NEWS

   1 Bison News
   2 ----------
   3
   4 Changes in version 2.3a+ (????-??-??):
   5
   6 * The -g and --graph options now output graphs in Graphviz DOT format,
   7   not VCG format.
   8
   9 * An experimental directive %language specifies the language of the
  10   generated parser, which can be C (the default) or C++.  This
  11   directive affects the skeleton used, and the names of the generated
  12   files if the grammar file's name ends in ".y".
  13
  14 * The grammar file may now specify the name of the parser header file using
  15   %defines.  For example:
  16
  17     %defines "parser.h"
  18
  19 * The `=' that used to be required in the following directives is now
  20   deprecated:
  21
  22     %file-prefix "parser"
  23     %name-prefix "c_"
  24     %output "parser.c"
  25
  26 * The quotes around NAME that used to be required in the following directive
  27   are now deprecated:
  28
  29     %define NAME "VALUE"
  30
  31 * Bison 2.3a provided a new set of directives as a more flexible alternative to
  32   the traditional Yacc prologue blocks.  Those have now been consolidated into
  33   a single %code directive with an optional qualifier field, which identifies
  34   the purpose of the code and thus the location(s) where Bison should generate
  35   it:
  36
  37     1. `%code          {CODE}' replaces `%after-header  {CODE}'
  38     2. `%code requires {CODE}' replaces `%start-header  {CODE}'
  39     3. `%code provides {CODE}' replaces `%end-header    {CODE}'
  40     4. `%code top      {CODE}' replaces `%before-header {CODE}'
  41
  42   See the %code entries in `Appendix A Bison Symbols' in the Bison manual for a
  43   summary of the new functionality.  See the new section `Prologue
  44   Alternatives' for a detailed discussion including the advantages of %code
  45   over the traditional Yacc prologues.
  46
  47   The prologue alternatives are experimental.  More user feedback will help to
  48   determine whether they should become permanent features.
  49
  50 * Revised warning: unset or unused mid-rule values
  51
  52   Since Bison 2.2, Bison has warned about mid-rule values that are set but not
  53   used within any of the actions of the parent rule.  For example, Bison warns
  54   about unused $2 in:
  55
  56     exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; };
  57
  58   Now, Bison also warns about mid-rule values that are used but not set.  For
  59   example, Bison warns about unset $$ in the mid-rule action in:
  60
  61     exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; };
  62
  63   However, Bison now disables both of these warnings by default since they
  64   sometimes prove to be false alarms in existing grammars employing the Yacc
  65   constructs $0 or $-N (where N is some positive integer).
  66
  67   To enable these warnings, specify the flag `--warnings=midrule-values' or
  68   `-W', which is a synonym for `--warnings=all'.
  69
  70 * Bison now recognizes two separate kinds of default %destructor's and
  71   %printer's:
  72
  73     1. Place `<*>' in a %destructor/%printer symbol list to define a default
  74        %destructor/%printer for all grammar symbols for which you have formally
  75        declared semantic type tags.
  76
  77     2. Place `<>' in a %destructor/%printer symbol list to define a default
  78        %destructor/%printer for all grammar symbols without declared semantic
  79        type tags.
  80
  81   Bison no longer supports the `%symbol-default' notation from Bison 2.3a.
  82   `<*>' and `<>' combined achieve the same effect with one exception: Bison no
  83   longer applies any %destructor to a mid-rule value if that mid-rule value is
  84   not actually ever referenced using either $$ or $n in a semantic action.
  85
  86   The default %destructor's and %printer's are experimental.  More user
  87   feedback will help to determine whether they should become permanent
  88   features.
  89
  90   See the section `Freeing Discarded Symbols' in the Bison manual for further
  91   details.
  92
  93 Changes in version 2.3a, 2006-09-13:
  94
  95 * Instead of %union, you can define and use your own union type
  96   YYSTYPE if your grammar contains at least one <type> tag.
  97   Your YYSTYPE need not be a macro; it can be a typedef.
  98   This change is for compatibility with other Yacc implementations,
  99   and is required by POSIX.
 100
 101 * Locations columns and lines start at 1.
 102   In accordance with the GNU Coding Standards and Emacs.
 103
 104 * You may now declare per-type and default %destructor's and %printer's:
 105
 106   For example:
 107
 108     %union { char *string; }
 109     %token <string> STRING1
 110     %token <string> STRING2
 111     %type  <string> string1
 112     %type  <string> string2
 113     %union { char character; }
 114     %token <character> CHR
 115     %type  <character> chr
 116     %destructor { free ($$); } %symbol-default
 117     %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1
 118     %destructor { } <character>
 119
 120   guarantees that, when the parser discards any user-defined symbol that has a
 121   semantic type tag other than `<character>', it passes its semantic value to
 122   `free'.  However, when the parser discards a `STRING1' or a `string1', it
 123   also prints its line number to `stdout'.  It performs only the second
 124   `%destructor' in this case, so it invokes `free' only once.
 125
 126   [Although we failed to mention this here in the 2.3a release, the default
 127   %destructor's and %printer's were experimental, and they were rewritten in
 128   future versions.]
 129
 130 * Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with `-y',
 131   `--yacc', or `%yacc'), Bison no longer generates #define statements for
 132   associating token numbers with token names.  Removing the #define statements
 133   helps to sanitize the global namespace during preprocessing, but POSIX Yacc
 134   requires them.  Bison still generates an enum for token names in all cases.
 135
 136 * Handling of traditional Yacc prologue blocks is now more consistent but
 137   potentially incompatible with previous releases of Bison.
 138
 139   As before, you declare prologue blocks in your grammar file with the
 140   `%{ ... %}' syntax.  To generate the pre-prologue, Bison concatenates all
 141   prologue blocks that you've declared before the first %union.  To generate
 142   the post-prologue, Bison concatenates all prologue blocks that you've
 143   declared after the first %union.
 144
 145   Previous releases of Bison inserted the pre-prologue into both the header
 146   file and the code file in all cases except for LALR(1) parsers in C.  In the
 147   latter case, Bison inserted it only into the code file.  For parsers in C++,
 148   the point of insertion was before any token definitions (which associate
 149   token numbers with names).  For parsers in C, the point of insertion was
 150   after the token definitions.
 151
 152   Now, Bison never inserts the pre-prologue into the header file.  In the code
 153   file, it always inserts it before the token definitions.
 154
 155 * Bison now provides a more flexible alternative to the traditional Yacc
 156   prologue blocks: %before-header, %start-header, %end-header, and
 157   %after-header.
 158
 159   For example, the following declaration order in the grammar file reflects the
 160   order in which Bison will output these code blocks.  However, you are free to
 161   declare these code blocks in your grammar file in whatever order is most
 162   convenient for you:
 163
 164     %before-header {
 165       /* Bison treats this block like a pre-prologue block: it inserts it into
 166        * the code file before the contents of the header file.  It does *not*
 167        * insert it into the header file.  This is a good place to put
 168        * #include's that you want at the top of your code file.  A common
 169        * example is `#include "system.h"'.  */
 170     }
 171     %start-header {
 172       /* Bison inserts this block into both the header file and the code file.
 173        * In both files, the point of insertion is before any Bison-generated
 174        * token, semantic type, location type, and class definitions.  This is a
 175        * good place to define %union dependencies, for example.  */
 176     }
 177     %union {
 178       /* Unlike the traditional Yacc prologue blocks, the output order for the
 179        * new %*-header blocks is not affected by their declaration position
 180        * relative to any %union in the grammar file.  */
 181     }
 182     %end-header {
 183       /* Bison inserts this block into both the header file and the code file.
 184        * In both files, the point of insertion is after the Bison-generated
 185        * definitions.  This is a good place to declare or define public
 186        * functions or data structures that depend on the Bison-generated
 187        * definitions.  */
 188     }
 189     %after-header {
 190       /* Bison treats this block like a post-prologue block: it inserts it into
 191        * the code file after the contents of the header file.  It does *not*
 192        * insert it into the header file.  This is a good place to declare or
 193        * define internal functions or data structures that depend on the
 194        * Bison-generated definitions.  */
 195     }
 196
 197   If you have multiple occurrences of any one of the above declarations, Bison
 198   will concatenate the contents in declaration order.
 199
 200   [Although we failed to mention this here in the 2.3a release, the prologue
 201   alternatives were experimental, and they were rewritten in future versions.]
 202
 203 * The option `--report=look-ahead' has been changed to `--report=lookahead'.
 204   The old spelling still works, but is not documented and may be removed
 205   in a future release.
 206
 207 Changes in version 2.3, 2006-06-05:
 208
 209 * GLR grammars should now use `YYRECOVERING ()' instead of `YYRECOVERING',
 210   for compatibility with LALR(1) grammars.
 211
 212 * It is now documented that any definition of YYSTYPE or YYLTYPE should
 213   be to a type name that does not contain parentheses or brackets.
 214
 215 Changes in version 2.2, 2006-05-19:
 216
 217 * The distribution terms for all Bison-generated parsers now permit
 218   using the parsers in nonfree programs.  Previously, this permission
 219   was granted only for Bison-generated LALR(1) parsers in C.
 220
 221 * %name-prefix changes the namespace name in C++ outputs.
 222
 223 * The C++ parsers export their token_type.
 224
 225 * Bison now allows multiple %union declarations, and concatenates
 226   their contents together.
 227
 228 * New warning: unused values
 229   Right-hand side symbols whose values are not used are reported,
 230   if the symbols have destructors.  For instance:
 231
 232      exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
 233         | exp "+" exp
 234         ;
 235
 236   will trigger a warning about $$ and $5 in the first rule, and $3 in
 237   the second ($1 is copied to $$ by the default rule).  This example
 238   most likely contains three errors, and could be rewritten as:
 239
 240      exp: exp "?" exp ":" exp
 241             { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
 242         | exp "+" exp
 243             { $$ = $1 ? $1 : $3; if ($1) free ($3); }
 244         ;
 245
 246   However, if the original actions were really intended, memory leaks
 247   and all, the warnings can be suppressed by letting Bison believe the
 248   values are used, e.g.:
 249
 250      exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
 251         | exp "+" exp         { $$ = $1; (void) $3; }
 252         ;
 253
 254   If there are mid-rule actions, the warning is issued if no action
 255   uses it.  The following triggers no warning: $1 and $3 are used.
 256
 257      exp: exp { push ($1); } '+' exp { push ($3); sum (); };
 258
 259   The warning is intended to help catching lost values and memory leaks.
 260   If a value is ignored, its associated memory typically is not reclaimed.
 261
 262 * %destructor vs. YYABORT, YYACCEPT, and YYERROR.
 263   Destructors are now called when user code invokes YYABORT, YYACCEPT,
 264   and YYERROR, for all objects on the stack, other than objects
 265   corresponding to the right-hand side of the current rule.
 266
 267 * %expect, %expect-rr
 268   Incorrect numbers of expected conflicts are now actual errors,
 269   instead of warnings.
 270
 271 * GLR, YACC parsers.
 272   The %parse-params are available in the destructors (and the
 273   experimental printers) as per the documentation.
 274
 275 * Bison now warns if it finds a stray `$' or `@' in an action.
 276
 277 * %require "VERSION"
 278   This specifies that the grammar file depends on features implemented
 279   in Bison version VERSION or higher.
 280
 281 * lalr1.cc: The token and value types are now class members.
 282   The tokens were defined as free form enums and cpp macros.  YYSTYPE
 283   was defined as a free form union.  They are now class members:
 284   tokens are enumerations of the `yy::parser::token' struct, and the
 285   semantic values have the `yy::parser::semantic_type' type.
 286
 287   If you do not want or can update to this scheme, the directive
 288   `%define "global_tokens_and_yystype" "1"' triggers the global
 289   definition of tokens and YYSTYPE.  This change is suitable both
 290   for previous releases of Bison, and this one.
 291
 292   If you wish to update, then make sure older version of Bison will
 293   fail using `%require "2.2"'.
 294
 295 * DJGPP support added.
 296 \f
 297 Changes in version 2.1, 2005-09-16:
 298
 299 * The C++ lalr1.cc skeleton supports %lex-param.
 300
 301 * Bison-generated parsers now support the translation of diagnostics like
 302   "syntax error" into languages other than English.  The default
 303   language is still English.  For details, please see the new
 304   Internationalization section of the Bison manual.  Software
 305   distributors should also see the new PACKAGING file.  Thanks to
 306   Bruno Haible for this new feature.
 307
 308 * Wording in the Bison-generated parsers has been changed slightly to
 309   simplify translation.  In particular, the message "memory exhausted"
 310   has replaced "parser stack overflow", as the old message was not
 311   always accurate for modern Bison-generated parsers.
 312
 313 * Destructors are now called when the parser aborts, for all symbols left
 314   behind on the stack.  Also, the start symbol is now destroyed after a
 315   successful parse.  In both cases, the behavior was formerly inconsistent.
 316
 317 * When generating verbose diagnostics, Bison-generated parsers no longer
 318   quote the literal strings associated with tokens.  For example, for
 319   a syntax error associated with '%token NUM "number"' they might
 320   print 'syntax error, unexpected number' instead of 'syntax error,
 321   unexpected "number"'.
 322 \f
 323 Changes in version 2.0, 2004-12-25:
 324
 325 * Possibly-incompatible changes
 326
 327   - Bison-generated parsers no longer default to using the alloca function
 328     (when available) to extend the parser stack, due to widespread
 329     problems in unchecked stack-overflow detection.  You can "#define
 330     YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
 331     the manual to determine safe values for YYMAXDEPTH in that case.
 332
 333   - Error token location.
 334     During error recovery, the location of the syntax error is updated
 335     to cover the whole sequence covered by the error token: it includes
 336     the shifted symbols thrown away during the first part of the error
 337     recovery, and the lookahead rejected during the second part.
 338
 339   - Semicolon changes:
 340     . Stray semicolons are no longer allowed at the start of a grammar.
 341     . Semicolons are now required after in-grammar declarations.
 342
 343   - Unescaped newlines are no longer allowed in character constants or
 344     string literals.  They were never portable, and GCC 3.4.0 has
 345     dropped support for them.  Better diagnostics are now generated if
 346     forget a closing quote.
 347
 348   - NUL bytes are no longer allowed in Bison string literals, unfortunately.
 349
 350 * New features
 351
 352   - GLR grammars now support locations.
 353
 354   - New directive: %initial-action.
 355     This directive allows the user to run arbitrary code (including
 356     initializing @$) from yyparse before parsing starts.
 357
 358   - A new directive "%expect-rr N" specifies the expected number of
 359     reduce/reduce conflicts in GLR parsers.
 360
 361   - %token numbers can now be hexadecimal integers, e.g., `%token FOO 0x12d'.
 362     This is a GNU extension.
 363
 364   - The option `--report=lookahead' was changed to `--report=look-ahead'.
 365     [However, this was changed back after 2.3.]
 366
 367   - Experimental %destructor support has been added to lalr1.cc.
 368
 369   - New configure option --disable-yacc, to disable installation of the
 370     yacc command and -ly library introduced in 1.875 for POSIX conformance.
 371
 372 * Bug fixes
 373
 374   - For now, %expect-count violations are now just warnings, not errors.
 375     This is for compatibility with Bison 1.75 and earlier (when there are
 376     reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
 377     are too many or too few shift/reduce conflicts).  However, in future
 378     versions of Bison we plan to improve the %expect machinery so that
 379     these violations will become errors again.
 380
 381   - Within Bison itself, numbers (e.g., goto numbers) are no longer
 382     arbitrarily limited to 16-bit counts.
 383
 384   - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
 385 \f
 386 Changes in version 1.875, 2003-01-01:
 387
 388 * The documentation license has been upgraded to version 1.2
 389   of the GNU Free Documentation License.
 390
 391 * syntax error processing
 392
 393   - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
 394     locations too.  This fixes bugs in error-location computation.
 395
 396   - %destructor
 397     It is now possible to reclaim the memory associated to symbols
 398     discarded during error recovery.  This feature is still experimental.
 399
 400   - %error-verbose
 401     This new directive is preferred over YYERROR_VERBOSE.
 402
 403   - #defining yyerror to steal internal variables is discouraged.
 404     It is not guaranteed to work forever.
 405
 406 * POSIX conformance
 407
 408   - Semicolons are once again optional at the end of grammar rules.
 409     This reverts to the behavior of Bison 1.33 and earlier, and improves
 410     compatibility with Yacc.
 411
 412   - `parse error' -> `syntax error'
 413     Bison now uniformly uses the term `syntax error'; formerly, the code
 414     and manual sometimes used the term `parse error' instead.  POSIX
 415     requires `syntax error' in diagnostics, and it was thought better to
 416     be consistent.
 417
 418   - The documentation now emphasizes that yylex and yyerror must be
 419     declared before use.  C99 requires this.
 420
 421   - Bison now parses C99 lexical constructs like UCNs and
 422     backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
 423
 424   - File names are properly escaped in C output.  E.g., foo\bar.y is
 425     output as "foo\\bar.y".
 426
 427   - Yacc command and library now available
 428     The Bison distribution now installs a `yacc' command, as POSIX requires.
 429     Also, Bison now installs a small library liby.a containing
 430     implementations of Yacc-compatible yyerror and main functions.
 431     This library is normally not useful, but POSIX requires it.
 432
 433   - Type clashes now generate warnings, not errors.
 434
 435   - If the user does not define YYSTYPE as a macro, Bison now declares it
 436     using typedef instead of defining it as a macro.
 437     For consistency, YYLTYPE is also declared instead of defined.
 438
 439 * Other compatibility issues
 440
 441   - %union directives can now have a tag before the `{', e.g., the
 442     directive `%union foo {...}' now generates the C code
 443     `typedef union foo { ... } YYSTYPE;'; this is for Yacc compatibility.
 444     The default union tag is `YYSTYPE', for compatibility with Solaris 9 Yacc.
 445     For consistency, YYLTYPE's struct tag is now `YYLTYPE' not `yyltype'.
 446     This is for compatibility with both Yacc and Bison 1.35.
 447
 448   - `;' is output before the terminating `}' of an action, for
 449     compatibility with Bison 1.35.
 450
 451   - Bison now uses a Yacc-style format for conflict reports, e.g.,
 452     `conflicts: 2 shift/reduce, 1 reduce/reduce'.
 453
 454   - `yystype' and `yyltype' are now obsolescent macros instead of being
 455     typedefs or tags; they are no longer documented and are planned to be
 456     withdrawn in a future release.
 457
 458 * GLR parser notes
 459
 460   - GLR and inline
 461     Users of Bison have to decide how they handle the portability of the
 462     C keyword `inline'.
 463
 464   - `parsing stack overflow...' -> `parser stack overflow'
 465     GLR parsers now report `parser stack overflow' as per the Bison manual.
 466
 467 * Bison now warns if it detects conflicting outputs to the same file,
 468   e.g., it generates a warning for `bison -d -o foo.h foo.y' since
 469   that command outputs both code and header to foo.h.
 470
 471 * #line in output files
 472   - --no-line works properly.
 473
 474 * Bison can no longer be built by a K&R C compiler; it requires C89 or
 475   later to be built.  This change originally took place a few versions
 476   ago, but nobody noticed until we recently asked someone to try
 477   building Bison with a K&R C compiler.
 478 \f
 479 Changes in version 1.75, 2002-10-14:
 480
 481 * Bison should now work on 64-bit hosts.
 482
 483 * Indonesian translation thanks to Tedi Heriyanto.
 484
 485 * GLR parsers
 486   Fix spurious parse errors.
 487
 488 * Pure parsers
 489   Some people redefine yyerror to steal yyparse' private variables.
 490   Reenable this trick until an official feature replaces it.
 491
 492 * Type Clashes
 493   In agreement with POSIX and with other Yaccs, leaving a default
 494   action is valid when $$ is untyped, and $1 typed:
 495
 496         untyped: ... typed;
 497
 498   but the converse remains an error:
 499
 500         typed: ... untyped;
 501
 502 * Values of mid-rule actions
 503   The following code:
 504
 505         foo: { ... } { $$ = $1; } ...
 506
 507   was incorrectly rejected: $1 is defined in the second mid-rule
 508   action, and is equal to the $$ of the first mid-rule action.
 509 \f
 510 Changes in version 1.50, 2002-10-04:
 511
 512 * GLR parsing
 513   The declaration
 514      %glr-parser
 515   causes Bison to produce a Generalized LR (GLR) parser, capable of handling
 516   almost any context-free grammar, ambiguous or not.  The new declarations
 517   %dprec and %merge on grammar rules allow parse-time resolution of
 518   ambiguities.  Contributed by Paul Hilfinger.
 519
 520   Unfortunately Bison 1.50 does not work properly on 64-bit hosts
 521   like the Alpha, so please stick to 32-bit hosts for now.
 522
 523 * Output Directory
 524   When not in Yacc compatibility mode, when the output file was not
 525   specified, running `bison foo/bar.y' created `foo/bar.c'.  It
 526   now creates `bar.c'.
 527
 528 * Undefined token
 529   The undefined token was systematically mapped to 2 which prevented
 530   the use of 2 by the user.  This is no longer the case.
 531
 532 * Unknown token numbers
 533   If yylex returned an out of range value, yyparse could die.  This is
 534   no longer the case.
 535
 536 * Error token
 537   According to POSIX, the error token must be 256.
 538   Bison extends this requirement by making it a preference: *if* the
 539   user specified that one of her tokens is numbered 256, then error
 540   will be mapped onto another number.
 541
 542 * Verbose error messages
 543   They no longer report `..., expecting error or...' for states where
 544   error recovery is possible.
 545
 546 * End token
 547   Defaults to `$end' instead of `$'.
 548
 549 * Error recovery now conforms to documentation and to POSIX
 550   When a Bison-generated parser encounters a syntax error, it now pops
 551   the stack until it finds a state that allows shifting the error
 552   token.  Formerly, it popped the stack until it found a state that
 553   allowed some non-error action other than a default reduction on the
 554   error token.  The new behavior has long been the documented behavior,
 555   and has long been required by POSIX.  For more details, please see
 556   Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
 557   <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
 558
 559 * Traces
 560   Popped tokens and nonterminals are now reported.
 561
 562 * Larger grammars
 563   Larger grammars are now supported (larger token numbers, larger grammar
 564   size (= sum of the LHS and RHS lengths), larger LALR tables).
 565   Formerly, many of these numbers ran afoul of 16-bit limits;
 566   now these limits are 32 bits on most hosts.
 567
 568 * Explicit initial rule
 569   Bison used to play hacks with the initial rule, which the user does
 570   not write.  It is now explicit, and visible in the reports and
 571   graphs as rule 0.
 572
 573 * Useless rules
 574   Before, Bison reported the useless rules, but, although not used,
 575   included them in the parsers.  They are now actually removed.
 576
 577 * Useless rules, useless nonterminals
 578   They are now reported, as a warning, with their locations.
 579
 580 * Rules never reduced
 581   Rules that can never be reduced because of conflicts are now
 582   reported.
 583
 584 * Incorrect `Token not used'
 585   On a grammar such as
 586
 587            %token useless useful
 588            %%
 589            exp: '0' %prec useful;
 590
 591   where a token was used to set the precedence of the last rule,
 592   bison reported both `useful' and `useless' as useless tokens.
 593
 594 * Revert the C++ namespace changes introduced in 1.31
 595   as they caused too many portability hassles.
 596
 597 * Default locations
 598   By an accident of design, the default computation of @$ was
 599   performed after another default computation was performed: @$ = @1.
 600   The latter is now removed: YYLLOC_DEFAULT is fully responsible of
 601   the computation of @$.
 602
 603 * Token end-of-file
 604   The token end of file may be specified by the user, in which case,
 605   the user symbol is used in the reports, the graphs, and the verbose
 606   error messages instead of `$end', which remains being the default.
 607   For instance
 608      %token MYEOF 0
 609   or
 610      %token MYEOF 0 "end of file"
 611
 612 * Semantic parser
 613   This old option, which has been broken for ages, is removed.
 614
 615 * New translations
 616   Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
 617   Croatian, thanks to Denis Lackovic.
 618
 619 * Incorrect token definitions
 620   When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
 621
 622 * Token definitions as enums
 623   Tokens are output both as the traditional #define's, and, provided
 624   the compiler supports ANSI C or is a C++ compiler, as enums.
 625   This lets debuggers display names instead of integers.
 626
 627 * Reports
 628   In addition to --verbose, bison supports --report=THINGS, which
 629   produces additional information:
 630   - itemset
 631     complete the core item sets with their closure
 632   - lookahead [changed to `look-ahead' in 1.875e through 2.3, but changed back]
 633     explicitly associate lookahead tokens to items
 634   - solved
 635     describe shift/reduce conflicts solving.
 636     Bison used to systematically output this information on top of
 637     the report.  Solved conflicts are now attached to their states.
 638
 639 * Type clashes
 640   Previous versions don't complain when there is a type clash on
 641   the default action if the rule has a mid-rule action, such as in:
 642
 643       %type <foo> bar
 644       %%
 645       bar: '0' {} '0';
 646
 647   This is fixed.
 648
 649 * GNU M4 is now required when using Bison.
 650 \f
 651 Changes in version 1.35, 2002-03-25:
 652
 653 * C Skeleton
 654   Some projects use Bison's C parser with C++ compilers, and define
 655   YYSTYPE as a class.  The recent adjustment of C parsers for data
 656   alignment and 64 bit architectures made this impossible.
 657
 658   Because for the time being no real solution for C++ parser
 659   generation exists, kludges were implemented in the parser to
 660   maintain this use.  In the future, when Bison has C++ parsers, this
 661   kludge will be disabled.
 662
 663   This kludge also addresses some C++ problems when the stack was
 664   extended.
 665 \f
 666 Changes in version 1.34, 2002-03-12:
 667
 668 * File name clashes are detected
 669   $ bison foo.y -d -o foo.x
 670   fatal error: header and parser would both be named `foo.x'
 671
 672 * A missing `;' at the end of a rule triggers a warning
 673   In accordance with POSIX, and in agreement with other
 674   Yacc implementations, Bison will mandate this semicolon in the near
 675   future.  This eases the implementation of a Bison parser of Bison
 676   grammars by making this grammar LALR(1) instead of LR(2).  To
 677   facilitate the transition, this release introduces a warning.
 678
 679 * Revert the C++ namespace changes introduced in 1.31, as they caused too
 680   many portability hassles.
 681
 682 * DJGPP support added.
 683
 684 * Fix test suite portability problems.
 685 \f
 686 Changes in version 1.33, 2002-02-07:
 687
 688 * Fix C++ issues
 689   Groff could not be compiled for the definition of size_t was lacking
 690   under some conditions.
 691
 692 * Catch invalid @n
 693   As is done with $n.
 694 \f
 695 Changes in version 1.32, 2002-01-23:
 696
 697 * Fix Yacc output file names
 698
 699 * Portability fixes
 700
 701 * Italian, Dutch translations
 702 \f
 703 Changes in version 1.31, 2002-01-14:
 704
 705 * Many Bug Fixes
 706
 707 * GNU Gettext and %expect
 708   GNU Gettext asserts 10 s/r conflicts, but there are 7.  Now that
 709   Bison dies on incorrect %expectations, we fear there will be
 710   too many bug reports for Gettext, so _for the time being_, %expect
 711   does not trigger an error when the input file is named `plural.y'.
 712
 713 * Use of alloca in parsers
 714   If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
 715   malloc exclusively.  Since 1.29, but was not NEWS'ed.
 716
 717   alloca is used only when compiled with GCC, to avoid portability
 718   problems as on AIX.
 719
 720 * yyparse now returns 2 if memory is exhausted; formerly it dumped core.
 721
 722 * When the generated parser lacks debugging code, YYDEBUG is now 0
 723   (as POSIX requires) instead of being undefined.
 724
 725 * User Actions
 726   Bison has always permitted actions such as { $$ = $1 }: it adds the
 727   ending semicolon.  Now if in Yacc compatibility mode, the semicolon
 728   is no longer output: one has to write { $$ = $1; }.
 729
 730 * Better C++ compliance
 731   The output parsers try to respect C++ namespaces.
 732   [This turned out to be a failed experiment, and it was reverted later.]
 733
 734 * Reduced Grammars
 735   Fixed bugs when reporting useless nonterminals.
 736
 737 * 64 bit hosts
 738   The parsers work properly on 64 bit hosts.
 739
 740 * Error messages
 741   Some calls to strerror resulted in scrambled or missing error messages.
 742
 743 * %expect
 744   When the number of shift/reduce conflicts is correct, don't issue
 745   any warning.
 746
 747 * The verbose report includes the rule line numbers.
 748
 749 * Rule line numbers are fixed in traces.
 750
 751 * Swedish translation
 752
 753 * Parse errors
 754   Verbose parse error messages from the parsers are better looking.
 755   Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
 756      Now: parse error: unexpected '/', expecting "number" or '-' or '('
 757
 758 * Fixed parser memory leaks.
 759   When the generated parser was using malloc to extend its stacks, the
 760   previous allocations were not freed.
 761
 762 * Fixed verbose output file.
 763   Some newlines were missing.
 764   Some conflicts in state descriptions were missing.
 765
 766 * Fixed conflict report.
 767   Option -v was needed to get the result.
 768
 769 * %expect
 770   Was not used.
 771   Mismatches are errors, not warnings.
 772
 773 * Fixed incorrect processing of some invalid input.
 774
 775 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
 776
 777 * Fixed some typos in the documentation.
 778
 779 * %token MY_EOF 0 is supported.
 780   Before, MY_EOF was silently renumbered as 257.
 781
 782 * doc/refcard.tex is updated.
 783
 784 * %output, %file-prefix, %name-prefix.
 785   New.
 786
 787 * --output
 788   New, aliasing `--output-file'.
 789 \f
 790 Changes in version 1.30, 2001-10-26:
 791
 792 * `--defines' and `--graph' have now an optional argument which is the
 793   output file name. `-d' and `-g' do not change; they do not take any
 794   argument.
 795
 796 * `%source_extension' and `%header_extension' are removed, failed
 797   experiment.
 798
 799 * Portability fixes.
 800 \f
 801 Changes in version 1.29, 2001-09-07:
 802
 803 * The output file does not define const, as this caused problems when used
 804   with common autoconfiguration schemes.  If you still use ancient compilers
 805   that lack const, compile with the equivalent of the C compiler option
 806   `-Dconst='.  autoconf's AC_C_CONST macro provides one way to do this.
 807
 808 * Added `-g' and `--graph'.
 809
 810 * The Bison manual is now distributed under the terms of the GNU FDL.
 811
 812 * The input and the output files has automatically a similar extension.
 813
 814 * Russian translation added.
 815
 816 * NLS support updated; should hopefully be less troublesome.
 817
 818 * Added the old Bison reference card.
 819
 820 * Added `--locations' and `%locations'.
 821
 822 * Added `-S' and `--skeleton'.
 823
 824 * `%raw', `-r', `--raw' is disabled.
 825
 826 * Special characters are escaped when output.  This solves the problems
 827   of the #line lines with path names including backslashes.
 828
 829 * New directives.
 830   `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
 831   `%debug', `%source_extension' and `%header_extension'.
 832
 833 * @$
 834   Automatic location tracking.
 835 \f
 836 Changes in version 1.28, 1999-07-06:
 837
 838 * Should compile better now with K&R compilers.
 839
 840 * Added NLS.
 841
 842 * Fixed a problem with escaping the double quote character.
 843
 844 * There is now a FAQ.
 845 \f
 846 Changes in version 1.27:
 847
 848 * The make rule which prevented bison.simple from being created on
 849   some systems has been fixed.
 850 \f
 851 Changes in version 1.26:
 852
 853 * Bison now uses automake.
 854
 855 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
 856
 857 * Token numbers now start at 257 as previously documented, not 258.
 858
 859 * Bison honors the TMPDIR environment variable.
 860
 861 * A couple of buffer overruns have been fixed.
 862
 863 * Problems when closing files should now be reported.
 864
 865 * Generated parsers should now work even on operating systems which do
 866   not provide alloca().
 867 \f
 868 Changes in version 1.25, 1995-10-16:
 869
 870 * Errors in the input grammar are not fatal; Bison keeps reading
 871 the grammar file, and reports all the errors found in it.
 872
 873 * Tokens can now be specified as multiple-character strings: for
 874 example, you could use "<=" for a token which looks like <=, instead
 875 of chosing a name like LESSEQ.
 876
 877 * The %token_table declaration says to write a table of tokens (names
 878 and numbers) into the parser file.  The yylex function can use this
 879 table to recognize multiple-character string tokens, or for other
 880 purposes.
 881
 882 * The %no_lines declaration says not to generate any #line preprocessor
 883 directives in the parser file.
 884
 885 * The %raw declaration says to use internal Bison token numbers, not
 886 Yacc-compatible token numbers, when token names are defined as macros.
 887
 888 * The --no-parser option produces the parser tables without including
 889 the parser engine; a project can now use its own parser engine.
 890 The actions go into a separate file called NAME.act, in the form of
 891 a switch statement body.
 892 \f
 893 Changes in version 1.23:
 894
 895 The user can define YYPARSE_PARAM as the name of an argument to be
 896 passed into yyparse.  The argument should have type void *.  It should
 897 actually point to an object.  Grammar actions can access the variable
 898 by casting it to the proper pointer type.
 899
 900 Line numbers in output file corrected.
 901 \f
 902 Changes in version 1.22:
 903
 904 --help option added.
 905 \f
 906 Changes in version 1.20:
 907
 908 Output file does not redefine const for C++.
 909
 910 Local Variables:
 911 mode: outline
 912 End:
 913
 914 -----
 915
 916 Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
 917 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
 918
 919 This file is part of Bison, the GNU Compiler Compiler.
 920
 921 Bison is free software; you can redistribute it and/or modify
 922 it under the terms of the GNU General Public License as published by
 923 the Free Software Foundation; either version 2, or (at your option)
 924 any later version.
 925
 926 Bison is distributed in the hope that it will be useful,
 927 but WITHOUT ANY WARRANTY; without even the implied warranty of
 928 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 929 GNU General Public License for more details.
 930
 931 You should have received a copy of the GNU General Public License
 932 along with autoconf; see the file COPYING.  If not, write to
 933 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
 934 Boston, MA 02110-1301, USA.