code{yyerror} is not called. @xref{Parser Function, ,The |
| 9971 Parser Function @code{yyparse}}. |
| 9972 |
| 9973 For Java parsers, this functionality is invoked using @code{return YYABORT;} |
| 9974 instead. |
| 9975 @end deffn |
| 9976 |
| 9977 @deffn {Macro} YYACCEPT |
| 9978 Macro to pretend that a complete utterance of the language has been |
| 9979 read, by making @code{yyparse} return 0 immediately. |
| 9980 @xref{Parser Function, ,The Parser Function @code{yyparse}}. |
| 9981 |
| 9982 For Java parsers, this functionality is invoked using @code{return YYACCEPT;} |
| 9983 instead. |
| 9984 @end deffn |
| 9985 |
| 9986 @deffn {Macro} YYBACKUP |
| 9987 Macro to discard a value from the parser stack and fake a lookahead |
| 9988 token. @xref{Action Features, ,Special Features for Use in Actions}. |
| 9989 @end deffn |
| 9990 |
| 9991 @deffn {Variable} yychar |
| 9992 External integer variable that contains the integer value of the |
| 9993 lookahead token. (In a pure parser, it is a local variable within |
| 9994 @code{yyparse}.) Error-recovery rule actions may examine this variable. |
| 9995 @xref{Action Features, ,Special Features for Use in Actions}. |
| 9996 @end deffn |
| 9997 |
| 9998 @deffn {Variable} yyclearin |
| 9999 Macro used in error-recovery rule actions. It clears the previous |
| 10000 lookahead token. @xref{Error Recovery}. |
| 10001 @end deffn |
| 10002 |
| 10003 @deffn {Macro} YYDEBUG |
| 10004 Macro to define to equip the parser with tracing code. @xref{Tracing, |
| 10005 ,Tracing Your Parser}. |
| 10006 @end deffn |
| 10007 |
| 10008 @deffn {Variable} yydebug |
| 10009 External integer variable set to zero by default. If @code{yydebug} |
| 10010 is given a nonzero value, the parser will output information on input |
| 10011 symbols and parser action. @xref{Tracing, ,Tracing Your Parser}. |
| 10012 @end deffn |
| 10013 |
| 10014 @deffn {Macro} yyerrok |
| 10015 Macro to cause parser to recover immediately to its normal mode |
| 10016 after a syntax error. @xref{Error Recovery}. |
| 10017 @end deffn |
| 10018 |
| 10019 @deffn {Macro} YYERROR |
| 10020 Macro to pretend that a syntax error has just been detected: call |
| 10021 @code{yyerror} and then perform normal error recovery if possible |
| 10022 (@pxref{Error Recovery}), or (if recovery is impossible) make |
| 10023 @code{yyparse} return 1. @xref{Error Recovery}. |
| 10024 |
| 10025 For Java parsers, this functionality is invoked using @code{return YYERROR;} |
| 10026 instead. |
| 10027 @end deffn |
| 10028 |
| 10029 @deffn {Function} yyerror |
| 10030 User-supplied function to be called by @code{yyparse} on error. |
| 10031 @xref{Error Reporting, ,The Error |
| 10032 Reporting Function @code{yyerror}}. |
| 10033 @end deffn |
| 10034 |
| 10035 @deffn {Macro} YYERROR_VERBOSE |
| 10036 An obsolete macro that you define with @code{#define} in the prologue |
| 10037 to request verbose, specific error message strings |
| 10038 when @code{yyerror} is called. It doesn't matter what definition you |
| 10039 use for @code{YYERROR_VERBOSE}, just whether you define it. Using |
| 10040 @code{%error-verbose} is preferred. |
| 10041 @end deffn |
| 10042 |
| 10043 @deffn {Macro} YYINITDEPTH |
| 10044 Macro for specifying the initial size of the parser stack. |
| 10045 @xref{Memory Management}. |
| 10046 @end deffn |
| 10047 |
| 10048 @deffn {Function} yylex |
| 10049 User-supplied lexical analyzer function, called with no arguments to get |
| 10050 the next token. @xref{Lexical, ,The Lexical Analyzer Function |
| 10051 @code{yylex}}. |
| 10052 @end deffn |
| 10053 |
| 10054 @deffn {Macro} YYLEX_PARAM |
| 10055 An obsolete macro for specifying an extra argument (or list of extra |
| 10056 arguments) for @code{yyparse} to pass to @code{yylex}. The use of this |
| 10057 macro is deprecated, and is supported only for Yacc like parsers. |
| 10058 @xref{Pure Calling,, Calling Conventions for Pure Parsers}. |
| 10059 @end deffn |
| 10060 |
| 10061 @deffn {Variable} yylloc |
| 10062 External variable in which @code{yylex} should place the line and column |
| 10063 numbers associated with a token. (In a pure parser, it is a local |
| 10064 variable within @code{yyparse}, and its address is passed to |
| 10065 @code{yylex}.) |
| 10066 You can ignore this variable if you don't use the @samp{@@} feature in the |
| 10067 grammar actions. |
| 10068 @xref{Token Locations, ,Textual Locations of Tokens}. |
| 10069 In semantic actions, it stores the location of the lookahead token. |
| 10070 @xref{Actions and Locations, ,Actions and Locations}. |
| 10071 @end deffn |
| 10072 |
| 10073 @deffn {Type} YYLTYPE |
| 10074 Data type of @code{yylloc}; by default, a structure with four |
| 10075 members. @xref{Location Type, , Data Types of Locations}. |
| 10076 @end deffn |
| 10077 |
| 10078 @deffn {Variable} yylval |
| 10079 External variable in which @code{yylex} should place the semantic |
| 10080 value associated with a token. (In a pure parser, it is a local |
| 10081 variable within @code{yyparse}, and its address is passed to |
| 10082 @code{yylex}.) |
| 10083 @xref{Token Values, ,Semantic Values of Tokens}. |
| 10084 In semantic actions, it stores the semantic value of the lookahead token. |
| 10085 @xref{Actions, ,Actions}. |
| 10086 @end deffn |
| 10087 |
| 10088 @deffn {Macro} YYMAXDEPTH |
| 10089 Macro for specifying the maximum size of the parser stack. @xref{Memory |
| 10090 Management}. |
| 10091 @end deffn |
| 10092 |
| 10093 @deffn {Variable} yynerrs |
| 10094 Global variable which Bison increments each time it reports a syntax error. |
| 10095 (In a pure parser, it is a local variable within @code{yyparse}. In a |
| 10096 pure push parser, it is a member of yypstate.) |
| 10097 @xref{Error Reporting, ,The Error Reporting Function @code{yyerror}}. |
| 10098 @end deffn |
| 10099 |
| 10100 @deffn {Function} yyparse |
| 10101 The parser function produced by Bison; call this function to start |
| 10102 parsing. @xref{Parser Function, ,The Parser Function @code{yyparse}}. |
| 10103 @end deffn |
| 10104 |
| 10105 @deffn {Function} yypstate_delete |
| 10106 The function to delete a parser instance, produced by Bison in push mode; |
| 10107 call this function to delete the memory associated with a parser. |
| 10108 @xref{Parser Delete Function, ,The Parser Delete Function |
| 10109 @code{yypstate_delete}}. |
| 10110 (The current push parsing interface is experimental and may evolve. |
| 10111 More user feedback will help to stabilize it.) |
| 10112 @end deffn |
| 10113 |
| 10114 @deffn {Function} yypstate_new |
| 10115 The function to create a parser instance, produced by Bison in push mode; |
| 10116 call this function to create a new parser. |
| 10117 @xref{Parser Create Function, ,The Parser Create Function |
| 10118 @code{yypstate_new}}. |
| 10119 (The current push parsing interface is experimental and may evolve. |
| 10120 More user feedback will help to stabilize it.) |
| 10121 @end deffn |
| 10122 |
| 10123 @deffn {Function} yypull_parse |
| 10124 The parser function produced by Bison in push mode; call this function to |
| 10125 parse the rest of the input stream. |
| 10126 @xref{Pull Parser Function, ,The Pull Parser Function |
| 10127 @code{yypull_parse}}. |
| 10128 (The current push parsing interface is experimental and may evolve. |
| 10129 More user feedback will help to stabilize it.) |
| 10130 @end deffn |
| 10131 |
| 10132 @deffn {Function} yypush_parse |
| 10133 The parser function produced by Bison in push mode; call this function to |
| 10134 parse a single token. @xref{Push Parser Function, ,The Push Parser Function |
| 10135 @code{yypush_parse}}. |
| 10136 (The current push parsing interface is experimental and may evolve. |
| 10137 More user feedback will help to stabilize it.) |
| 10138 @end deffn |
| 10139 |
| 10140 @deffn {Macro} YYPARSE_PARAM |
| 10141 An obsolete macro for specifying the name of a parameter that |
| 10142 @code{yyparse} should accept. The use of this macro is deprecated, and |
| 10143 is supported only for Yacc like parsers. @xref{Pure Calling,, Calling |
| 10144 Conventions for Pure Parsers}. |
| 10145 @end deffn |
| 10146 |
| 10147 @deffn {Macro} YYRECOVERING |
| 10148 The expression @code{YYRECOVERING ()} yields 1 when the parser |
| 10149 is recovering from a syntax error, and 0 otherwise. |
| 10150 @xref{Action Features, ,Special Features for Use in Actions}. |
| 10151 @end deffn |
| 10152 |
| 10153 @deffn {Macro} YYSTACK_USE_ALLOCA |
| 10154 Macro used to control the use of @code{alloca} when the C |
| 10155 @acronym{LALR}(1) parser needs to extend its stacks. If defined to 0, |
| 10156 the parser will use @code{malloc} to extend its stacks. If defined to |
| 10157 1, the parser will use @code{alloca}. Values other than 0 and 1 are |
| 10158 reserved for future Bison extensions. If not defined, |
| 10159 @code{YYSTACK_USE_ALLOCA} defaults to 0. |
| 10160 |
| 10161 In the all-too-common case where your code may run on a host with a |
| 10162 limited stack and with unreliable stack-overflow checking, you should |
| 10163 set @code{YYMAXDEPTH} to a value that cannot possibly result in |
| 10164 unchecked stack overflow on any of your target hosts when |
| 10165 @code{alloca} is called. You can inspect the code that Bison |
| 10166 generates in order to determine the proper numeric values. This will |
| 10167 require some expertise in low-level implementation details. |
| 10168 @end deffn |
| 10169 |
| 10170 @deffn {Type} YYSTYPE |
| 10171 Data type of semantic values; @code{int} by default. |
| 10172 @xref{Value Type, ,Data Types of Semantic Values}. |
| 10173 @end deffn |
| 10174 |
| 10175 @node Glossary |
| 10176 @appendix Glossary |
| 10177 @cindex glossary |
| 10178 |
| 10179 @table @asis |
| 10180 @item Backus-Naur Form (@acronym{BNF}; also called ``Backus Normal Form'') |
| 10181 Formal method of specifying context-free grammars originally proposed |
| 10182 by John Backus, and slightly improved by Peter Naur in his 1960-01-02 |
| 10183 committee document contributing to what became the Algol 60 report. |
| 10184 @xref{Language and Grammar, ,Languages and Context-Free Grammars}. |
| 10185 |
| 10186 @item Context-free grammars |
| 10187 Grammars specified as rules that can be applied regardless of context. |
| 10188 Thus, if there is a rule which says that an integer can be used as an |
| 10189 expression, integers are allowed @emph{anywhere} an expression is |
| 10190 permitted. @xref{Language and Grammar, ,Languages and Context-Free |
| 10191 Grammars}. |
| 10192 |
| 10193 @item Dynamic allocation |
| 10194 Allocation of memory that occurs during execution, rather than at |
| 10195 compile time or on entry to a function. |
| 10196 |
| 10197 @item Empty string |
| 10198 Analogous to the empty set in set theory, the empty string is a |
| 10199 character string of length zero. |
| 10200 |
| 10201 @item Finite-state stack machine |
| 10202 A ``machine'' that has discrete states in which it is said to exist at |
| 10203 each instant in time. As input to the machine is processed, the |
| 10204 machine moves from state to state as specified by the logic of the |
| 10205 machine. In the case of the parser, the input is the language being |
| 10206 parsed, and the states correspond to various stages in the grammar |
| 10207 rules. @xref{Algorithm, ,The Bison Parser Algorithm}. |
| 10208 |
| 10209 @item Generalized @acronym{LR} (@acronym{GLR}) |
| 10210 A parsing algorithm that can handle all context-free grammars, including those |
| 10211 that are not @acronym{LALR}(1). It resolves situations that Bison's |
| 10212 usual @acronym{LALR}(1) |
| 10213 algorithm cannot by effectively splitting off multiple parsers, trying all |
| 10214 possible parsers, and discarding those that fail in the light of additional |
| 10215 right context. @xref{Generalized LR Parsing, ,Generalized |
| 10216 @acronym{LR} Parsing}. |
| 10217 |
| 10218 @item Grouping |
| 10219 A language construct that is (in general) grammatically divisible; |
| 10220 for example, `expression' or `declaration' in C@. |
| 10221 @xref{Language and Grammar, ,Languages and Context-Free Grammars}. |
| 10222 |
| 10223 @item Infix operator |
| 10224 An arithmetic operator that is placed between the operands on which it |
| 10225 performs some operation. |
| 10226 |
| 10227 @item Input stream |
| 10228 A continuous flow of data between devices or programs. |
| 10229 |
| 10230 @item Language construct |
| 10231 One of the typical usage schemas of the language. For example, one of |
| 10232 the constructs of the C language is the @code{if} statement. |
| 10233 @xref{Language and Grammar, ,Languages and Context-Free Grammars}. |
| 10234 |
| 10235 @item Left associativity |
| 10236 Operators having left associativity are analyzed from left to right: |
| 10237 @samp{a+b+c} first computes @samp{a+b} and then combines with |
| 10238 @samp{c}. @xref{Precedence, ,Operator Precedence}. |
| 10239 |
| 10240 @item Left recursion |
| 10241 A rule whose result symbol is also its first component symbol; for |
| 10242 example, @samp{expseq1 : expseq1 ',' exp;}. @xref{Recursion, ,Recursive |
| 10243 Rules}. |
| 10244 |
| 10245 @item Left-to-right parsing |
| 10246 Parsing a sentence of a language by analyzing it token by token from |
| 10247 left to right. @xref{Algorithm, ,The Bison Parser Algorithm}. |
| 10248 |
| 10249 @item Lexical analyzer (scanner) |
| 10250 A function that reads an input stream and returns tokens one by one. |
| 10251 @xref{Lexical, ,The Lexical Analyzer Function @code{yylex}}. |
| 10252 |
| 10253 @item Lexical tie-in |
| 10254 A flag, set by actions in the grammar rules, which alters the way |
| 10255 tokens are parsed. @xref{Lexical Tie-ins}. |
| 10256 |
| 10257 @item Literal string token |
| 10258 A token which consists of two or more fixed characters. @xref{Symbols}. |
| 10259 |
| 10260 @item Lookahead token |
| 10261 A token already read but not yet shifted. @xref{Lookahead, ,Lookahead |
| 10262 Tokens}. |
| 10263 |
| 10264 @item @acronym{LALR}(1) |
| 10265 The class of context-free grammars that Bison (like most other parser |
| 10266 generators) can handle; a subset of @acronym{LR}(1). @xref{Mystery |
| 10267 Conflicts, ,Mysterious Reduce/Reduce Conflicts}. |
| 10268 |
| 10269 @item @acronym{LR}(1) |
| 10270 The class of context-free grammars in which at most one token of |
| 10271 lookahead is needed to disambiguate the parsing of any piece of input. |
| 10272 |
| 10273 @item Nonterminal symbol |
| 10274 A grammar symbol standing for a grammatical construct that can |
| 10275 be expressed through rules in terms of smaller constructs; in other |
| 10276 words, a construct that is not a token. @xref{Symbols}. |
| 10277 |
| 10278 @item Parser |
| 10279 A function that recognizes valid sentences of a language by analyzing |
| 10280 the syntax structure of a set of tokens passed to it from a lexical |
| 10281 analyzer. |
| 10282 |
| 10283 @item Postfix operator |
| 10284 An arithmetic operator that is placed after the operands upon which it |
| 10285 performs some operation. |
| 10286 |
| 10287 @item Reduction |
| 10288 Replacing a string of nonterminals and/or terminals with a single |
| 10289 nonterminal, according to a grammar rule. @xref{Algorithm, ,The Bison |
| 10290 Parser Algorithm}. |
| 10291 |
| 10292 @item Reentrant |
| 10293 A reentrant subprogram is a subprogram which can be in invoked any |
| 10294 number of times in parallel, without interference between the various |
| 10295 invocations. @xref{Pure Decl, ,A Pure (Reentrant) Parser}. |
| 10296 |
| 10297 @item Reverse polish notation |
| 10298 A language in which all operators are postfix operators. |
| 10299 |
| 10300 @item Right recursion |
| 10301 A rule whose result symbol is also its last component symbol; for |
| 10302 example, @samp{expseq1: exp ',' expseq1;}. @xref{Recursion, ,Recursive |
| 10303 Rules}. |
| 10304 |
| 10305 @item Semantics |
| 10306 In computer languages, the semantics are specified by the actions |
| 10307 taken for each instance of the language, i.e., the meaning of |
| 10308 each statement. @xref{Semantics, ,Defining Language Semantics}. |
| 10309 |
| 10310 @item Shift |
| 10311 A parser is said to shift when it makes the choice of analyzing |
| 10312 further input from the stream rather than reducing immediately some |
| 10313 already-recognized rule. @xref{Algorithm, ,The Bison Parser Algorithm}. |
| 10314 |
| 10315 @item Single-character literal |
| 10316 A single character that is recognized and interpreted as is. |
| 10317 @xref{Grammar in Bison, ,From Formal Rules to Bison Input}. |
| 10318 |
| 10319 @item Start symbol |
| 10320 The nonterminal symbol that stands for a complete valid utterance in |
| 10321 the language being parsed. The start symbol is usually listed as the |
| 10322 first nonterminal symbol in a language specification. |
| 10323 @xref{Start Decl, ,The Start-Symbol}. |
| 10324 |
| 10325 @item Symbol table |
| 10326 A data structure where symbol names and associated data are stored |
| 10327 during parsing to allow for recognition and use of existing |
| 10328 information in repeated uses of a symbol. @xref{Multi-function Calc}. |
| 10329 |
| 10330 @item Syntax error |
| 10331 An error encountered during parsing of an input stream due to invalid |
| 10332 syntax. @xref{Error Recovery}. |
| 10333 |
| 10334 @item Token |
| 10335 A basic, grammatically indivisible unit of a language. The symbol |
| 10336 that describes a token in the grammar is a terminal symbol. |
| 10337 The input of the Bison parser is a stream of tokens which comes from |
| 10338 the lexical analyzer. @xref{Symbols}. |
| 10339 |
| 10340 @item Terminal symbol |
| 10341 A grammar symbol that has no rules in the grammar and therefore is |
| 10342 grammatically indivisible. The piece of text it represents is a token. |
| 10343 @xref{Language and Grammar, ,Languages and Context-Free Grammars}. |
| 10344 @end table |
| 10345 |
| 10346 @node Copying This Manual |
| 10347 @appendix Copying This Manual |
| 10348 @include fdl.texi |
| 10349 |
| 10350 @node Index |
| 10351 @unnumbered Index |
| 10352 |
| 10353 @printindex cp |
| 10354 |
| 10355 @bye |
| 10356 |
| 10357 @c LocalWords: texinfo setfilename settitle setchapternewpage finalout |
| 10358 @c LocalWords: ifinfo smallbook shorttitlepage titlepage GPL FIXME iftex |
| 10359 @c LocalWords: akim fn cp syncodeindex vr tp synindex dircategory direntry |
| 10360 @c LocalWords: ifset vskip pt filll insertcopying sp ISBN Etienne Suvasa |
| 10361 @c LocalWords: ifnottex yyparse detailmenu GLR RPN Calc var Decls Rpcalc |
| 10362 @c LocalWords: rpcalc Lexer Expr ltcalc mfcalc yylex |
| 10363 @c LocalWords: yyerror pxref LR yylval cindex dfn LALR samp gpl BNF xref |
| 10364 @c LocalWords: const int paren ifnotinfo AC noindent emph expr stmt findex |
| 10365 @c LocalWords: glr YYSTYPE TYPENAME prog dprec printf decl init stmtMerge |
| 10366 @c LocalWords: pre STDC GNUC endif yy YY alloca lf stddef stdlib YYDEBUG |
| 10367 @c LocalWords: NUM exp subsubsection kbd Ctrl ctype EOF getchar isdigit |
| 10368 @c LocalWords: ungetc stdin scanf sc calc ulator ls lm cc NEG prec yyerrok |
| 10369 @c LocalWords: longjmp fprintf stderr yylloc YYLTYPE cos ln |
| 10370 @c LocalWords: smallexample symrec val tptr FNCT fnctptr func struct sym |
| 10371 @c LocalWords: fnct putsym getsym fname arith fncts atan ptr malloc sizeof |
| 10372 @c LocalWords: strlen strcpy fctn strcmp isalpha symbuf realloc isalnum |
| 10373 @c LocalWords: ptypes itype YYPRINT trigraphs yytname expseq vindex dtype |
| 10374 @c LocalWords: Rhs YYRHSLOC LE nonassoc op deffn typeless yynerrs |
| 10375 @c LocalWords: yychar yydebug msg YYNTOKENS YYNNTS YYNRULES YYNSTATES |
| 10376 @c LocalWords: cparse clex deftypefun NE defmac YYACCEPT YYABORT param |
| 10377 @c LocalWords: strncmp intval tindex lvalp locp llocp typealt YYBACKUP |
| 10378 @c LocalWords: YYEMPTY YYEOF YYRECOVERING yyclearin GE def UMINUS maybeword |
| 10379 @c LocalWords: Johnstone Shamsa Sadaf Hussain Tomita TR uref YYMAXDEPTH |
| 10380 @c LocalWords: YYINITDEPTH stmnts ref stmnt initdcl maybeasm notype |
| 10381 @c LocalWords: hexflag STR exdent itemset asis DYYDEBUG YYFPRINTF args |
| 10382 @c LocalWords: infile ypp yxx outfile itemx tex leaderfill |
| 10383 @c LocalWords: hbox hss hfill tt ly yyin fopen fclose ofirst gcc ll |
| 10384 @c LocalWords: nbar yytext fst snd osplit ntwo strdup AST |
| 10385 @c LocalWords: YYSTACK DVI fdl printindex |
OLD | NEW |