newsBison - News


Bison 3.6 released

Item posted by Akim Demaille <akim> on Sat 09 May 2020 08:48:38 AM UTC.

We are extremely happy to announce the release of Bison 3.6:

- the developer can forge syntax error messages the way she wants.

- token string aliases can be internationalized, and UTF-8 sequences
  are properly preserved.

- push parsers can ask at any moment for the list of "expected tokens",
  which can be used to provide syntax-driven autocompletion.

- yylex may now tell the parser to enter error-recovery without issuing an
  error message (when the error was already reported by the scanner).

- several new examples were added, in particular "bistromathic" demonstrates
  almost all the existing bells and whistles, including interactive
  autocompletion on top of GNU readline.

Please see the much more detailed release notes below.

Many thanks to testers, bug reporters, contributors and feature requesters:
Adrian Vogelsgesang, Ahcheong Lee, Alexandre Duret-Lutz, Andy Fiddaman,
Angelo Borsotti, Arthur Schwarz, Christian Schoenebeck, Dagobert Michelsen,
Denis Excoffier, Dennis Clarke, Don Macpherson, Evan Lavelle, Frank
Heckenbach, Horst von Brand, Jannick, Nikki Valen, Paolo Bonzini, Paul
Eggert, Pramod Kumbhar and Victor Morales Cayuela.  The author also thanks
an anonymous reviewer for his precious comments.

Special thanks to Bruno Haible for his investment into making Bison

Happy parsing!


PS/ The experimental back-end for the D programming language is still
looking for active support from the D community.


GNU Bison is a general-purpose parser generator that converts an annotated
context-free grammar into a deterministic LR or generalized LR (GLR) parser
employing LALR(1) parser tables.  Bison can also generate IELR(1) or
canonical LR(1) parser tables.  Once you are proficient with Bison, you can
use it to develop a wide range of language parsers, from those used in
simple desk calculators to complex programming languages.

Bison is upward compatible with Yacc: all properly-written Yacc grammars
work with Bison with no change.  Anyone familiar with Yacc should be able to
use Bison with little trouble.  You need to be fluent in C, C++ or Java
programming in order to use Bison.

Bison and the parsers it generates are portable, they do not require any
specific compilers.

GNU Bison's home page is


Here are the compressed sources:   (5.1MB)   (3.1MB)

Here are the GPG detached signatures[*]:

Use a mirror for higher download bandwidth:

[*] Use a .sig file to verify that the corresponding file (without the
.sig suffix) is intact.  First, be sure to download both the .sig file
and the corresponding tarball.  Then, run a command like this:

  gpg --verify bison-3.6.tar.gz.sig

If that command fails because you don't have the required public key,
then run this command to import it:

  gpg --keyserver --recv-keys 0DDCAA3278D5264E

and rerun the 'gpg --verify' command.

This release was bootstrapped with the following tools:
  Autoconf 2.69
  Automake 1.16.2
  Flex 2.6.4
  Gnulib v0.1-3382-g2ac33b29f


* Noteworthy changes in release 3.6 (2020-05-08) [stable]

** Backward incompatible changes

  TL;DR: replace "#define YYERROR_VERBOSE 1" by "%define parse.error verbose".

  The YYERROR_VERBOSE macro is no longer supported; the parsers that still
  depend on it will now produce Yacc-like error messages (just "syntax
  error").  It was superseded by the "%error-verbose" directive in Bison
  1.875 (2003-01-01).  Bison 2.6 (2012-07-19) clearly announced that support
  for YYERROR_VERBOSE would be removed.  Note that since Bison 3.0
  (2013-07-25), "%error-verbose" is deprecated in favor of "%define
  parse.error verbose".

** Deprecated features

  The YYPRINT macro, which works only with yacc.c and only for tokens, was
  obsoleted long ago by %printer, introduced in Bison 1.50 (November 2002).
  It is deprecated and its support will be removed eventually.

** New features

*** Improved syntax error messages

  Two new values for the %define parse.error variable offer more control to
  the user.  Available in all the skeletons (C, C++, Java).

**** %define parse.error detailed

  The behavior of "%define parse.error detailed" is closely resembling that
  of "%define parse.error verbose" with a few exceptions.  First, it is safe
  to use non-ASCII characters in token aliases (with 'verbose', the result
  depends on the locale with which bison was run).  Second, a yysymbol_name
  function is exposed to the user, instead of the yytnamerr function and the
  yytname table.  Third, token internationalization is supported (see

**** %define parse.error custom

  With this directive, the user forges and emits the syntax error message
  herself by defining the yyreport_syntax_error function.  A new type,
  yypcontext_t, captures the circumstances of the error, and provides the
  user with functions to get details, such as yypcontext_expected_tokens to
  get the list of expected token kinds.

  A possible implementation of yyreport_syntax_error is:

    yyreport_syntax_error (const yypcontext_t *ctx)
      int res = 0;
      YY_LOCATION_PRINT (stderr, *yypcontext_location (ctx));
      fprintf (stderr, ": syntax error");
      // Report the tokens expected at this point.
        enum { TOKENMAX = 10 };
        yysymbol_kind_t expected[TOKENMAX];
        int n = yypcontext_expected_tokens (ctx, expected, TOKENMAX);
        if (n < 0)
          // Forward errors to yyparse.
          res = n;
          for (int i = 0; i < n; ++i)
            fprintf (stderr, "%s %s",
                     i == 0 ? ": expected" : " or", yysymbol_name (expected[i]));
      // Report the unexpected token.
        yysymbol_kind_t lookahead = yypcontext_token (ctx);
        if (lookahead != YYSYMBOL_YYEMPTY)
          fprintf (stderr, " before %s", yysymbol_name (lookahead));
      fprintf (stderr, "\n");
      return res;

**** Token aliases internationalization

  When the %define variable parse.error is set to `custom` or `detailed`,
  one may specify which token aliases are to be translated using _().  For

        PLUS   "+"
        MINUS  "-"
        NUM _("number")
        FUN _("function")
        VAR _("variable")

  In that case the user must define _() and N_(), and yysymbol_name returns
  the translated symbol (i.e., it returns '_("variable")' rather that
  '"variable"').  In Java, the user must provide an i18n() function.

*** List of expected tokens (yacc.c)

  Push parsers may invoke yypstate_expected_tokens at any point during
  parsing (including even before submitting the first token) to get the list
  of possible tokens.  This feature can be used to propose autocompletion
  (see below the "bistromathic" example).

  It makes little sense to use this feature without enabling LAC (lookahead

*** Returning the error token

  When the scanner returns an invalid token or the undefined token
  (YYUNDEF), the parser generates an error message and enters error
  recovery.  Because of that error message, most scanners that find lexical
  errors generate an error message, and then ignore the invalid input
  without entering the error-recovery.

  The scanners may now return YYerror, the error token, to enter the
  error-recovery mode without triggering an additional error message.  See
  the bistromathic for an example.

*** Deep overhaul of the symbol and token kinds

  To avoid the confusion with types in programming languages, we now refer
  to token and symbol "kinds" instead of token and symbol "types".  The
  documentation and error messages have been revised.

  All the skeletons have been updated to use dedicated enum types rather
  than integral types.  Special symbols are now regular citizens, instead of
  being declared in ad hoc ways.

**** Token kinds

  The "token kind" is what is returned by the scanner, e.g., PLUS, NUMBER,
  LPAREN, etc.  While backward compatibility is of course ensured, users are
  nonetheless invited to replace their uses of "enum yytokentype" by

  This type now also includes tokens that were previously hidden: YYEOF (end
  of input), YYUNDEF (undefined token), and YYerror (error token).  They
  now have string aliases, internationalized when internationalization is
  enabled.  Therefore, by default, error messages now refer to "end of file"
  (internationalized) rather than the cryptic "$end", or to "invalid token"
  rather than "$undefined".

  Therefore in most cases it is now useless to define the end-of-line token
  as follows:

    %token T_EOF 0 "end of file"

  Rather simply use "YYEOF" in your scanner.

**** Symbol kinds

  The "symbol kinds" is what the parser actually uses.  (Unless the
  api.token.raw %define variable is used, the symbol kind of a terminal
  differs from the corresponding token kind.)

  They are now exposed as a enum, "yysymbol_kind_t".

  This allows users to tailor the error messages the way they want, or to
  process some symbols in a specific way in autocompletion (see the
  bistromathic example below).

*** Modernize display of explanatory statements in diagnostics

  Since Bison 2.7, output was indented four spaces for explanatory
  statements.  For example:

    input.y:2.7-13: error: %type redeclaration for exp
    input.y:1.7-11:     previous declaration

  Since the introduction of caret-diagnostics, it became less clear.  This
  indentation has been removed and submessages are displayed similarly as in

    input.y:2.7-13: error: %type redeclaration for exp
        2 | %type <float> exp
          |       ^~~~~~~
    input.y:1.7-11: note: previous declaration
        1 | %type <int> exp
          |       ^~~~~

  Contributed by Victor Morales Cayuela.

*** C++

  The token and symbol kinds are yy::parser::token_kind_type and

  The symbol_type::kind() member function allows to get the kind of a
  symbol.  This can be used to write unit tests for scanners, e.g.,

    yy::parser::symbol_type t = make_NUMBER ("123");
    assert (t.kind () == yy::parser::symbol_kind::S_NUMBER);
    assert (<int> () == 123);

** Documentation

*** User Manual

  In order to avoid ambiguities with "type" as in "typing", we now refer to
  the "token kind" (e.g., `PLUS`, `NUMBER`, etc.) rather than the "token
  type".  We now also refer to the "symbol type" (e.g., `PLUS`, `expr`,

*** Examples

  There are now examples/java: a very simple calculator, and a more complete
  one (push-parser, location tracking, and debug traces).

  The lexcalc example (a simple example in C based on Flex and Bison) now
  also demonstrates location tracking.

  A new C example, bistromathic, is a fully featured interactive calculator
  using many Bison features: pure interface, push parser, autocompletion
  based on the current parser state (using yypstate_expected_tokens),
  location tracking, internationalized custom error messages, lookahead
  correction, rich debug traces, etc.

  It shows how to depend on the symbol kinds to tailor autocompletion.  For
  instance it recognizes the symbol kind "VARIABLE" to propose
  autocompletion on the existing variables, rather than of the word

Back to the top

Powered by Savane 3.13-b921.
Corresponding source code