version 3.5.91

* NEWS: Record release date.

.prev-version

@@ -1 +1 @@
-3.5
+3.5.90

.travis.yml

@@ -67,14 +67,13 @@ jobs:
     ## First.  ##
     ## ------- ##
 
-    # Start with three completely different environments, to get
-    # errors asap.
+    # Start with three completely different environments, to get errors asap.
 
     - name: "GCC 9 -O3"
       stage: check
       os: linux
       dist: bionic
-      addons:
+      addons: &gcc9
         apt:
           sources:
             # See https://github.com/travis-ci/apt-source-safelist/issues/410.
@@ -136,6 +135,49 @@ jobs:
           packages:
           - intel-oneapi-icc
 
+    ## ------- ##
+    ## First.  ##
+    ## ------- ##
+
+    # Start with three completely different environments, to get
+    # errors asap.
+
+    - name: "ARM64: GCC 9 -O3 part 1"
+      stage: check
+      os: linux
+      arch: arm64
+      dist: bionic
+      addons: *gcc9
+      env:
+        - CC=gcc-9
+        - CXX=g++-9
+        - CONFIGUREFLAGS='CPPFLAGS=-DNDEBUG CFLAGS=-O3 CXXFLAGS=-O3'
+        - PART=1
+
+    - name: "PPC64le: GCC 9 part 1"
+      stage: check
+      os: linux
+      arch: ppc64le
+      dist: bionic
+      addons: *gcc9
+      env:
+        - CC=gcc-9
+        - CXX=g++-9
+        - CONFIGUREFLAGS='CFLAGS=-O3 CXXFLAGS=-O3'
+        - PART=1
+
+    - name: "s390x: GCC 9 part 1"
+      stage: check
+      os: linux
+      arch: s390x
+      dist: bionic
+      addons: *gcc9
+      env:
+        - CC=gcc-9
+        - CXX=g++-9
+        - CONFIGUREFLAGS='CFLAGS=-O3 CXXFLAGS=-O3'
+        - PART=1
+
     ## ----- ##
     ## GCC.  ##
     ## ----- ##
@@ -437,11 +479,11 @@ script:
   - sudo apt-get install -qq doxygen flex m4
   # Install and activate dmd.
   - mkdir -p ~/dlang && wget https://dlang.org/install.sh -O ~/dlang/install.sh
-  - source $(source ~/dlang/install.sh dmd -a)
+  - source $(source ~/dlang/install.sh dmd -a) || true
 
   - $CC --version
   - $CXX --version
-  - dmd --version
+  - dmd --version || true
   - doxygen --version
   - flex --version
   - ld --version

Makefile.am

@@ -42,7 +42,7 @@ dependencies = $(BISON_IN) $(src_bison_SOURCES) $(dist_skeletons_DATA)
 BISON = $(top_builddir)/tests/bison
 BISON_IN = $(top_srcdir)/tests/bison.in
 YACC = $(BISON) -o y.tab.c
-AM_YFLAGS_WITH_LINES = --defines -Werror -Wall --report=all
+AM_YFLAGS_WITH_LINES = --defines -Werror -Wall,dangling-alias --report=all
 AM_YFLAGS = $(AM_YFLAGS_WITH_LINES) --no-lines
 
 # Initialization before completion by local.mk's.

NEWS

@@ -1,5 +1,275 @@
 GNU Bison NEWS
 
+* Noteworthy changes in release 3.5.91 (2020-04-29) [stable]
+
+** New features
+
+*** 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.
+
+*** The bistromathic features internationalization
+
+  Its way to build the error message is more general and is easy to use in
+  other projects.
+
+* Noteworthy changes in release 3.5.90 (2020-04-18) [beta]
+
+** 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
+  below).
+
+**** %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:
+
+    int
+    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;
+        else
+          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
+  instance
+
+    %token
+        PLUS   "+"
+        MINUS  "-"
+      <double>
+        NUM _("number")
+      <symrec*>
+        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
+  correction).
+
+*** 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
+  "yytoken_kind_t".
+
+  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
+  GCC:
+
+    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
+  yy::parser::symbol_kind_type.
+
+  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 (t.value.as<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`,
+  etc.).
+
+*** Examples
+
+  There are now two examples in examples/java: a very simple calculator, and
+  one that tracks locations to provide accurate error messages.
+
+  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
+  "variable".
+
+* Noteworthy changes in release 3.5.4 (2020-04-05) [stable]
+
+** WARNING: Future backward-incompatibilities!
+
+  TL;DR: replace "#define YYERROR_VERBOSE 1" by "%define parse.error verbose".
+
+  Bison 3.6 will no longer support the YYERROR_VERBOSE macro; the parsers
+  that still depend on it will 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".
+
+** Bug fixes
+
+  Fix portability issues of the package itself on old compilers.
+
+  Fix api.token.raw support in Java.
+
+* Noteworthy changes in release 3.5.3 (2020-03-08) [stable]
+
+** Bug fixes
+
+  Error messages could quote lines containing zero-width characters (such as
+  \005) with incorrect styling.  Fixes for similar issues with unexpectedly
+  short lines (e.g., the file was changed between parsing and diagnosing).
+
+  Several unlikely crashes found by fuzzing have been fixed.
+
+* Noteworthy changes in release 3.5.2 (2020-02-13) [stable]
+
+** Bug fixes
+
+  Portability issues and minor cosmetic issues.
+
+  The lalr1.cc skeleton properly rejects unsupported values for parse.lac
+  (as yacc.c does).
+
 * Noteworthy changes in release 3.5.1 (2020-01-19) [stable]
 
 ** Bug fixes
@@ -3867,7 +4137,11 @@ along with this program.  If not, see <http://www.gnu.org/licenses/>.
  LocalWords:  Wdeprecated yytext Variadic variadic yyrhs yyphrs RCS README
  LocalWords:  noexcept constexpr ispell american deprecations backend Teoh
  LocalWords:  YYPRINT Mangold Bonzini's Wdangling exVal baz checkable gcc
- LocalWords:  fsanitize Vogelsgesang lis redeclared stdint automata
+ LocalWords:  fsanitize Vogelsgesang lis redeclared stdint automata yytname
+ LocalWords:  yysymbol yytnamerr yyreport ctx ARGMAX yysyntax stderr LPAREN
+ LocalWords:  symrec yypcontext TOKENMAX yyexpected YYEMPTY yypstate YYEOF
+ LocalWords:  autocompletion bistromathic submessages Cayuela lexcalc hoc
+ LocalWords:  yytoken YYUNDEF YYerror
 
 Local Variables:
 ispell-dictionary: "american"

README

@@ -1,4 +1,19 @@
-This package contains the GNU Bison parser generator.
+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 https://gnu.org/software/bison/.
 
 # Installation
 ## Build from git
@@ -17,13 +32,18 @@ See the file INSTALL for generic compilation and installation instructions.
 Bison requires GNU m4 1.4.6 or later.  See
 https://ftp.gnu.org/gnu/m4/m4-1.4.6.tar.gz.
 
+## Running a non installed bison
+Once you ran `make`, you might want to toy with this fresh bison before
+installing it.  In that case, do not use `src/bison`: it would use the
+*installed* files (skeletons, etc.), not the local ones.  Use `tests/bison`.
+
 ## Colored diagnostics
 As an experimental feature, diagnostics are now colored, controlled by the
 `--color` and `--style` options.
 
 To use them, install the libtextstyle library before configuring Bison.  It
 is available from https://alpha.gnu.org/gnu/gettext/, for instance
-https://alpha.gnu.org/gnu/gettext/libtextstyle-0.8.tar.gz.
+https://alpha.gnu.org/pub/gnu/gettext/libtextstyle-0.20.5.tar.gz.
 
 The option --color supports the following arguments:
 - always, yes: Enable colors.
@@ -85,18 +105,12 @@ Software Foundation, Inc.
 
 This file is part of GNU bison, the GNU Compiler Compiler.
 
-This program is free software: you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation, either version 3 of the License, or
-(at your option) any later version.
-
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
 
 Local Variables:
 mode: markdown
@@ -104,7 +118,8 @@ fill-column: 76
 ispell-dictionary: "american"
 End:
 
-LocalWords:  parsers ngettext Texinfo pdf html YYYY ZZZZ ispell american
-LocalWords:  MERCHANTABILITY
+LocalWords:  parsers ngettext Texinfo pdf html YYYY ZZZZ ispell american md
+LocalWords:  MERCHANTABILITY GLR LALR IELR submodule init README src bw
+LocalWords:  Relocatability symlinks symlink
 
 -->

README-hacking.md

@@ -5,7 +5,8 @@ Everything related to the development of Bison is on Savannah:
 http://savannah.gnu.org/projects/bison/.
 
 
-# Administrivia
+Administrivia
+=============
 
 ## If you incorporate a change from somebody on the net:
 First, if it is a large change, you must make sure they have signed the
@@ -23,9 +24,11 @@ demonstrates the bug.  Then fix the bug, re-run the test suite, and check
 everything in.
 
 
-# Hacking
 
-## Visible changes
+Hacking
+=======
+
+## Visible Changes
 Which include serious bug fixes, must be mentioned in NEWS.
 
 ## Translations
@@ -34,15 +37,121 @@ Only user visible strings are to be translated: error messages, bits of the
 assert/abort), and all the --trace output which is meant for the maintainers
 only.
 
-## Horizontal tabs
+## Syntax highlighting
+It's quite nice to be in C++ mode when editing lalr1.cc for instance.
+However tools such as Emacs will be fooled by the fact that braces and
+parens do not nest, as in `[[}]]`.  As a consequence you might be misguided
+by its visual pairing to parens.  The m4-mode is safer.  Unfortunately the
+m4-mode is also fooled by `#` which is sees as a comment, stops pairing with
+parens/brackets that are inside...
+
+## Coding Style
 Do not add horizontal tab characters to any file in Bison's repository
 except where required.  For example, do not use tabs to format C code.
 However, make files, ChangeLog, and some regular expressions require tabs.
 Also, test cases might need to contain tabs to check that Bison properly
 processes tabs in its input.
 
+Prefer "res" as the name of the local variable that will be "return"ed by
+the function.
 
-# Working from the repository
+### Bison
+Follow the GNU Coding Standards.
+
+Don't reinvent the wheel: we use gnulib, which features many components.
+Actually, Bison has legacy code that we should replace with gnulib modules
+(e.g., many ad hoc implementations of lists).
+
+### Skeletons
+We try to use the "typical" coding style for each language.
+
+#### CPP
+We indent the CPP directives this way:
+
+```
+#if FOO
+# if BAR
+#  define BAZ
+# endif
+#endif
+```
+
+Don't indent with leading spaces in the skeletons (it's ok in the grammar
+files though, e.g., in `%code {...}` blocks).
+
+On occasions, use `cppi -c` to see where we stand.  We don't aim at full
+correctness: depending `-d`, some bits can be in the *.c file, or the *.h
+file within the double-inclusion cpp-guards.  In that case, favor the case
+of the *.h file, but don't waste time on this.
+
+Don't hesitate to leave a comment on the `#endif` (e.g., `#endif /* FOO
+*/`), especially for long blocks.
+
+There is no conistency on `! defined` vs. `!defined`.  The day gnulib
+decides, we'll follow them.
+
+#### C/C++
+Follow the GNU Coding Standards.
+
+The `glr.c` skeleton was implemented with `camlCase`.  We are migrating it
+to `snake_case`.  Because we are standardizing the code, it is currently
+inconsistent.
+
+Use `YYFOO` and `yyfoo` for entities that are exposed to the user.  They are
+part of our contract with the users wrt backward compatibility.
+
+Use `YY_FOO` and `yy_foo` for private matters.  Users should not use them,
+we are free to change them without fear of backward compatibility issues.
+
+Use `*_t` for types, especially for `yy*_t` in which case we shouldn't worry
+about the C standard introducing such a name.
+
+#### C++
+Follow the C++ Core Guidelines
+(http://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines).  The Google
+ones may be interesting too
+(https://google.github.io/styleguide/cppguide.html).
+
+Our enumerators, such as the kinds (symbol and token kinds), should be lower
+case, but it was too late to follow that track for token kinds, and symbol
+kind enumerators are made to be consistent with them.
+
+Use `*_type` for type aliases.  Use `foo_get()` and `foo_set(v)` for
+accessors, or simply `foo()` and `foo(v)`.
+
+Use the `yy` prefix for private stuff, but there's no need for it in the
+public API.  The `yy` prefix is already taken care of via the namespace.
+
+#### Java
+We follow https://www.oracle.com/technetwork/java/codeconventions-150003.pdf
+and https://google.github.io/styleguide/javaguide.html.  Unfortunately at
+some point some GNU Coding Style was installed in Java, but it's an error.
+So we should for instance stop putting spaces in function calls.  Because we
+are standardizing the code, it is currently inconsistent.
+
+Use a 2-space indentation (Google) rather than 4 (Oracle).
+
+Don't use the "yy" prefix for public members: "getExpectedTokens", not
+"yyexpectedTokens" or "yygetExpectedTokens".
+
+## Commit Messages
+Imitate the style we use.  Use `git log` to get sources of inspiration.
+
+If the changes have a small impact on Bison's generated parser, embed these
+changes in the commit itself.  If the impact is large, first push all the
+changes except those about src/parse-gram.[ch], and then another commit
+named "regen" which is only about them.
+
+## Debugging
+Bison supports tracing of its various steps, via the `--trace` option.
+Since it is not meant for the end user, it is not displayed by `bison
+--help`, nor is it documented in the manual.  Instead, run `bison
+--trace=help`.
+
+
+
+Working from the Repository
+===========================
 
 These notes intend to help people working on the checked-out sources.  These
 requirements do not apply when building from a distribution tarball.
@@ -185,7 +294,36 @@ These files don't change very often in Autoconf, so it should be relatively
 straight-forward to examine the differences in order to decide whether to
 update.
 
-# Test Suite
+## Troubleshooting
+
+Bison is self-hosted: its parser is generated by Bison.  We don't force
+ourselves to use the previous release of Bison, we use the current git
+master for several reasons:
+- dogfooding: let Bison be its first user
+- monitoring: seeing the diff on the generated parsers with git is very
+  helpful, as it allows to see easily the impact of the changes on a real
+  case parser.
+
+If you are unlucky the generated files, src/parse-gram.[ch], may be older
+than their source, src/parse-gram.y.  And your current version of Bison
+might not be able to grok it.  In that case, first refresh the generated
+files:
+
+    $ touch src/parse-gram.[ch]
+
+Then proceed.
+
+In case you wrecked your current copy of the parser, get back the previous
+version, compile bison, then force it to recreate the files:
+
+    $ git checkout HEAD^ src/parse-gram.[ch]
+    $ make -C _build
+    $ touch src/parse-gram.y
+    $ make -C _build
+
+
+Test Suite
+==========
 
 ## make check
 Consume without moderation.  It is composed of two kinds of tests: the
@@ -265,6 +403,12 @@ again, since some early failures may stop latter tests from being run.  Yet
 at some point, you'll have to fix remaining issues by hand...
 
 
+## Running Java parsers
+Use the `javaexec.sh` script.  For instance to run the parser of test case
+504:
+
+    $ sh ./_build/javaexec.sh -cp ./_build/tests/testsuite.dir/504 Calc
+
 ## make maintainer-check-valgrind
 This target uses valgrind both to check bison, and the generated parsers.
 
@@ -339,172 +483,13 @@ re-run the tests, run:
     ./configure -C CC='gcc-mp-8 -fsanitize=undefined' CFLAGS='-ggdb'
     make check
 
-# Release Procedure
-This section needs to be updated to take into account features from gnulib.
-In particular, be sure to read README-release.
 
-## Update the submodules.  See above.
 
-## Update maintainer tools, such as Autoconf.  See above.
-
-## Try to get the *.pot files to the Translation Project at least one
-week before a stable release, to give them time to translate them.  Before
-generating the *.pot files, make sure that po/POTFILES.in and
-runtime-po/POTFILES.in list all files with translatable strings.  This
-helps: `grep -l '\<_(' *`.
-
-## Tests
-See above.
-
-## Update the foreign files
-Running `./bootstrap` in the top level should update them all for you.  This
-covers PO files too.  Sometimes a PO file contains problems that causes it
-to be rejected by recent Gettext releases; please report these to the
-Translation Project.
-
-## Update README
-Make sure the information in README is current.  Most notably, make sure it
-recommends a version of GNU M4 that is compatible with the latest Bison
-sources.
-
-## Check copyright years.
-We update years in copyright statements throughout Bison once at the start
-of every year by running `make update-copyright`.  However, before a
-release, it's good to verify that it's actually been run.  Besides the
-copyright statement for each Bison file, check the copyright statements that
-the skeletons insert into generated parsers, and check all occurrences of
-`PACKAGE_COPYRIGHT_YEAR` in configure.ac.
-
-## Update NEWS, commit and tag.
-See do-release-commit-and-tag in README-release.  For a while, we used beta
-names such as `2.6_rc1`.  Now that we use gnulib in the release procedure,
-we must use `2.5.90`, which has the additional benefit of being properly
-sorted in `git tag -l`.
-
-## make alpha, beta, or stable
-See README-release.
-
-## Upload
-There are two ways to upload the tarballs to the GNU servers: using gnupload
-(from gnulib), or by hand.  Obviously prefer the former.  But in either
-case, be sure to read the following paragraph.
-
-### Setup
-You need `gnupg`.
-
-Make sure your public key has been uploaded at least to keys.gnupg.net.  You
-can upload it with:
-
-  gpg --keyserver keys.gnupg.net --send-keys F125BDF3
-
-where F125BDF3 should be replaced with your key ID.
-
-### Using gnupload
-You need `ncftp`.
-
-At the end `make stable` (or alpha/beta) will display the procedure to run.
-Just copy and paste it in your shell.
-
-### By hand
-
-The generic GNU upload procedure is at
-http://www.gnu.org/prep/maintain/maintain.html#Automated-FTP-Uploads.
-
-Follow the instructions there to register your information so you're permitted
-to upload.
-
-Here's a brief reminder of how to roll the tarballs and upload them:
-
-### make distcheck
-### gpg -b bison-2.3b.tar.gz
-### In a file named `bison-2.3b.tar.gz.directive`, type:
-
-    version: 1.1
-    directory: bison
-    filename: bison-2.3b.tar.gz
-
-### gpg --clearsign bison-2.3b.tar.gz.directive
-### ftp ftp-upload.gnu.org # Log in as anonymous.
-### cd /incoming/alpha # cd /incoming/ftp for full release.
-### put bison-2.3b.tar.gz # This can take a while.
-### put bison-2.3b.tar.gz.sig
-### put bison-2.3b.tar.gz.directive.asc
-### Repeat all these steps for bison-2.3b.tar.xz.
-
-## Update Bison manual on www.gnu.org.
-
-The instructions below are obsolete, and left in case one would like to run
-the commands by hand.  Today, one just needs to run
-
-    $ make web-manual-update
+Release Procedure
+=================
 
 See README-release.
 
-### You need a non-anonymous checkout of the web pages directory.
-
-    $ cvs -d YOUR_USERID@cvs.savannah.gnu.org:/web/bison checkout bison
-
-### Get familiar with the instructions for web page maintainers.
-http://www.gnu.org/server/standards/readme_index.html
-http://www.gnu.org/server/standards/README.software.html
-especially the note about symlinks.
-
-### Build the web pages.
-Assuming BISON_CHECKOUT refers to a checkout of the Bison dir, and
-BISON_WWW_CHECKOUT refers to the web directory created above, do:
-
-    $ cd $BISON_CHECKOUT/doc
-    $ make stamp-vti
-    $ ../build-aux/gendocs.sh -o "$BISON_WWW_CHECKOUT/manual" \
-      bison "Bison - GNU parser generator"
-    $ cd $BISON_WWW_CHECKOUT
-
-Verify that the result looks sane.
-
-### Commit the modified and the new files.
-
-### Remove old files.
-Find the files which have not been overwritten (because they belonged to
-sections that have been removed or renamed):
-
-     $ cd manual/html_node
-     $ ls -lt
-
-Remove these files and commit their removal to CVS.  For each of these
-files, add a line to the file .symlinks.  This will ensure that hyperlinks
-to the removed files will redirect to the entire manual; this is better than
-a 404 error.
-
-## Announce
-The "make release" command just created a template,
-`$HOME/announce-bison-X.Y`.  Otherwise, to generate it, run:
-
-    make RELEASE_TYPE=alpha gpg_key_ID=F125BDF3 announcement
-
-where alpha can be replaced by `beta` or `table` and F125BDF3 should be
-replaced with your key ID.
-
-Complete/fix the announcement file.  The generated list of recipients
-(info-gnu@gnu.org, bison-announce@gnu.org, bug-bison@gnu.org,
-help-bison@gnu.org, bison-patches@gnu.org, and
-coordinator@translationproject.org) is appropriate for a stable release or a
-"serious beta".  For any other release, drop at least info-gnu@gnu.org.  For
-an example of how to fill out the rest of the template, search the mailing
-list archives for the most recent release announcement.
-
-For a stable release, send the same announcement on the comp.compilers
-newsgroup by sending email to compilers@iecc.com.  Do not make any Cc as the
-moderator will throw away anything cross-posted or Cc'ed.  It really needs
-to be a separate message.
-
-## Prepare NEWS
-So that developers don't accidentally add new items to the old NEWS entry,
-create a new empty entry in line 3 (without the two leading spaces):
-
-  * Noteworthy changes in release ?.? (????-??-??) [?]
-
-Push these changes.
-
 <!--
 
 Copyright (C) 2002-2005, 2007-2015, 2018-2020 Free Software Foundation,
@@ -536,9 +521,11 @@ LocalWords:  submodule init cd distcheck ChangeLog valgrind sigreturn sudo
 LocalWords:  UC gcc DGNULIB POSIXCHECK xml XSLT glr lalr README po runtime rc
 LocalWords:  gnupload gnupg gpg keyserver BDF ncftp filename clearsign cvs dir
 LocalWords:  symlinks vti html lt POSIX Cc'ed Graphviz Texinfo autoconf jN
-LocalWords:  automake autopoint graphviz texinfo PROG Wother parsers
-LocalWords:  TESTSUITEFLAGS deprec struct gnulib's getopt config ggdb
-LocalWords:  bitset fsanitize symlink CFLAGS MERCHANTABILITY ispell
-LocalWords:  american
+LocalWords:  automake autopoint graphviz texinfo PROG Wother parsers YYFOO
+LocalWords:  TESTSUITEFLAGS deprec struct gnulib's getopt config ggdb yyfoo
+LocalWords:  bitset fsanitize symlink CFLAGS MERCHANTABILITY ispell wrt YY
+LocalWords:  american Administrivia camlCase yy accessors namespace src
+LocalWords:  getExpectedTokens yyexpectedTokens yygetExpectedTokens
+LocalWords:  regen dogfooding Autotest testsuite
 
 -->

THANKS

@@ -4,8 +4,10 @@ it is today without the invaluable help of these people:
 Aaro Koskinen             [email protected]
 Аскар Сафин               [email protected]
 Adam Sampson              [email protected]
+Adrian Vogelsgesang       [email protected]
+Ahcheong Lee              [email protected]
 Airy Andre                [email protected]
-Akim Demaille             akim@lrde.epita.fr
+Akim Demaille             akim@gnu.org
 Albert Chin-A-Young       [email protected]
 Alexander Belopolsky      [email protected]
 Alexandre Duret-Lutz      [email protected]
@@ -60,6 +62,7 @@ Enrico Scholz             [email protected]
 Eric Blake                [email protected]
 Eric S. Raymond           [email protected]
 Étienne Renault           [email protected]
+Evan Lavelle              [email protected]
 Evan Nemerson             [email protected]
 Evgeny Stambulchik        [email protected]
 Fabrice Bauzac            [email protected]
@@ -130,6 +133,7 @@ Nick Bowler               [email protected]
 Nicolas Bedon             [email protected]
 Nicolas Burrus            [email protected]
 Nicolas Tisserand         [email protected]
+Nikki Valen               [email protected]
 Noah Friedman             [email protected]
 Odd Arild Olsen           [email protected]
 Oleg Smolsky              [email protected]

TODO

@@ -1,121 +1,78 @@
 * Bison 3.6
+** Questions
+*** Java
+- Should i18n be part of the Lexer?  Currently it's a static method of
+  Lexer.
+
+- is there a migration path that would allow to use TokenKinds in
+  yylex?
+
+- define the tokens as an enum too.
+
+- promote YYEOF rather than EOF.
+
+*** D
+- is there a way to attach yysymbol_name to the enum itself?  As we did
+  in Java.
+
+- It would be better to have TokenKind as return value.  Can we use
+  reflection to support both output types?
+
+** YYerror
+https://git.savannah.gnu.org/gitweb/?p=gettext.git;a=blob;f=gettext-runtime/intl/plural.y;h=a712255af4f2f739c93336d4ff6556d932a426a5;hb=HEAD
+
+should be updated to not use YYERRCODE.  Returning an undef token is good
+enough.
+
+** Java
+*** Examples
+Have an example with a push parser.  Use autocompletion in that case.
+
+*** calc.at
+Stop hard-coding "Calc".  Adjust local.at (look for FIXME).
+
+* Bison 3.7
+** Counter example generation
+See https://github.com/akimd/bison/pull/15.
+
+** Clean up
+Rename user_token_number for tokens as "code".  It's not a "user number",
+it's the token code, and the user can control it, but this code always
+exists.
+
+See also USER_NUMBER_UNDEFINED and NUMBER_UNDEFINED.
+
+Rename endtoken as eoftoken.
+
+Also do it in data/skeletons.
+
+Don't rename in Bison 3.6 (it would be logical to do so) because that
+would probably create many conflicts in Vincent's work (see previous point).
+
+** A dev warning for b4_
+Maybe we should check for m4_ and b4_ leaking out of the m4 processing, as
+Autoconf does.  It would have caught overquotation issues.
+
 ** doc
-I feel its ugly to use the GNU style to declare functions in the doc.  It
+I feel it's ugly to use the GNU style to declare functions in the doc.  It
 generates tons of white space in the page, and may contribute to bad page
 breaks.
 
-Also, we seem to teach YYPRINT very early on, although it should be
-considered deprecated: %printer is superior.
-
-** improve syntax errors (UTF-8, internationalization)
-Bison depends on the current locale.  For instance:
-
-%define parse.error verbose
-%code top {
- #include <stdio.h>
- #include <stdlib.h>
- void yyerror(const char* msg) { fprintf(stderr, "%s\n", msg); }
- int yylex() { return 0; }
-}
-%%
-exp: "↦" | "🎅🐃" | '\n'
-%%
-int main() { return yyparse(); }
-
-gives different results with/without LC_ALL=C.
-
-$ LC_ALL=C /opt/local/bin/bison /tmp/mangle.y -o ascii.c
-$ /opt/local/bin/bison /tmp/mangle.y -o utf8.c
-$ diff -u ascii.c utf8.c -I#line
---- ascii.c     2019-01-12 08:15:35.878010093 +0100
-+++ utf8.c      2019-01-12 08:15:38.856495929 +0100
-@@ -415,9 +415,8 @@
-    First, the terminals, then, starting at YYNTOKENS, nonterminals.  */
- static const char *const yytname[] =
- {
--  "$end", "error", "$undefined", "\"\\342\\206\\246\"",
--  "\"\\360\\237\\216\\205\\360\\237\\220\\203\"", "'\\n'", "$accept",
--  "exp", YY_NULLPTR
-+  "$end", "error", "$undefined", "\"↦\"", "\"🎅🐃\"", "'\\n'",
-+  "$accept", "exp", YY_NULLPTR
- };
- #endif
-
-$ gcc ascii.c -o ascii && ./ascii
-syntax error, unexpected $end, expecting "\342\206\246" or "\360\237\216\205\360\237\220\203" or '\n'
-$ gcc utf8.c -o utf8 && ./utf8
-syntax error, unexpected $end, expecting ↦ or 🎅🐃 or '\n'
-
-
-While at it, we should stop using "$end" by default, in favor of "end of
-file", or "end of input", whatever.  See how lalr1.java does that.
-
 ** consistency
 token vs terminal, variable vs non terminal.
 
-** Stop indentation in diagnostics
-Before Bison 2.7, we printed "flatly" the dependencies in long diagnostics:
-
-    input.y:2.7-12: %type redeclaration for exp
-    input.y:1.7-12: previous declaration
-
-In Bison 2.7, we indented them
-
-    input.y:2.7-12: error: %type redeclaration for exp
-    input.y:1.7-12:     previous declaration
-
-Later we quoted the source in the diagnostics, and today we have:
-
-    /tmp/foo.y:1.12-14: warning: symbol FOO redeclared [-Wother]
-        1 | %token FOO FOO
-          |            ^~~
-    /tmp/foo.y:1.8-10:      previous declaration
-        1 | %token FOO FOO
-          |        ^~~
-
-The indentation is no longer helping.  We should probably get rid of it, or
-maybe keep it only when -fno-caret. GCC displays this as a "note":
-
-    $ g++-mp-9 -Wall /tmp/foo.c -c
-    /tmp/foo.c:1:10: error: redefinition of 'int foo'
-        1 | int foo, foo;
-          |          ^~~
-    /tmp/foo.c:1:5: note: 'int foo' previously declared here
-        1 | int foo, foo;
-          |     ^~~
-
-Likewise for Clang, contrary to what I believed (because "note:" is written
-in black, so it doesn't show in my terminal :-)
-
-    $ clang++-mp-8.0 -Wall /tmp/foo.c -c
-    clang: warning: treating 'c' input as 'c++' when in C++ mode, this behavior is deprecated [-Wdeprecated]
-    /tmp/foo.c:1:10: error: redefinition of 'foo'
-    int foo, foo;
-             ^
-    /tmp/foo.c:1:5: note: previous definition is here
-    int foo, foo;
-        ^
-    1 error generated.
-
-See also the item "Complaint submessage indentation" below.
-
 ** api.token.raw
-Maybe we should exhibit the YYUNDEFTOK token.  It could also be assigned a
-semantic value so that yyerror could be used to report invalid lexemes.
-See also the item "$undefined" below.
+The YYUNDEFTOK could be assigned a semantic value so that yyerror could be
+used to report invalid lexemes.
 
-** C++
-Move to int everywhere instead of unsigned?  stack_size, etc.  The parser
-itself uses int (for yylen for instance), yet stack is based on size_t.
+** push parsers
+Consider deprecating impure push parsers.  They add a lot of complexity, for
+a bad feature.  On the other hand, that would make it much harder to sit
+push parsers on top of pull parser.  Which is currently not relevant, since
+push parsers are measurably slower.
 
-Maybe locations should also move to ints.
 
-Paul Eggert already covered most of this.  But before publishing these
-changes, we need to ask our C++ users if they agree with that change, or if
-we need some migration path.  Could be a %define variable, or simply
-%require "3.5".
-
-* Bison 3.7
+* Bison 3.8
 ** Unit rules / Injection rules (Akim Demaille)
 Maybe we could expand unit rules (or "injections", see
 https://homepages.cwi.nl/~daybuild/daily-books/syntax/2-sdf/sdf.html), i.e.,
@@ -170,6 +127,16 @@ functions.
 states/nstates, rules/nrules, ..., ritem/nritems
 Fix the latter.
 
+* Better error messages
+The users are not provided with enough tools to forge their error messages.
+See for instance "Is there an option to change the message produced by
+YYERROR_VERBOSE?" by Simon Sobisch, on bison-help.
+
+See also
+https://www.cs.tufts.edu/~nr/cs257/archive/clinton-jefferey/lr-error-messages.pdf
+https://research.swtch.com/yyerror
+http://gallium.inria.fr/~fpottier/publis/fpottier-reachability-cc2016.pdf
+
 * Modernization
 Fix data/skeletons/yacc.c so that it defines YYPTRDIFF_T properly for modern
 and older C++ compilers.  Currently the code defaults to defining it to
@@ -179,10 +146,10 @@ define it to the same type as the C ptrdiff_t type.
 * Completion
 Several features are not available in all the backends.
 
-- push parsers: glr.cc, lalr1.cc
-- ielr: C++ and Java
-- glr: Java
-- token constructors: Java and C
+- push parsers: glr.c, glr.cc, lalr1.cc
+- lac: D, Java
+- glr: D, Java
+- token constructors: Java, C, D
 
 * Bugs
 ** Autotest has quotation issues
@@ -198,8 +165,8 @@ $ ./tests/testsuite -l | grep errors | sed q
 Besides yytoknum is wrong when api.token.raw is defined.
 
 ** Better design for diagnostics
-The current implementation of diagnostics is adhoc, it grew organically.  It
-works as a series of calls to several functions, with dependency of the
+The current implementation of diagnostics is ad hoc, it grew organically.
+It works as a series of calls to several functions, with dependency of the
 latter calls on the former.  For instance:
 
       complain (&sym->location,
@@ -314,52 +281,7 @@ It would be a very nice source of inspiration for the other languages.
 
 Valentin Tolmer is working on this.
 
-** YYERRCODE
-Defined to 256, but not used, not documented.  Probably the token
-number for the error token, which POSIX wants to be 256, but which
-Bison might renumber if the user used number 256.  Keep fix and doc?
-Throw away?
-
-Also, why don't we output the token name of the error token in the
-output?  It is explicitly skipped:
-
-      /* Skip error token and tokens without identifier.  */
-      if (sym != errtoken && id)
-
-Of course there are issues with name spaces, but if we disable we have
-something which seems to be more simpler and more consistent instead
-of the special case YYERRCODE.
-
-   enum yytokentype {
-     error = 256,
-     // ...
-   };
-
-
-We could (should?) also treat the case of the undef_token, which is
-numbered 257 for yylex, and 2 internal.  Both appear for instance in
-toknum:
-
-  const unsigned short
-  parser::yytoken_number_[] =
-  {
-       0,   256,   257,   258,   259,   260,   261,   262,   263,   264,
-
-while here
-
-   enum yytokentype {
-     TOK_EOF = 0,
-     TOK_EQ = 258,
-
-so both 256 and 257 are "mysterious".
-
-  const char*
-  const parser::yytname_[] =
-  {
-  "\"end of command\"", "error", "$undefined", "\"=\"", "\"break\"",
-
-
-** yychar == yyempty_
+** yychar == YYEMPTY
 The code in yyerrlab reads:
 
       if (yychar <= YYEOF)
@@ -391,14 +313,6 @@ We need trustworthy benchmarks for Bison, for all our backends.  Akim has a
 few things scattered around; we need to put them in the repo, and make them
 more useful.
 
-** yysyntax_error
-The code bw glr.c and yacc.c is really alike, we can certainly factor
-some parts.
-
-This should be worked on when we also address the expected improvements for
-error generation (e.g., i18n).
-
-
 * Report
 
 ** Figures
@@ -436,6 +350,11 @@ An Experimental Ambiguity Detection Tool ∗ Sylvain Schmitz
 LORIA, INRIA Nancy - Grand Est, Nancy, France
 
 * Extensions
+** More languages?
+Well, only if there is really some demand for it.
+
+https://github.com/scfc/bison-php/blob/master/data/lalr1.php
+
 ** Multiple start symbols
 Would be very useful when parsing closely related languages.  The idea is to
 declare several start symbols, for instance
@@ -458,15 +377,6 @@ happen with yy_start: stmt | expr).  Then adjust the skeletons so that this
 initial token (YY_START_STMT, YY_START_EXPR) be shifted first in the
 corresponding parse function.
 
-** Better error messages
-The users are not provided with enough tools to forge their error messages.
-See for instance "Is there an option to change the message produced by
-YYERROR_VERBOSE?" by Simon Sobisch, on bison-help.
-
-See also
-https://www.cs.tufts.edu/~nr/cs257/archive/clinton-jefferey/lr-error-messages.pdf
-and https://research.swtch.com/yyerror.
-
 ** %include
 This is a popular demand.  We already made many changes in the parser that
 should make this reasonably easy to implement.
@@ -563,15 +473,6 @@ move to partial orders (sounds like series/parallel orders to me).
 
 This is a prerequisite for modules.
 
-* $undefined
-From Hans:
-- If the Bison generated parser experiences an undefined number in the
-character range, that character is written out in diagnostic messages, an
-addition to the $undefined value.
-
-Suggest: Change the name $undefined to undefined; looks better in outputs.
-
-
 * Pre and post actions.
 From: Florian Krohm <[email protected]>
 Subject: YYACT_EPILOGUE

bootstrap

@@ -1,6 +1,6 @@
 #! /bin/sh
 # Print a version string.
-scriptversion=2019-01-04.17; # UTC
+scriptversion=2020-04-13.15; # UTC
 
 # Bootstrap this package from checked-out sources.
 
@@ -970,7 +970,7 @@ bootstrap_post_import_hook \
 # Uninitialized submodules are listed with an initial dash.
 if $use_git && git submodule | grep '^-' >/dev/null; then
   die "some git submodules are not initialized. "     \
-      "Run 'git submodule init' and bootstrap again."
+      "Run 'git submodule update --init' and bootstrap again."
 fi
 
 # Remove any dangling symlink matching "*.m4" or "*.[ch]" in some

bootstrap.conf

@@ -37,6 +37,7 @@ gnulib_modules='
   obstack-printf
   perror progname
   quote quotearg
+  readline
   readme-release
   realloc-posix
   relocatable-prog relocatable-script

build-aux/cross-options.pl

@@ -17,7 +17,7 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
 #
-# Written by Akimn Demaille.
+# Written by Akim Demaille.
 
 use warnings;
 use 5.005;

build-aux/update-test

@@ -102,7 +102,7 @@ def update(at_file, logfile):
 def process(logfile):
     log = contents(logfile)
     # Look for the file to update.
-    m = re.search(r'^\d+\. ([-\w]+\.at):\d+: ', log, re.MULTILINE)
+    m = re.search(r'^\d+\. ([\-\+\w]+\.at):\d+: ', log, re.MULTILINE)
     if not m:
         trace("no diff found:", logfile)
         return

cfg.mk

@@ -156,7 +156,7 @@ exclude = \
 $(call exclude,                                                                 \
   bindtextdomain=^lib/main.c$$                                                  \
   cast_of_argument_to_free=^src/muscle-tab.c$$                                  \
-  po_check=(^po/POTFILES.in|.md)$$		                                \
+  po_check=^tests|(^po/POTFILES.in|.md)$$                                       \
   preprocessor_indentation=^data/|^lib/|^src/parse-gram.[ch]$$                  \
   program_name=^lib/main.c$$                                                    \
   prohibit_always-defined_macros=^data/skeletons/yacc.c$$                       \
@@ -165,7 +165,7 @@ $(call exclude,
   prohibit_doubled_word=^tests/named-refs.at$$                                  \
   prohibit_magic_number_exit=^doc/bison.texi$$                                  \
   prohibit_magic_number_exit+=?|^tests/(conflicts|regression).at$$              \
-  prohibit_strcmp=^doc/bison\.texi|examples|tests/local\.at$$                   \
+  prohibit_strcmp=^doc/bison\.texi|examples|tests                               \
   prohibit_tab_based_indentation=install-icc.sh|\.(am|mk)$$|^\.git|tests/input.at|Makefile$$   \
   require_config_h=^(lib/yyerror|data/skeletons/(glr|yacc))\.c$$                \
   require_config_h_first=^(lib/yyerror|data/skeletons/(glr|yacc))\.c$$          \

data/README.md

@@ -83,18 +83,20 @@ field), where field can `has_id`, `id`, etc.: see
 
 The macro `b4_symbol(NUM, FIELD)` gives access to the following FIELDS:
 
-- `has_id`: 0 or 1.
-
+- `has_id`: 0 or 1
   Whether the symbol has an id.
 
 - `id`: string
   If has_id, the id (prefixed by api.token.prefix if defined), otherwise
-  defined as empty.  Guaranteed to be usable as a C identifier.
+  defined as empty.  Guaranteed to be usable as a C identifier.  This is
+  used to define the token kind (i.e., the enum used by the return value of
+  yylex).
 
-- `tag`: string.
-  A representation of the symbol.  Can be 'foo', 'foo.id', '"foo"' etc.
+- `tag`: string
+  A human representation of the symbol.  Can be 'foo', 'foo.id', '"foo"' etc.
 
 - `user_number`: integer
+  The code associated to the `id`.
   The external number as used by yylex.  Can be ASCII code when a character,
   some number chosen by bison, or some user number in the case of
   %token FOO <NUM>.  Corresponds to yychar in yacc.c.
@@ -102,7 +104,12 @@ The macro `b4_symbol(NUM, FIELD)` gives access to the following FIELDS:
 - `is_token`: 0 or 1
   Whether this is a terminal symbol.
 
+- `kind`: string
+  The symbol kind, i.e., the enumerator of this symbol (token or nonterminal)
+  which is mapping to its `number`.
+
 - `number`: integer
+  The code associated to the `kind`.
   The internal number (computed from the external number by yytranslate).
   Corresponds to yytoken in yacc.c.  This is the same number that serves as
   key in b4_symbol(NUM, FIELD).

data/bison-default.css

@@ -29,3 +29,6 @@
 
 /* "Sections" in traces (--trace).  */
 .trace0    { color: green; }
+
+.expected   { color: green; }
+.unexpected { color: red; }

data/skeletons/README-D.txt

@@ -27,7 +27,7 @@ public interface Lexer
    * to the next token and prepares to return the semantic value
    * and beginning/ending positions of the token.
    * @return the token identifier corresponding to the next token. */
-  YYTokenType yylex ();
+  TokenKind yylex ();
 
   /**
    * Entry point for error reporting.  Emits an error
@@ -39,7 +39,7 @@ public interface Lexer
    void yyerror (YYLocation loc, string s);
 }
 
-- semantic types are handled by D usions (same as for C/C++ parsers)
+- semantic types are handled by D unions (same as for C/C++ parsers)
 
 - the following (non-standard) %defines are supported:
 

data/skeletons/bison.m4

@@ -72,8 +72,9 @@ version 2.2 of Bison.])
 # -------------
 # Issue a warning about private implementation details.
 m4_define([b4_disclaimer],
-[b4_comment([Undocumented macros, especially those whose name start with YY_,
-are private implementation details.  Do not rely on them.])
+[b4_comment([DO NOT RELY ON FEATURES THAT ARE NOT DOCUMENTED in the manual,
+especially those whose name start with YY_ or yy_.  They are
+private implementation details that can be changed or removed.])
 ])
 
 
@@ -369,6 +370,7 @@ m4_define([b4_$3_if],
 # Expand IF-TRUE, if FLAG is true, IF-FALSE otherwise.
 b4_define_flag_if([defines])            # Whether headers are requested.
 b4_define_flag_if([glr])                # Whether a GLR parser is requested.
+b4_define_flag_if([has_translations])   # Whether some tokens are internationalized.
 b4_define_flag_if([nondeterministic])   # Whether conflicts should be handled.
 b4_define_flag_if([token_table])        # Whether yytoken_table is demanded.
 b4_define_flag_if([yacc])               # Whether POSIX Yacc is emulated.
@@ -384,37 +386,64 @@ m4_define([b4_glr_cc_if],
 ## Symbols.  ##
 ## --------- ##
 
-# For a description of the Symbol handling, see README.
+# For a description of the Symbol handling, see README.md.
 #
 # The following macros provide access to symbol related values.
 
 # __b4_symbol(NUM, FIELD)
 # -----------------------
-# Recover a FIELD about symbol #NUM.  Thanks to m4_indir, fails if
-# undefined.
+# Fetch FIELD of symbol #NUM.  Fail if undefined.
 m4_define([__b4_symbol],
 [m4_indir([b4_symbol($1, $2)])])
 
 
 # _b4_symbol(NUM, FIELD)
 # ----------------------
-# Recover a FIELD about symbol #NUM (or "orig NUM").  Fails if
-# undefined.
+# Fetch FIELD of symbol #NUM (or "orig NUM", see README.md).
+# Fail if undefined.
 m4_define([_b4_symbol],
 [m4_ifdef([b4_symbol($1, number)],
           [__b4_symbol(m4_indir([b4_symbol($1, number)]), $2)],
           [__b4_symbol([$1], [$2])])])
 
 
+# b4_symbol_token_kind(NUM)
+# -------------------------
+# The token kind of this symbol.
+m4_define([b4_symbol_token_kind],
+[b4_percent_define_get([api.token.prefix])dnl
+_b4_symbol([$1], [id])])
+
+
+# b4_symbol_kind(NUM)
+# -------------------
+# Build the name of the kind of this symbol.  It must always exist,
+# otherwise some symbols might not be represented in the enum, which
+# might be compiled into too small a type to contain all the symbol
+# numbers.
+m4_define([b4_symbol_prefix], [b4_percent_define_get([api.symbol.prefix])])
+m4_define([b4_symbol_kind],
+[b4_percent_define_get([api.symbol.prefix])dnl
+m4_case([$1],
+  [-2],                             [[YYEMPTY]],
+  [0],                              [[YYEOF]],
+  [1],                              [[YYerror]],
+  [2],                              [[YYUNDEF]],
+  [m4_case(b4_symbol([$1], [tag]),
+      [$accept],                    [[YYACCEPT]],
+      [b4_symbol_if([$1], [has_id], _b4_symbol([$1], [id]),
+                                    [m4_bpatsubst([$1-][]_b4_symbol([$1], [tag]), [[^a-zA-Z_0-9]+], [_])])])])])
+
 
 # b4_symbol(NUM, FIELD)
 # ---------------------
-# Recover a FIELD about symbol #NUM (or "orig NUM").  Fails if
-# undefined.  If FIELD = id, prepend the token prefix.
+# Fetch FIELD of symbol #NUM (or "orig NUM").  Fail if undefined.
+#
+# If FIELD = id, prepend the token prefix.
 m4_define([b4_symbol],
 [m4_case([$2],
-         [id],    [m4_do([b4_percent_define_get([api.token.prefix])],
-                         [_b4_symbol([$1], [id])])],
+         [id],    [b4_symbol_token_kind([$1])],
+         [kind],  [b4_symbol_kind([$1])],
          [_b4_symbol($@)])])
 
 
@@ -437,9 +466,9 @@ m4_define([b4_symbol_tag_comment],
 ])
 
 
-# b4_symbol_action(SYMBOL-NUM, KIND)
-# ----------------------------------
-# Run the action KIND (destructor or printer) for SYMBOL-NUM.
+# b4_symbol_action(SYMBOL-NUM, ACTION)
+# ------------------------------------
+# Run the action ACTION ("destructor" or "printer") for SYMBOL-NUM.
 m4_define([b4_symbol_action],
 [b4_symbol_if([$1], [has_$2],
 [b4_dollar_pushdef([(*yyvaluep)],
@@ -463,21 +492,21 @@ m4_define([b4_symbol_destructor], [b4_symbol_action([$1], [destructor])])
 m4_define([b4_symbol_printer],    [b4_symbol_action([$1], [printer])])
 
 
-# b4_symbol_actions(KIND, [TYPE = yytype])
-# ----------------------------------------
-# Emit the symbol actions for KIND ("printer" or "destructor").
-# Dispatch on TYPE.
+# b4_symbol_actions(ACTION, [KIND = yykind])
+# ------------------------------------------
+# Emit the symbol actions for ACTION ("destructor" or "printer").
+# Dispatch on KIND.
 m4_define([b4_symbol_actions],
 [m4_pushdef([b4_actions_], m4_expand([b4_symbol_foreach([b4_symbol_$1])]))dnl
 m4_ifval(m4_defn([b4_actions_]),
-[switch (m4_default([$2], [yytype]))
+[switch (m4_default([$2], [yykind]))
     {
 m4_defn([b4_actions_])[]dnl
       default:
         break;
     }dnl
 ],
-[YYUSE (m4_default([$2], [yytype]));])dnl
+[YYUSE (m4_default([$2], [yykind]));])dnl
 m4_popdef([b4_actions_])dnl
 ])
 
@@ -507,13 +536,14 @@ m4_define([b4_symbol_map],
 
 # b4_token_visible_if(NUM, IF-TRUE, IF-FALSE)
 # -------------------------------------------
-# Whether NUM denotes a token that has an exported definition (i.e.,
-# shows in enum yytokentype).
+# Whether NUM denotes a token kind that has an exported definition
+# (i.e., shows in enum yytokentype).
 m4_define([b4_token_visible_if],
 [b4_symbol_if([$1], [is_token],
               [b4_symbol_if([$1], [has_id], [$2], [$3])],
               [$3])])
 
+
 # b4_token_has_definition(NUM)
 # ----------------------------
 # 1 if NUM is visible, nothing otherwise.
@@ -530,13 +560,27 @@ m4_define([b4_any_token_visible_if],
 
 # b4_token_format(FORMAT, NUM)
 # ----------------------------
+# If token NUM has a visible ID, format FORMAT with ID, USER_NUMBER.
 m4_define([b4_token_format],
 [b4_token_visible_if([$2],
-[m4_quote(m4_format([$1],
-                     [b4_symbol([$2], [id])],
-                     [b4_symbol([$2], b4_api_token_raw_if([[number]], [[user_number]]))]))])])
+[m4_format([[$1]],
+           m4_expand(b4_symbol([$2], [id])),
+           m4_expand(b4_symbol([$2], b4_api_token_raw_if([[number]], [[user_number]]))))])])
 
 
+# b4_last_enum_token
+# ------------------
+# The code of the last token visible token.
+m4_define([_b4_last_enum_token],
+[b4_token_visible_if([$1],
+   [m4_define([b4_last_enum_token], [$1])])])
+b4_symbol_foreach([_b4_last_enum_token])
+
+# b4_last_symbol
+# --------------
+# The code of the last symbol.
+m4_define([b4_last_symbol], m4_eval(b4_tokens_number + b4_nterms_number - 1))
+
 ## ------- ##
 ## Types.  ##
 ## ------- ##
@@ -603,7 +647,7 @@ m4_define([b4_sync_end],   [ b4_comment([$2:$1])]
 # This generates dependencies on the Bison skeletons hence lots of
 # useless 'git diff'.  This location is useless for the regular
 # user (who does not care about the skeletons) and is actually not
-# useful for Bison developpers too (I, Akim, never used this to locate
+# useful for Bison developers too (I, Akim, never used this to locate
 # the code in skeletons that generated output).  So disable it
 # completely.  If someone thinks this was actually useful, a %define
 # variable should be provided to control the level of verbosity of
@@ -1016,21 +1060,23 @@ m4_define([b4_bison_locations_if],
 [b4_locations_if([b4_percent_define_ifdef([[api.location.type]], [], [$1])])])
 
 
-# b4_error_verbose_if([IF-ERRORS-ARE-VERBOSE], [IF-NOT])
+
+# %define parse.error "(custom|detailed|simple|verbose)"
 # ------------------------------------------------------
-# Map %define parse.error "(simple|verbose)" to b4_error_verbose_if and
-# b4_error_verbose_flag.
 b4_percent_define_default([[parse.error]], [[simple]])
 b4_percent_define_check_values([[[[parse.error]],
-                                 [[simple]], [[verbose]]]])
-m4_define([b4_error_verbose_flag],
-          [m4_case(b4_percent_define_get([[parse.error]]),
-                   [simple],  [[0]],
-                   [verbose], [[1]])])
-b4_define_flag_if([error_verbose])
+                                 [[custom]], [[detailed]], [[simple]], [[verbose]]]])
+
+# b4_parse_error_case(CASE1, THEN1, CASE2, THEN2, ..., ELSE)
+# ----------------------------------------------------------
+m4_define([b4_parse_error_case],
+[m4_case(b4_percent_define_get([[parse.error]]), $@)])
+
+# b4_parse_error_bmatch(PATTERN1, THEN1, PATTERN2, THEN2, ..., ELSE)
+# ------------------------------------------------------------------
+m4_define([b4_parse_error_bmatch],
+[m4_bmatch(b4_percent_define_get([[parse.error]]), $@)])
 
-# yytoken_table is needed to support verbose errors.
-b4_error_verbose_if([m4_define([b4_token_table_flag], [1])])
 
 
 # b4_variant_if([IF-VARIANT-ARE-USED], [IF-NOT])
@@ -1116,3 +1162,11 @@ b4_percent_define_ifdef([api.value.type],
 
 # api.value.union.name.
 b4_percent_define_check_kind([api.value.union.name], [keyword])
+
+# parse.error (custom|detailed) >< token-table.
+b4_token_table_if(
+[b4_parse_error_bmatch([custom\|detailed],
+[b4_complain_at(b4_percent_define_get_loc([parse.error]),
+                [['%s' and '%s' cannot be used together]],
+                [%token-table],
+                [%define parse.error (custom|detailed)])])])

data/skeletons/c++.m4

@@ -22,6 +22,8 @@ b4_percent_define_ifdef([[api.value.union.name]],
   [b4_complain_at(b4_percent_define_get_loc([[api.value.union.name]]),
                   [named %union is invalid in C++])])
 
+b4_percent_define_default([[api.symbol.prefix]], [[S_]])
+
 m4_include(b4_skeletonsdir/[c.m4])
 
 b4_percent_define_check_kind([api.namespace], [code], [deprecated])
@@ -160,20 +162,42 @@ m4_bpatsubst(m4_dquote(m4_bpatsubst(m4_dquote(b4_namespace_ref[ ]),
              [::\([^][:]\|:[^:]\)*], [} ])[} // ]b4_namespace_ref])])
 
 
+## ------------- ##
+## Token kinds.  ##
+## ------------- ##
+
+
 # b4_token_enums
 # --------------
-# Output the definition of the tokens as enums.
+# Output the definition of the token kinds.
 m4_define([b4_token_enums],
-[[enum yytokentype
+[[enum token_kind_type
       {
-        ]m4_join([,
-        ],
-                 b4_symbol_map([b4_token_enum]))[
-      };]dnl
+        ]b4_symbol([-2], [id])[ = -2,
+]b4_symbol_foreach([b4_token_enum])dnl
+[      };]dnl
 ])
 
 
 
+## -------------- ##
+## Symbol kinds.  ##
+## -------------- ##
+
+# b4_declare_symbol_enum
+# ----------------------
+# The definition of the symbol internal numbers as an enum.
+# Defining YYEMPTY here is important: it forces the compiler
+# to use a signed type, which matters for yytoken.
+m4_define([b4_declare_symbol_enum],
+[[enum symbol_kind_type
+      {
+        YYNTOKENS = ]b4_tokens_number[, ///< Number of tokens.
+        ]b4_symbol_kind([-2])[ = -2,
+]b4_symbol_foreach([      b4_symbol_enum])dnl
+[      };]])
+
+
 
 ## ----------------- ##
 ## Semantic Values.  ##
@@ -231,23 +255,31 @@ m4_define([b4_public_types_declare],
       location_type location;])[
     };
 
-    /// Tokens.
+    /// Token kinds.
     struct token
     {
       ]b4_token_enums[
+      /// Backward compatibility alias.
+      typedef token_kind_type yytokentype;
     };
 
-    /// (External) token type, as returned by yylex.
-    typedef token::yytokentype token_type;
+    /// Token kind, as returned by yylex.
+    typedef token::yytokentype token_kind_type;
 
-    /// Symbol type: an internal symbol number.
-    typedef int symbol_number_type;
+    /// Backward compatibility alias.
+    typedef token_kind_type token_type;
 
-    /// The symbol type number to denote an empty symbol.
-    enum { empty_symbol = -2 };
+    /// Symbol kinds.
+    struct symbol_kind
+    {
+      ]b4_declare_symbol_enum[
+    };
 
-    /// Internal symbol number for tokens (subsumed by symbol_number_type).
-    typedef ]b4_int_type_for([b4_translate])[ token_number_type;
+    /// (Internal) symbol kind.
+    typedef symbol_kind::symbol_kind_type symbol_kind_type;
+
+    /// The number of tokens.
+    static const symbol_kind_type YYNTOKENS = symbol_kind::YYNTOKENS;
 ]])
 
 
@@ -258,8 +290,8 @@ m4_define([b4_public_types_declare],
 m4_define([b4_symbol_type_define],
 [[    /// A complete symbol.
     ///
-    /// Expects its Base type to provide access to the symbol type
-    /// via type_get ().
+    /// Expects its Base type to provide access to the symbol kind
+    /// via kind ().
     ///
     /// Provide access to semantic value]b4_locations_if([ and location])[.
     template <typename Base>
@@ -276,7 +308,14 @@ m4_define([b4_symbol_type_define],
 
 #if 201103L <= YY_CPLUSPLUS
       /// Move constructor.
-      basic_symbol (basic_symbol&& that);
+      basic_symbol (basic_symbol&& that)
+        : Base (std::move (that))
+        , value (]b4_variant_if([], [std::move (that.value)]))b4_locations_if([
+        , location (std::move (that.location))])[
+      {]b4_variant_if([
+        b4_symbol_variant([this->kind ()], [value], [move],
+                          [std::move (that.value)])
+      ])[}
 #endif
 
       /// Copy constructor.
@@ -303,21 +342,24 @@ m4_define([b4_symbol_type_define],
       void clear ()
       {]b4_variant_if([[
         // User destructor.
-        symbol_number_type yytype = this->type_get ();
+        symbol_kind_type yykind = this->kind ();
         basic_symbol<Base>& yysym = *this;
         (void) yysym;
-        switch (yytype)
+        switch (yykind)
         {
 ]b4_symbol_foreach([b4_symbol_destructor])dnl
 [       default:
           break;
         }
 
-        // Type destructor.
-]b4_symbol_variant([[yytype]], [[value]], [[template destroy]])])[
+        // Value type destructor.
+]b4_symbol_variant([[yykind]], [[value]], [[template destroy]])])[
         Base::clear ();
       }
 
+      /// Backward compatibility.
+      symbol_kind_type type_get () const YY_NOEXCEPT;
+
       /// Whether empty.
       bool empty () const YY_NOEXCEPT;
 
@@ -338,46 +380,45 @@ m4_define([b4_symbol_type_define],
     };
 
     /// Type access provider for token (enum) based symbols.
-    struct by_type
+    struct by_kind
     {
       /// Default constructor.
-      by_type ();
+      by_kind ();
 
 #if 201103L <= YY_CPLUSPLUS
       /// Move constructor.
-      by_type (by_type&& that);
+      by_kind (by_kind&& that);
 #endif
 
       /// Copy constructor.
-      by_type (const by_type& that);
+      by_kind (const by_kind& that);
 
-      /// The symbol type as needed by the constructor.
-      typedef token_type kind_type;
+      /// The symbol kind as needed by the constructor.
+      typedef token_kind_type kind_type;
 
       /// Constructor from (external) token numbers.
-      by_type (kind_type t);
+      by_kind (kind_type t);
 
       /// Record that this symbol is empty.
       void clear ();
 
-      /// Steal the symbol type from \a that.
-      void move (by_type& that);
+      /// Steal the symbol kind from \a that.
+      void move (by_kind& that);
 
       /// The (internal) type number (corresponding to \a type).
       /// \a empty when empty.
-      symbol_number_type type_get () const YY_NOEXCEPT;
+      symbol_kind_type kind () const YY_NOEXCEPT;
 
-      /// The symbol type.
-      /// \a empty_symbol when empty.
-      /// An int, not token_number_type, to be able to store empty_symbol.
-      int type;
+      /// The symbol kind.
+      /// \a ]b4_symbol_prefix[YYEMPTY when empty.
+      symbol_kind_type kind_;
     };
 
     /// "External" symbols: returned by the scanner.
-    struct symbol_type : basic_symbol<by_type>
+    struct symbol_type : basic_symbol<by_kind>
     {]b4_variant_if([[
       /// Superclass.
-      typedef basic_symbol<by_type> super_type;
+      typedef basic_symbol<by_kind> super_type;
 
       /// Empty symbol.
       symbol_type () {}
@@ -393,25 +434,13 @@ m4_define([b4_symbol_type_define],
 # Provide the implementation needed by the public types.
 m4_define([b4_public_types_define],
 [[  // basic_symbol.
-#if 201103L <= YY_CPLUSPLUS
-  template <typename Base>
-  ]b4_parser_class[::basic_symbol<Base>::basic_symbol (basic_symbol&& that)
-    : Base (std::move (that))
-    , value (]b4_variant_if([], [std::move (that.value)]))b4_locations_if([
-    , location (std::move (that.location))])[
-  {]b4_variant_if([
-    b4_symbol_variant([this->type_get ()], [value], [move],
-                      [std::move (that.value)])
-  ])[}
-#endif
-
   template <typename Base>
   ]b4_parser_class[::basic_symbol<Base>::basic_symbol (const basic_symbol& that)
     : Base (that)
     , value (]b4_variant_if([], [that.value]))b4_locations_if([
     , location (that.location)])[
   {]b4_variant_if([
-    b4_symbol_variant([this->type_get ()], [value], [copy],
+    b4_symbol_variant([this->kind ()], [value], [copy],
                       [YY_MOVE (that.value)])
   ])[}
 
@@ -436,13 +465,20 @@ m4_define([b4_public_types_define],
     , location (YY_MOVE (l))])[
   {]b4_variant_if([[
     (void) v;
-    ]b4_symbol_variant([this->type_get ()], [value], [YY_MOVE_OR_COPY], [YY_MOVE (v)])])[}]])[
+    ]b4_symbol_variant([this->kind ()], [value], [YY_MOVE_OR_COPY], [YY_MOVE (v)])])[}]])[
+
+  template <typename Base>
+  ]b4_parser_class[::symbol_kind_type
+  ]b4_parser_class[::basic_symbol<Base>::type_get () const YY_NOEXCEPT
+  {
+    return this->kind ();
+  }
 
   template <typename Base>
   bool
   ]b4_parser_class[::basic_symbol<Base>::empty () const YY_NOEXCEPT
   {
-    return Base::type_get () == empty_symbol;
+    return this->kind () == symbol_kind::]b4_symbol_prefix[YYEMPTY;
   }
 
   template <typename Base>
@@ -450,50 +486,50 @@ m4_define([b4_public_types_define],
   ]b4_parser_class[::basic_symbol<Base>::move (basic_symbol& s)
   {
     super_type::move (s);
-    ]b4_variant_if([b4_symbol_variant([this->type_get ()], [value], [move],
+    ]b4_variant_if([b4_symbol_variant([this->kind ()], [value], [move],
                                       [YY_MOVE (s.value)])],
                    [value = YY_MOVE (s.value);])[]b4_locations_if([
     location = YY_MOVE (s.location);])[
   }
 
-  // by_type.
-  ]b4_inline([$1])b4_parser_class[::by_type::by_type ()
-    : type (empty_symbol)
+  // by_kind.
+  ]b4_inline([$1])b4_parser_class[::by_kind::by_kind ()
+    : kind_ (symbol_kind::]b4_symbol_prefix[YYEMPTY)
   {}
 
 #if 201103L <= YY_CPLUSPLUS
-  ]b4_inline([$1])b4_parser_class[::by_type::by_type (by_type&& that)
-    : type (that.type)
+  ]b4_inline([$1])b4_parser_class[::by_kind::by_kind (by_kind&& that)
+    : kind_ (that.kind_)
   {
     that.clear ();
   }
 #endif
 
-  ]b4_inline([$1])b4_parser_class[::by_type::by_type (const by_type& that)
-    : type (that.type)
+  ]b4_inline([$1])b4_parser_class[::by_kind::by_kind (const by_kind& that)
+    : kind_ (that.kind_)
   {}
 
-  ]b4_inline([$1])b4_parser_class[::by_type::by_type (token_type t)
-    : type (yytranslate_ (t))
+  ]b4_inline([$1])b4_parser_class[::by_kind::by_kind (token_kind_type t)
+    : kind_ (yytranslate_ (t))
   {}
 
   ]b4_inline([$1])[void
-  ]b4_parser_class[::by_type::clear ()
+  ]b4_parser_class[::by_kind::clear ()
   {
-    type = empty_symbol;
+    kind_ = symbol_kind::]b4_symbol_prefix[YYEMPTY;
   }
 
   ]b4_inline([$1])[void
-  ]b4_parser_class[::by_type::move (by_type& that)
+  ]b4_parser_class[::by_kind::move (by_kind& that)
   {
-    type = that.type;
+    kind_ = that.kind_;
     that.clear ();
   }
 
-  ]b4_inline([$1])[int
-  ]b4_parser_class[::by_type::type_get () const YY_NOEXCEPT
+  ]b4_inline([$1])[]b4_parser_class[::symbol_kind_type
+  ]b4_parser_class[::by_kind::kind () const YY_NOEXCEPT
   {
-    return type;
+    return kind_;
   }
 ]])
 
@@ -510,15 +546,15 @@ m4_define([b4_token_constructor_define], [])
 # Define yytranslate_.  Sometimes used in the header file ($1=hh),
 # sometimes in the cc file.
 m4_define([b4_yytranslate_define],
-[  b4_inline([$1])b4_parser_class[::token_number_type
+[  b4_inline([$1])b4_parser_class[::symbol_kind_type
   ]b4_parser_class[::yytranslate_ (int t)
   {
 ]b4_api_token_raw_if(
-[[    return static_cast<token_number_type> (t);]],
+[[    return static_cast<symbol_kind_type> (t);]],
 [[    // YYTRANSLATE[TOKEN-NUM] -- Symbol number corresponding to
     // TOKEN-NUM as returned by yylex.
     static
-    const token_number_type
+    const ]b4_int_type_for([b4_translate])[
     translate_table[] =
     {
   ]b4_translate[
@@ -526,11 +562,11 @@ m4_define([b4_yytranslate_define],
     const int user_token_number_max_ = ]b4_user_token_number_max[;
 
     if (t <= 0)
-      return yyeof_;
+      return symbol_kind::]b4_symbol_prefix[YYEOF;
     else if (t <= user_token_number_max_)
-      return translate_table[t];
+      return YY_CAST (symbol_kind_type, translate_table[t]);
     else
-      return yy_undef_token_;]])[
+      return symbol_kind::]b4_symbol_prefix[YYUNDEF;]])[
   }
 ]])
 

data/skeletons/c.m4

@@ -101,6 +101,7 @@ m4_define_default([b4_prefix], [b4_api_prefix])
 b4_percent_define_default([[api.value.union.name]],
                           [b4_api_PREFIX[][STYPE]])
 
+b4_percent_define_default([[api.symbol.prefix]], [[YYSYMBOL_]])
 
 ## ------------------------ ##
 ## Pure/impure interfaces.  ##
@@ -129,6 +130,13 @@ m4_define([b4_user_args],
 [m4_ifset([b4_parse_param], [, b4_args(b4_parse_param)])])
 
 
+# b4_user_formals
+# ---------------
+# The possible parse-params formal arguments preceded by a comma.
+m4_define([b4_user_formals],
+[m4_ifset([b4_parse_param], [, b4_formals(b4_parse_param)])])
+
+
 # b4_parse_param
 # --------------
 # If defined, b4_parse_param arrives double quoted, but below we prefer
@@ -418,9 +426,22 @@ static const b4_int_type_for([$2]) yy$1[[]] =
 ])
 
 
-## ------------------------- ##
-## Assigning token numbers.  ##
-## ------------------------- ##
+## ------------- ##
+## Token kinds.  ##
+## ------------- ##
+
+# Because C enums are not scoped, because tokens are exposed in the
+# header, and because these tokens are common to all the parsers, we
+# need to make sure their names don't collide: use the api.prefix.
+# YYEOF is special, since the user may give it a different name.
+m4_define([b4_symbol(-2, id)],  [b4_api_PREFIX[][EMPTY]])
+m4_define([b4_symbol(-2, tag)], [[No symbol.]])
+
+m4_if(b4_symbol(0, id), [YYEOF],
+     [m4_define([b4_symbol(0, id)],  [b4_api_PREFIX[][EOF]])])
+m4_define([b4_symbol(1, id)],  [b4_api_PREFIX[][error]])
+m4_define([b4_symbol(2, id)],  [b4_api_PREFIX[][UNDEF]])
+
 
 # b4_token_define(TOKEN-NUM)
 # --------------------------
@@ -432,7 +453,7 @@ m4_define([b4_token_define],
 # ----------------
 # Output the definition of the tokens.
 m4_define([b4_token_defines],
-[b4_any_token_visible_if([/* Tokens.  */
+[b4_any_token_visible_if([/* Token kinds.  */
 m4_join([
 ], b4_symbol_map([b4_token_define]))
 ])])
@@ -442,34 +463,81 @@ m4_join([
 # ------------------------
 # Output the definition of this token as an enum.
 m4_define([b4_token_enum],
-[b4_token_format([%s = %s], [$1])])
+[b4_token_visible_if([$1],
+    [m4_format([    %-30s %s],
+               m4_format([[%s = %s%s%s]],
+                         b4_symbol([$1], [id]),
+                         b4_symbol([$1], b4_api_token_raw_if([[number]], [[user_number]])),
+                         m4_if([$1], b4_last_enum_token, [], [[,]])),
+               [b4_symbol_tag_comment([$1])])])])
 
 
 # b4_token_enums
 # --------------
-# Output the definition of the tokens (if there are) as enums.
+# The definition of the token kinds.
 m4_define([b4_token_enums],
-[b4_any_token_visible_if([[/* Token type.  */
+[b4_any_token_visible_if([[/* Token kinds.  */
 #ifndef ]b4_api_PREFIX[TOKENTYPE
 # define ]b4_api_PREFIX[TOKENTYPE
   enum ]b4_api_prefix[tokentype
   {
-    ]m4_join([,
-    ],
-             b4_symbol_map([b4_token_enum]))[
-  };
+    ]b4_symbol([-2], [id])[ = -2,
+]b4_symbol_foreach([b4_token_enum])dnl
+[  };
+  typedef enum ]b4_api_prefix[tokentype ]b4_api_prefix[token_kind_t;
 #endif
 ]])])
 
 
 # b4_token_enums_defines
 # ----------------------
-# Output the definition of the tokens (if there are any) as enums and,
+# The definition of the tokens (if there are any) as enums and,
 # if POSIX Yacc is enabled, as #defines.
 m4_define([b4_token_enums_defines],
 [b4_token_enums[]b4_yacc_if([b4_token_defines])])
 
 
+# b4_symbol_translate(STRING)
+# ---------------------------
+# Used by "bison" in the array of symbol names to mark those that
+# require translation.
+m4_define([b4_symbol_translate],
+[[N_($1)]])
+
+
+
+## -------------- ##
+## Symbol kinds.  ##
+## -------------- ##
+
+# b4_symbol_enum(SYMBOL-NUM)
+# --------------------------
+# Output the definition of this symbol as an enum.
+m4_define([b4_symbol_enum],
+[m4_format([  %-40s %s],
+           m4_format([[%s = %s%s%s]],
+                     b4_symbol([$1], [kind]),
+                     [$1],
+                     m4_if([$1], b4_last_symbol, [], [[,]])),
+           [b4_symbol_tag_comment([$1])])])
+
+
+# b4_declare_symbol_enum
+# ----------------------
+# The definition of the symbol internal numbers as an enum.
+# Defining YYEMPTY here is important: it forces the compiler
+# to use a signed type, which matters for yytoken.
+m4_define([b4_declare_symbol_enum],
+[[/* Symbol kind.  */
+enum yysymbol_kind_t
+{
+  ]b4_symbol_kind([-2])[ = -2,
+]b4_symbol_foreach([b4_symbol_enum])dnl
+[};
+typedef enum yysymbol_kind_t yysymbol_kind_t;
+]])])
+
+
 ## ----------------- ##
 ## Semantic Values.  ##
 ## ----------------- ##
@@ -493,15 +561,6 @@ m4_define([b4_symbol_value],
 ## ---------------------- ##
 
 
-# b4_function_define(NAME, RETURN-VALUE, [DECL1, NAME1], ...)
-# -----------------------------------------------------------
-# Declare the function NAME in C.
-m4_define([b4_function_define],
-[$2
-$1 (b4_formals(m4_shift2($@)))[]dnl
-])
-
-
 # b4_formals([DECL1, NAME1], ...)
 # -------------------------------
 # The formal arguments of a C function definition.
@@ -515,21 +574,6 @@ m4_define([b4_formal],
 
 
 
-## ----------------------- ##
-## Declaring C functions.  ##
-## ----------------------- ##
-
-
-# b4_function_declare(NAME, RETURN-VALUE, [DECL1, NAME1], ...)
-# ------------------------------------------------------------
-# Declare the function NAME.
-m4_define([b4_function_declare],
-[$2 $1 (b4_formals(m4_shift2($@)));[]dnl
-])
-
-
-
-
 ## --------------------- ##
 ## Calling C functions.  ##
 ## --------------------- ##
@@ -593,18 +637,15 @@ m4_define_default([b4_yydestruct_define],
 | Release the memory associated to this symbol.  |
 `-----------------------------------------------*/
 
-]b4_function_define([yydestruct],
-    [static void],
-    [[const char *yymsg],    [yymsg]],
-    [[int yytype],           [yytype]],
-    [[YYSTYPE *yyvaluep],    [yyvaluep]][]dnl
-b4_locations_if(            [, [[YYLTYPE *yylocationp], [yylocationp]]])[]dnl
-m4_ifset([b4_parse_param], [, b4_parse_param]))[
+static void
+yydestruct (const char *yymsg,
+            yysymbol_kind_t yykind, YYSTYPE *yyvaluep]b4_locations_if(dnl
+[[, YYLTYPE *yylocationp]])[]b4_user_formals[)
 {
 ]b4_parse_param_use([yyvaluep], [yylocationp])dnl
 [  if (!yymsg)
     yymsg = "Deleting";
-  YY_SYMBOL_PRINT (yymsg, yytype, yyvaluep, yylocationp);
+  YY_SYMBOL_PRINT (yymsg, yykind, yyvaluep, yylocationp);
 
   YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN
   ]b4_symbol_actions([destructor])[
@@ -622,13 +663,10 @@ m4_define_default([b4_yy_symbol_print_define],
 | Print this symbol's value on YYO.  |
 `-----------------------------------*/
 
-]b4_function_define([yy_symbol_value_print],
-    [static void],
-               [[FILE *yyo],                            [yyo]],
-               [[int yytype],                           [yytype]],
-               [[YYSTYPE const * const yyvaluep],       [yyvaluep]][]dnl
-b4_locations_if([, [[YYLTYPE const * const yylocationp], [yylocationp]]])[]dnl
-m4_ifset([b4_parse_param], [, b4_parse_param]))[
+static void
+yy_symbol_value_print (FILE *yyo,
+                       yysymbol_kind_t yykind, YYSTYPE const * const yyvaluep]b4_locations_if(dnl
+[[, YYLTYPE const * const yylocationp]])[]b4_user_formals[)
 {
   FILE *yyoutput = yyo;
 ]b4_parse_param_use([yyoutput], [yylocationp])dnl
@@ -637,8 +675,8 @@ m4_ifset([b4_parse_param], [, b4_parse_param]))[
 dnl glr.c does not feature yytoknum.
 m4_if(b4_skeleton, ["yacc.c"],
 [[# ifdef YYPRINT
-  if (yytype < YYNTOKENS)
-    YYPRINT (yyo, yytoknum[yytype], *yyvaluep);
+  if (yykind < YYNTOKENS)
+    YYPRINT (yyo, yytoknum[yykind], *yyvaluep);
 # endif
 ]])dnl
 b4_percent_code_get([[pre-printer]])dnl
@@ -653,21 +691,18 @@ b4_percent_code_get([[post-printer]])dnl
 | Print this symbol on YYO.  |
 `---------------------------*/
 
-]b4_function_define([yy_symbol_print],
-    [static void],
-               [[FILE *yyo],                            [yyo]],
-               [[int yytype],                           [yytype]],
-               [[YYSTYPE const * const yyvaluep],       [yyvaluep]][]dnl
-b4_locations_if([, [[YYLTYPE const * const yylocationp], [yylocationp]]])[]dnl
-m4_ifset([b4_parse_param], [, b4_parse_param]))[
+static void
+yy_symbol_print (FILE *yyo,
+                 yysymbol_kind_t yykind, YYSTYPE const * const yyvaluep]b4_locations_if(dnl
+[[, YYLTYPE const * const yylocationp]])[]b4_user_formals[)
 {
   YYFPRINTF (yyo, "%s %s (",
-             yytype < YYNTOKENS ? "token" : "nterm", yytname[yytype]);
+             yykind < YYNTOKENS ? "token" : "nterm", yysymbol_name (yykind));
 
 ]b4_locations_if([  YY_LOCATION_PRINT (yyo, *yylocationp);
   YYFPRINTF (yyo, ": ");
 ])dnl
-[  yy_symbol_value_print (yyo, yytype, yyvaluep]dnl
+[  yy_symbol_value_print (yyo, yykind, yyvaluep]dnl
 b4_locations_if([, yylocationp])[]b4_user_args[);
   YYFPRINTF (yyo, ")");
 }]dnl
@@ -693,11 +728,11 @@ m4_define([b4_symbol_type_register],
 [m4_define([b4_symbol($1, type_tag)],
            [b4_symbol_if([$1], [has_id],
                          [b4_symbol([$1], [id])],
-                         [yytype_[]b4_symbol([$1], [number])])])dnl
+                         [yykind_[]b4_symbol([$1], [number])])])dnl
 m4_append([b4_union_members],
-m4_expand([
-  b4_symbol_tag_comment([$1])dnl
-  b4_symbol([$1], [type]) b4_symbol([$1], [type_tag]);]))
+m4_expand([m4_format([  %-40s %s],
+                     m4_expand([b4_symbol([$1], [type]) b4_symbol([$1], [type_tag]);]),
+                     [b4_symbol_tag_comment([$1])])]))
 ])
 
 
@@ -930,16 +965,14 @@ m4_define([b4_yy_location_print_define],
    This macro was not mandated originally: define only if we know
    we won't break user code: when these are the locations we know.  */
 
-#ifndef YY_LOCATION_PRINT
-# if defined ]b4_api_PREFIX[LTYPE_IS_TRIVIAL && ]b4_api_PREFIX[LTYPE_IS_TRIVIAL
+# ifndef YY_LOCATION_PRINT
+#  if defined ]b4_api_PREFIX[LTYPE_IS_TRIVIAL && ]b4_api_PREFIX[LTYPE_IS_TRIVIAL
 
 /* Print *YYLOCP on YYO.  Private, do not rely on its existence. */
 
 YY_ATTRIBUTE_UNUSED
-]b4_function_define([yy_location_print_],
-    [static int],
-               [[FILE *yyo],                    [yyo]],
-               [[YYLTYPE const * const yylocp], [yylocp]])[
+static int
+yy_location_print_ (FILE *yyo, YYLTYPE const * const yylocp)
 {
   int res = 0;
   int end_col = 0 != yylocp->last_column ? yylocp->last_column - 1 : 0;
@@ -963,17 +996,17 @@ YY_ATTRIBUTE_UNUSED
   return res;
  }
 
-#  define YY_LOCATION_PRINT(File, Loc)          \
+#   define YY_LOCATION_PRINT(File, Loc)          \
   yy_location_print_ (File, &(Loc))
 
-# else
-#  define YY_LOCATION_PRINT(File, Loc) ((void) 0)
-# endif
-#endif]],
+#  else
+#   define YY_LOCATION_PRINT(File, Loc) ((void) 0)
+#  endif
+# endif /* !defined YY_LOCATION_PRINT */]],
 [[/* This macro is provided for backward compatibility. */
-#ifndef YY_LOCATION_PRINT
-# define YY_LOCATION_PRINT(File, Loc) ((void) 0)
-#endif]])
+# ifndef YY_LOCATION_PRINT
+#  define YY_LOCATION_PRINT(File, Loc) ((void) 0)
+# endif]])
 ])
 
 # b4_yyloc_default

data/skeletons/d.m4

@@ -148,9 +148,9 @@ private static immutable b4_int_type_for([$2])[[]] yy$1_ =
 ])
 
 
-## ------------------------- ##
-## Assigning token numbers.  ##
-## ------------------------- ##
+## ------------- ##
+## Token kinds.  ##
+## ------------- ##
 
 # b4_token_enum(TOKEN-NAME, TOKEN-NUMBER)
 # ---------------------------------------
@@ -159,19 +159,53 @@ m4_define([b4_token_enum],
 [b4_token_format([  %s = %s,
 ], [$1])])
 
-# b4_token_enums(LIST-OF-PAIRS-TOKEN-NAME-TOKEN-NUMBER)
-# -----------------------------------------------------
+# b4_token_enums
+# --------------
 # Output the definition of the tokens as enums.
 m4_define([b4_token_enums],
-[/* Tokens.  */
-public enum YYTokenType {
-
-  /** Token returned by the scanner to signal the end of its input.  */
-  EOF = 0,
-b4_symbol_foreach([b4_token_enum])
+[/* Token kinds.  */
+public enum TokenKind {
+  ]b4_symbol_kind([-2])[ = -2,
+b4_symbol_foreach([b4_token_enum])dnl
 }
 ])
 
+
+
+## -------------- ##
+## Symbol kinds.  ##
+## -------------- ##
+
+b4_percent_define_default([[api.symbol.prefix]], [[S_]])
+
+
+# b4_symbol_enum(SYMBOL-NUM)
+# --------------------------
+# Output the definition of this symbol as an enum.
+m4_define([b4_symbol_enum],
+[m4_format([    %-30s %s],
+           m4_format([[%s = %s,]],
+                     b4_symbol([$1], [kind]),
+                     [$1]),
+           [b4_symbol_tag_comment([$1])])])
+
+
+# b4_declare_symbol_enum
+# ----------------------
+# The definition of the symbol internal numbers as an enum.
+# Defining YYEMPTY here is important: it forces the compiler
+# to use a signed type, which matters for yytoken.
+m4_define([b4_declare_symbol_enum],
+[[  /* Symbol kinds.  */
+  public enum SymbolKind
+  {
+    ]b4_symbol_kind([-2])[ = -2,  /* No symbol.  */
+]b4_symbol_foreach([b4_symbol_enum])dnl
+[  };
+]])])
+
+
+
 # b4-case(ID, CODE)
 # -----------------
 m4_define([b4_case], [    case $1:

data/skeletons/glr.cc

@@ -66,30 +66,29 @@ m4_defn([b4_parse_param]))],
            [[b4_namespace_ref::b4_parser_class[& yyparser], [[yyparser]]]])
 ])
 
+# b4_declare_symbol_enum
+# ----------------------
+m4_append([b4_declare_symbol_enum],
+[[typedef symbol_kind_type yysymbol_kind_t;
+]])
+
 
 # b4_yy_symbol_print_define
 # -------------------------
 # Bypass the default implementation to generate the "yy_symbol_print"
 # and "yy_symbol_value_print" functions.
 m4_define([b4_yy_symbol_print_define],
-[[
-/*--------------------.
+[[/*--------------------.
 | Print this symbol.  |
 `--------------------*/
 
-]b4_function_define([yy_symbol_print],
-    [static void],
-    [[FILE *],      []],
-    [[int yytype],  [yytype]],
-    [[const ]b4_namespace_ref::b4_parser_class[::semantic_type *yyvaluep],
-                    [yyvaluep]][]dnl
-b4_locations_if([,
-    [[const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp],
-                    [yylocationp]]]),
-    b4_parse_param)[
+static void
+yy_symbol_print (FILE *, ]b4_namespace_ref::b4_parser_class[::symbol_kind_type yytoken,
+                 const ]b4_namespace_ref::b4_parser_class[::semantic_type *yyvaluep]b4_locations_if([[,
+                 const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp]])[]b4_user_formals[)
 {
 ]b4_parse_param_use[]dnl
-[  yyparser.yy_symbol_print_ (yytype, yyvaluep]b4_locations_if([, yylocationp])[);
+[  yyparser.yy_symbol_print_ (yytoken, yyvaluep]b4_locations_if([, yylocationp])[);
 }
 ]])[
 
@@ -101,45 +100,30 @@ m4_defn([b4_initial_action])]))])[
 # Hijack the post prologue to declare yyerror.
 ]m4_append([b4_post_prologue],
 [b4_syncline([@oline@], [@ofile@])dnl
-b4_function_declare([yyerror],
-    [static void],b4_locations_if([
-    [[const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp],
-                        [yylocationp]],])
-    b4_parse_param,
-    [[const char* msg], [msg]])])[
+[static void
+yyerror (]b4_locations_if([[const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp,
+         ]])[]m4_ifset([b4_parse_param], [b4_formals(b4_parse_param),
+         ])[const char* msg);]])[
 
-
-#undef yynerrs
-#undef yychar
-#undef yylval]b4_locations_if([
-#undef yylloc])
-
-m4_if(b4_prefix, [yy], [],
-[[/* Substitute the variable and function names.  */
-#define yyparse ]b4_prefix[parse
-#define yylex   ]b4_prefix[lex
-#define yyerror ]b4_prefix[error
-#define yydebug ]b4_prefix[debug]]b4_pure_if([], [[
-#define yylval  ]b4_prefix[lval
-#define yychar  ]b4_prefix[char
-#define yynerrs ]b4_prefix[nerrs]b4_locations_if([[
-#define yylloc  ]b4_prefix[lloc]])]))
+]b4_percent_define_flag_if([[global_tokens_and_yystype]], [],
+[m4_define([b4_pre_epilogue],
+[[/* The user is using the C++ token kind, not the C one. */
+#undef ]b4_symbol(0, [id])
+])])[
 
 # Hijack the epilogue to define implementations (yyerror, parser member
 # functions etc.).
-m4_append([b4_epilogue],
+]m4_append([b4_epilogue],
 [b4_syncline([@oline@], [@ofile@])dnl
 [
 /*------------------.
 | Report an error.  |
 `------------------*/
 
-]b4_function_define([yyerror],
-    [static void],b4_locations_if([
-    [[const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp],
-                        [yylocationp]],])
-    b4_parse_param,
-    [[const char* msg], [msg]])[
+static void
+yyerror (]b4_locations_if([[const ]b4_namespace_ref::b4_parser_class[::location_type *yylocationp,
+         ]])[]m4_ifset([b4_parse_param], [b4_formals(b4_parse_param),
+         ])[const char* msg)
 {
 ]b4_parse_param_use[]dnl
 [  yyparser.error (]b4_locations_if([[*yylocationp, ]])[msg);
@@ -181,9 +165,9 @@ m4_pushdef([b4_parse_param], m4_defn([b4_parse_param_orig]))dnl
   `--------------------*/
 
   void
-  ]b4_parser_class[::yy_symbol_value_print_ (int yytype,
+  ]b4_parser_class[::yy_symbol_value_print_ (symbol_kind_type yykind,
                            const semantic_type* yyvaluep]b4_locations_if([[,
-                           const location_type* yylocationp]])[)
+                           const location_type* yylocationp]])[) const
   {]b4_locations_if([[
     YYUSE (yylocationp);]])[
     YYUSE (yyvaluep);
@@ -195,14 +179,14 @@ m4_pushdef([b4_parse_param], m4_defn([b4_parse_param_orig]))dnl
 
 
   void
-  ]b4_parser_class[::yy_symbol_print_ (int yytype,
+  ]b4_parser_class[::yy_symbol_print_ (symbol_kind_type yykind,
                            const semantic_type* yyvaluep]b4_locations_if([[,
-                           const location_type* yylocationp]])[)
+                           const location_type* yylocationp]])[) const
   {
-    *yycdebug_ << (yytype < YYNTOKENS ? "token" : "nterm")
-               << ' ' << yytname[yytype] << " ("]b4_locations_if([[
+    *yycdebug_ << (yykind < YYNTOKENS ? "token" : "nterm")
+               << ' ' << yytname[yykind] << " ("]b4_locations_if([[
                << *yylocationp << ": "]])[;
-    yy_symbol_value_print_ (yytype, yyvaluep]b4_locations_if([[, yylocationp]])[);
+    yy_symbol_value_print_ (yykind, yyvaluep]b4_locations_if([[, yylocationp]])[);
     *yycdebug_ << ')';
   }
 
@@ -328,21 +312,21 @@ b4_percent_code_get([[requires]])[
 # if ]b4_api_PREFIX[DEBUG
   public:
     /// \brief Report a symbol value on the debug stream.
-    /// \param yytype       The token type.
+    /// \param yykind       The symbol kind.
     /// \param yyvaluep     Its semantic value.]b4_locations_if([[
     /// \param yylocationp  Its location.]])[
-    virtual void yy_symbol_value_print_ (int yytype,
+    virtual void yy_symbol_value_print_ (symbol_kind_type yykind,
                                          const semantic_type* yyvaluep]b4_locations_if([[,
-                                         const location_type* yylocationp]])[);
+                                         const location_type* yylocationp]])[) const;
     /// \brief Report a symbol on the debug stream.
-    /// \param yytype       The token type.
+    /// \param yykind       The symbol kind.
     /// \param yyvaluep     Its semantic value.]b4_locations_if([[
     /// \param yylocationp  Its location.]])[
-    virtual void yy_symbol_print_ (int yytype,
+    virtual void yy_symbol_print_ (symbol_kind_type yykind,
                                    const semantic_type* yyvaluep]b4_locations_if([[,
-                                   const location_type* yylocationp]])[);
+                                   const location_type* yylocationp]])[) const;
   private:
-    // Debugging.
+    /// Debug stream.
     std::ostream* yycdebug_;
 #endif
 
@@ -351,8 +335,18 @@ b4_percent_code_get([[requires]])[
 
 ]dnl Redirections for glr.c.
 b4_percent_define_flag_if([[global_tokens_and_yystype]],
-[b4_token_defines])
-[
+[b4_token_defines
+])[
+]b4_namespace_close[
+
+]dnl Map the name used in c.m4 to the one used in c++.m4.
+[#undef ]b4_symbol(-2, [id])[
+#define ]b4_symbol(-2, [id])[ ]b4_namespace_ref[::]b4_parser_class[::token::]b4_symbol(-2, [id])[
+#undef ]b4_symbol(0, [id])[
+#define ]b4_symbol(0, [id])[ ]b4_namespace_ref[::]b4_parser_class[::token::]b4_symbol(0, [id])[
+#undef ]b4_symbol(1, [id])[
+#define ]b4_symbol(1, [id])[ ]b4_namespace_ref[::]b4_parser_class[::token::]b4_symbol(1, [id])[
+
 #ifndef ]b4_api_PREFIX[STYPE
 # define ]b4_api_PREFIX[STYPE ]b4_namespace_ref[::]b4_parser_class[::semantic_type
 #endif
@@ -360,7 +354,13 @@ b4_percent_define_flag_if([[global_tokens_and_yystype]],
 # define ]b4_api_PREFIX[LTYPE ]b4_namespace_ref[::]b4_parser_class[::location_type
 #endif
 
-]b4_namespace_close[
+]m4_define([b4_declare_symbol_enum],
+[[typedef ]b4_namespace_ref[::]b4_parser_class[::symbol_kind_type yysymbol_kind_t;
+#define ]b4_symbol_prefix[YYEMPTY  ]b4_namespace_ref[::]b4_parser_class[::symbol_kind::]b4_symbol_prefix[YYEMPTY
+#define ]b4_symbol_prefix[YYerror  ]b4_namespace_ref[::]b4_parser_class[::symbol_kind::]b4_symbol_prefix[YYerror
+#define ]b4_symbol_prefix[YYEOF    ]b4_namespace_ref[::]b4_parser_class[::symbol_kind::]b4_symbol_prefix[YYEOF
+#define ]b4_symbol_prefix[YYUNDEF  ]b4_namespace_ref[::]b4_parser_class[::symbol_kind::]b4_symbol_prefix[YYUNDEF
+]])[
 ]b4_percent_code_get([[provides]])[
 ]m4_popdef([b4_parse_param])dnl
 ])

data/skeletons/java.m4

@@ -106,10 +106,13 @@ m4_define([b4_null], [null])
 
 # b4_typed_parser_table_define(TYPE, NAME, DATA, COMMENT)
 # -------------------------------------------------------
+# We use intermediate functions (e.g., yypact_init) to work around the
+# 64KB limit for JVM methods.  See
+# https://lists.gnu.org/r/help-bison/2008-11/msg00004.html.
 m4_define([b4_typed_parser_table_define],
 [m4_ifval([$4], [b4_comment([$4])
   ])dnl
-[private static final ]$1[ yy$2_[] = yy$2_init();
+[private static final ]$1[[] yy$2_ = yy$2_init();
   private static final ]$1[[] yy$2_init()
   {
     return new ]$1[[]
@@ -125,40 +128,151 @@ m4_define([b4_integral_parser_table_define],
 [b4_typed_parser_table_define([b4_int_type_for([$2])], [$1], [$2], [$3])])
 
 
-## ------------------------- ##
-## Assigning token numbers.  ##
-## ------------------------- ##
+## ------------- ##
+## Token kinds.  ##
+## ------------- ##
+
 
 # b4_token_enum(TOKEN-NUM)
 # ------------------------
 # Output the definition of this token as an enum.
 m4_define([b4_token_enum],
-[b4_token_format([    /** Token number, to be returned by the scanner.  */
-    static final int %s = %s;
-], [$1])])
+[b4_token_visible_if([$1],
+    [m4_format([[    /** Token %s, to be returned by the scanner.  */
+    static final int %s = %s%s;
+]],
+               b4_symbol([$1], [tag]),
+               b4_symbol([$1], [id]),
+               b4_symbol([$1], b4_api_token_raw_if([[number]], [[user_number]])))])])
+
 
 # b4_token_enums
 # --------------
 # Output the definition of the tokens (if there are) as enums.
 m4_define([b4_token_enums],
-[b4_any_token_visible_if([/* Tokens.  */
+[b4_any_token_visible_if([    /* Token kinds.  */
 b4_symbol_foreach([b4_token_enum])])])
 
-# b4-case(ID, CODE)
+
+
+## -------------- ##
+## Symbol kinds.  ##
+## -------------- ##
+
+# b4_symbol_enum(SYMBOL-NUM)
+# --------------------------
+# Output the definition of this symbol as an enum.
+m4_define([b4_symbol_enum],
+[m4_format([    %-30s %s],
+           m4_format([[%s(%s)%s]],
+                     b4_symbol([$1], [kind]),
+                     [$1],
+                     m4_if([$1], b4_last_symbol, [[;]], [[,]])),
+           [b4_symbol_tag_comment([$1])])])
+
+
+# b4_declare_symbol_enum
+# ----------------------
+# The definition of the symbol internal numbers as an enum.
+m4_define([b4_declare_symbol_enum],
+[[  public enum SymbolKind
+  {
+]b4_symbol_foreach([b4_symbol_enum])[
+
+    private final int yycode_;
+
+    SymbolKind (int n) {
+      this.yycode_ = n;
+    }
+
+    private static final SymbolKind[] values_ = {
+      ]m4_map_args_sep([b4_symbol_kind(], [)], [,
+      ], b4_symbol_numbers)[
+    };
+
+    static final SymbolKind get(int code) {
+      return values_[code];
+    }
+
+    public final int getCode() {
+      return this.yycode_;
+    }
+
+]b4_parse_error_bmatch(
+[simple\|verbose],
+[[    /* Return YYSTR after stripping away unnecessary quotes and
+       backslashes, so that it's suitable for yyerror.  The heuristic is
+       that double-quoting is unnecessary unless the string contains an
+       apostrophe, a comma, or backslash (other than backslash-backslash).
+       YYSTR is taken from yytname.  */
+    private static String yytnamerr_(String yystr)
+    {
+      if (yystr.charAt (0) == '"')
+        {
+          StringBuffer yyr = new StringBuffer();
+          strip_quotes: for (int i = 1; i < yystr.length(); i++)
+            switch (yystr.charAt(i))
+              {
+              case '\'':
+              case ',':
+                break strip_quotes;
+
+              case '\\':
+                if (yystr.charAt(++i) != '\\')
+                  break strip_quotes;
+                /* Fall through.  */
+              default:
+                yyr.append(yystr.charAt(i));
+                break;
+
+              case '"':
+                return yyr.toString();
+              }
+        }
+      return yystr;
+    }
+
+    /* YYTNAME[SYMBOL-NUM] -- String name of the symbol SYMBOL-NUM.
+       First, the terminals, then, starting at \a YYNTOKENS_, nonterminals.  */
+    ]b4_typed_parser_table_define([String], [tname], [b4_tname])[
+
+    /* The user-facing name of this symbol.  */
+    public final String getName() {
+      return yytnamerr_(yytname_[yycode_]);
+    }
+]],
+[custom\|detailed],
+[[    /* YYNAMES_[SYMBOL-NUM] -- String name of the symbol SYMBOL-NUM.
+       First, the terminals, then, starting at \a YYNTOKENS_, nonterminals.  */
+    ]b4_typed_parser_table_define([String], [names], [b4_symbol_names])[
+
+    /* The user-facing name of this symbol.  */
+    public final String getName() {
+      return yynames_[yycode_];
+    }]])[
+  };
+]])])
+
+
+
+# b4_case(ID, CODE)
 # -----------------
 # We need to fool Java's stupid unreachable code detection.
-m4_define([b4_case], [  case $1:
+m4_define([b4_case],
+[  case $1:
   if (yyn == $1)
     $2;
   break;
-    ])
+])
+
 
 # b4_predicate_case(LABEL, CONDITIONS)
 # ------------------------------------
-m4_define([b4_predicate_case], [  case $1:
+m4_define([b4_predicate_case],
+[  case $1:
      if (! ($2)) YYERROR;
     break;
-    ])
+])
 
 
 ## -------- ##
@@ -183,6 +297,7 @@ b4_percent_define_check_kind([[throws]],            [code], [deprecated])
 
 m4_define([b4_yystype], [b4_percent_define_get([[api.value.type]])])
 b4_percent_define_default([[api.value.type]], [[Object]])
+b4_percent_define_default([[api.symbol.prefix]], [[S_]])
 
 # b4_api_prefix, b4_api_PREFIX
 # ----------------------------
@@ -222,6 +337,22 @@ m4_define([b4_position_type], [b4_percent_define_get([[api.position.type]])])
 ## ----------------- ##
 
 
+# b4_symbol_translate(STRING)
+# ---------------------------
+# Used by "bison" in the array of symbol names to mark those that
+# require translation.
+m4_define([b4_symbol_translate],
+[[i18n($1)]])
+
+
+# b4_trans(STRING)
+# ----------------
+# Translate a string if i18n is enabled.  Avoid collision with b4_translate.
+m4_define([b4_trans],
+[b4_has_translations_if([i18n($1)], [$1])])
+
+
+
 # b4_symbol_value(VAL, [SYMBOL-NUM], [TYPE-TAG])
 # ----------------------------------------------
 # See README.
@@ -361,4 +492,4 @@ m4_define([b4_var_decl],
 # -----------------------
 # Expand to either an empty string or "throws THROWS".
 m4_define([b4_maybe_throws],
-          [m4_ifval($1, [throws $1])])
+          [m4_ifval($1, [ throws $1])])

data/skeletons/lalr1.cc

@@ -23,12 +23,22 @@ m4_define([b4_value_type_setup_variant])
 # Check the value of %define parse.lac, where LAC stands for lookahead
 # correction.
 b4_percent_define_default([[parse.lac]], [[none]])
+b4_percent_define_check_values([[[[parse.lac]], [[full]], [[none]]]])
 b4_define_flag_if([lac])
 m4_define([b4_lac_flag],
           [m4_if(b4_percent_define_get([[parse.lac]]),
                  [none], [[0]], [[1]])])
 
 
+# b4_tname_if(TNAME-NEEDED, TNAME-NOT-NEEDED)
+# -------------------------------------------
+m4_define([b4_tname_if],
+[m4_case(b4_percent_define_get([[parse.error]]),
+         [verbose],         [$1],
+         [b4_token_table_if([$1],
+                            [$2])])])
+
+
 # b4_integral_parser_table_declare(TABLE-NAME, CONTENT, COMMENT)
 # --------------------------------------------------------------
 # Declare "parser::yy<TABLE-NAME>_" whose contents is CONTENT.
@@ -225,7 +235,25 @@ m4_define([b4_shared_declarations],
     void error (const syntax_error& err);
 
 ]b4_token_constructor_define[
+]b4_parse_error_bmatch([custom\|detailed\|verbose], [[
+    class context
+    {
+    public:
+      context (const ]b4_parser_class[& yyparser, const symbol_type& yyla);
+      const symbol_type& lookahead () const { return yyla_; }
+      symbol_kind_type token () const { return yyla_.kind (); }]b4_locations_if([[
+      const location_type& location () const { return yyla_.location; }
+]])[
+      /// Put in YYARG at most YYARGN of the expected tokens, and return the
+      /// number of tokens stored in YYARG.  If YYARG is null, return the
+      /// number of expected tokens (guaranteed to be less than YYNTOKENS).
+      int expected_tokens (symbol_kind_type yyarg[], int yyargn) const;
 
+    private:
+      const ]b4_parser_class[& yyparser_;
+      const symbol_type& yyla_;
+    };
+]])[
   private:
     /// This class is not copyable.
     ]b4_parser_class[ (const ]b4_parser_class[&);
@@ -233,10 +261,10 @@ m4_define([b4_shared_declarations],
 
     /// Check the lookahead yytoken.
     /// \returns  true iff the token will be eventually shifted.
-    bool yy_lac_check_ (int yytoken) const;
+    bool yy_lac_check_ (symbol_kind_type yytoken) const;
     /// Establish the initial context if no initial context currently exists.
     /// \returns  true iff the token will be eventually shifted.
-    bool yy_lac_establish_ (int yytoken);
+    bool yy_lac_establish_ (symbol_kind_type yytoken);
     /// Discard any previous initial lookahead context because of event.
     /// \param event  the event which caused the lookahead to be discarded.
     ///               Only used for debbuging output.
@@ -244,13 +272,19 @@ m4_define([b4_shared_declarations],
 
     /// Stored state numbers (used for stacks).
     typedef ]b4_int_type(0, m4_eval(b4_states_number - 1))[ state_type;
+]b4_parse_error_bmatch(
+      [custom], [[
+    /// Report a syntax error
+    /// \param yyctx     the context in which the error occurred.
+    void report_syntax_error (const context& yyctx) const;]],
+      [detailed\|verbose], [[
+    /// The arguments of the error message.
+    int yy_syntax_error_arguments_ (const context& yyctx,
+                                    symbol_kind_type yyarg[], int yyargn) const;
 
     /// Generate an error message.
-    /// \param yystate   the state where the error occurred.
-    /// \param yyla      the lookahead token.
-    virtual std::string yysyntax_error_ (state_type yystate,
-                                         const symbol_type& yyla) const;
-
+    /// \param yyctx     the context in which the error occurred.
+    virtual std::string yysyntax_error_ (const context& yyctx) const;]])[
     /// Compute post-reduction state.
     /// \param yystate   the current state
     /// \param yysym     the nonterminal to push on the stack
@@ -267,34 +301,55 @@ m4_define([b4_shared_declarations],
     static const ]b4_int_type(b4_pact_ninf, b4_pact_ninf)[ yypact_ninf_;
     static const ]b4_int_type(b4_table_ninf, b4_table_ninf)[ yytable_ninf_;
 
-    /// Convert a scanner token number \a t to a symbol number.
-    /// In theory \a t should be a token_type, but character literals
+    /// Convert a scanner token kind \a t to a symbol kind.
+    /// In theory \a t should be a token_kind_type, but character literals
     /// are valid, yet not members of the token_type enum.
-    static token_number_type yytranslate_ (int t);
+    static symbol_kind_type yytranslate_ (int t);
 
-    // Tables.
-]b4_parser_tables_declare[]b4_error_verbose_if([
+]b4_parse_error_bmatch(
+[custom\|detailed],
+[[    /// The user-facing name of the symbol whose (internal) number is
+    /// YYSYMBOL.  No bounds checking.
+    static const char *symbol_name (symbol_kind_type yysymbol);]],
+[simple],
+[[#if ]b4_api_PREFIX[DEBUG || ]b4_token_table_flag[
+    /// The user-facing name of the symbol whose (internal) number is
+    /// YYSYMBOL.  No bounds checking.
+    static const char *symbol_name (symbol_kind_type yysymbol);
 
-    /// Convert the symbol name \a n to a form suitable for a diagnostic.
-    static std::string yytnamerr_ (const char *n);])[
-
-]b4_token_table_if([], [[#if ]b4_api_PREFIX[DEBUG]])[
     /// For a symbol, its name in clear.
     static const char* const yytname_[];
-]b4_token_table_if([[#if ]b4_api_PREFIX[DEBUG]])[
+#endif // #if ]b4_api_PREFIX[DEBUG || ]b4_token_table_flag[
+]],
+[verbose],
+[[    /// The user-facing name of the symbol whose (internal) number is
+    /// YYSYMBOL.  No bounds checking.
+    static std::string symbol_name (symbol_kind_type yysymbol);
+
+    /// Convert the symbol name \a n to a form suitable for a diagnostic.
+    static std::string yytnamerr_ (const char *yystr);
+
+    /// For a symbol, its name in clear.
+    static const char* const yytname_[];
+]])[
+
+    // Tables.
+]b4_parser_tables_declare[
+
+#if ]b4_api_PREFIX[DEBUG
 ]b4_integral_parser_table_declare([rline], [b4_rline],
      [[YYRLINE[YYN] -- Source line where rule number YYN was defined.]])[
     /// Report on the debug stream that the rule \a r is going to be reduced.
-    virtual void yy_reduce_print_ (int r);
+    virtual void yy_reduce_print_ (int r) const;
     /// Print the state stack on the debug stream.
-    virtual void yystack_print_ ();
+    virtual void yy_stack_print_ () const;
 
     /// Debugging level.
     int yydebug_;
     /// Debug stream.
     std::ostream* yycdebug_;
 
-    /// \brief Display a symbol type, value and location.
+    /// \brief Display a symbol kind, value and location.
     /// \param yyo    The output stream.
     /// \param yysym  The symbol.
     template <typename Base>
@@ -315,7 +370,7 @@ m4_define([b4_shared_declarations],
       /// Default constructor.
       by_state () YY_NOEXCEPT;
 
-      /// The symbol type as needed by the constructor.
+      /// The symbol kind as needed by the constructor.
       typedef state_type kind_type;
 
       /// Constructor.
@@ -327,12 +382,12 @@ m4_define([b4_shared_declarations],
       /// Record that this symbol is empty.
       void clear () YY_NOEXCEPT;
 
-      /// Steal the symbol type from \a that.
+      /// Steal the symbol kind from \a that.
       void move (by_state& that);
 
-      /// The (internal) type number (corresponding to \a state).
-      /// \a empty_symbol when empty.
-      symbol_number_type type_get () const YY_NOEXCEPT;
+      /// The symbol kind (corresponding to \a state).
+      /// \a ]b4_symbol(-2, kind)[ when empty.
+      symbol_kind_type kind () const YY_NOEXCEPT;
 
       /// The state number used to denote an empty symbol.
       /// We use the initial state, as it does not have a value.
@@ -400,18 +455,12 @@ m4_define([b4_shared_declarations],
     /// Pop \a n symbols from the stack.
     void yypop_ (int n = 1);
 
-    /// Some specific tokens.
-    static const token_number_type yy_error_token_ = 1;
-    static const token_number_type yy_undef_token_ = ]b4_undef_token_number[;
-
     /// Constants.
     enum
     {
-      yyeof_ = 0,
       yylast_ = ]b4_last[,     ///< Last index in yytable_.
       yynnts_ = ]b4_nterms_number[,  ///< Number of nonterminal symbols.
-      yyfinal_ = ]b4_final_state_number[, ///< Termination state number.
-      yyntokens_ = ]b4_tokens_number[  ///< Number of tokens.
+      yyfinal_ = ]b4_final_state_number[ ///< Termination state number.
     };
 
 ]b4_parse_param_vars[
@@ -483,6 +532,11 @@ m4_if(b4_prefix, [yy], [],
 #  define YY_(msgid) msgid
 # endif
 #endif
+]b4_has_translations_if([
+#ifndef N_
+# define N_(Msgid) Msgid
+#endif
+])[
 
 // Whether we are compiled with exception support.
 #ifndef YY_EXCEPTIONS
@@ -522,7 +576,7 @@ m4_if(b4_prefix, [yy], [],
 # define YY_STACK_PRINT()               \
   do {                                  \
     if (yydebug_)                       \
-      yystack_print_ ();                \
+      yy_stack_print_ ();                \
   } while (false)
 
 #else // !]b4_api_PREFIX[DEBUG
@@ -542,9 +596,38 @@ m4_if(b4_prefix, [yy], [],
 #define YYERROR         goto yyerrorlab
 #define YYRECOVERING()  (!!yyerrstatus_)
 
-]b4_namespace_open[]b4_error_verbose_if([[
-
-  /* Return YYSTR after stripping away unnecessary quotes and
+]b4_namespace_open[
+]b4_parse_error_bmatch([custom\|detailed],
+[[  const char *
+  ]b4_parser_class[::symbol_name (symbol_kind_type yysymbol)
+  {
+    static const char *const yy_sname[] =
+    {
+    ]b4_symbol_names[
+    };]b4_has_translations_if([[
+    /* YYTRANSLATABLE[SYMBOL-NUM] -- Whether YY_SNAME[SYMBOL-NUM] is
+       internationalizable.  */
+    static ]b4_int_type_for([b4_translatable])[ yytranslatable[] =
+    {
+    ]b4_translatable[
+    };
+    return (yysymbol < YYNTOKENS && yytranslatable[yysymbol]
+            ? _(yy_sname[yysymbol])
+            : yy_sname[yysymbol]);]], [[
+    return yy_sname[yysymbol];]])[
+  }
+]],
+[simple],
+[[#if ]b4_api_PREFIX[DEBUG || ]b4_token_table_flag[
+  const char *
+  ]b4_parser_class[::symbol_name (symbol_kind_type yysymbol)
+  {
+    return yytname_[yysymbol];
+  }
+#endif // #if ]b4_api_PREFIX[DEBUG || ]b4_token_table_flag[
+]],
+[verbose],
+[[  /* Return YYSTR after stripping away unnecessary quotes and
      backslashes, so that it's suitable for yyerror.  The heuristic is
      that double-quoting is unnecessary unless the string contains an
      apostrophe, a comma, or backslash (other than backslash-backslash).
@@ -583,6 +666,12 @@ m4_if(b4_prefix, [yy], [],
 
     return yystr;
   }
+
+  std::string
+  ]b4_parser_class[::symbol_name (symbol_kind_type yysymbol)
+  {
+    return yytnamerr_ (yytname_[yysymbol]);
+  }
 ]])[
 
   /// Build a parser object.
@@ -603,7 +692,7 @@ m4_if(b4_prefix, [yy], [],
   {}
 
   /*---------------.
-  | Symbol types.  |
+  | symbol kinds.  |
   `---------------*/
 
 ]b4_token_ctor_if([], [b4_public_types_define([cc])])[
@@ -634,13 +723,13 @@ m4_if(b4_prefix, [yy], [],
     : state (s)
   {}
 
-  ]b4_parser_class[::symbol_number_type
-  ]b4_parser_class[::by_state::type_get () const YY_NOEXCEPT
+  ]b4_parser_class[::symbol_kind_type
+  ]b4_parser_class[::by_state::kind () const YY_NOEXCEPT
   {
     if (state == empty_state)
-      return empty_symbol;
+      return symbol_kind::]b4_symbol(-2, kind)[;
     else
-      return yystos_[+state];
+      return YY_CAST (symbol_kind_type, yystos_[+state]);
   }
 
   ]b4_parser_class[::stack_symbol_type::stack_symbol_type ()
@@ -649,7 +738,7 @@ m4_if(b4_prefix, [yy], [],
   ]b4_parser_class[::stack_symbol_type::stack_symbol_type (YY_RVREF (stack_symbol_type) that)
     : super_type (YY_MOVE (that.state)]b4_variant_if([], [, YY_MOVE (that.value)])b4_locations_if([, YY_MOVE (that.location)])[)
   {]b4_variant_if([
-    b4_symbol_variant([that.type_get ()],
+    b4_symbol_variant([that.kind ()],
                       [value], [YY_MOVE_OR_COPY], [YY_MOVE (that.value)])])[
 #if 201103L <= YY_CPLUSPLUS
     // that is emptied.
@@ -660,10 +749,10 @@ m4_if(b4_prefix, [yy], [],
   ]b4_parser_class[::stack_symbol_type::stack_symbol_type (state_type s, YY_MOVE_REF (symbol_type) that)
     : super_type (s]b4_variant_if([], [, YY_MOVE (that.value)])[]b4_locations_if([, YY_MOVE (that.location)])[)
   {]b4_variant_if([
-    b4_symbol_variant([that.type_get ()],
+    b4_symbol_variant([that.kind ()],
                       [value], [move], [YY_MOVE (that.value)])])[
     // that is emptied.
-    that.type = empty_symbol;
+    that.kind_ = symbol_kind::]b4_symbol(-2, kind)[;
   }
 
 #if YY_CPLUSPLUS < 201103L
@@ -671,7 +760,7 @@ m4_if(b4_prefix, [yy], [],
   ]b4_parser_class[::stack_symbol_type::operator= (const stack_symbol_type& that)
   {
     state = that.state;
-    ]b4_variant_if([b4_symbol_variant([that.type_get ()],
+    ]b4_variant_if([b4_symbol_variant([that.kind ()],
                                       [value], [copy], [that.value])],
                    [[value = that.value;]])[]b4_locations_if([
     location = that.location;])[
@@ -682,7 +771,7 @@ m4_if(b4_prefix, [yy], [],
   ]b4_parser_class[::stack_symbol_type::operator= (stack_symbol_type& that)
   {
     state = that.state;
-    ]b4_variant_if([b4_symbol_variant([that.type_get ()],
+    ]b4_variant_if([b4_symbol_variant([that.kind ()],
                                       [value], [move], [that.value])],
                    [[value = that.value;]])[]b4_locations_if([
     location = that.location;])[
@@ -700,29 +789,27 @@ m4_if(b4_prefix, [yy], [],
       YY_SYMBOL_PRINT (yymsg, yysym);]b4_variant_if([], [
 
     // User destructor.
-    b4_symbol_actions([destructor], [yysym.type_get ()])])[
+    b4_symbol_actions([destructor], [yysym.kind ()])])[
   }
 
 #if ]b4_api_PREFIX[DEBUG
   template <typename Base>
   void
-  ]b4_parser_class[::yy_print_ (std::ostream& yyo,
-                                     const basic_symbol<Base>& yysym) const
+  ]b4_parser_class[::yy_print_ (std::ostream& yyo, const basic_symbol<Base>& yysym) const
   {
     std::ostream& yyoutput = yyo;
     YYUSE (yyoutput);
-    symbol_number_type yytype = yysym.type_get ();
-#if defined __GNUC__ && ! defined __clang__ && ! defined __ICC && __GNUC__ * 100 + __GNUC_MINOR__ <= 408
-    // Avoid a (spurious) G++ 4.8 warning about "array subscript is
-    // below array bounds".
     if (yysym.empty ())
-      std::abort ();
-#endif
-    yyo << (yytype < yyntokens_ ? "token" : "nterm")
-        << ' ' << yytname_[yytype] << " ("]b4_locations_if([
-        << yysym.location << ": "])[;
-    ]b4_symbol_actions([printer])[
-    yyo << ')';
+      yyo << "empty symbol";
+    else
+      {
+        symbol_kind_type yykind = yysym.kind ();
+        yyo << (yykind < YYNTOKENS ? "token" : "nterm")
+            << ' ' << symbol_name (yykind) << " ("]b4_locations_if([
+            << yysym.location << ": "])[;
+        ]b4_symbol_actions([printer])[
+        yyo << ')';
+      }
   }
 #endif
 
@@ -781,11 +868,11 @@ m4_if(b4_prefix, [yy], [],
   ]b4_parser_class[::state_type
   ]b4_parser_class[::yy_lr_goto_state_ (state_type yystate, int yysym)
   {
-    int yyr = yypgoto_[yysym - yyntokens_] + yystate;
+    int yyr = yypgoto_[yysym - YYNTOKENS] + yystate;
     if (0 <= yyr && yyr <= yylast_ && yycheck_[yyr] == yystate)
       return yytable_[yyr];
     else
-      return yydefgoto_[yysym - yyntokens_];
+      return yydefgoto_[yysym - YYNTOKENS];
   }
 
   bool
@@ -853,6 +940,7 @@ b4_dollar_popdef])[]dnl
   `-----------------------------------------------*/
   yynewstate:
     YYCDEBUG << "Entering state " << int (yystack_[0].state) << '\n';
+    YY_STACK_PRINT ();
 
     // Accept?
     if (yystack_[0].state == yyfinal_)
@@ -873,14 +961,14 @@ b4_dollar_popdef])[]dnl
     // Read a lookahead token.
     if (yyla.empty ())
       {
-        YYCDEBUG << "Reading a token: ";
+        YYCDEBUG << "Reading a token\n";
 #if YY_EXCEPTIONS
         try
 #endif // YY_EXCEPTIONS
           {]b4_token_ctor_if([[
             symbol_type yylookahead (]b4_lex[);
             yyla.move (yylookahead);]], [[
-            yyla.type = yytranslate_ (]b4_lex[);]])[
+            yyla.kind_ = yytranslate_ (]b4_lex[);]])[
           }
 #if YY_EXCEPTIONS
         catch (const syntax_error& yyexc)
@@ -893,12 +981,22 @@ b4_dollar_popdef])[]dnl
       }
     YY_SYMBOL_PRINT ("Next token is", yyla);
 
+    if (yyla.kind () == ]symbol_kind::b4_symbol(1, kind)[)
+    {
+      // The scanner already issued an error message, process directly
+      // to error recovery.  But do not keep the error token as
+      // lookahead, it is too special and may lead us to an endless
+      // loop in error recovery. */
+      yyla.kind_ = ]symbol_kind::b4_symbol(2, kind)[;
+      goto yyerrlab1;
+    }
+
     /* If the proper action on seeing token YYLA.TYPE is to reduce or
        to detect an error, take that action.  */
-    yyn += yyla.type_get ();
-    if (yyn < 0 || yylast_ < yyn || yycheck_[yyn] != yyla.type_get ())
+    yyn += yyla.kind ();
+    if (yyn < 0 || yylast_ < yyn || yycheck_[yyn] != yyla.kind ())
       {]b4_lac_if([[
-        if (!yy_lac_establish_ (yyla.type_get ()))
+        if (!yy_lac_establish_ (yyla.kind ()))
            goto yyerrlab;]])[
         goto yydefault;
       }
@@ -909,7 +1007,7 @@ b4_dollar_popdef])[]dnl
       {
         if (yy_table_value_is_error_ (yyn))
           goto yyerrlab;]b4_lac_if([[
-        if (!yy_lac_establish_ (yyla.type_get ()))
+        if (!yy_lac_establish_ (yyla.kind ()))
            goto yyerrlab;
 ]])[
         yyn = -yyn;
@@ -991,7 +1089,6 @@ b4_dollar_popdef])[]dnl
       YY_SYMBOL_PRINT ("-> $$ =", yylhs);
       yypop_ (yylen);
       yylen = 0;
-      YY_STACK_PRINT ();
 
       // Shift the result of the reduction.
       yypush_ (YY_NULLPTR, YY_MOVE (yylhs));
@@ -1006,9 +1103,17 @@ b4_dollar_popdef])[]dnl
     // If not already recovering from an error, report this error.
     if (!yyerrstatus_)
       {
-        ++yynerrs_;
-        error (]b4_join(b4_locations_if([yyla.location]),
-                        [[yysyntax_error_ (yystack_[0].state, yyla)]])[);
+        ++yynerrs_;]b4_parse_error_case(
+                  [simple], [[
+        std::string msg = YY_("syntax error");
+        error (]b4_join(b4_locations_if([yyla.location]), [[YY_MOVE (msg)]])[);]],
+                  [custom], [[
+        context yyctx (*this, yyla);
+        report_syntax_error (yyctx);]],
+                  [[
+        context yyctx (*this, yyla);
+        std::string msg = yysyntax_error_ (yyctx);
+        error (]b4_join(b4_locations_if([yyla.location]), [[YY_MOVE (msg)]])[);]])[
       }
 
 ]b4_locations_if([[
@@ -1019,7 +1124,7 @@ b4_dollar_popdef])[]dnl
            error, discard it.  */
 
         // Return failure if at end of input.
-        if (yyla.type_get () == yyeof_)
+        if (yyla.kind () == symbol_kind::]b4_symbol_prefix[YYEOF)
           YYABORT;
         else if (!yyla.empty ())
           {
@@ -1045,6 +1150,7 @@ b4_dollar_popdef])[]dnl
        this YYERROR.  */
     yypop_ (yylen);
     yylen = 0;
+    YY_STACK_PRINT ();
     goto yyerrlab1;
 
 
@@ -1053,31 +1159,33 @@ b4_dollar_popdef])[]dnl
   `-------------------------------------------------------------*/
   yyerrlab1:
     yyerrstatus_ = 3;   // Each real token shifted decrements this.
+    // Pop stack until we find a state that shifts the error token.
+    for (;;)
+      {
+        yyn = yypact_[+yystack_[0].state];
+        if (!yy_pact_value_is_default_ (yyn))
+          {
+            yyn += symbol_kind::]b4_symbol(1, kind)[;
+            if (0 <= yyn && yyn <= yylast_
+                && yycheck_[yyn] == symbol_kind::]b4_symbol(1, kind)[)
+              {
+                yyn = yytable_[yyn];
+                if (0 < yyn)
+                  break;
+              }
+          }
+
+        // Pop the current state because it cannot handle the error token.
+        if (yystack_.size () == 1)
+          YYABORT;
+]b4_locations_if([[
+        yyerror_range[1].location = yystack_[0].location;]])[
+        yy_destroy_ ("Error: popping", yystack_[0]);
+        yypop_ ();
+        YY_STACK_PRINT ();
+      }
     {
       stack_symbol_type error_token;
-      for (;;)
-        {
-          yyn = yypact_[+yystack_[0].state];
-          if (!yy_pact_value_is_default_ (yyn))
-            {
-              yyn += yy_error_token_;
-              if (0 <= yyn && yyn <= yylast_ && yycheck_[yyn] == yy_error_token_)
-                {
-                  yyn = yytable_[yyn];
-                  if (0 < yyn)
-                    break;
-                }
-            }
-
-          // Pop the current state because it cannot handle the error token.
-          if (yystack_.size () == 1)
-            YYABORT;
-]b4_locations_if([[
-          yyerror_range[1].location = yystack_[0].location;]])[
-          yy_destroy_ ("Error: popping", yystack_[0]);
-          yypop_ ();
-          YY_STACK_PRINT ();
-        }
 ]b4_locations_if([[
       yyerror_range[2].location = yyla.location;
       YYLLOC_DEFAULT (error_token.location, yyerror_range, 2);]])[
@@ -1116,6 +1224,7 @@ b4_dollar_popdef])[]dnl
     /* Do not reclaim the symbols of the rule whose action triggered
        this YYABORT or YYACCEPT.  */
     yypop_ (yylen);
+    YY_STACK_PRINT ();
     while (1 < yystack_.size ())
       {
         yy_destroy_ ("Cleanup: popping", yystack_[0]);
@@ -1148,17 +1257,78 @@ b4_dollar_popdef])[]dnl
   {
     error (]b4_join(b4_locations_if([yyexc.location]),
                     [[yyexc.what ()]])[);
-  }]b4_lac_if([[
+  }]b4_parse_error_bmatch([custom\|detailed\|verbose], [[
 
+  // ]b4_parser_class[::context.
+  ]b4_parser_class[::context::context (const ]b4_parser_class[& yyparser, const symbol_type& yyla)
+    : yyparser_ (yyparser)
+    , yyla_ (yyla)
+  {}
+
+  int
+  ]b4_parser_class[::context::expected_tokens (symbol_kind_type yyarg[], int yyargn) const
+  {
+    // Actual number of expected tokens
+    int yycount = 0;
+]b4_lac_if([[
+#if ]b4_api_PREFIX[DEBUG
+    // Execute LAC once. We don't care if it is successful, we
+    // only do it for the sake of debugging output.
+    if (!yyparser_.yy_lac_established_)
+      yyparser_.yy_lac_check_ (yyla_.kind ());
+#endif
+
+    for (int yyx = 0; yyx < YYNTOKENS; ++yyx)
+      {
+        symbol_kind_type yysym = YY_CAST (symbol_kind_type, yyx);
+        if (yysym != symbol_kind::]b4_symbol(1, kind)[
+            && yysym != symbol_kind::]b4_symbol_prefix[YYUNDEF
+            && yyparser_.yy_lac_check_ (yysym))
+          {
+            if (!yyarg)
+              ++yycount;
+            else if (yycount == yyargn)
+              return 0;
+            else
+              yyarg[yycount++] = yysym;
+          }
+      }]], [[
+    int yyn = yypact_[+yyparser_.yystack_[0].state];
+    if (!yy_pact_value_is_default_ (yyn))
+      {
+        /* Start YYX at -YYN if negative to avoid negative indexes in
+           YYCHECK.  In other words, skip the first -YYN actions for
+           this state because they are default actions.  */
+        int yyxbegin = yyn < 0 ? -yyn : 0;
+        // Stay within bounds of both yycheck and yytname.
+        int yychecklim = yylast_ - yyn + 1;
+        int yyxend = yychecklim < YYNTOKENS ? yychecklim : YYNTOKENS;
+        for (int yyx = yyxbegin; yyx < yyxend; ++yyx)
+          if (yycheck_[yyx + yyn] == yyx && yyx != symbol_kind::]b4_symbol(1, kind)[
+              && !yy_table_value_is_error_ (yytable_[yyx + yyn]))
+            {
+              if (!yyarg)
+                ++yycount;
+              else if (yycount == yyargn)
+                return 0;
+              else
+                yyarg[yycount++] = YY_CAST (symbol_kind_type, yyx);
+            }
+      }
+]])[
+    return yycount;
+  }
+
+]])b4_lac_if([[
   bool
-  ]b4_parser_class[::yy_lac_check_ (int yytoken) const
+  ]b4_parser_class[::yy_lac_check_ (symbol_kind_type yytoken) const
   {
     // Logically, the yylac_stack's lifetime is confined to this function.
     // Clear it, to get rid of potential left-overs from previous call.
     yylac_stack_.clear ();
     // Reduce until we encounter a shift and thereby accept the token.
 #if ]b4_api_PREFIX[DEBUG
-    YYCDEBUG << "LAC: checking lookahead " << yytname_[yytoken] << ':';
+    YYCDEBUG << "LAC: checking lookahead " << symbol_name (yytoken) << ':';
 #endif
     std::ptrdiff_t lac_top = 0;
     while (true)
@@ -1229,7 +1399,7 @@ b4_dollar_popdef])[]dnl
 
   // Establish the initial context if no initial context currently exists.
   bool
-  ]b4_parser_class[::yy_lac_establish_ (int yytoken)
+  ]b4_parser_class[::yy_lac_establish_ (symbol_kind_type yytoken)
   {
     /* Establish the initial context for the current lookahead if no initial
        context is currently established.
@@ -1258,7 +1428,7 @@ b4_dollar_popdef])[]dnl
       {
 #if ]b4_api_PREFIX[DEBUG
         YYCDEBUG << "LAC: initial context established for "
-                 << yytname_[yytoken] << '\n';
+                 << symbol_name (yytoken) << '\n';
 #endif
         yy_lac_established_ = true;
         return yy_lac_check_ (yytoken);
@@ -1287,22 +1457,12 @@ b4_dollar_popdef])[]dnl
                  << evt << '\n';
         yy_lac_established_ = false;
       }
-  }]])[
-
-  // Generate an error message.
-  std::string
-  ]b4_parser_class[::yysyntax_error_ (]dnl
-b4_error_verbose_if([state_type yystate, const symbol_type& yyla],
-                    [state_type, const symbol_type&])[) const
-  {]b4_error_verbose_if([[
-    // Number of reported tokens (one for the "unexpected", one per
-    // "expected").
-    std::ptrdiff_t yycount = 0;
-    // Its maximum.
-    enum { YYERROR_VERBOSE_ARGS_MAXIMUM = 5 };
-    // Arguments of yyformat.
-    char const *yyarg[YYERROR_VERBOSE_ARGS_MAXIMUM];
+  }]])b4_parse_error_bmatch([detailed\|verbose], [[
 
+  int
+  ]b4_parser_class[::yy_syntax_error_arguments_ (const context& yyctx,
+                                                 symbol_kind_type yyarg[], int yyargn) const
+  {
     /* There are many possibilities here to consider:
        - If this state is a consistent state with a default action, then
          the only way this function was invoked is if the default action
@@ -1333,46 +1493,26 @@ b4_error_verbose_if([state_type yystate, const symbol_type& yyla],
          one exception: it will still contain any token that will not be
          accepted due to an error action in a later state.]])[
     */
-    if (!yyla.empty ())
+
+    if (!yyctx.lookahead ().empty ())
       {
-        symbol_number_type yytoken = yyla.type_get ();
-        yyarg[yycount++] = yytname_[yytoken];]b4_lac_if([[
-
-#if ]b4_api_PREFIX[DEBUG
-        // Execute LAC once. We don't care if it is succesful, we
-        // only do it for the sake of debugging output.
-        if (!yy_lac_established_)
-          yy_lac_check_ (yytoken);
-#endif]])[
-
-        int yyn = yypact_[+yystate];
-        if (!yy_pact_value_is_default_ (yyn))
-          {]b4_lac_if([[
-            for (int yyx = 0; yyx < yyntokens_; ++yyx)
-              if (yyx != yy_error_token_ && yyx != yy_undef_token_
-                  && yy_lac_check_ (yyx))
-                {]], [[
-            /* Start YYX at -YYN if negative to avoid negative indexes in
-               YYCHECK.  In other words, skip the first -YYN actions for
-               this state because they are default actions.  */
-            int yyxbegin = yyn < 0 ? -yyn : 0;
-            // Stay within bounds of both yycheck and yytname.
-            int yychecklim = yylast_ - yyn + 1;
-            int yyxend = yychecklim < yyntokens_ ? yychecklim : yyntokens_;
-            for (int yyx = yyxbegin; yyx < yyxend; ++yyx)
-              if (yycheck_[yyx + yyn] == yyx && yyx != yy_error_token_
-                  && !yy_table_value_is_error_ (yytable_[yyx + yyn]))
-                {]])[
-                  if (yycount == YYERROR_VERBOSE_ARGS_MAXIMUM)
-                    {
-                      yycount = 1;
-                      break;
-                    }
-                  else
-                    yyarg[yycount++] = yytname_[yyx];
-                }
-          }
+        if (yyarg)
+          yyarg[0] = yyctx.token ();
+        int yyn = yyctx.expected_tokens (yyarg ? yyarg + 1 : yyarg, yyargn - 1);
+        return yyn + 1;
       }
+    return 0;
+  }
+
+  // Generate an error message.
+  std::string
+  ]b4_parser_class[::yysyntax_error_ (const context& yyctx) const
+  {
+    // Its maximum.
+    enum { YYARGS_MAX = 5 };
+    // Arguments of yyformat.
+    symbol_kind_type yyarg[YYARGS_MAX];
+    int yycount = yy_syntax_error_arguments_ (yyctx, yyarg, YYARGS_MAX);
 
     char const* yyformat = YY_NULLPTR;
     switch (yycount)
@@ -1397,14 +1537,13 @@ b4_error_verbose_if([state_type yystate, const symbol_type& yyla],
     for (char const* yyp = yyformat; *yyp; ++yyp)
       if (yyp[0] == '%' && yyp[1] == 's' && yyi < yycount)
         {
-          yyres += yytnamerr_ (yyarg[yyi++]);
+          yyres += symbol_name (yyarg[yyi++]);
           ++yyp;
         }
       else
         yyres += *yyp;
-    return yyres;]], [[
-    return YY_("syntax error");]])[
-  }
+    return yyres;
+  }]])[
 
 
   const ]b4_int_type(b4_pact_ninf, b4_pact_ninf) b4_parser_class::yypact_ninf_ = b4_pact_ninf[;
@@ -1413,21 +1552,23 @@ b4_error_verbose_if([state_type yystate, const symbol_type& yyla],
 
 ]b4_parser_tables_define[
 
-]b4_token_table_if([], [[#if ]b4_api_PREFIX[DEBUG]])[
+]b4_parse_error_bmatch([simple\|verbose],
+[[#if ]b4_api_PREFIX[DEBUG]b4_tname_if([[ || 1]])[
   // YYTNAME[SYMBOL-NUM] -- String name of the symbol SYMBOL-NUM.
-  // First, the terminals, then, starting at \a yyntokens_, nonterminals.
+  // First, the terminals, then, starting at \a YYNTOKENS, nonterminals.
   const char*
   const ]b4_parser_class[::yytname_[] =
   {
   ]b4_tname[
   };
+#endif
+]])[
 
-]b4_token_table_if([[#if ]b4_api_PREFIX[DEBUG]])[
+#if ]b4_api_PREFIX[DEBUG][
 ]b4_integral_parser_table_define([rline], [b4_rline])[
 
-  // Print the state stack on the debug stream.
   void
-  ]b4_parser_class[::yystack_print_ ()
+  ]b4_parser_class[::yy_stack_print_ () const
   {
     *yycdebug_ << "Stack now";
     for (stack_type::const_iterator
@@ -1438,9 +1579,8 @@ b4_error_verbose_if([state_type yystate, const symbol_type& yyla],
     *yycdebug_ << '\n';
   }
 
-  // Report on the debug stream that the rule \a yyrule is going to be reduced.
   void
-  ]b4_parser_class[::yy_reduce_print_ (int yyrule)
+  ]b4_parser_class[::yy_reduce_print_ (int yyrule) const
   {
     int yylno = yyrline_[yyrule];
     int yynrhs = yyr2_[yyrule];

data/skeletons/lalr1.d

@@ -21,7 +21,7 @@ m4_include(b4_skeletonsdir/[d.m4])
 b4_output_begin([b4_parser_file_name])
 b4_copyright([Skeleton implementation for Bison LALR(1) parsers in D],
              [2007-2012, 2019-2020])[
-
+]b4_disclaimer[
 ]b4_percent_define_ifdef([package], [module b4_percent_define_get([package]);
 ])[
 version(D_Version2) {
@@ -182,11 +182,13 @@ b4_user_union_members
 };],
 [m4_if(b4_tag_seen_flag, 0,
 [[private alias int YYSemanticType;]])])[
-]b4_token_enums(b4_tokens)[
+]b4_token_enums[
 ]b4_parser_class_declaration[
 {
   ]b4_identification[
 
+]b4_declare_symbol_enum[
+
 ]b4_locations_if([[
   private final ]b4_location_type[ yylloc_from_stack (ref YYStack rhs, int n)
   {
@@ -261,6 +263,11 @@ b4_user_union_members
   public final void setDebugLevel(int level) { yydebug = level; }
 
   protected final void yycdebug (string s) {
+    if (0 < yydebug)
+      yyDebugStream.write (s);
+  }
+
+  protected final void yycdebugln (string s) {
     if (0 < yydebug)
       yyDebugStream.writeln (s);
   }
@@ -335,7 +342,8 @@ b4_user_union_members
     }
 
 ]b4_parse_trace_if([[
-    yy_symbol_print ("-> $$ =", yyr1_[yyn], yyval]b4_locations_if([, yyloc])[);]])[
+    import std.conv : to;
+    yy_symbol_print ("-> $$ =", to!SymbolKind (yyr1_[yyn]), yyval]b4_locations_if([, yyloc])[);]])[
 
     yystack.pop (yylen);
     yylen = 0;
@@ -393,21 +401,21 @@ b4_user_union_members
   | Print this symbol on YYOUTPUT.  |
   `--------------------------------*/
 
-  private final void yy_symbol_print (string s, int yytype,
+  private final void yy_symbol_print (string s, SymbolKind yykind,
     ref ]b4_yystype[ yyvaluep]dnl
 b4_locations_if([, ref ]b4_location_type[ yylocationp])[)
   {
     if (0 < yydebug)
     {
-      string message = s ~ (yytype < yyntokens_ ? " token " : " nterm ")
-              ~ yytname_[yytype] ~ " ("]b4_locations_if([
+      string message = s ~ (yykind < yyntokens_ ? " token " : " nterm ")
+              ~ yytname_[yykind] ~ " ("]b4_locations_if([
               ~ yylocationp.toString() ~ ": "])[;
       static if (__traits(compiles, message ~= yyvaluep.toString ()))
               message ~= yyvaluep.toString ();
       else
               message ~= format ("%s", &yyvaluep);
       message ~= ")";
-      yycdebug (message);
+      yycdebugln (message);
     }
   }
 ]])[
@@ -421,8 +429,8 @@ b4_locations_if([, ref ]b4_location_type[ yylocationp])[)
   public bool parse ()
   {
     /// Lookahead and lookahead in internal form.
-    int yychar = yyempty_;
-    int yytoken = 0;
+    int yychar = TokenKind.YYEMPTY;
+    SymbolKind yytoken = SymbolKind.]b4_symbol(-2, kind)[;
 
     /* State.  */
     int yyn = 0;
@@ -445,9 +453,9 @@ b4_locations_if([, ref ]b4_location_type[ yylocationp])[)
     /// Semantic value of the lookahead.
     ]b4_yystype[ yylval;
 
-    int yyresult;]b4_parse_trace_if([[
+    bool yyresult;]b4_parse_trace_if([[
 
-    yycdebug ("Starting parse\n");]])[
+    yycdebugln ("Starting parse");]])[
     yyerrstatus_ = 0;
 
 ]m4_ifdef([b4_initial_action], [
@@ -468,7 +476,7 @@ m4_popdef([b4_at_dollar])])dnl
         /* New state.  Unlike in the C/C++ skeletons, the state is already
            pushed when we come here.  */
       case YYNEWSTATE:]b4_parse_trace_if([[
-        yycdebug (format("Entering state %d\n", yystate));
+        yycdebugln (format("Entering state %d", yystate));
         if (0 < yydebug)
           yystack.print (yyDebugStream);]])[
 
@@ -485,9 +493,9 @@ m4_popdef([b4_at_dollar])])dnl
         }
 
         /* Read a lookahead token.  */
-        if (yychar == yyempty_)
+        if (yychar == TokenKind.YYEMPTY)
         {]b4_parse_trace_if([[
-          yycdebug ("Reading a token: ");]])[
+          yycdebugln ("Reading a token");]])[
           yychar = yylex ();]b4_locations_if([[
           static if (yy_location_is_class) {
             yylloc = new ]b4_location_type[(yylexer.startPos, yylexer.endPos);
@@ -499,43 +507,55 @@ m4_popdef([b4_at_dollar])])dnl
 
         /* Convert token to internal form.  */
         yytoken = yytranslate_ (yychar);]b4_parse_trace_if([[
-        yy_symbol_print ("Next token is",
-                         yytoken, yylval]b4_locations_if([, yylloc])[);]])[
+        yy_symbol_print ("Next token is", yytoken, yylval]b4_locations_if([, yylloc])[);]])[
 
-        /* If the proper action on seeing token YYTOKEN is to reduce or to
-           detect an error, take that action.  */
-        yyn += yytoken;
-        if (yyn < 0 || yylast_ < yyn || yycheck_[yyn] != yytoken)
-          label = YYDEFAULT;
-
-        /* <= 0 means reduce or error.  */
-        else if ((yyn = yytable_[yyn]) <= 0)
+        if (yytoken == SymbolKind.]b4_symbol(1, kind)[)
         {
-          if (yy_table_value_is_error_ (yyn))
-            label = YYERRLAB;
-          else
-          {
-            yyn = -yyn;
-            label = YYREDUCE;
-          }
+          // The scanner already issued an error message, process directly
+          // to error recovery.  But do not keep the error token as
+          // lookahead, it is too special and may lead us to an endless
+          // loop in error recovery. */
+          yychar = TokenKind.YYUNDEF;
+          yytoken = SymbolKind.]b4_symbol_prefix[YYUNDEF;]b4_locations_if([[
+          yyerrloc = yylloc;]])[
+          label = YYERRLAB1;
         }
         else
         {
-          /* Shift the lookahead token.  */]b4_parse_trace_if([[
-          yy_symbol_print ("Shifting", yytoken,
-                            yylval]b4_locations_if([, yylloc])[);]])[
+          /* If the proper action on seeing token YYTOKEN is to reduce or to
+             detect an error, take that action.  */
+          yyn += yytoken;
+          if (yyn < 0 || yylast_ < yyn || yycheck_[yyn] != yytoken)
+            label = YYDEFAULT;
 
-          /* Discard the token being shifted.  */
-          yychar = yyempty_;
+          /* <= 0 means reduce or error.  */
+          else if ((yyn = yytable_[yyn]) <= 0)
+          {
+            if (yy_table_value_is_error_ (yyn))
+              label = YYERRLAB;
+            else
+            {
+              yyn = -yyn;
+              label = YYREDUCE;
+            }
+          }
+          else
+          {
+            /* Shift the lookahead token.  */]b4_parse_trace_if([[
+            yy_symbol_print ("Shifting", yytoken, yylval]b4_locations_if([, yylloc])[);]])[
 
-          /* Count tokens shifted since error; after three, turn off error
-           * status.  */
-          if (yyerrstatus_ > 0)
-            --yyerrstatus_;
+            /* Discard the token being shifted.  */
+            yychar = TokenKind.YYEMPTY;
 
-          yystate = yyn;
-          yystack.push (yystate, yylval]b4_locations_if([, yylloc])[);
-          label = YYNEWSTATE;
+            /* Count tokens shifted since error; after three, turn off error
+             * status.  */
+            if (yyerrstatus_ > 0)
+              --yyerrstatus_;
+
+            yystate = yyn;
+            yystack.push (yystate, yylval]b4_locations_if([, yylloc])[);
+            label = YYNEWSTATE;
+          }
         }
         break;
 
@@ -559,33 +579,33 @@ m4_popdef([b4_at_dollar])])dnl
         yystate = yystack.stateAt (0);
         break;
 
-      /*------------------------------------.
-      | yyerrlab -- here on detecting error |
-      `------------------------------------*/
+      /*--------------------------------------.
+      | yyerrlab -- here on detecting error.  |
+      `--------------------------------------*/
       case YYERRLAB:
         /* If not already recovering from an error, report this error.  */
         if (yyerrstatus_ == 0)
         {
           ++yynerrs_;
-          if (yychar == yyempty_)
-            yytoken = yyempty_;
+          if (yychar == TokenKind.YYEMPTY)
+            yytoken = SymbolKind.]b4_symbol(-2, kind)[;
           yyerror (]b4_locations_if([yylloc, ])[yysyntax_error (yystate, yytoken));
         }
-
-]b4_locations_if([        yyerrloc = yylloc;])[
+]b4_locations_if([
+        yyerrloc = yylloc;])[
         if (yyerrstatus_ == 3)
         {
           /* If just tried and failed to reuse lookahead token after an
            * error, discard it.  */
 
-          if (yychar <= YYTokenType.EOF)
+          if (yychar <= TokenKind.]b4_symbol(0, [id])[)
           {
             /* Return failure if at end of input.  */
-            if (yychar == YYTokenType.EOF)
+            if (yychar == TokenKind.]b4_symbol(0, [id])[)
              return false;
           }
           else
-            yychar = yyempty_;
+            yychar = TokenKind.YYEMPTY;
         }
 
         /* Else will try to reuse lookahead token after shifting the error
@@ -612,13 +632,14 @@ m4_popdef([b4_at_dollar])])dnl
       case YYERRLAB1:
         yyerrstatus_ = 3;       /* Each real token shifted decrements this.  */
 
+        // Pop stack until we find a state that shifts the error token.
         for (;;)
         {
           yyn = yypact_[yystate];
           if (!yy_pact_value_is_default_ (yyn))
           {
-            yyn += yy_error_token_;
-            if (0 <= yyn && yyn <= yylast_ && yycheck_[yyn] == yy_error_token_)
+            yyn += SymbolKind.]b4_symbol(1, kind)[;
+            if (0 <= yyn && yyn <= yylast_ && yycheck_[yyn] == SymbolKind.]b4_symbol(1, kind)[)
             {
               yyn = yytable_[yyn];
               if (0 < yyn)
@@ -645,9 +666,8 @@ m4_popdef([b4_at_dollar])])dnl
         yystack.pop (2);])[
 
         /* Shift the error token.  */]b4_parse_trace_if([[
-        yy_symbol_print ("Shifting", yystos_[yyn],
-        yylval]b4_locations_if([, yyloc])[);]])[
-
+        import std.conv : to;
+        yy_symbol_print ("Shifting", to!SymbolKind (yystos_[yyn]), yylval]b4_locations_if([, yyloc])[);]])[
         yystate = yyn;
         yystack.push (yyn, yylval]b4_locations_if([, yyloc])[);
         label = YYNEWSTATE;
@@ -655,17 +675,26 @@ m4_popdef([b4_at_dollar])])dnl
 
       /* Accept.  */
       case YYACCEPT:
-        return true;
+        yyresult = true;
+        label = YYRETURN;
+        break;
 
       /* Abort.  */
       case YYABORT:
-        return false;
+        yyresult = false;
+        label = YYRETURN;
+        break;
+
+      case YYRETURN:]b4_parse_trace_if([[
+        if (0 < yydebug)
+          yystack.print (yyDebugStream);]])[
+        return yyresult;
     }
   }
 
   // Generate an error message.
-  private final string yysyntax_error (int yystate, int tok)
-  {]b4_error_verbose_if([[
+  private final string yysyntax_error (int yystate, SymbolKind tok)
+  {]b4_parse_error_case([verbose], [[
     /* There are many possibilities here to consider:
        - Assume YYFAIL is not used.  It's too flawed to consider.
          See
@@ -697,7 +726,7 @@ m4_popdef([b4_at_dollar])])dnl
          will still contain any token that will not be accepted due
          to an error action in a later state.
       */
-    if (tok != yyempty_)
+    if (tok != SymbolKind.]b4_symbol(-2, kind)[)
     {
       // FIXME: This method of building the message is not compatible
       // with internationalization.
@@ -716,14 +745,14 @@ m4_popdef([b4_at_dollar])])dnl
         int yyxend = yychecklim < yyntokens_ ? yychecklim : yyntokens_;
         int count = 0;
         for (int x = yyxbegin; x < yyxend; ++x)
-          if (yycheck_[x + yyn] == x && x != yy_error_token_
+          if (yycheck_[x + yyn] == x && x != SymbolKind.]b4_symbol(1, kind)[
               && !yy_table_value_is_error_ (yytable_[x + yyn]))
              ++count;
           if (count < 5)
           {
              count = 0;
              for (int x = yyxbegin; x < yyxend; ++x)
-               if (yycheck_[x + yyn] == x && x != yy_error_token_
+               if (yycheck_[x + yyn] == x && x != SymbolKind.]b4_symbol(1, kind)[
                    && !yy_table_value_is_error_ (yytable_[x + yyn]))
                {
                   res ~= count++ == 0 ? ", expecting " : " or ";
@@ -788,47 +817,45 @@ m4_popdef([b4_at_dollar])])dnl
     int yylno = yyrline_[yyrule];
     int yynrhs = yyr2_[yyrule];
     /* Print the symbols being reduced, and their result.  */
-    yycdebug (format("Reducing stack by rule %d (line %d), ",
-              yyrule - 1, yylno));
+    yycdebugln (format("Reducing stack by rule %d (line %d):",
+                yyrule - 1, yylno));
 
     /* The symbols being reduced.  */
+    import std.conv : to;
     for (int yyi = 0; yyi < yynrhs; yyi++)
       yy_symbol_print (format("   $%d =", yyi + 1),
-                       yystos_[yystack.stateAt(yynrhs - (yyi + 1))],
+                       to!SymbolKind (yystos_[yystack.stateAt(yynrhs - (yyi + 1))]),
                        ]b4_rhs_value(yynrhs, yyi + 1)b4_locations_if([,
                        b4_rhs_location(yynrhs, yyi + 1)])[);
   }
 ]])[
 
-  private static token_number_type yytranslate_ (int t)
+  private static SymbolKind yytranslate_ (int t)
   {
 ]b4_api_token_raw_if(
 [[    import std.conv : to;
-    return to!byte (t);]],
+    return to!SymbolKind (t);]],
 [[    /* YYTRANSLATE(YYLEX) -- Bison symbol number corresponding to YYLEX.  */
-    immutable token_number_type[] translate_table =
+    immutable ]b4_int_type_for([b4_translate])[[] translate_table =
     @{
   ]b4_translate[
     @};
 
     immutable int user_token_number_max_ = ]b4_user_token_number_max[;
-    immutable token_number_type undef_token_ = ]b4_undef_token_number[;
 
     if (t <= 0)
-      return YYTokenType.EOF;
+      return SymbolKind.]b4_symbol_prefix[YYEOF;
     else if (t <= user_token_number_max_)
-      return translate_table[t];
+      {
+        import std.conv : to;
+        return to!SymbolKind (translate_table[t]);
+      }
     else
-      return undef_token_;]])[
+      return SymbolKind.]b4_symbol_prefix[YYUNDEF;]])[
   }
 
-  alias ]b4_int_type_for([b4_translate])[ token_number_type;
-
-  private static immutable token_number_type yy_error_token_ = 1;
-
   private static immutable int yylast_ = ]b4_last[;
   private static immutable int yynnts_ = ]b4_nterms_number[;
-  private static immutable int yyempty_ = -2;
   private static immutable int yyfinal_ = ]b4_final_state_number[;
   private static immutable int yyntokens_ = ]b4_tokens_number[;
 
@@ -887,9 +914,8 @@ m4_popdef([b4_at_dollar])])dnl
       stream.writeln ();
     }]])[
   }
-
-  /* User implementation code.  */
 ]b4_percent_code_get[
 }
-]b4_epilogue[]dnl
+]b4_percent_code_get([[epilogue]])[]dnl
+b4_epilogue[]dnl
 b4_output_end

data/skeletons/lalr1.java

@@ -49,13 +49,13 @@ b4_use_push_for_pull_if([
   b4_push_if([m4_define([b4_use_push_for_pull_flag], [[0]])],
              [m4_define([b4_push_flag], [[1]])])])
 
-# Define a macro to encapsulate the parse state variables.
-# This allows them to be defined either in parse() when doing
-# pull parsing, or as class instance variable when doing push parsing.
+# Define a macro to encapsulate the parse state variables.  This
+# allows them to be defined either in parse() when doing pull parsing,
+# or as class instance variable when doing push parsing.
 m4_define([b4_define_state],[[
     /* Lookahead and lookahead in internal form.  */
-    int yychar = yyempty_;
-    int yytoken = 0;
+    int yychar = YYEMPTY_;
+    SymbolKind yytoken = null;
 
     /* State.  */
     int yyn = 0;
@@ -64,9 +64,8 @@ m4_define([b4_define_state],[[
     YYStack yystack = new YYStack ();
     int label = YYNEWSTATE;
 
-    /* Error handling.  */
-    int yynerrs_ = 0;
-    ]b4_locations_if([[/* The location where the error started.  */
+]b4_locations_if([[
+    /* The location where the error started.  */
     ]b4_location_type[ yyerrloc = null;
 
     /* Location. */
@@ -79,10 +78,12 @@ m4_define([b4_define_state],[[
 ]b4_output_begin([b4_parser_file_name])[
 ]b4_copyright([Skeleton implementation for Bison LALR(1) parsers in Java],
               [2007-2015, 2018-2020])[
+]b4_disclaimer[
 ]b4_percent_define_ifdef([package], [package b4_percent_define_get([package]);[
 ]])[
 ]b4_user_pre_prologue[
 ]b4_user_post_prologue[
+import java.text.MessageFormat;
 ]b4_percent_code_get([[imports]])[
 /**
  * A Bison parser, automatically generated from <tt>]m4_bpatsubst(b4_file_name, [^"\(.*\)"$], [\1])[</tt>.
@@ -91,8 +92,10 @@ m4_define([b4_define_state],[[
  */
 ]b4_parser_class_declaration[
 {
-  ]b4_identification[
-]b4_error_verbose_if([[
+]b4_identification[
+][
+]b4_parse_error_bmatch(
+           [detailed\|verbose], [[
   /**
    * True if verbose error messages are enabled.
    */
@@ -109,9 +112,9 @@ m4_define([b4_define_state],[[
    */
   public final void setErrorVerbose(boolean verbose)
   { yyErrorVerbose = verbose; }
-]])
+]])[
 
-b4_locations_if([[
+]b4_locations_if([[
   /**
    * A class defining a pair of positions.  Positions, defined by the
    * <code>]b4_position_type[</code> class, denote a point in the input.
@@ -161,28 +164,27 @@ b4_locations_if([[
     }
   }
 
-]])[
-
-]b4_locations_if([[
-  private ]b4_location_type[ yylloc (YYStack rhs, int n)
+  private ]b4_location_type[ yylloc(YYStack rhs, int n)
   {
     if (0 < n)
-      return new ]b4_location_type[ (rhs.locationAt (n-1).begin, rhs.locationAt (0).end);
+      return new ]b4_location_type[(rhs.locationAt(n-1).begin, rhs.locationAt(0).end);
     else
-      return new ]b4_location_type[ (rhs.locationAt (0).end);
+      return new ]b4_location_type[(rhs.locationAt(0).end);
   }]])[
 
+]b4_declare_symbol_enum[
+
   /**
    * Communication interface between the scanner and the Bison-generated
    * parser <tt>]b4_parser_class[</tt>.
    */
   public interface Lexer {
-    /** Token returned by the scanner to signal the end of its input.  */
-    public static final int EOF = 0;
-
 ]b4_token_enums[
+    /** Deprecated, use ]b4_symbol(0, id)[ instead.  */
+    public static final int EOF = ]b4_symbol(0, id)[;
 
-    ]b4_locations_if([[/**
+]b4_locations_if([[
+    /**
      * Method to retrieve the beginning position of the last scanned token.
      * @@return the position at which the last scanned token starts.
      */
@@ -206,17 +208,25 @@ b4_locations_if([[
      * ]b4_locations_if([and beginning/ending positions ])[of the token.
      * @@return the token identifier corresponding to the next token.
      */
-    int yylex () ]b4_maybe_throws([b4_lex_throws])[;
+    int yylex ()]b4_maybe_throws([b4_lex_throws])[;
 
     /**
      * Entry point for error reporting.  Emits an error
      * ]b4_locations_if([referring to the given location ])[in a user-defined way.
      *
-     * ]b4_locations_if([[@@param loc The location of the element to which the
+     *]b4_locations_if([[ @@param loc The location of the element to which the
      *                error message is related]])[
      * @@param msg The string for the error message.
      */
      void yyerror (]b4_locations_if([b4_location_type[ loc, ]])[String msg);
+
+]b4_parse_error_bmatch(
+           [custom], [[
+    /**
+     * Build and emit a syntax error message.
+     */
+     void reportSyntaxError (][Context yyctx);
+]])[
   }
 
 ]b4_lexer_if([[
@@ -224,22 +234,23 @@ b4_locations_if([[
 ]b4_percent_code_get([[lexer]])[
   }
 
-  ]])[/**
+]])[
+  /**
    * The object doing lexical analysis for us.
    */
   private Lexer yylexer;
 
-  ]b4_parse_param_vars[
+]b4_parse_param_vars[
 
 ]b4_lexer_if([[
   /**
    * Instantiates the Bison-generated parser.
    */
-  public ]b4_parser_class (b4_parse_param_decl([b4_lex_param_decl])[) ]b4_maybe_throws([b4_init_throws])[
+  public ]b4_parser_class (b4_parse_param_decl([b4_lex_param_decl])[)]b4_maybe_throws([b4_init_throws])[
   {
-    ]b4_percent_code_get([[init]])[
-    this.yylexer = new YYLexer(]b4_lex_param_call[);
-    ]b4_parse_param_cons[
+]b4_percent_code_get([[init]])[
+    this.yylexer = new YYLexer (]b4_lex_param_call[);
+]b4_parse_param_cons[
   }
 ]])[
 
@@ -247,11 +258,11 @@ b4_locations_if([[
    * Instantiates the Bison-generated parser.
    * @@param yylexer The scanner that will supply tokens to the parser.
    */
-  ]b4_lexer_if([[protected]], [[public]]) b4_parser_class[ (]b4_parse_param_decl([[Lexer yylexer]])[) ]b4_maybe_throws([b4_init_throws])[
+  ]b4_lexer_if([[protected]], [[public]]) b4_parser_class[ (]b4_parse_param_decl([[Lexer yylexer]])[)]b4_maybe_throws([b4_init_throws])[
   {
-    ]b4_percent_code_get([[init]])[
+]b4_percent_code_get([[init]])[
     this.yylexer = yylexer;
-    ]b4_parse_param_cons[
+]b4_parse_param_cons[
   }
 
 ]b4_parse_trace_if([[
@@ -266,7 +277,7 @@ b4_locations_if([[
    * Set the <tt>PrintStream</tt> on which the debug output is printed.
    * @@param s The stream that is used for debugging output.
    */
-  public final void setDebugStream(java.io.PrintStream s) { yyDebugStream = s; }
+  public final void setDebugStream (java.io.PrintStream s) { yyDebugStream = s; }
 
   private int yydebug = 0;
 
@@ -274,16 +285,23 @@ b4_locations_if([[
    * Answer the verbosity of the debugging output; 0 means that all kinds of
    * output from the parser are suppressed.
    */
-  public final int getDebugLevel() { return yydebug; }
+  public final int getDebugLevel () { return yydebug; }
 
   /**
    * Set the verbosity of the debugging output; 0 means that all kinds of
    * output from the parser are suppressed.
    * @@param level The verbosity level for debugging output.
    */
-  public final void setDebugLevel(int level) { yydebug = level; }
+  public final void setDebugLevel (int level) { yydebug = level; }
 ]])[
 
+  private int yynerrs = 0;
+
+  /**
+   * The number of syntax errors so far.
+   */
+  public final int getNumberOfErrors () { return yynerrs; }
+
   /**
    * Print an error message via the lexer.
    *]b4_locations_if([[ Use a <code>null</code> location.]])[
@@ -320,22 +338,20 @@ b4_locations_if([[
   }]])[
 
   private final class YYStack {
-    private int[] stateStack = new int[16];
-    ]b4_locations_if([[private ]b4_location_type[[] locStack = new ]b4_location_type[[16];]])[
+    private int[] stateStack = new int[16];]b4_locations_if([[
+    private ]b4_location_type[[] locStack = new ]b4_location_type[[16];]])[
     private ]b4_yystype[[] valueStack = new ]b4_yystype[[16];
 
     public int size = 16;
     public int height = -1;
 
-    public final void push (int state, ]b4_yystype[ value]dnl
-                            b4_locations_if([, ]b4_location_type[ loc])[) {
+    public final void push (int state, ]b4_yystype[ value]b4_locations_if([, ]b4_location_type[ loc])[) {
       height++;
       if (size == height)
         {
           int[] newStateStack = new int[size * 2];
           System.arraycopy (stateStack, 0, newStateStack, 0, height);
-          stateStack = newStateStack;
-          ]b4_locations_if([[
+          stateStack = newStateStack;]b4_locations_if([[
           ]b4_location_type[[] newLocStack = new ]b4_location_type[[size * 2];
           System.arraycopy (locStack, 0, newLocStack, 0, height);
           locStack = newLocStack;]])
@@ -347,8 +363,8 @@ b4_locations_if([[
           size *= 2;
         }
 
-      stateStack[height] = state;
-      ]b4_locations_if([[locStack[height] = loc;]])[
+      stateStack[height] = state;]b4_locations_if([[
+      locStack[height] = loc;]])[
       valueStack[height] = value;
     }
 
@@ -359,8 +375,8 @@ b4_locations_if([[
     public final void pop (int num) {
       // Avoid memory leaks... garbage collection is a white lie!
       if (0 < num) {
-        java.util.Arrays.fill (valueStack, height - num + 1, height + 1, null);
-        ]b4_locations_if([[java.util.Arrays.fill (locStack, height - num + 1, height + 1, null);]])[
+        java.util.Arrays.fill (valueStack, height - num + 1, height + 1, null);]b4_locations_if([[
+        java.util.Arrays.fill (locStack, height - num + 1, height + 1, null);]])[
       }
       height -= num;
     }
@@ -368,12 +384,13 @@ b4_locations_if([[
     public final int stateAt (int i) {
       return stateStack[height - i];
     }
+]b4_locations_if([[
 
-    ]b4_locations_if([[public final ]b4_location_type[ locationAt (int i) {
+    public final ]b4_location_type[ locationAt (int i) {
       return locStack[height - i];
     }
-
-    ]])[public final ]b4_yystype[ valueAt (int i) {
+]])[
+    public final ]b4_yystype[ valueAt (int i) {
       return valueStack[height - i];
     }
 
@@ -445,14 +462,14 @@ b4_locations_if([[
    */
   private int yyLRGotoState (int yystate, int yysym)
   {
-    int yyr = yypgoto_[yysym - yyntokens_] + yystate;
-    if (0 <= yyr && yyr <= yylast_ && yycheck_[yyr] == yystate)
+    int yyr = yypgoto_[yysym - YYNTOKENS_] + yystate;
+    if (0 <= yyr && yyr <= YYLAST_ && yycheck_[yyr] == yystate)
       return yytable_[yyr];
     else
-      return yydefgoto_[yysym - yyntokens_];
+      return yydefgoto_[yysym - YYNTOKENS_];
   }
 
-  private int yyaction (int yyn, YYStack yystack, int yylen) ]b4_maybe_throws([b4_throws])[
+  private int yyaction (int yyn, YYStack yystack, int yylen)]b4_maybe_throws([b4_throws])[
   {
     /* If YYLEN is nonzero, implement the default value of the action:
        '$$ = $1'.  Otherwise, use the top of the stack.
@@ -460,8 +477,8 @@ b4_locations_if([[
        Otherwise, the following line sets YYVAL to garbage.
        This behavior is undocumented and Bison
        users should not rely upon it.  */
-    ]b4_yystype[ yyval = (0 < yylen) ? yystack.valueAt (yylen - 1) : yystack.valueAt (0);
-    ]b4_locations_if([b4_location_type[ yyloc = yylloc (yystack, yylen);]])[]b4_parse_trace_if([[
+    ]b4_yystype[ yyval = (0 < yylen) ? yystack.valueAt (yylen - 1) : yystack.valueAt (0);]b4_locations_if([[
+    ]b4_location_type[ yyloc = yylloc (yystack, yylen);]])[]b4_parse_trace_if([[
 
     yyReducePrint (yyn, yystack);]])[
 
@@ -471,66 +488,29 @@ b4_locations_if([[
         default: break;
       }]b4_parse_trace_if([[
 
-    yySymbolPrint ("-> $$ =", yyr1_[yyn], yyval]b4_locations_if([, yyloc])[);]])[
+    yySymbolPrint ("-> $$ =", SymbolKind.get (yyr1_[yyn]), yyval]b4_locations_if([, yyloc])[);]])[
 
     yystack.pop (yylen);
     yylen = 0;
-
     /* Shift the result of the reduction.  */
     int yystate = yyLRGotoState (yystack.stateAt (0), yyr1_[yyn]);
     yystack.push (yystate, yyval]b4_locations_if([, yyloc])[);
     return YYNEWSTATE;
   }
 
-]b4_error_verbose_if([[
-  /* Return YYSTR after stripping away unnecessary quotes and
-     backslashes, so that it's suitable for yyerror.  The heuristic is
-     that double-quoting is unnecessary unless the string contains an
-     apostrophe, a comma, or backslash (other than backslash-backslash).
-     YYSTR is taken from yytname.  */
-  private final String yytnamerr_ (String yystr)
-  {
-    if (yystr.charAt (0) == '"')
-      {
-        StringBuffer yyr = new StringBuffer ();
-        strip_quotes: for (int i = 1; i < yystr.length (); i++)
-          switch (yystr.charAt (i))
-            {
-            case '\'':
-            case ',':
-              break strip_quotes;
-
-            case '\\':
-              if (yystr.charAt(++i) != '\\')
-                break strip_quotes;
-              /* Fall through.  */
-            default:
-              yyr.append (yystr.charAt (i));
-              break;
-
-            case '"':
-              return yyr.toString ();
-            }
-      }
-    else if (yystr.equals ("$end"))
-      return "end of input";
-
-    return yystr;
-  }
-]])[
 ]b4_parse_trace_if([[
   /*--------------------------------.
   | Print this symbol on YYOUTPUT.  |
   `--------------------------------*/
 
-  private void yySymbolPrint (String s, int yytype,
-                             ]b4_yystype[ yyvaluep]dnl
-                              b4_locations_if([, Object yylocationp])[)
+  private void yySymbolPrint(String s, SymbolKind yykind,
+                             ]b4_yystype[ yyvalue]b4_locations_if([, ]b4_location_type[ yylocation])[)
   {
-    yycdebug (s + (yytype < yyntokens_ ? " token " : " nterm ")
-              + yytname_[yytype] + " ("]b4_locations_if([
-              + yylocationp + ": "])[
-              + (yyvaluep == null ? "(null)" : yyvaluep.toString ()) + ")");
+    yycdebug (s
+              + (yykind.getCode() < YYNTOKENS_ ? " token " : " nterm ")
+              + yykind.getName() + " ("]b4_locations_if([
+              + yylocation + ": "])[
+              + (yyvalue == null ? "(null)" : yyvalue.toString()) + ")");
   }]])[
 
 ]b4_push_if([],[[
@@ -541,29 +521,29 @@ b4_locations_if([[
    * @@return <tt>true</tt> if the parsing succeeds.  Note that this does not
    *          imply that there were no syntax errors.
    */
-  public boolean parse () ]b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])[]])[
+  public boolean parse ()]b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])[]])[
 ]b4_push_if([
   /**
    * Push Parse input from external lexer
    *
    * @@param yylextoken current token
-   * @@param yylexval current lval
-]b4_locations_if([   * @@param yylexloc current position])[
+   * @@param yylexval current lval]b4_locations_if([[
+   * @@param yylexloc current position]])[
    *
    * @@return <tt>YYACCEPT, YYABORT, YYPUSH_MORE</tt>
    */
-  public int push_parse (int yylextoken, b4_yystype yylexval[]b4_locations_if([, b4_location_type yylexloc]))
-      b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])])[
-  {
-    ]b4_locations_if([/* @@$.  */
-    b4_location_type yyloc;])[
+  public int push_parse (int yylextoken, b4_yystype yylexval[]b4_locations_if([, b4_location_type yylexloc]))b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])])[
+  {]b4_locations_if([[
+    /* @@$.  */
+    ]b4_location_type[ yyloc;]])[
 ]b4_push_if([],[[
 ]b4_define_state[]b4_parse_trace_if([[
-    yycdebug ("Starting parse\n");]])[
+    yycdebug ("Starting parse");]])[
     yyerrstatus_ = 0;
+    yynerrs = 0;
 
     /* Initialize the stack.  */
-    yystack.push (yystate, yylval ]b4_locations_if([, yylloc])[);
+    yystack.push (yystate, yylval]b4_locations_if([, yylloc])[);
 ]m4_ifdef([b4_initial_action], [
 b4_dollar_pushdef([yylval], [], [], [yylloc])dnl
     b4_user_initial_action
@@ -579,7 +559,7 @@ b4_dollar_pushdef([yylval], [], [], [yylloc])dnl
     b4_user_initial_action
 b4_dollar_popdef[]dnl
 ])[]b4_parse_trace_if([[
-        yycdebug ("Starting parse\n");]])[
+        yycdebug ("Starting parse");]])[
         yyerrstatus_ = 0;
       } else
         label = YYGETTOKEN;
@@ -592,12 +572,12 @@ b4_dollar_popdef[]dnl
         /* New state.  Unlike in the C/C++ skeletons, the state is already
            pushed when we come here.  */
       case YYNEWSTATE:]b4_parse_trace_if([[
-        yycdebug ("Entering state " + yystate + "\n");
+        yycdebug ("Entering state " + yystate);
         if (0 < yydebug)
           yystack.print (yyDebugStream);]])[
 
         /* Accept?  */
-        if (yystate == yyfinal_)
+        if (yystate == YYFINAL_)
           ]b4_push_if([{label = YYACCEPT; break;}],
                       [return true;])[
 
@@ -612,17 +592,17 @@ b4_dollar_popdef[]dnl
 
       case YYGETTOKEN:])[
         /* Read a lookahead token.  */
-        if (yychar == yyempty_)
+        if (yychar == YYEMPTY_)
           {
 ]b4_push_if([[
             if (!push_token_consumed)
               return YYPUSH_MORE;]b4_parse_trace_if([[
-            yycdebug ("Reading a token: ");]])[
+            yycdebug ("Reading a token");]])[
             yychar = yylextoken;
             yylval = yylexval;]b4_locations_if([
             yylloc = yylexloc;])[
             push_token_consumed = false;]], [b4_parse_trace_if([[
-            yycdebug ("Reading a token: ");]])[
+            yycdebug ("Reading a token");]])[
             yychar = yylexer.yylex ();
             yylval = yylexer.getLVal ();]b4_locations_if([
             yylloc = new b4_location_type (yylexer.getStartPos (),
@@ -635,41 +615,55 @@ b4_dollar_popdef[]dnl
         yySymbolPrint ("Next token is", yytoken,
                        yylval]b4_locations_if([, yylloc])[);]])[
 
-        /* If the proper action on seeing token YYTOKEN is to reduce or to
-           detect an error, take that action.  */
-        yyn += yytoken;
-        if (yyn < 0 || yylast_ < yyn || yycheck_[yyn] != yytoken)
-          label = YYDEFAULT;
-
-        /* <= 0 means reduce or error.  */
-        else if ((yyn = yytable_[yyn]) <= 0)
+        if (yytoken == SymbolKind.]b4_symbol_prefix[YYerror)
           {
-            if (yyTableValueIsError (yyn))
-              label = YYERRLAB;
-            else
-              {
-                yyn = -yyn;
-                label = YYREDUCE;
-              }
+            // The scanner already issued an error message, process directly
+            // to error recovery.  But do not keep the error token as
+            // lookahead, it is too special and may lead us to an endless
+            // loop in error recovery. */
+            yychar = Lexer.]b4_percent_define_get([api.token.prefix])[YYUNDEF;
+            yytoken = SymbolKind.]b4_symbol_prefix[YYUNDEF;]b4_locations_if([[
+            yyerrloc = yylloc;]])[
+            label = YYERRLAB1;
           }
-
         else
           {
-            /* Shift the lookahead token.  */]b4_parse_trace_if([[
-            yySymbolPrint ("Shifting", yytoken,
-                           yylval]b4_locations_if([, yylloc])[);
+            /* If the proper action on seeing token YYTOKEN is to reduce or to
+               detect an error, take that action.  */
+            yyn += yytoken.getCode ();
+            if (yyn < 0 || YYLAST_ < yyn || yycheck_[yyn] != yytoken.getCode ())
+              label = YYDEFAULT;
+
+            /* <= 0 means reduce or error.  */
+            else if ((yyn = yytable_[yyn]) <= 0)
+              {
+                if (yyTableValueIsError (yyn))
+                  label = YYERRLAB;
+                else
+                  {
+                    yyn = -yyn;
+                    label = YYREDUCE;
+                  }
+              }
+
+            else
+              {
+                /* Shift the lookahead token.  */]b4_parse_trace_if([[
+                yySymbolPrint ("Shifting", yytoken,
+                               yylval]b4_locations_if([, yylloc])[);
 ]])[
-            /* Discard the token being shifted.  */
-            yychar = yyempty_;
+                /* Discard the token being shifted.  */
+                yychar = YYEMPTY_;
 
-            /* Count tokens shifted since error; after three, turn off error
-               status.  */
-            if (yyerrstatus_ > 0)
-              --yyerrstatus_;
+                /* Count tokens shifted since error; after three, turn off error
+                   status.  */
+                if (yyerrstatus_ > 0)
+                  --yyerrstatus_;
 
-            yystate = yyn;
-            yystack.push (yystate, yylval]b4_locations_if([, yylloc])[);
-            label = YYNEWSTATE;
+                yystate = yyn;
+                yystack.push (yystate, yylval]b4_locations_if([, yylloc])[);
+                label = YYNEWSTATE;
+              }
           }
         break;
 
@@ -700,26 +694,26 @@ b4_dollar_popdef[]dnl
         /* If not already recovering from an error, report this error.  */
         if (yyerrstatus_ == 0)
           {
-            ++yynerrs_;
-            if (yychar == yyempty_)
-              yytoken = yyempty_;
-            yyerror (]b4_locations_if([yylloc, ])[yysyntax_error (yystate, yytoken));
+            ++yynerrs;
+            if (yychar == YYEMPTY_)
+              yytoken = null;
+            yyreportSyntaxError (new Context (yystack, yytoken]b4_locations_if([[, yylloc]])[));
           }
-
-        ]b4_locations_if([yyerrloc = yylloc;])[
+]b4_locations_if([[
+        yyerrloc = yylloc;]])[
         if (yyerrstatus_ == 3)
           {
             /* If just tried and failed to reuse lookahead token after an
                error, discard it.  */
 
-            if (yychar <= Lexer.EOF)
+            if (yychar <= Lexer.]b4_symbol(0, id)[)
               {
                 /* Return failure if at end of input.  */
-                if (yychar == Lexer.EOF)
+                if (yychar == Lexer.]b4_symbol(0, id)[)
                   ]b4_push_if([{label = YYABORT; break;}], [return false;])[
               }
             else
-              yychar = yyempty_;
+              yychar = YYEMPTY_;
           }
 
         /* Else will try to reuse lookahead token after shifting the error
@@ -730,8 +724,8 @@ b4_dollar_popdef[]dnl
       /*-------------------------------------------------.
       | errorlab -- error raised explicitly by YYERROR.  |
       `-------------------------------------------------*/
-      case YYERROR:
-        ]b4_locations_if([yyerrloc = yystack.locationAt (yylen - 1);])[
+      case YYERROR:]b4_locations_if([[
+        yyerrloc = yystack.locationAt (yylen - 1);]])[
         /* Do not reclaim the symbols of the rule which action triggered
            this YYERROR.  */
         yystack.pop (yylen);
@@ -746,13 +740,15 @@ b4_dollar_popdef[]dnl
       case YYERRLAB1:
         yyerrstatus_ = 3;       /* Each real token shifted decrements this.  */
 
+        // Pop stack until we find a state that shifts the error token.
         for (;;)
           {
             yyn = yypact_[yystate];
             if (!yyPactValueIsDefault (yyn))
               {
-                yyn += yy_error_token_;
-                if (0 <= yyn && yyn <= yylast_ && yycheck_[yyn] == yy_error_token_)
+                yyn += SymbolKind.]b4_symbol(1, kind)[.getCode ();
+                if (0 <= yyn && yyn <= YYLAST_
+                    && yycheck_[yyn] == SymbolKind.]b4_symbol(1, kind)[.getCode ())
                   {
                     yyn = yytable_[yyn];
                     if (0 < yyn)
@@ -765,7 +761,8 @@ b4_dollar_popdef[]dnl
             if (yystack.height == 0)
               ]b4_push_if([{label = YYABORT; break;}],[return false;])[
 
-            ]b4_locations_if([yyerrloc = yystack.locationAt (0);])[
+]b4_locations_if([[
+            yyerrloc = yystack.locationAt (0);]])[
             yystack.pop ();
             yystate = yystack.stateAt (0);]b4_parse_trace_if([[
             if (0 < yydebug)
@@ -773,18 +770,18 @@ b4_dollar_popdef[]dnl
           }
 
         if (label == YYABORT)
-            /* Leave the switch.  */
-            break;
+          /* Leave the switch.  */
+          break;
 
-]b4_locations_if([
+]b4_locations_if([[
         /* Muck with the stack to setup for yylloc.  */
         yystack.push (0, null, yylloc);
         yystack.push (0, null, yyerrloc);
         yyloc = yylloc (yystack, 2);
-        yystack.pop (2);])[
+        yystack.pop (2);]])[
 
         /* Shift the error token.  */]b4_parse_trace_if([[
-        yySymbolPrint ("Shifting", yystos_[yyn],
+        yySymbolPrint ("Shifting", SymbolKind.get (yystos_[yyn]),
                        yylval]b4_locations_if([, yyloc])[);]])[
 
         yystate = yyn;
@@ -809,11 +806,11 @@ b4_dollar_popdef[]dnl
     /**
      * (Re-)Initialize the state of the push parser.
      */
-  public void push_parse_initialize()
+  public void push_parse_initialize ()
   {
     /* Lookahead and lookahead in internal form.  */
-    this.yychar = yyempty_;
-    this.yytoken = 0;
+    this.yychar = YYEMPTY_;
+    this.yytoken = null;
 
     /* State.  */
     this.yyn = 0;
@@ -823,10 +820,10 @@ b4_dollar_popdef[]dnl
     this.label = YYNEWSTATE;
 
     /* Error handling.  */
-    this.yynerrs_ = 0;
-    ]b4_locations_if([/* The location where the error started.  */
+    this.yynerrs = 0;]b4_locations_if([[
+    /* The location where the error started.  */
     this.yyerrloc = null;
-    this.yylloc = new b4_location_type (null, null);])[
+    this.yylloc = new ]b4_location_type[ (null, null);]])[
 
     /* Semantic value of the lookahead.  */
     this.yylval = null;
@@ -846,8 +843,7 @@ b4_dollar_popdef[]dnl
    *
    * @@return <tt>YYACCEPT, YYABORT, YYPUSH_MORE</tt>
    */
-  public int push_parse (int yylextoken, b4_yystype yylexval, b4_position_type yylexpos)
-      b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])
+  public int push_parse (int yylextoken, b4_yystype yylexval, b4_position_type yylexpos)b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])
   {
     return push_parse (yylextoken, yylexval, new b4_location_type (yylexpos));
   }
@@ -862,7 +858,7 @@ b4_dollar_popdef[]dnl
    * @@return <tt>true</tt> if the parsing succeeds.  Note that this does not
    *          imply that there were no syntax errors.
    */
-  public boolean parse () ]b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])[
+  public boolean parse ()]b4_maybe_throws([b4_list2([b4_lex_throws], [b4_throws])])[
   {
     if (yylexer == null)
       throw new NullPointerException("Null Lexer");
@@ -872,85 +868,167 @@ b4_dollar_popdef[]dnl
       ]b4_yystype[ lval = yylexer.getLVal();
 ]b4_locations_if([dnl
       b4_location_type yyloc = new b4_location_type (yylexer.getStartPos (),
-                                            yylexer.getEndPos ());])[
-      ]b4_locations_if([status = push_parse(token,lval,yyloc);],[
-      status = push_parse(token,lval);])[
+                                            yylexer.getEndPos ());])[]b4_locations_if([[
+      status = push_parse(token,lval,yyloc);]], [[
+      status = push_parse(token,lval);]])[
     } while (status == YYPUSH_MORE);
     return (status == YYACCEPT);
   }
 ]])[
 
-  // Generate an error message.
-  private String yysyntax_error (int yystate, int tok)
-  {]b4_error_verbose_if([[
+  /**
+   * Information needed to get the list of expected tokens and to forge
+   * a syntax error diagnostic.
+   */
+  public static final class Context
+  {
+    Context (YYStack stack, SymbolKind token]b4_locations_if([[, ]b4_location_type[ loc]])[)
+    {
+      yystack = stack;
+      yytoken = token;]b4_locations_if([[
+      yylocation = loc;]])[
+    }
+
+    private YYStack yystack;
+
+
+    /**
+     * The symbol kind of the lookahead token.
+     */
+    public final SymbolKind getToken ()
+    {
+      return yytoken;
+    }
+
+    private SymbolKind yytoken;]b4_locations_if([[
+
+    /**
+     * The location of the lookahead.
+     */
+    public final ]b4_location_type[ getLocation ()
+    {
+      return yylocation;
+    }
+
+    private ]b4_location_type[ yylocation;]])[
+    static final int NTOKENS = ]b4_parser_class[.YYNTOKENS_;
+
+    /**
+     * Put in YYARG at most YYARGN of the expected tokens given the
+     * current YYCTX, and return the number of tokens stored in YYARG.  If
+     * YYARG is null, return the number of expected tokens (guaranteed to
+     * be less than YYNTOKENS).
+     */
+    int getExpectedTokens (SymbolKind yyarg[], int yyargn)
+    {
+      return getExpectedTokens (yyarg, 0, yyargn);
+    }
+
+    int getExpectedTokens (SymbolKind yyarg[], int yyoffset, int yyargn)
+    {
+      int yycount = yyoffset;
+      int yyn = yypact_[this.yystack.stateAt (0)];
+      if (!yyPactValueIsDefault (yyn))
+        {
+          /* Start YYX at -YYN if negative to avoid negative
+             indexes in YYCHECK.  In other words, skip the first
+             -YYN actions for this state because they are default
+             actions.  */
+          int yyxbegin = yyn < 0 ? -yyn : 0;
+          /* Stay within bounds of both yycheck and yytname.  */
+          int yychecklim = YYLAST_ - yyn + 1;
+          int yyxend = yychecklim < NTOKENS ? yychecklim : NTOKENS;
+          for (int yyx = yyxbegin; yyx < yyxend; ++yyx)
+            if (yycheck_[yyx + yyn] == yyx && yyx != SymbolKind.]b4_symbol(1, kind)[.getCode ()
+                && !yyTableValueIsError (yytable_[yyx + yyn]))
+              {
+                if (yyarg == null)
+                  yycount += 1;
+                else if (yycount == yyargn)
+                  return 0; // FIXME: this is incorrect.
+                else
+                  yyarg[yycount++] = SymbolKind.get (yyx);
+              }
+        }
+      return yycount - yyoffset;
+    }
+  }
+
+]b4_parse_error_bmatch(
+[detailed\|verbose], [[
+  private int yysyntaxErrorArguments (Context yyctx, SymbolKind[] yyarg, int yyargn)
+  {
+    /* There are many possibilities here to consider:
+       - If this state is a consistent state with a default action,
+         then the only way this function was invoked is if the
+         default action is an error action.  In that case, don't
+         check for expected tokens because there are none.
+       - The only way there can be no lookahead present (in tok) is
+         if this state is a consistent state with a default action.
+         Thus, detecting the absence of a lookahead is sufficient to
+         determine that there is no unexpected or expected token to
+         report.  In that case, just report a simple "syntax error".
+       - Don't assume there isn't a lookahead just because this
+         state is a consistent state with a default action.  There
+         might have been a previous inconsistent state, consistent
+         state with a non-default action, or user semantic action
+         that manipulated yychar.  (However, yychar is currently out
+         of scope during semantic actions.)
+       - Of course, the expected token list depends on states to
+         have correct lookahead information, and it depends on the
+         parser not to perform extra reductions after fetching a
+         lookahead from the scanner and before detecting a syntax
+         error.  Thus, state merging (from LALR or IELR) and default
+         reductions corrupt the expected token list.  However, the
+         list is correct for canonical LR with one exception: it
+         will still contain any token that will not be accepted due
+         to an error action in a later state.
+    */
+    int yycount = 0;
+    if (yyctx.getToken () != null)
+      {
+        if (yyarg != null)
+          yyarg[yycount] = yyctx.getToken ();
+        yycount += 1;
+        yycount += yyctx.getExpectedTokens (yyarg, 1, yyargn);
+      }
+    return yycount;
+  }
+]])[
+
+/**
+   * Report a syntax error.
+   */
+  private void yyreportSyntaxError (Context yyctx)
+  {]b4_parse_error_bmatch(
+[custom], [[
+    yylexer.reportSyntaxError (yyctx);]],
+[detailed\|verbose], [[
     if (yyErrorVerbose)
       {
-        /* There are many possibilities here to consider:
-           - If this state is a consistent state with a default action,
-             then the only way this function was invoked is if the
-             default action is an error action.  In that case, don't
-             check for expected tokens because there are none.
-           - The only way there can be no lookahead present (in tok) is
-             if this state is a consistent state with a default action.
-             Thus, detecting the absence of a lookahead is sufficient to
-             determine that there is no unexpected or expected token to
-             report.  In that case, just report a simple "syntax error".
-           - Don't assume there isn't a lookahead just because this
-             state is a consistent state with a default action.  There
-             might have been a previous inconsistent state, consistent
-             state with a non-default action, or user semantic action
-             that manipulated yychar.  (However, yychar is currently out
-             of scope during semantic actions.)
-           - Of course, the expected token list depends on states to
-             have correct lookahead information, and it depends on the
-             parser not to perform extra reductions after fetching a
-             lookahead from the scanner and before detecting a syntax
-             error.  Thus, state merging (from LALR or IELR) and default
-             reductions corrupt the expected token list.  However, the
-             list is correct for canonical LR with one exception: it
-             will still contain any token that will not be accepted due
-             to an error action in a later state.
-        */
-        if (tok != yyempty_)
+        final int argmax = 5;
+        SymbolKind[] yyarg = new SymbolKind[argmax];
+        int yycount = yysyntaxErrorArguments (yyctx, yyarg, argmax);
+        String[] yystr = new String[yycount];
+        for (int yyi = 0; yyi < yycount; ++yyi)
+          yystr[yyi] = yyarg[yyi].getName();
+        String yyformat;
+        switch (yycount)
           {
-            /* FIXME: This method of building the message is not compatible
-               with internationalization.  */
-            StringBuffer res =
-              new StringBuffer ("syntax error, unexpected ");
-            res.append (yytnamerr_ (yytname_[tok]));
-            int yyn = yypact_[yystate];
-            if (!yyPactValueIsDefault (yyn))
-              {
-                /* Start YYX at -YYN if negative to avoid negative
-                   indexes in YYCHECK.  In other words, skip the first
-                   -YYN actions for this state because they are default
-                   actions.  */
-                int yyxbegin = yyn < 0 ? -yyn : 0;
-                /* Stay within bounds of both yycheck and yytname.  */
-                int yychecklim = yylast_ - yyn + 1;
-                int yyxend = yychecklim < yyntokens_ ? yychecklim : yyntokens_;
-                int count = 0;
-                for (int x = yyxbegin; x < yyxend; ++x)
-                  if (yycheck_[x + yyn] == x && x != yy_error_token_
-                      && !yyTableValueIsError (yytable_[x + yyn]))
-                    ++count;
-                if (count < 5)
-                  {
-                    count = 0;
-                    for (int x = yyxbegin; x < yyxend; ++x)
-                      if (yycheck_[x + yyn] == x && x != yy_error_token_
-                          && !yyTableValueIsError (yytable_[x + yyn]))
-                        {
-                          res.append (count++ == 0 ? ", expecting " : " or ");
-                          res.append (yytnamerr_ (yytname_[x]));
-                        }
-                  }
-              }
-            return res.toString ();
+            default:
+            case 0: yyformat = ]b4_trans(["syntax error"])[; break;
+            case 1: yyformat = ]b4_trans(["syntax error, unexpected {0}"])[; break;
+            case 2: yyformat = ]b4_trans(["syntax error, unexpected {0}, expecting {1}"])[; break;
+            case 3: yyformat = ]b4_trans(["syntax error, unexpected {0}, expecting {1} or {2}"])[; break;
+            case 4: yyformat = ]b4_trans(["syntax error, unexpected {0}, expecting {1} or {2} or {3}"])[; break;
+            case 5: yyformat = ]b4_trans(["syntax error, unexpected {0}, expecting {1} or {2} or {3} or {4}"])[; break;
           }
+        yyerror (]b4_locations_if([[yyctx.yylocation, ]])[new MessageFormat (yyformat).format (yystr));
+        return;
       }
-]])[
-    return "syntax error";
+    yyerror (]b4_locations_if([[yyctx.yylocation, ]])["syntax error");]],
+[simple], [[
+    yyerror (]b4_locations_if([[yyctx.yylocation, ]])["syntax error");]])[
   }
 
   /**
@@ -975,11 +1053,7 @@ b4_dollar_popdef[]dnl
   private static final ]b4_int_type_for([b4_pact])[ yypact_ninf_ = ]b4_pact_ninf[;
   private static final ]b4_int_type_for([b4_table])[ yytable_ninf_ = ]b4_table_ninf[;
 
-  ]b4_parser_tables_define[
-
-  /* YYTNAME[SYMBOL-NUM] -- String name of the symbol SYMBOL-NUM.
-     First, the terminals, then, starting at \a yyntokens_, nonterminals.  */
-  ]b4_typed_parser_table_define([String], [tname], [b4_tname])[
+]b4_parser_tables_define[
 
 ]b4_parse_trace_if([[
   ]b4_integral_parser_table_define([rline], [b4_rline],
@@ -996,50 +1070,43 @@ b4_dollar_popdef[]dnl
     int yynrhs = yyr2_[yyrule];
     /* Print the symbols being reduced, and their result.  */
     yycdebug ("Reducing stack by rule " + (yyrule - 1)
-              + " (line " + yylno + "), ");
+              + " (line " + yylno + "):");
 
     /* The symbols being reduced.  */
     for (int yyi = 0; yyi < yynrhs; yyi++)
       yySymbolPrint ("   $" + (yyi + 1) + " =",
-                     yystos_[yystack.stateAt(yynrhs - (yyi + 1))],
+                     SymbolKind.get (yystos_[yystack.stateAt (yynrhs - (yyi + 1))]),
                      ]b4_rhs_data(yynrhs, yyi + 1)b4_locations_if([,
                      b4_rhs_location(yynrhs, yyi + 1)])[);
   }]])[
 
   /* YYTRANSLATE_(TOKEN-NUM) -- Symbol number corresponding to TOKEN-NUM
      as returned by yylex, with out-of-bounds checking.  */
-  private static final ]b4_int_type_for([b4_translate])[ yytranslate_ (int t)
+  private static final SymbolKind yytranslate_ (int t)
 ]b4_api_token_raw_if(dnl
 [[  {
-    return t;
+    return SymbolKind.get (t);
   }
 ]],
 [[  {
     int user_token_number_max_ = ]b4_user_token_number_max[;
-    ]b4_int_type_for([b4_translate])[ undef_token_ = ]b4_undef_token_number[;
-
     if (t <= 0)
-      return Lexer.EOF;
+      return SymbolKind.]b4_symbol_prefix[YYEOF;
     else if (t <= user_token_number_max_)
-      return yytranslate_table_[t];
+      return SymbolKind.get (yytranslate_table_[t]);
     else
-      return undef_token_;
+      return SymbolKind.]b4_symbol_prefix[YYUNDEF;
   }
   ]b4_integral_parser_table_define([translate_table], [b4_translate])[
 ]])[
 
-  private static final ]b4_int_type_for([b4_translate])[ yy_error_token_ = 1;
-
-  private static final int yylast_ = ]b4_last[;
-  private static final int yynnts_ = ]b4_nterms_number[;
-  private static final int yyempty_ = -2;
-  private static final int yyfinal_ = ]b4_final_state_number[;
-  private static final int yyntokens_ = ]b4_tokens_number[;
-
-/* User implementation code.  */
-]b4_percent_code_get[]dnl
+  private static final int YYLAST_ = ]b4_last[;
+  private static final int YYEMPTY_ = -2;
+  private static final int YYFINAL_ = ]b4_final_state_number[;
+  private static final int YYNTOKENS_ = ]b4_tokens_number[;
 
+]b4_percent_code_get[
 }
-
+]b4_percent_code_get([[epilogue]])[]dnl
 b4_epilogue[]dnl
 b4_output_end

data/skeletons/stack.hh

@@ -32,8 +32,8 @@ m4_define([b4_stack_define],
     {
     public:
       // Hide our reversed order.
-      typedef typename S::reverse_iterator iterator;
-      typedef typename S::const_reverse_iterator const_iterator;
+      typedef typename S::iterator iterator;
+      typedef typename S::const_iterator const_iterator;
       typedef typename S::size_type size_type;
       typedef typename std::ptrdiff_t index_type;
 
@@ -91,24 +91,18 @@ m4_define([b4_stack_define],
         return index_type (seq_.size ());
       }
 
-      std::ptrdiff_t
-      ssize () const YY_NOEXCEPT
-      {
-        return std::ptrdiff_t (size ());
-      }
-
       /// Iterator on top of the stack (going downwards).
       const_iterator
       begin () const YY_NOEXCEPT
       {
-        return seq_.rbegin ();
+        return seq_.begin ();
       }
 
       /// Bottom of the stack.
       const_iterator
       end () const YY_NOEXCEPT
       {
-        return seq_.rend ();
+        return seq_.end ();
       }
 
       /// Present a slice of the top of a stack.

doc/local.mk

@@ -20,36 +20,36 @@ AM_MAKEINFOFLAGS =                                              \
   --set-customization-variable=AVOID_MENU_REDUNDANCY=true       \
   --set-customization-variable=ICONS=true
 
-info_TEXINFOS = doc/bison.texi
-doc_bison_TEXINFOS =                            \
+info_TEXINFOS = %D%/bison.texi
+%C%_bison_TEXINFOS =                            \
   $(CROSS_OPTIONS_TEXI)                         \
-  doc/fdl.texi                                  \
-  doc/gpl-3.0.texi                              \
-  doc/relocatable.texi
+  %D%/fdl.texi                                  \
+  %D%/gpl-3.0.texi                              \
+  %D%/relocatable.texi
 
 # Cannot express dependencies directly on file names because of Automake.
 # Obfuscate with a variable.
-doc_bison = doc/bison
-$(doc_bison).dvi: $(FIGS_GV:.gv=.eps)
-$(doc_bison).info: $(FIGS_GV:.gv=.txt)
-$(doc_bison).pdf: $(FIGS_GV:.gv=.pdf)
-$(doc_bison).html: $(FIGS_GV:.gv=.svg)
+%C%_bison = %D%/bison
+$(%C%_bison).dvi: $(FIGS_GV:.gv=.eps)
+$(%C%_bison).info: $(FIGS_GV:.gv=.txt)
+$(%C%_bison).pdf: $(FIGS_GV:.gv=.pdf)
+$(%C%_bison).html: $(FIGS_GV:.gv=.svg)
 
-TEXI2DVI = texi2dvi --build-dir=doc/bison.t2d -I doc
-CLEANDIRS += doc/bison.t2d
+TEXI2DVI = texi2dvi --build-dir=%D%/bison.t2d -I %D%
+CLEANDIRS += %D%/bison.t2d
 
-MOSTLYCLEANFILES += $(top_srcdir)/doc/*.tmp
+MOSTLYCLEANFILES += $(top_srcdir)/%D%/*.tmp
 
 CROSS_OPTIONS_PL = $(top_srcdir)/build-aux/cross-options.pl
-CROSS_OPTIONS_TEXI = $(top_srcdir)/doc/cross-options.texi
-$(CROSS_OPTIONS_TEXI): doc/bison.help $(CROSS_OPTIONS_PL)
+CROSS_OPTIONS_TEXI = $(top_srcdir)/%D%/cross-options.texi
+$(CROSS_OPTIONS_TEXI): %D%/bison.help $(CROSS_OPTIONS_PL)
 # Create $@~ which is the previous contents.  Don't use 'mv' here so
 # that even if we are interrupted, the file is still available for
 # diff in the next run.  Note that $@ might not exist yet.
 	$(AM_V_GEN){ test ! -f $@ || cat $@; } >$@~
 	$(AM_V_at)test ! -f $@.tmp || rm -f $@.tmp
 	$(AM_V_at)$(PERL) $(CROSS_OPTIONS_PL) $(top_srcdir)/src/scan-gram.l \
-	  <$(top_srcdir)/doc/bison.help >$@.tmp
+	  <$(top_srcdir)/%D%/bison.help >$@.tmp
 	$(AM_V_at)diff -u $@~ $@.tmp || true
 	$(AM_V_at)mv $@.tmp $@
 MAINTAINERCLEANFILES = $(CROSS_OPTIONS_TEXI)
@@ -57,34 +57,34 @@ MAINTAINERCLEANFILES = $(CROSS_OPTIONS_TEXI)
 
 # Fix Info's @code in @deftype
 # https://lists.gnu.org/archive/html/help-texinfo/2019-11/msg00004.html
-all: $(srcdir)/$(doc_bison).info.bak
-$(srcdir)/$(doc_bison).info.bak: $(srcdir)/$(doc_bison).info
+all: $(srcdir)/$(%C%_bison).info.bak
+$(srcdir)/$(%C%_bison).info.bak: $(srcdir)/$(%C%_bison).info
 	$(AM_V_GEN) $(PERL) -pi.bak -0777	\
 	  -e 's{(^ --.*\n(?: {10}.*\n)*)}'	\
 	  -e '{'				\
 	  -e '  $$def = $$1;'			\
 	  -e '  $$def =~ s/‘|’//g;'		\
 	  -e '  $$def;'				\
-	  -e '}gem;' $(srcdir)/$(doc_bison).info
+	  -e '}gem;' $(srcdir)/$(%C%_bison).info
 	@ touch $@
-EXTRA_DIST += $(srcdir)/$(doc_bison).info.bak
-MAINTAINERCLEANFILES += $(srcdir)/$(doc_bison).info.bak
+EXTRA_DIST += $(srcdir)/$(%C%_bison).info.bak
+MAINTAINERCLEANFILES += $(srcdir)/$(%C%_bison).info.bak
 
 
 ## ---------- ##
 ## Ref card.  ##
 ## ---------- ##
 
-EXTRA_DIST += doc/refcard.tex
-CLEANFILES += doc/refcard.pdf
+EXTRA_DIST += %D%/refcard.tex
+CLEANFILES += %D%/refcard.pdf
 
-doc/refcard.pdf: doc/refcard.tex
-	$(AM_V_GEN) cd doc && pdftex $(abs_top_srcdir)/doc/refcard.tex
+%D%/refcard.pdf: %D%/refcard.tex
+	$(AM_V_GEN) cd %D% && pdftex $(abs_top_srcdir)/%D%/refcard.tex
 
 
 
 ## ---------------- ##
-## doc/bison.help.  ##
+## %D%/bison.help.  ##
 ## ---------------- ##
 
 # Some of our targets (cross-options.texi, bison.1) use "bison --help".
@@ -106,18 +106,19 @@ doc/refcard.pdf: doc/refcard.tex
 # compilations of cross-options.texi and bison.1.  At the cost of
 # repeated builds of bison.help.
 
-EXTRA_DIST += $(top_srcdir)/doc/bison.help
+EXTRA_DIST += $(top_srcdir)/%D%/bison.help
 if ! CROSS_COMPILING
-MAINTAINERCLEANFILES += $(top_srcdir)/doc/bison.help
-$(top_srcdir)/doc/bison.help: src/bison$(EXEEXT)
-	$(AM_V_GEN)LC_ALL=C tests/bison --version >doc/bison.help.tmp
+MAINTAINERCLEANFILES += $(top_srcdir)/%D%/bison.help
+$(top_srcdir)/%D%/bison.help: src/bison$(EXEEXT)
+	$(AM_V_GEN)$(MKDIR_P) %D%
+	$(AM_V_at) LC_ALL=C tests/bison --version >%D%/bison.help.tmp
 	$(AM_V_at) LC_ALL=C tests/bison --help | \
 ## Avoid depending on the path to Bison.
 	  sed -e 's,^Usage: .*/bison \[OPTION\],Usage: bison [OPTION],g' \
 ## Avoid variations in the output depending on whether we are
 ## on a glibc system.
-	      -e '/translation bugs/d'  >>doc/bison.help.tmp
-	$(AM_V_at)$(top_srcdir)/build-aux/move-if-change doc/bison.help.tmp $@
+	      -e '/translation bugs/d'  >>%D%/bison.help.tmp
+	$(AM_V_at)$(top_srcdir)/build-aux/move-if-change %D%/bison.help.tmp $@
 endif ! CROSS_COMPILING
 
 
@@ -125,7 +126,7 @@ endif ! CROSS_COMPILING
 ## Man Pages.  ##
 ## ----------- ##
 
-dist_man_MANS = $(top_srcdir)/doc/bison.1
+dist_man_MANS = $(top_srcdir)/%D%/bison.1
 
 EXTRA_DIST += $(dist_man_MANS:.1=.x)
 MAINTAINERCLEANFILES += $(dist_man_MANS)
@@ -136,12 +137,12 @@ remove_time_stamp = \
 
 # Depend on configure to get version number changes.
 if ! CROSS_COMPILING
-MAN_DEPS = doc/bison.help doc/bison.x $(top_srcdir)/configure
+MAN_DEPS = %D%/bison.help %D%/bison.x $(top_srcdir)/configure
 endif
 
-$(top_srcdir)/doc/bison.1: $(MAN_DEPS)
+$(top_srcdir)/%D%/bison.1: $(MAN_DEPS)
 	$(AM_V_GEN)$(HELP2MAN)			\
-	    --include=$(top_srcdir)/doc/bison.x	\
+	    --include=$(top_srcdir)/%D%/bison.x	\
 	    --output=$@.tmp tests/bison
 	$(AM_V_at)if $(remove_time_stamp) $@ >$@a.tmp 2>/dev/null &&		\
 	   $(remove_time_stamp) $@.tmp | cmp $@a.tmp - >/dev/null 2>&1; then	\
@@ -152,7 +153,7 @@ $(top_srcdir)/doc/bison.1: $(MAN_DEPS)
 	$(AM_V_at)rm -f $@*.tmp
 
 if ENABLE_YACC
-nodist_man_MANS = doc/yacc.1
+nodist_man_MANS = %D%/yacc.1
 endif
 
 ## ----------------------------- ##
@@ -161,8 +162,8 @@ endif
 
 CLEANFILES += $(FIGS_GV:.gv=.eps) $(FIGS_GV:.gv=.pdf) $(FIGS_GV:.gv=.svg)
 FIGS_GV =                                               \
-  doc/figs/example.gv                                   \
-  doc/figs/example-reduce.gv doc/figs/example-shift.gv
+  %D%/figs/example.gv                                   \
+  %D%/figs/example-reduce.gv %D%/figs/example-shift.gv
 EXTRA_DIST +=                                                   \
   $(FIGS_GV) $(FIGS_GV:.gv=.txt)                                \
   $(FIGS_GV:.gv=.eps) $(FIGS_GV:.gv=.pdf) $(FIGS_GV:.gv=.svg)
@@ -193,8 +194,8 @@ DOXYGEN = doxygen
 
 doc: html
 
-html-local: doc/Doxyfile
-	$(AM_V_GEN) $(DOXYGEN) doc/Doxyfile
+html-local: %D%/Doxyfile
+	$(AM_V_GEN) $(DOXYGEN) %D%/Doxyfile
 
 edit = sed -e 's,@PACKAGE_NAME\@,$(PACKAGE_NAME),g' \
 	   -e 's,@PACKAGE_VERSION\@,$(PACKAGE_VERSION),g' \
@@ -202,11 +203,11 @@ edit = sed -e 's,@PACKAGE_NAME\@,$(PACKAGE_NAME),g' \
 	   -e 's,@top_builddir\@,$(top_builddir),g' \
 	   -e 's,@top_srcdir\@,$(top_srcdir),g'
 
-EXTRA_DIST += doc/Doxyfile.in
-CLEANFILES += doc/Doxyfile
+EXTRA_DIST += %D%/Doxyfile.in
+CLEANFILES += %D%/Doxyfile
 # Sed is used to generate Doxyfile from Doxyfile.in instead of
 # configure, because the former is way faster than the latter.
-doc/Doxyfile: $(top_srcdir)/doc/Doxyfile.in
-	$(AM_V_GEN) $(edit) $(top_srcdir)/doc/Doxyfile.in >doc/Doxyfile
+%D%/Doxyfile: $(top_srcdir)/%D%/Doxyfile.in
+	$(AM_V_GEN) $(edit) $(top_srcdir)/%D%/Doxyfile.in >%D%/Doxyfile
 
-CLEANDIRS += doc/html
+CLEANDIRS += %D%/html

examples/c++/README.md

@@ -8,7 +8,7 @@ afterwards.
 ## simple.yy - Simple example in C++14
 A very simple example in C++, based on variants and symbol constructors.
 Variants allow to use any C++ type as semantic value type, and symbol
-constructors ensure consistency between declared token type and effective
+constructors ensure consistency between declared token kind and effective
 semantic value.
 
 Run as `./simple`.
@@ -55,5 +55,5 @@ Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
 Texts.  A copy of the license is included in the "GNU Free
 Documentation License" file as part of this distribution.
 
-# LocalWords:  mfcalc calc parsers yy ispell american
+LocalWords:  mfcalc calc parsers yy ispell american
 --->

examples/c++/calc++/README.md

@@ -38,20 +38,12 @@ End:
 
 Copyright (C) 2018-2020 Free Software Foundation, Inc.
 
-This file is part of Bison, the GNU Compiler Compiler.
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
 
-This program is free software: you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation, either version 3 of the License, or
-(at your option) any later version.
-
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-# LocalWords:  calc parsers yy MERCHANTABILITY Ctrl ispell american
+LocalWords:  calc parsers yy MERCHANTABILITY Ctrl ispell american
 --->

examples/c++/calc++/calc++.test

@@ -44,11 +44,18 @@ EOF
 run 0 9
 
 
+cat >input <<EOF
+1 +
+EOF
+run 1 'err: -:2.1: syntax error, unexpected end of file, expecting ( or identifier or number'
+
+
+# LAC finds many more tokens.
 cat >input <<EOF
 a := 1
 d := a + b * c
 EOF
-run 1 'err: -:3.1: syntax error, unexpected end of file, expecting ( or identifier or number'
+run 1 'err: -:3.1: syntax error, unexpected end of file'
 
 
 cat >input <<EOF

examples/c/README.md

@@ -28,7 +28,8 @@ Extracted from the documentation: "Multi-Function Calculator: mfcalc".
 https://www.gnu.org/software/bison/manual/html_node/Multi_002dfunction-Calc.html
 
 ## lexcalc - calculator with Flex and Bison
-The calculator, redux.  This time using a scanner generated by Flex.
+The calculator with precedence directives and location tracking.  It uses
+Flex to generate the scanner.
 
 ## reccalc - recursive calculator with Flex and Bison
 The example builds on top of the previous one to provide a reentrant parser.
@@ -37,6 +38,32 @@ recursively.  To demonstrate this feature, expressions in parentheses are
 tokenized as strings, and then recursively parsed from the parser.  So
 `(((1)+(2))*((3)+(4)))` uses eight parsers, with a depth of four.
 
+## pushcalc - calculator implemented with a push parser
+All the previous examples are so called "pull parsers": the user invokes the
+parser once, which repeatedly calls the scanner until the input is drained.
+
+This example demonstrates the "push parsers": the user calls the scanner to
+fetch the next token, passes it to the parser, and repeats the operation
+until the input is drained.
+
+This example is a straightforward conversion of the 'calc' example to the
+push-parser model.
+
+## bistromathic - all the bells and whistles
+This example demonstrates best practices when using Bison.
+- Its hand-written scanner tracks locations.
+- Its interface is pure.
+- It uses the `error` token to get error recovery.
+- Its interface is "incremental", well suited for interaction: it uses the
+  push-parser API to feed the parser with the incoming tokens.
+- It features an interactive command line with completion based on the
+  parser state, based on `yyexpected_tokens`.
+- It uses Bison's standard catalogue for internationalization of generated
+  messages.
+- It uses a custom syntax error with location, lookahead correction and
+  token internationalization.
+- It supports debug traces with semantic values.
+- It uses named references instead of the traditional $1, $2, etc.
 
 <!---
 
@@ -47,6 +74,8 @@ End:
 
 Copyright (C) 2018-2020 Free Software Foundation, Inc.
 
+This file is part of GNU bison, the GNU Compiler Compiler.
+
 Permission is granted to copy, distribute and/or modify this document
 under the terms of the GNU Free Documentation License, Version 1.3 or
 any later version published by the Free Software Foundation; with no
@@ -54,6 +83,8 @@ Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
 Texts.  A copy of the license is included in the "GNU Free
 Documentation License" file as part of this distribution.
 
-# LocalWords:  mfcalc calc parsers yy rpcalc lexcalc redux reccalc ispell
-# LocalWords:  reentrant tokenized american postfix
---->
+LocalWords:  mfcalc calc parsers yy rpcalc lexcalc redux reccalc ispell
+LocalWords:  reentrant tokenized american postfix pushcalc bistromathic
+LocalWords:  lookahead
+
+-->

examples/c/bistromathic/Makefile

@@ -0,0 +1,40 @@
+# This Makefile is designed to be simple and readable.  It does not
+# aim at portability.  It requires GNU Make.
+
+BASE = bistromathic
+BISON = bison
+XSLTPROC = xsltproc
+
+# We need to find the headers and libs for readline and intl.
+# You probably need to customize this for your own environment.
+CPPFLAGS = -I/opt/local/include
+LDFLAGS = -L/opt/local/lib
+
+# Find the translation catalogue for Bison's generated messagess.
+BISON_LOCALEDIR = $(shell $(BISON) $(BISON_FLAGS) --print-localedir)
+CPPFLAGS += -DENABLE_NLS -DBISON_LOCALEDIR='"$(BISON_LOCALEDIR)"'
+
+LIBS = -lreadline -lintl
+
+all: $(BASE)
+
+%.c %.h %.xml %.gv: %.y
+	$(BISON) $(BISONFLAGS) --defines --xml --graph=$*.gv -o $*.c $<
+
+$(BASE): parse.o
+	$(CC) $(CPPFLAGS) $(CFLAGS) -o $@ $^ $(LDFLAGS) $(LIBS)
+
+run: $(BASE)
+	@echo "Type bistromathic expressions.  Quit with ctrl-d."
+	./$<
+
+html: $(BASE).html
+%.html: %.xml
+	$(XSLTPROC) $(XSLTPROCFLAGS) -o $@ $$($(BISON) --print-datadir)/xslt/xml2xhtml.xsl $<
+
+CLEANFILES =						\
+  $(BASE) *.o						\
+  parse.[ch] parse.output parse.xml parse.html parse.gv
+
+clean:
+	rm -f $(CLEANFILES)

examples/c/bistromathic/README.md

@@ -0,0 +1,37 @@
+# bistromathic - all the bells and whistles
+This example demonstrates best practices when using Bison.
+- Its hand-written scanner tracks locations.
+- Its interface is pure.
+- It uses the `error` token to get error recovery.
+- Its interface is "incremental", well suited for interaction: it uses the
+  push-parser API to feed the parser with the incoming tokens.
+- It features an interactive command line with completion based on the
+  parser state, based on `yyexpected_tokens`.
+- It uses Bison's standard catalogue for internationalization of generated
+  messages.
+- It uses a custom syntax error with location, lookahead correction and
+  token internationalization.
+- It supports debug traces with semantic values.
+- It uses named references instead of the traditional $1, $2, etc.
+
+<!---
+Local Variables:
+fill-column: 76
+ispell-dictionary: "american"
+End:
+
+Copyright (C) 2020 Free Software Foundation, Inc.
+
+This file is part of Bison, the GNU Compiler Compiler.
+
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
+
+LocalWords:  bistromathic yyexpected lookahead ispell american
+LocalWords:  MERCHANTABILITY
+
+--->

examples/c/bistromathic/bistromathic.test

@@ -0,0 +1,265 @@
+#! /bin/sh
+
+# Copyright (C) 2020 Free Software Foundation, Inc.
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# macOS' version of readline does not repeat stdin on stdout in
+# non-interactive mode.
+if ! echo '1-1' | prog | grep '1-1' >/dev/null; then
+   strip_prompt=true
+fi
+
+cat >input <<EOF
+1+2*3
+EOF
+run 0 '> 1+2*3
+7
+> '
+
+cat >input <<EOF
+(1+2) * 3
+EOF
+run 0  '> (1+2) * 3
+9
+> '
+run -noerr 0 '> (1+2) * 3
+9
+> ' -p
+
+cat >input <<EOF
+a = 256
+sqrt (a)
+EOF
+run 0 '> a = 256
+256
+> sqrt (a)
+16
+> '
+
+cat >input <<EOF
+a = .16
+b = 10 ^ 2
+sqrt (a * b)
+EOF
+run 0 '> a = .16
+0.16
+> b = 10 ^ 2
+100
+> sqrt (a * b)
+4
+> '
+
+cat >input <<EOF
+*
+EOF
+run 0 '> *
+> ''
+err: 1.1: syntax error: expected end of file or - or ( or exit or number or function or variable before *'
+
+cat >input <<EOF
+1 + 2 * * 3
+EOF
+run 0 '> 1 + 2 * * 3
+> ''
+err: 1.9: syntax error: expected - or ( or number or function or variable before *'
+
+cat >input <<EOF
+1 / 0
+EOF
+run 0 '> 1 / 0
+> ''
+err: 1.1-5: error: division by zero'
+
+
+## ---------------- ##
+## Error recovery.  ##
+## ---------------- ##
+
+cat >input <<EOF
+((1 ++ 2) ** 3)
+(1 ++ 2) + (3 ** 4)
+EOF
+run 0 '> ((1 ++ 2) ** 3)
+666
+> (1 ++ 2) + (3 ** 4)
+1332
+> ''
+err: 1.6: syntax error: expected - or ( or number or function or variable before +
+err: 2.5: syntax error: expected - or ( or number or function or variable before +
+err: 2.16: syntax error: expected - or ( or number or function or variable before *'
+
+# The rule "( error )" should work even if there are no tokens between "(" and ")".
+cat >input <<EOF
+()
+EOF
+run 0 '> ()
+666
+> ''
+err: 1.2: syntax error: expected - or ( or number or function or variable before )'
+
+
+cat >input <<EOF
+100% + 10
+EOF
+run 0 '> 100% + 10
+> ''
+err: 1.4: syntax error: invalid character: %'
+
+# Traces.  This allows to check the location of the error.  If we
+# forget to map YYerror to YYUNDEF, error recovery enters an endless
+# loop with this input.
+cat >input <<EOF
+(+_)
+EOF
+run 0 '> (+_)
+666
+> ''
+err: Starting parse
+err: Entering state 0
+err: Stack now 0
+err: Reading a token
+err: Next token is token ( (1.1: )
+err: Shifting token ( (1.1: )
+err: Entering state 2
+err: Stack now 0 2
+err: Return for a new token:
+err: Reading a token
+err: Next token is token + (1.2: )
+err: LAC: initial context established for +
+err: LAC: checking lookahead +: Err
+err: LAC: checking lookahead end of file: Err
+err: LAC: checking lookahead +: Err
+err: LAC: checking lookahead -: S1
+err: LAC: checking lookahead *: Err
+err: LAC: checking lookahead /: Err
+err: LAC: checking lookahead ^: Err
+err: LAC: checking lookahead (: S2
+err: LAC: checking lookahead ): Err
+err: LAC: checking lookahead =: Err
+err: LAC: checking lookahead exit: Err
+err: LAC: checking lookahead number: S4
+err: LAC: checking lookahead function: S5
+err: LAC: checking lookahead variable: S6
+err: LAC: checking lookahead NEG: Err
+err: 1.2: syntax error: expected - or ( or number or function or variable before +
+err: LAC: initial context discarded due to error recovery
+err: Shifting token error (1.2: )
+err: Entering state 10
+err: Stack now 0 2 10
+err: Next token is token + (1.2: )
+err: LAC: initial context established for +
+err: LAC: checking lookahead +: Err
+err: Error: discarding token + (1.2: )
+err: Error: popping token error (1.2: )
+err: Stack now 0 2
+err: LAC: initial context discarded due to error recovery
+err: Shifting token error (1.2: )
+err: Entering state 10
+err: Stack now 0 2 10
+err: Return for a new token:
+err: 1.3: syntax error: invalid character: _
+err: Reading a token
+err: Error: popping token error (1.2: )
+err: Stack now 0 2
+err: Shifting token error (1.2-3: )
+err: Entering state 10
+err: Stack now 0 2 10
+err: Next token is token invalid token (1.3: )
+err: LAC: initial context established for invalid token
+err: LAC: checking lookahead invalid token: Always Err
+err: Error: discarding token invalid token (1.3: )
+err: Error: popping token error (1.2-3: )
+err: Stack now 0 2
+err: LAC: initial context discarded due to error recovery
+err: Shifting token error (1.2-3: )
+err: Entering state 10
+err: Stack now 0 2 10
+err: Return for a new token:
+err: Reading a token
+err: Next token is token ) (1.4: )
+err: Shifting token ) (1.4: )
+err: Entering state 20
+err: Stack now 0 2 10 20
+err: Reducing stack by rule 15 (line 151):
+err:    $1 = token ( (1.1: )
+err:    $2 = token error (1.2-3: )
+err:    $3 = token ) (1.4: )
+err: -> $$ = nterm exp (1.1-4: 666)
+err: Entering state 7
+err: Stack now 0 7
+err: Return for a new token:
+err: Reading a token
+err: Now at end of input.
+err: LAC: initial context established for end of file
+err: LAC: checking lookahead end of file: R2 G8 S19
+err: Reducing stack by rule 2 (line 126):
+err:    $1 = nterm exp (1.1-4: 666)
+err: -> $$ = nterm input (1.1-4: )
+err: Entering state 8
+err: Stack now 0 8
+err: Now at end of input.
+err: Shifting token end of file (1.5: )
+err: LAC: initial context discarded due to shift
+err: Entering state 19
+err: Stack now 0 8 19
+err: Stack now 0 8 19
+err: Cleanup: popping token end of file (1.5: )
+err: Cleanup: popping nterm input (1.1-4: )' -p
+
+
+
+## ------------ ##
+## Completion.  ##
+## ------------ ##
+
+# From now on, the differences between versions of GNU Readline are
+# too painful to try to cope with.
+if $strip_prompt; then
+    echo "SKIP: this is not the GNU Readline we expect"
+    if $exit; then
+        exit 0
+    else
+        exit 1
+    fi
+fi
+
+# Check completion after an operator.
+sed -e 's/\\t/	/g' >input <<EOF
+(1+\t\t
+EOF
+run 0 '> (1+
+(       -       atan    cos     exp     ln      number  sin     sqrt
+> (1+
+> ''
+err: 1.4: syntax error: expected - or ( or number or function or variable before end of file'
+
+# Check the completion of a word.
+sed -e 's/\\t/	/g' >input <<EOF
+(at\t\t
+EOF
+run 0 '> (atan ( ''
+> ''
+err: 1.9: syntax error: expected - or ( or number or function or variable before end of file'
+
+# Check the completion at the very beginning.
+sed -e 's/\\t/	/g' >input <<EOF
+e\t\t
+EOF
+run -n 0 '> e
+end of file  exit         exp          ''
+> e
+0
+> ''
+err: '

examples/c/bistromathic/local.mk

@@ -0,0 +1,37 @@
+## Copyright (C) 2020 Free Software Foundation, Inc.
+##
+## This program is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+bistromathicdir = $(docdir)/%D%
+
+## --------------- ##
+## Bistromathics.  ##
+## --------------- ##
+
+check_PROGRAMS += %D%/bistromathic
+TESTS += %D%/bistromathic.test
+EXTRA_DIST += %D%/bistromathic.test
+nodist_%C%_bistromathic_SOURCES = %D%/parse.y %D%/parse.h
+%D%/parse.c: $(dependencies)
+
+# Don't use gnulib's system headers.
+%C%_bistromathic_CPPFLAGS =			\
+  -DBISON_LOCALEDIR='"$(localdir)"'		\
+  -DLOCALEDIR='"$(localdir)"'			\
+  -I$(top_srcdir)/%D% -I$(top_builddir)/%D%
+%C%_bistromathic_LDADD = -lm -lreadline $(LIBINTL)
+
+dist_bistromathic_DATA = %D%/parse.y %D%/Makefile %D%/README.md
+CLEANFILES += %D%/parse.[ch] %D%/parse.output
+CLEANDIRS += %D%/*.dSYM

examples/c/bistromathic/parse.y

@@ -0,0 +1,571 @@
+%require "3.6"
+
+%code top {
+  #include <ctype.h>  // isdigit
+  #include <math.h>   // cos, sin, etc.
+  #include <stdarg.h> // va_start
+  #include <stdio.h>  // printf
+  #include <stdlib.h> // calloc
+  #include <string.h> // strcmp
+
+  #include <readline/readline.h>
+  #include <readline/history.h>
+
+  #if defined ENABLE_NLS && ENABLE_NLS
+  // Unable the translation of Bison's generated messages.
+  # define YYENABLE_NLS 1
+  # include <libintl.h>
+  // Unless specified otherwise, we expect bistromathic's own
+  // catalogue to be installed in the same tree as Bison's catalogue.
+  # ifndef LOCALEDIR
+  #  define LOCALEDIR BISON_LOCALEDIR
+  # endif
+  #endif
+}
+
+%code requires {
+  // Function type.
+  typedef double (func_t) (double);
+
+  // Data type for links in the chain of symbols.
+  typedef struct symrec symrec;
+  struct symrec
+  {
+    char *name;  // name of symbol
+    int type;    // type of symbol: either VAR or FUN
+    union
+    {
+      double var;    // value of a VAR
+      func_t *fun;   // value of a FUN
+    } value;
+    symrec *next;  // link field
+  };
+
+  symrec *putsym (char const *name, int sym_type);
+  symrec *getsym (char const *name);
+}
+
+%code provides {
+# ifndef __attribute__
+#  ifndef __GNUC__
+#   define __attribute__(Spec) /* empty */
+#  endif
+# endif
+  int yylex (const char **line, YYSTYPE *yylval, YYLTYPE *yylloc);
+  void yyerror (YYLTYPE *loc, char const *format, ...)
+    __attribute__ ((__format__ (__printf__, 2, 3)));
+}
+
+%code {
+  #if defined ENABLE_NLS && ENABLE_NLS
+  # define _(Msgid)  gettext (Msgid)
+  #else
+  # define _(Msgid)  (Msgid)
+  #endif
+
+  // Whether to quit.
+  int done = 0;
+}
+
+// Don't share global variables between the scanner and the parser.
+%define api.pure full
+
+// Generate a push parser.
+%define api.push-pull push
+
+// To avoid name clashes (e.g., with C's EOF) prefix token definitions
+// with TOK_ (e.g., TOK_EOF).
+%define api.token.prefix {TOK_}
+
+// Customized syntax error messages (see yyreport_syntax_error).
+%define parse.error custom
+
+// with locations.
+%locations
+
+// and accurate list of expected tokens.
+%define parse.lac full
+
+// Generate the parser description file (calc.output).
+%verbose
+
+// Generate YYSTYPE from the types assigned to symbols.
+%define api.value.type union
+%token
+    PLUS   "+"
+    MINUS  "-"
+    STAR   "*"
+    SLASH  "/"
+    CARET  "^"
+    LPAREN "("
+    RPAREN ")"
+    EQUAL  "="
+    EXIT   "exit"
+  <double>
+    NUM _("number")
+  <symrec*>
+    FUN _("function")
+    VAR _("variable")
+
+%nterm <double>  exp
+
+// Enable run-time traces (yydebug).
+%define parse.trace
+
+// Formatting semantic values in debug traces.
+%printer { fprintf (yyo, "%s", $$->name); } VAR;
+%printer { fprintf (yyo, "%s()", $$->name); } FUN;
+%printer { fprintf (yyo, "%g", $$); } <double>;
+
+
+// Precedence (from lowest to highest) and associativity.
+%precedence "="
+%left "+" "-"
+%left "*" "/"
+%precedence NEG // negation--unary minus
+%right "^"      // exponentiation
+
+%% // The grammar follows.
+input:
+  %empty
+| exp     { printf ("%.10g\n", $exp); }
+| "exit"  { done = 1; }
+;
+
+exp:
+  NUM
+| VAR               { $$ = $VAR->value.var; }
+| VAR "=" exp       { $$ = $3; $VAR->value.var = $3; }
+| FUN "(" exp ")"   { $$ = $FUN->value.fun ($3); }
+| exp[l] "+" exp[r] { $$ = $l + $r; }
+| exp[l] "-" exp[r] { $$ = $l - $r; }
+| exp[l] "*" exp[r] { $$ = $l * $r; }
+| exp[l] "/" exp[r]
+  {
+    if ($r == 0)
+      {
+        yyerror (&@$, "error: division by zero");
+        YYERROR;
+      }
+    else
+      $$ = $l / $r;
+  }
+| "-" exp  %prec NEG { $$ = -$2; }
+| exp[l] "^" exp[r]  { $$ = pow ($l, $r); }
+| "(" exp ")"        { $$ = $2; }
+| "(" error ")"      { $$ = 666; }
+;
+
+// End of grammar.
+%%
+
+/*------------.
+| Functions.  |
+`------------*/
+
+struct init
+{
+  char const *name;
+  func_t *fun;
+};
+
+static struct init const funs[] =
+{
+  { "atan", atan },
+  { "cos",  cos  },
+  { "exp",  exp  },
+  { "ln",   log  },
+  { "sin",  sin  },
+  { "sqrt", sqrt },
+  { 0, 0 },
+};
+
+// The symbol table: a chain of 'struct symrec'.
+static symrec *sym_table;
+
+// Put functions in table.
+static void
+init_table (void)
+{
+  for (int i = 0; funs[i].name; i++)
+    {
+      symrec *ptr = putsym (funs[i].name, TOK_FUN);
+      ptr->value.fun = funs[i].fun;
+    }
+}
+
+symrec *
+putsym (char const *name, int sym_type)
+{
+  symrec *res = (symrec *) malloc (sizeof (symrec));
+  res->name = strdup (name);
+  res->type = sym_type;
+  res->value.var = 0; // Set value to 0 even if fun.
+  res->next = sym_table;
+  sym_table = res;
+  return res;
+}
+
+symrec *
+getsym (char const *name)
+{
+  for (symrec *p = sym_table; p; p = p->next)
+    if (strcmp (p->name, name) == 0)
+      return p;
+  return NULL;
+}
+
+// How many symbols are registered.
+int
+symbol_count (void)
+{
+  int res = 0;
+  for (symrec *p = sym_table; p; p = p->next)
+    ++res;
+  return res;
+}
+
+
+/*----------.
+| Scanner.  |
+`----------*/
+
+int
+yylex (const char **line, YYSTYPE *yylval, YYLTYPE *yylloc)
+{
+  int c;
+
+  // Ignore white space, get first nonwhite character.
+  do {
+    // Move the first position onto the last.
+    yylloc->first_line = yylloc->last_line;
+    yylloc->first_column = yylloc->last_column;
+
+    yylloc->last_column += 1;
+    c = *((*line)++);
+  } while (c == ' ' || c == '\t');
+
+  switch (c)
+    {
+    case '+': return TOK_PLUS;
+    case '-': return TOK_MINUS;
+    case '*': return TOK_STAR;
+    case '/': return TOK_SLASH;
+    case '^': return TOK_CARET;
+    case '=': return TOK_EQUAL;
+    case '(': return TOK_LPAREN;
+    case ')': return TOK_RPAREN;
+
+    case '!': return TOK_YYUNDEF;
+
+    case '\0': return TOK_YYEOF;
+
+      // Numbers.
+    case '.':
+    case '0': case '1': case '2': case '3': case '4':
+    case '5': case '6': case '7': case '8': case '9':
+      {
+        int nchars = 0;
+        sscanf (*line - 1, "%lf%n", &yylval->TOK_NUM, &nchars);
+        *line += nchars - 1;
+        yylloc->last_column += nchars - 1;
+        return TOK_NUM;
+      }
+
+      // Identifiers.
+    case 'a': case 'b': case 'c': case 'd': case 'e':
+    case 'f': case 'g': case 'h': case 'i': case 'j':
+    case 'k': case 'l': case 'm': case 'n': case 'o':
+    case 'p': case 'q': case 'r': case 's': case 't':
+    case 'u': case 'v': case 'w': case 'x': case 'y':
+    case 'z':
+      {
+        int nchars = 0;
+        char buf[100];
+        sscanf (*line - 1, "%99[a-z]%n", buf, &nchars);
+        *line += nchars - 1;
+        yylloc->last_column += nchars - 1;
+        if (strcmp (buf, "exit") == 0)
+          return TOK_EXIT;
+        else
+          {
+            symrec *s = getsym (buf);
+            if (!s)
+              s = putsym (buf, TOK_VAR);
+            yylval->TOK_VAR = s;
+            return s->type;
+          }
+      }
+
+      // Stray characters.
+    default:
+      yyerror (yylloc, "syntax error: invalid character: %c", c);
+      return TOK_YYerror;
+    }
+}
+
+
+/*---------.
+| Parser.  |
+`---------*/
+
+
+const char *
+error_format_string (int argc)
+{
+  switch (argc)
+    {
+    default: /* Avoid compiler warnings. */
+    case 0: return _("%@: syntax error");
+    case 1: return _("%@: syntax error: unexpected %u");
+      // TRANSLATORS: '%@' is a location in a file, '%u' is an
+      // "unexpected token", and '%0e', '%1e'... are expected tokens
+      // at this point.
+      //
+      // For instance on the expression "1 + * 2", you'd get
+      //
+      // 1.5: syntax error: expected - or ( or number or function or variable before *
+    case 2: return _("%@: syntax error: expected %0e before %u");
+    case 3: return _("%@: syntax error: expected %0e or %1e before %u");
+    case 4: return _("%@: syntax error: expected %0e or %1e or %2e before %u");
+    case 5: return _("%@: syntax error: expected %0e or %1e or %2e or %3e before %u");
+    case 6: return _("%@: syntax error: expected %0e or %1e or %2e or %3e or %4e before %u");
+    case 7: return _("%@: syntax error: expected %0e or %1e or %2e or %3e or %4e or %5e before %u");
+    case 8: return _("%@: syntax error: expected %0e or %1e or %2e or %3e or %4e or %5e or %6e before %u");
+    }
+}
+
+
+int
+yyreport_syntax_error (const yypcontext_t *ctx)
+{
+  enum { ARGS_MAX = 7 };
+  yysymbol_kind_t arg[ARGS_MAX];
+  int argsize = yypcontext_expected_tokens (ctx, arg, ARGS_MAX);
+  if (argsize < 0)
+    return argsize;
+  const char *format = error_format_string (1 + argsize);
+
+  while (*format)
+    // %@: location.
+    if (format[0] == '%' && format[1] == '@')
+      {
+        YY_LOCATION_PRINT (stderr, *yypcontext_location (ctx));
+        format += 2;
+      }
+    // %u: unexpected token.
+    else if (format[0] == '%' && format[1] == 'u')
+      {
+        fputs (yysymbol_name (yypcontext_token (ctx)), stderr);
+        format += 2;
+      }
+    // %0e, %1e...: expected token.
+    else if (format[0] == '%'
+             && isdigit (format[1])
+             && format[2] == 'e'
+             && (format[1] - '0') < argsize)
+      {
+        int i = format[1] - '0';
+        fputs (yysymbol_name (arg[i]), stderr);
+        format += 3;
+      }
+    else
+      {
+        fputc (*format, stderr);
+        ++format;
+      }
+  fputc ('\n', stderr);
+  return 0;
+}
+
+
+// Called by yyparse on error.
+void yyerror (YYLTYPE *loc, char const *format, ...)
+{
+  YY_LOCATION_PRINT (stderr, *loc);
+  fputs (": ", stderr);
+  va_list args;
+  va_start (args, format);
+  vfprintf (stderr, format, args);
+  va_end (args);
+  putc ('\n', stderr);
+}
+
+
+/*-----------.
+| Readline.  |
+`-----------*/
+
+// Parse (and execute) this line.
+int process_line (YYLTYPE *lloc, const char *line)
+{
+  yypstate *ps = yypstate_new ();
+  int status = 0;
+  do {
+    YYSTYPE lval;
+    status = yypush_parse (ps, yylex (&line, &lval, lloc), &lval, lloc);
+  } while (status == YYPUSH_MORE);
+  yypstate_delete (ps);
+  lloc->last_line++;
+  lloc->last_column = 1;
+  return status;
+}
+
+// Get the list of possible tokens after INPUT was read.
+int
+expected_tokens (const char *input,
+                 int *tokens, int ntokens)
+{
+  YYDPRINTF ((stderr, "expected_tokens (\"%s\")", input));
+
+  // Parse the current state of the line.
+  yypstate *ps = yypstate_new ();
+  int status = 0;
+  do {
+    YYLTYPE lloc;
+    YYSTYPE lval;
+    int token = yylex (&input, &lval, &lloc);
+    // Don't let the parse know when we reach the end of input.
+    if (!token)
+      break;
+    status = yypush_parse (ps, token, &lval, &lloc);
+  } while (status == YYPUSH_MORE);
+
+  // Then query for the accepted tokens at this point.
+  int res = yypstate_expected_tokens (ps, tokens, ntokens);
+  yypstate_delete (ps);
+  return res;
+}
+
+/* Attempt to complete on the contents of TEXT.  START and END bound the
+   region of rl_line_buffer that contains the word to complete.  TEXT is
+   the word to complete.  We can use the entire contents of rl_line_buffer
+   in case we want to do some simple parsing.  Return the array of matches,
+   or NULL if there aren't any. */
+char **
+completion (const char *text, int start, int end)
+{
+  YYDPRINTF ((stderr, "completion (\"%.*s[%.*s]%s\")\n",
+              start, rl_line_buffer,
+              end - start, rl_line_buffer + start,
+              rl_line_buffer + end));
+
+  // Get list of token numbers.
+  int tokens[YYNTOKENS];
+  char *line = strndup (rl_line_buffer, start);
+  int ntokens = expected_tokens (line, tokens, YYNTOKENS);
+  free (line);
+
+  // Build MATCHES, the list of possible completions.
+  const int len = strlen (text);
+  // Need initial prefix and final NULL.
+  char **matches = calloc (ntokens + symbol_count () + 2, sizeof *matches);
+  int match = 1;
+  for (int i = 0; i < ntokens; ++i)
+    switch (tokens[i])
+      {
+      case YYSYMBOL_FUN:
+        for (symrec *s = sym_table; s; s = s->next)
+          if (s->type == TOK_FUN && strncmp (text, s->name, len) == 0)
+            matches[match++] = strdup (s->name);
+        break;
+      case YYSYMBOL_VAR:
+        for (symrec *s = sym_table; s; s = s->next)
+          if (s->type == TOK_VAR && strncmp (text, s->name, len) == 0)
+            matches[match++] = strdup (s->name);
+        break;
+      default:
+        {
+          const char* token = yysymbol_name (tokens[i]);
+          if (strncmp (text, token, len) == 0)
+            matches[match++] = strdup (token);
+          break;
+        }
+      }
+
+  // Find the longest common prefix, and install it in matches[0], as
+  // required by readline.
+  if (match == 1)
+    matches[0] = strdup (text);
+  else
+    {
+      int lcplen = strlen (matches[1]);
+      for (int i = 2; i < match && lcplen; ++i)
+        for (int j = 0; j < lcplen; ++j)
+          if (matches[1][j] != matches[i][j])
+            lcplen = j;
+      matches[0] = strndup (matches[1], lcplen);
+    }
+
+  if (yydebug)
+    {
+      fprintf (stderr, "completion (\"%.*s[%.*s]%s\") = ",
+               start, rl_line_buffer,
+               end - start, rl_line_buffer + start,
+               rl_line_buffer + end);
+      for (int i = 1; matches[i]; ++i)
+        fprintf (stderr, "%s%s",
+                 i == 1 ? "{" : ", ",
+                 matches[i]);
+      fprintf (stderr, "}\n");
+    }
+
+  // Don't fall back to proposing file names.
+  rl_attempted_completion_over = 1;
+  return matches;
+}
+
+void init_readline (void)
+{
+  // Allow conditional parsing of the ~/.inputrc file.
+  rl_readline_name = "bistromathic";
+
+  // Tell the completer that we want a crack first.
+  rl_attempted_completion_function = completion;
+
+  // The basic list of characters that signal a break between words
+  // for the completer routine.
+  rl_basic_word_break_characters = " \t\n\"\\'`@$><=;|&{(+-*/^)";
+}
+
+
+
+/*-------.
+| Main.  |
+`-------*/
+
+int main (int argc, char const* argv[])
+{
+#if defined ENABLE_NLS && ENABLE_NLS
+  // Set up internationalization.
+  setlocale (LC_ALL, "");
+  // Use Bison's standard translation catalogue for error messages
+  // (the generated messages).
+  bindtextdomain ("bison-runtime", BISON_LOCALEDIR);
+  // The translation catalogue of bistromathic is actually included in
+  // Bison's.  In your own project, use the name of your project.
+  bindtextdomain ("bison", LOCALEDIR);
+  textdomain ("bison");
+#endif
+
+  // Enable parse traces on option -p.
+  if (1 < argc && strcmp (argv[1], "-p") == 0)
+    yydebug = 1;
+  init_table ();
+  init_readline ();
+  YYLTYPE lloc = {1, 1, 1, 1};
+  while (!done)
+    {
+      char *line = readline ("> ");
+      if (!line)
+        {
+          // Finish the line started by the prompt.
+          putchar ('\n');
+          break;
+        }
+      if (*line)
+        add_history (line);
+      process_line (&lloc, line);
+      free (line);
+    }
+}

examples/c/calc/README.md

@@ -13,16 +13,11 @@ Copyright (C) 2019-2020 Free Software Foundation, Inc.
 
 This file is part of Bison, the GNU Compiler Compiler.
 
-This program is free software: you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation, either version 3 of the License, or
-(at your option) any later version.
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
 
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
 --->

examples/c/calc/calc.test

@@ -34,4 +34,10 @@ cat >input <<EOF
 (1+2) * 3
 EOF
 run 0 9
+# Check the traces.
 run -noerr 0 9 -p
+
+cat >input <<EOF
+1++2
+EOF
+run 0 "err: syntax error, unexpected '+', expecting number or '('"

examples/c/calc/calc.y

@@ -8,16 +8,22 @@
 }
 
 %define api.header.include {"calc.h"}
-%define api.value.type union /* Generate YYSTYPE from these types:  */
+
+/* Generate YYSTYPE from the types used in %token and %type.  */
+%define api.value.type union
 %token <double> NUM "number"
 %type  <double> expr term fact
 
-/* Generate the parser description file.  */
+/* Generate the parser description file (calc.output).  */
 %verbose
+
+/* Nice error messages with details. */
+%define parse.error detailed
+
 /* Enable run-time traces (yydebug).  */
 %define parse.trace
 
-/* Formatting semantic values.  */
+/* Formatting semantic values in debug traces.  */
 %printer { fprintf (yyo, "%g", $$); } <double>;
 
 %% /* The grammar follows.  */

examples/c/lexcalc/README.md

@@ -3,6 +3,8 @@
 This directory contains lexcalc, the traditional example of using Flex and
 Bison to build a simple calculator.
 
+It features detailed syntax errors with locations.
+
 <!---
 Local Variables:
 fill-column: 76
@@ -13,16 +15,10 @@ Copyright (C) 2018-2020 Free Software Foundation, Inc.
 
 This file is part of Bison, the GNU Compiler Compiler.
 
-This program is free software: you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation, either version 3 of the License, or
-(at your option) any later version.
-
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
 --->

examples/c/lexcalc/lexcalc.test

@@ -29,4 +29,9 @@ run -noerr 0 9 -p
 cat >input <<EOF
 (1+2) *
 EOF
-run 1 'err: syntax error, unexpected end-of-line, expecting ( or number'
+run 1 'err: 1.8-2.0: syntax error, unexpected end of line, expecting ( or number'
+
+cat >input <<EOF
+1 / (2 - 2)
+EOF
+run 1 'err: 1.1-11: error: division by zero'

examples/c/lexcalc/parse.y

@@ -6,10 +6,10 @@
 {
   // Tell Flex the expected prototype of yylex.
 #define YY_DECL                                 \
-  enum yytokentype yylex (YYSTYPE* yylval, int *nerrs)
+  yytoken_kind_t yylex (YYSTYPE* yylval, YYLTYPE *yylloc, int *nerrs)
   YY_DECL;
 
-  void yyerror (int *nerrs, const char *msg);
+  void yyerror (YYLTYPE *loc, int *nerrs, const char *msg);
 }
 
 // Emitted on top of the implementation file.
@@ -19,12 +19,27 @@
 #include <stdlib.h> // getenv.
 }
 
+// Don't share global variables between the scanner and the parser.
 %define api.pure full
+
+// To avoid name clashes (e.g., with C's EOF) prefix token definitions
+// with TOK_ (e.g., TOK_EOF).
 %define api.token.prefix {TOK_}
+
+// %token and %type use genuine types (e.g., "%token <int>").  Let
+// %bison define YYSTYPE as a union of all these types.
 %define api.value.type union
-%define parse.error verbose
+
+// Generate detailed error messages.
+%define parse.error detailed
+
+// with locations.
+%locations
+
+// Enable debug traces (see yydebug in main).
 %define parse.trace
- // Error count, exchanged between main, yyparse and yylex.
+
+// Error count, exchanged between main, yyparse and yylex.
 %param {int *nerrs}
 
 %token
@@ -34,8 +49,7 @@
   SLASH  "/"
   LPAREN "("
   RPAREN ")"
-  EOL    "end-of-line"
-  EOF 0  "end-of-file"
+  EOL    "end of line"
 ;
 
 %token <int> NUM "number"
@@ -66,7 +80,7 @@ exp:
   {
     if ($3 == 0)
       {
-        yyerror (nerrs, "invalid division by zero");
+        yyerror (&@$, nerrs, "error: division by zero");
         YYERROR;
       }
     else
@@ -77,9 +91,11 @@ exp:
 ;
 %%
 // Epilogue (C code).
-void yyerror (int *nerrs, const char *msg)
+
+void yyerror (YYLTYPE *loc, int *nerrs, const char *msg)
 {
-  fprintf (stderr, "%s\n", msg);
+  YY_LOCATION_PRINT (stderr, *loc);
+  fprintf (stderr, ": %s\n", msg);
   ++*nerrs;
 }
 

examples/c/lexcalc/scan.l

@@ -1,4 +1,4 @@
-/* Prologue (directives).   -*- C++ -*- */
+/* Prologue (directives).   -*- C -*- */
 
 /* Disable Flex features we don't need, to avoid warnings. */
 %option nodefault noinput nounput noyywrap
@@ -9,9 +9,24 @@
 #include <stdlib.h> /* strtol */
 
 #include "parse.h"
+
+  // Each time a rule is matched, advance the end cursor/position.
+#define YY_USER_ACTION                          \
+  yylloc->last_column += yyleng;
+
+  // Move the first position onto the last.
+#define LOCATION_STEP()                         \
+  do {                                          \
+    yylloc->first_line = yylloc->last_line;     \
+    yylloc->first_column = yylloc->last_column; \
+  } while (0)
 %}
 
 %%
+%{
+  // Each time yylex is called, move the head position to the end one.
+  LOCATION_STEP ();
+%}
  /* Rules.  */
 
 "+"      return TOK_PLUS;
@@ -27,18 +42,18 @@
   errno = 0;
   long n = strtol (yytext, NULL, 10);
   if (! (INT_MIN <= n && n <= INT_MAX && errno != ERANGE))
-    yyerror (nerrs, "integer is out of range");
+    yyerror (yylloc, nerrs, "integer is out of range");
   yylval->TOK_NUM = (int) n;
   return TOK_NUM;
 }
 
+"\n"     yylloc->last_line++; yylloc->last_column = 1; return TOK_EOL;
+
  /* Ignore white spaces. */
-[ \t]+   continue;
+[ \t]+   LOCATION_STEP (); continue;
 
-"\n"     return TOK_EOL;
+.        yyerror (yylloc, nerrs, "syntax error, invalid character"); continue;
 
-.        yyerror (nerrs, "syntax error, invalid character");
-
-<<EOF>>  return TOK_EOF;
+<<EOF>>  return TOK_YYEOF;
 %%
 /* Epilogue (C code). */

examples/c/local.mk

@@ -16,8 +16,10 @@
 cdir = $(docdir)/%D%
 dist_c_DATA = %D%/README.md
 
+include %D%/bistromathic/local.mk
 include %D%/calc/local.mk
 include %D%/lexcalc/local.mk
 include %D%/mfcalc/local.mk
+include %D%/pushcalc/local.mk
 include %D%/reccalc/local.mk
 include %D%/rpcalc/local.mk

examples/c/mfcalc/mfcalc.test

@@ -25,3 +25,15 @@ cat >input <<EOF
 EOF
 run 0 9
 run -noerr 0 9 -p
+
+cat >input <<EOF
+a = 256
+sqrt (a)
+EOF
+run 0 '256
+16'
+
+cat >input <<EOF
+16 == 12
+EOF
+run 0 'err: syntax error'

examples/c/pushcalc/Makefile

@@ -0,0 +1,28 @@
+# This Makefile is designed to be simple and readable.  It does not
+# aim at portability.  It requires GNU Make.
+
+BASE = calc
+BISON = bison
+XSLTPROC = xsltproc
+
+all: $(BASE)
+
+%.c %.h %.xml %.gv: %.y
+	$(BISON) $(BISONFLAGS) --defines --xml --graph=$*.gv -o $*.c $<
+
+$(BASE): $(BASE).o
+	$(CC) $(CFLAGS) -o $@ $^
+
+run: $(BASE)
+	@echo "Type arithmetic expressions.  Quit with ctrl-d."
+	./$<
+
+html: $(BASE).html
+%.html: %.xml
+	$(XSLTPROC) $(XSLTPROCFLAGS) -o $@ $$($(BISON) --print-datadir)/xslt/xml2xhtml.xsl $<
+
+CLEANFILES =									\
+  $(BASE) *.o $(BASE).[ch] $(BASE).output $(BASE).xml $(BASE).html $(BASE).gv
+
+clean:
+	rm -f $(CLEANFILES)

examples/c/pushcalc/README.md

@@ -0,0 +1,32 @@
+# pushcalc - push parser with Bison
+
+This directory contains pushcalc, the traditional calculator, implemented as
+a push parser.
+
+Traditionally Bison is used to create so called "pull parsers": the user
+invokes the parser once, which repeatedly calls (pulls) the scanner until
+the input is drained.
+
+This example demonstrates the "push parsers": the user calls scanner to
+fetch the next token, passes (pushes) it to the parser, and repeats the
+operation until the input is drained.
+
+This example is a straightforward conversion of the 'calc' example to the
+push-parser model.
+
+<!---
+Local Variables:
+fill-column: 76
+ispell-dictionary: "american"
+End:
+
+Copyright (C) 2020 Free Software Foundation, Inc.
+
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
+
+--->

examples/c/pushcalc/calc.test

@@ -0,0 +1,42 @@
+#! /bin/sh
+
+# Copyright (C) 2020 Free Software Foundation, Inc.
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+cat >input <<EOF
+1+2*3
+EOF
+run 0 7
+
+cat >input <<EOF
+1 - 2 - 3
+EOF
+run 0 -4
+
+cat >input <<EOF
+8 / 2 / 2
+EOF
+run 0 2
+
+cat >input <<EOF
+(1+2) * 3
+EOF
+run 0 9
+run -noerr 0 9 -p
+
+cat >input <<EOF
+1++2
+EOF
+run 0 "err: syntax error, unexpected '+', expecting number or '('"

examples/c/pushcalc/calc.y

@@ -0,0 +1,114 @@
+%code top {
+  #include <ctype.h>  /* isdigit. */
+  #include <stdbool.h>
+  #include <stdio.h>  /* For printf, etc. */
+  #include <string.h> /* strcmp. */
+}
+
+%code {
+  int yylex (YYSTYPE *yylval);
+  void yyerror (char const *);
+}
+
+%define api.header.include {"calc.h"}
+
+/* Generate YYSTYPE from the types used in %token and %type.  */
+%define api.value.type union
+%token <double> NUM "number"
+%type  <double> expr term fact
+
+/* Don't share global variables between the scanner and the parser.  */
+%define api.pure full
+/* Generate a push parser.  */
+%define api.push-pull push
+
+/* Nice error messages with details. */
+%define parse.error detailed
+
+/* Generate the parser description file (calc.output).  */
+%verbose
+
+ /* Enable run-time traces (yydebug).  */
+%define parse.trace
+
+/* Formatting semantic values in debug traces.  */
+%printer { fprintf (yyo, "%g", $$); } <double>;
+
+%% /* The grammar follows.  */
+input:
+  %empty
+| input line
+;
+
+line:
+  '\n'
+| expr '\n'  { printf ("%.10g\n", $1); }
+| error '\n' { yyerrok; }
+;
+
+expr:
+  expr '+' term { $$ = $1 + $3; }
+| expr '-' term { $$ = $1 - $3; }
+| term
+;
+
+term:
+  term '*' fact { $$ = $1 * $3; }
+| term '/' fact { $$ = $1 / $3; }
+| fact
+;
+
+fact:
+  "number"
+| '(' expr ')' { $$ = $expr; }
+;
+
+%%
+
+int
+yylex (YYSTYPE *yylval)
+{
+  int c;
+
+  /* Ignore white space, get first nonwhite character.  */
+  while ((c = getchar ()) == ' ' || c == '\t')
+    continue;
+
+  if (c == EOF)
+    return 0;
+
+  /* Char starts a number => parse the number.         */
+  if (c == '.' || isdigit (c))
+    {
+      ungetc (c, stdin);
+      scanf ("%lf", &yylval->NUM);
+      return NUM;
+    }
+
+  /* Any other character is a token by itself.        */
+  return c;
+}
+
+/* Called by yyparse on error.  */
+void
+yyerror (char const *s)
+{
+  fprintf (stderr, "%s\n", s);
+}
+
+int
+main (int argc, char const* argv[])
+{
+  /* Enable parse traces on option -p.  */
+  for (int i = 1; i < argc; ++i)
+    if (!strcmp (argv[i], "-p"))
+      yydebug = 1;
+  int status;
+  yypstate *ps = yypstate_new ();
+  do {
+    YYSTYPE lval;
+    status = yypush_parse (ps, yylex (&lval), &lval);
+  } while (status == YYPUSH_MORE);
+  yypstate_delete (ps);
+  return status;
+}

examples/c/pushcalc/local.mk

@@ -0,0 +1,33 @@
+## Copyright (C) 2020 Free Software Foundation, Inc.
+##
+## This program is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+pushcalcdir = $(docdir)/%D%
+
+## ------ ##
+## Calc.  ##
+## ------ ##
+
+check_PROGRAMS += %D%/calc
+TESTS += %D%/calc.test
+EXTRA_DIST += %D%/calc.test
+nodist_%C%_calc_SOURCES = %D%/calc.y
+%D%/calc.c: $(dependencies)
+
+# Don't use gnulib's system headers.
+%C%_calc_CPPFLAGS = -I$(top_srcdir)/%D% -I$(top_builddir)/%D%
+
+dist_pushcalc_DATA = %D%/calc.y %D%/Makefile %D%/README.md
+CLEANFILES += %D%/calc.[ch] %D%/calc.output
+CLEANDIRS += %D%/*.dSYM

examples/c/reccalc/README.md

@@ -28,18 +28,11 @@ End:
 
 Copyright (C) 2018-2020 Free Software Foundation, Inc.
 
-This file is part of Bison, the GNU Compiler Compiler.
+Permission is granted to copy, distribute and/or modify this document
+under the terms of the GNU Free Documentation License, Version 1.3 or
+any later version published by the Free Software Foundation; with no
+Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
+Texts.  A copy of the license is included in the "GNU Free
+Documentation License" file as part of this distribution.
 
-This program is free software: you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation, either version 3 of the License, or
-(at your option) any later version.
-
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
 --->

examples/c/reccalc/parse.y

@@ -4,7 +4,11 @@
 // Emitted in the header file, before the definition of YYSTYPE.
 %code requires
 {
+  #ifndef YY_TYPEDEF_YY_SCANNER_T
+  # define YY_TYPEDEF_YY_SCANNER_T
   typedef void* yyscan_t;
+  #endif
+
   typedef struct
   {
     // Whether to print the intermediate results.
@@ -22,7 +26,7 @@
   // Tell Flex the expected prototype of yylex.
   // The scanner argument must be named yyscanner.
 #define YY_DECL                                                         \
-  enum yytokentype yylex (YYSTYPE* yylval, yyscan_t yyscanner, result *res)
+  yytoken_kind_t yylex (YYSTYPE* yylval, yyscan_t yyscanner, result *res)
   YY_DECL;
 
   void yyerror (yyscan_t scanner, result *res, const char *msg, ...);
@@ -132,7 +136,6 @@ exp:
 
 %%
 // Epilogue (C code).
-
 #include "scan.h"
 
 result

examples/c/reccalc/scan.l

@@ -1,4 +1,4 @@
-/* Prologue (directives).   -*- C++ -*- */
+/* Prologue (directives).   -*- C -*- */
 
 /* Disable Flex features we don't need, to avoid warnings. */
 %option nodefault noinput nounput noyywrap
@@ -17,10 +17,14 @@
 
 %%
 %{
+  // Number of opened parentheses.
   int nesting = 0;
+  // A buffer storing the text inside the outer parentheses.
   char *str = NULL;
-  int size = 0;
+  // Its allocated size.
   int capacity = 0;
+  // Its used size.
+  int size = 0;
 #define STR_APPEND()                                    \
   do {                                                  \
     if (capacity < size + yyleng + 1)                   \

examples/d/calc.y

@@ -99,16 +99,15 @@ class CalcLexer(R) : Lexer
     while (!input.empty && input.front != '\n' && isWhite (input.front))
       input.popFront;
 
-    // Handle EOF.
     if (input.empty)
-      return YYTokenType.EOF;
+      return TokenKind.YYEOF;
 
     // Numbers.
     if (input.front.isNumber)
       {
         import std.conv : parse;
         semanticVal_.ival = input.parse!int;
-        return YYTokenType.NUM;
+        return TokenKind.NUM;
       }
 
     // Individual characters
@@ -116,16 +115,15 @@ class CalcLexer(R) : Lexer
     input.popFront;
     switch (ch)
       {
-      case EOF: return YYTokenType.EOF;
-      case '=': return YYTokenType.EQ;
-      case '+': return YYTokenType.PLUS;
-      case '-': return YYTokenType.MINUS;
-      case '*': return YYTokenType.STAR;
-      case '/': return YYTokenType.SLASH;
-      case '(': return YYTokenType.LPAR;
-      case ')': return YYTokenType.RPAR;
-      case '\n': return YYTokenType.EOL;
-      default:  assert(0);
+      case '=':  return TokenKind.EQ;
+      case '+':  return TokenKind.PLUS;
+      case '-':  return TokenKind.MINUS;
+      case '*':  return TokenKind.STAR;
+      case '/':  return TokenKind.SLASH;
+      case '(':  return TokenKind.LPAR;
+      case ')':  return TokenKind.RPAR;
+      case '\n': return TokenKind.EOL;
+      default: assert(0);
       }
   }
 }

examples/java/Calc.y

@@ -1,168 +0,0 @@
-%language "Java"
-
-%define api.parser.class {Calc}
-%define api.parser.public
-
-%define parse.error verbose
-%locations
-
-%code imports {
-  import java.io.IOException;
-  import java.io.InputStream;
-  import java.io.InputStreamReader;
-  import java.io.Reader;
-  import java.io.StreamTokenizer;
-}
-
-%code {
-  public static void main (String args[]) throws IOException
-  {
-    CalcLexer l = new CalcLexer (System.in);
-    Calc p = new Calc (l);
-    p.parse ();
-  }
-}
-
-/* Bison Declarations */
-%token <Integer> NUM "number"
-%type  <Integer> exp
-
-%nonassoc '='       /* comparison            */
-%left '-' '+'
-%left '*' '/'
-%precedence NEG     /* negation--unary minus */
-%right '^'          /* exponentiation        */
-
-/* Grammar follows */
-%%
-input:
-  line
-| input line
-;
-
-line:
-  '\n'
-| exp '\n'           { System.out.println ($exp); }
-| error '\n'
-;
-
-exp:
-  NUM                { $$ = $1;                                             }
-| exp '=' exp
-  {
-    if ($1.intValue () != $3.intValue ())
-      yyerror (@$, "calc: error: " + $1 + " != " + $3);
-  }
-| exp '+' exp        { $$ = new Integer ($1.intValue () + $3.intValue ());  }
-| exp '-' exp        { $$ = new Integer ($1.intValue () - $3.intValue ());  }
-| exp '*' exp        { $$ = new Integer ($1.intValue () * $3.intValue ());  }
-| exp '/' exp        { $$ = new Integer ($1.intValue () / $3.intValue ());  }
-| '-' exp  %prec NEG { $$ = new Integer (-$2.intValue ());                  }
-| exp '^' exp        { $$ = new Integer ((int)
-                                         Math.pow ($1.intValue (),
-                                                   $3.intValue ()));        }
-| '(' exp ')'        { $$ = $2;                                             }
-| '(' error ')'      { $$ = new Integer (1111);                             }
-| '!'                { $$ = new Integer (0); return YYERROR;                }
-| '-' error          { $$ = new Integer (0); return YYERROR;                }
-;
-
-
-%%
-class CalcLexer implements Calc.Lexer {
-
-  StreamTokenizer st;
-
-  public CalcLexer (InputStream is)
-  {
-    st = new StreamTokenizer (new InputStreamReader (is));
-    st.resetSyntax ();
-    st.eolIsSignificant (true);
-    st.whitespaceChars ('\t', '\t');
-    st.whitespaceChars (' ', ' ');
-    st.wordChars ('0', '9');
-  }
-
-
-  Position yypos = new Position (1, 0);
-
-  public Position getStartPos () {
-    return yypos;
-  }
-
-  public Position getEndPos () {
-    return yypos;
-  }
-
-  public void yyerror (Calc.Location l, String s)
-  {
-    if (l == null)
-      System.err.println (s);
-    else
-      System.err.println (l + ": " + s);
-  }
-
-
-  Integer yylval;
-
-  public Object getLVal () {
-    return yylval;
-  }
-
-  public int yylex () throws IOException {
-    int ttype = st.nextToken ();
-    yypos = new Position (yypos.lineno (), yypos.token () + 1);
-    if (ttype == st.TT_EOF)
-      return EOF;
-    else if (ttype == st.TT_EOL)
-      {
-        yypos = new Position (yypos.lineno () + 1, 0);
-        return (int) '\n';
-      }
-    else if (ttype == st.TT_WORD)
-      {
-        yylval = new Integer (st.sval);
-        return NUM;
-      }
-    else
-      return st.ttype;
-  }
-}
-
-
-class Position {
-  public int line;
-  public int token;
-
-  public Position ()
-  {
-    line = 0;
-    token = 0;
-  }
-
-  public Position (int l, int t)
-  {
-    line = l;
-    token = t;
-  }
-
-  public boolean equals (Position l)
-  {
-    return l.line == line && l.token == token;
-  }
-
-  public String toString ()
-  {
-    return Integer.toString (line) + "." + Integer.toString(token);
-  }
-
-  public int lineno ()
-  {
-    return line;
-  }
-
-  public int token ()
-  {
-    return token;
-  }
-}

examples/java/README.md

@@ -5,8 +5,11 @@ This directory contains examples of Bison grammar files in Java.
 You can run `make` to compile these examples.  And `make clean` to tidy
 afterwards.
 
-## java/Calc.y
-The usual calculator.
+## simple/Calc.y
+The usual calculator, a very simple version.
+
+## calc/Calc.y
+The calculator, but with location tracking and debug traces.
 
 <!---
 

examples/java/calc/Calc.test

@@ -0,0 +1,38 @@
+#! /bin/sh
+
+# Copyright (C) 2018-2020 Free Software Foundation, Inc.
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+cat >input <<EOF
+1 + 2 * 3
+EOF
+run 0 7
+
+cat >input <<EOF
+1 + 2 * 3 = 7
+(1 + 2) * 3 = 9
+EOF
+run 0 '7
+9'
+
+cat >input <<EOF
+1 + 2 * * 3
+EOF
+run 0 "err: 1.9-1.10: syntax error: expected ! or - or ( or number before *"
+
+cat >input <<EOF
+12   222
+EOF
+run 0 "err: 1.6-1.9: syntax error: expected + or - or * or / or ^ or = or end of line before number"

examples/java/calc/Calc.y

@@ -0,0 +1,263 @@
+%language "Java"
+
+%define api.parser.class {Calc}
+%define api.parser.public
+
+%define parse.error custom
+%define parse.trace
+
+%locations
+
+%code imports {
+  import java.io.BufferedReader;
+  import java.io.IOException;
+  import java.io.InputStream;
+  import java.io.InputStreamReader;
+  import java.io.Reader;
+  import java.io.StreamTokenizer;
+  import java.nio.CharBuffer;
+}
+
+%code {
+  public static void main(String[] args) throws IOException {
+    CalcLexer l = new CalcLexer(System.in);
+    Calc p = new Calc(l);
+    for (String arg : args)
+      if (arg.equals("-p"))
+        p.setDebugLevel(1);
+    if (!p.parse())
+      System.exit(1);
+  }
+
+  static String i18n(String s) {
+    return s;
+  }
+}
+
+/* Bison Declarations */
+%token
+    BANG   "!"
+    PLUS   "+"
+    MINUS  "-"
+    STAR   "*"
+    SLASH  "/"
+    CARET  "^"
+    LPAREN "("
+    RPAREN ")"
+    EQUAL  "="
+    EOL    _("end of line")
+  <Integer>
+    NUM    _("number")
+%type  <Integer> exp
+
+%nonassoc "="       /* comparison            */
+%left "-" "+"
+%left "*" "/"
+%precedence NEG     /* negation--unary minus */
+%right "^"          /* exponentiation        */
+
+/* Grammar follows */
+%%
+input:
+  line
+| input line
+;
+
+line:
+  EOL
+| exp EOL            { System.out.println($exp); }
+| error EOL
+;
+
+exp:
+  NUM                { $$ = $1; }
+| exp "=" exp
+  {
+    if ($1.intValue() != $3.intValue())
+      yyerror(@$, "calc: error: " + $1 + " != " + $3);
+  }
+| exp "+" exp        { $$ = $1 + $3;  }
+| exp "-" exp        { $$ = $1 - $3;  }
+| exp "*" exp        { $$ = $1 * $3;  }
+| exp "/" exp        { $$ = $1 / $3;  }
+| "-" exp  %prec NEG { $$ = -$2; }
+| exp "^" exp        { $$ = (int) Math.pow($1, $3); }
+| "(" exp ")"        { $$ = $2; }
+| "(" error ")"      { $$ = 1111; }
+| "!"                { $$ = 0; return YYERROR; }
+| "-" error          { $$ = 0; return YYERROR; }
+;
+
+%%
+class CalcLexer implements Calc.Lexer {
+
+  StreamTokenizer st;
+  PositionReader reader;
+
+  public CalcLexer(InputStream is) {
+    reader = new PositionReader(new InputStreamReader(is));
+    st = new StreamTokenizer(reader);
+    st.resetSyntax();
+    st.eolIsSignificant(true);
+    st.wordChars('0', '9');
+  }
+
+  Position start = new Position(1, 0);
+  Position end = new Position(1, 0);
+
+  public Position getStartPos() {
+    return new Position(start);
+  }
+
+  public Position getEndPos() {
+    return new Position(end);
+  }
+
+  public void reportSyntaxError(Calc.Context ctx) {
+    System.err.print(ctx.getLocation() + ": syntax error");
+    {
+      final int TOKENMAX = 10;
+      Calc.SymbolKind[] arg = new Calc.SymbolKind[TOKENMAX];
+      int n = ctx.getExpectedTokens(arg, TOKENMAX);
+      for (int i = 0; i < n; ++i)
+        System.err.print((i == 0 ? ": expected " : " or ")
+                         + arg[i].getName());
+    }
+    {
+      Calc.SymbolKind lookahead = ctx.getToken();
+      if (lookahead != null)
+        System.err.print(" before " + lookahead.getName());
+    }
+    System.err.println("");
+  }
+
+  public void yyerror(Calc.Location l, String s) {
+    if (l == null)
+      System.err.println(s);
+    else
+      System.err.println(l + ": " + s);
+  }
+
+  Integer yylval;
+
+  public Object getLVal() {
+    return yylval;
+  }
+
+  public int yylex() throws IOException {
+    start.set(reader.getPosition());
+    int ttype = st.nextToken();
+    end.set(reader.getPosition());
+    switch (ttype) {
+    case StreamTokenizer.TT_EOF:
+      return YYEOF;
+    case StreamTokenizer.TT_EOL:
+      end.line += 1;
+      end.column = 0;
+      return EOL;
+    case StreamTokenizer.TT_WORD:
+      yylval = new Integer(st.sval);
+      end.set(reader.getPreviousPosition());
+      return NUM;
+    case ' ': case '\t':
+      return yylex();
+    case '!':
+      return BANG;
+    case '+':
+      return PLUS;
+    case '-':
+      return MINUS;
+    case '*':
+      return STAR;
+    case '/':
+      return SLASH;
+    case '^':
+      return CARET;
+    case '(':
+      return LPAREN;
+    case ')':
+      return RPAREN;
+    case '=':
+      return EQUAL;
+    default:
+      throw new AssertionError("invalid character: " + ttype);
+    }
+  }
+}
+
+
+class Position {
+  public int line = 1;
+  public int column = 1;
+
+  public Position() {
+    line = 1;
+    column = 1;
+  }
+
+  public Position(int l, int t) {
+    line = l;
+    column = t;
+  }
+
+  public Position(Position p) {
+    line = p.line;
+    column = p.column;
+  }
+
+  public void set(Position p) {
+    line = p.line;
+    column = p.column;
+  }
+
+  public boolean equals(Position l) {
+    return l.line == line && l.column == column;
+  }
+
+  public String toString() {
+    return Integer.toString(line) + "." + Integer.toString(column);
+  }
+
+  public int line() {
+    return line;
+  }
+
+  public int column() {
+    return column;
+  }
+}
+
+class PositionReader extends BufferedReader {
+
+  private Position position = new Position();
+  // Position before the latest call to "read", i.e. position
+  // of the last character of the current token.
+  private Position previousPosition = new Position();
+
+  public PositionReader(Reader reader) {
+    super(reader);
+  }
+
+  public int read() throws IOException {
+    previousPosition.set(position);
+    int res = super.read();
+    if (res > -1) {
+      char c = (char) res;
+      if (c == '\r' || c == '\n') {
+        position.line += 1;
+        position.column = 1;
+      } else {
+        position.column += 1;
+      }
+    }
+    return res;
+  }
+
+  public Position getPosition() {
+    return position;
+  }
+
+  public Position getPreviousPosition() {
+    return previousPosition;
+  }
+}

examples/java/calc/local.mk

@@ -0,0 +1,36 @@
+## Copyright (C) 2018-2020 Free Software Foundation, Inc.
+##
+## This program is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+java_calcdir = $(docdir)/%D%
+
+## ------ ##
+## Calc.  ##
+## ------ ##
+
+if ENABLE_JAVA
+  check_SCRIPTS += %D%/Calc.class
+  TESTS += %D%/Calc.test
+endif
+EXTRA_DIST += %D%/Calc.test
+
+%D%/Calc.java: %D%/Calc.y $(dependencies)
+	$(AM_V_GEN)$(MKDIR_P) %D%
+	$(AM_V_at)$(BISON) $(srcdir)/%D%/Calc.y -o $@
+
+%D%/Calc.class: %D%/Calc.java
+	$(AM_V_GEN) $(SHELL) $(top_builddir)/javacomp.sh %D%/Calc.java
+
+dist_java_calc_DATA = %D%/Calc.y %D%/Makefile
+CLEANFILES += %D%/*.class %D%/Calc.java

examples/java/local.mk

@@ -1,4 +1,4 @@
-## Copyright (C) 2018-2020 Free Software Foundation, Inc.
+## Copyright (C) 2020 Free Software Foundation, Inc.
 ##
 ## This program is free software: you can redistribute it and/or modify
 ## it under the terms of the GNU General Public License as published by
@@ -14,23 +14,7 @@
 ## along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 javadir = $(docdir)/%D%
+dist_java_DATA = %D%/README.md
 
-## ------ ##
-## Calc.  ##
-## ------ ##
-
-if ENABLE_JAVA
-  check_SCRIPTS += %D%/Calc.class
-  TESTS += %D%/Calc.test
-endif
-EXTRA_DIST += %D%/Calc.test
-
-%D%/Calc.java: %D%/Calc.y $(dependencies)
-	$(AM_V_GEN)$(MKDIR_P) %D%
-	$(AM_V_at)$(BISON) $(srcdir)/%D%/Calc.y -o $@
-
-%D%/Calc.class: %D%/Calc.java
-	$(AM_V_GEN) $(SHELL) $(top_builddir)/javacomp.sh %D%/Calc.java
-
-dist_java_DATA = %D%/Calc.y %D%/Makefile %D%/README.md
-CLEANFILES += %D%/*.class %D%/Calc.java
+include %D%/calc/local.mk
+include %D%/simple/local.mk

examples/java/simple/Calc.test

@@ -20,7 +20,14 @@ cat >input <<EOF
 EOF
 run 0 7
 
+cat >input <<EOF
+1 + 2 * 3 = 7
+(1 + 2) * 3 = 9
+EOF
+run 0 '7
+9'
+
 cat >input <<EOF
 1 + 2 * * 3
 EOF
-run 0 "err: 1.5: syntax error, unexpected '*', expecting number or '-' or '(' or '!'"
+run 0 "err: syntax error, unexpected '*', expecting number or '-' or '(' or '!'"

examples/java/simple/Calc.y

@@ -0,0 +1,106 @@
+%language "Java"
+
+%define api.parser.class {Calc}
+%define api.parser.public
+
+%define parse.error verbose
+
+%code imports {
+  import java.io.IOException;
+  import java.io.InputStream;
+  import java.io.InputStreamReader;
+  import java.io.Reader;
+  import java.io.StreamTokenizer;
+}
+
+%code {
+  public static void main(String[] args) throws IOException {
+    CalcLexer l = new CalcLexer(System.in);
+    Calc p = new Calc(l);
+    if (!p.parse())
+      System.exit(1);
+  }
+}
+
+/* Bison Declarations */
+%token <Integer> NUM "number"
+%type  <Integer> exp
+
+%nonassoc '='       /* comparison            */
+%left '-' '+'
+%left '*' '/'
+%precedence NEG     /* negation--unary minus */
+%right '^'          /* exponentiation        */
+
+/* Grammar follows */
+%%
+input:
+  line
+| input line
+;
+
+line:
+  '\n'
+| exp '\n'           { System.out.println($exp); }
+| error '\n'
+;
+
+exp:
+  NUM                { $$ = $1; }
+| exp '=' exp
+  {
+    if ($1.intValue() != $3.intValue())
+      yyerror("calc: error: " + $1 + " != " + $3);
+  }
+| exp '+' exp        { $$ = $1 + $3;  }
+| exp '-' exp        { $$ = $1 - $3;  }
+| exp '*' exp        { $$ = $1 * $3;  }
+| exp '/' exp        { $$ = $1 / $3;  }
+| '-' exp  %prec NEG { $$ = -$2; }
+| exp '^' exp        { $$ = (int) Math.pow($1, $3); }
+| '(' exp ')'        { $$ = $2; }
+| '(' error ')'      { $$ = 1111; }
+| '!'                { $$ = 0; return YYERROR; }
+| '-' error          { $$ = 0; return YYERROR; }
+;
+
+
+%%
+class CalcLexer implements Calc.Lexer {
+
+  StreamTokenizer st;
+
+  public CalcLexer(InputStream is) {
+    st = new StreamTokenizer(new InputStreamReader(is));
+    st.resetSyntax();
+    st.eolIsSignificant(true);
+    st.whitespaceChars('\t', '\t');
+    st.whitespaceChars(' ', ' ');
+    st.wordChars('0', '9');
+  }
+
+  public void yyerror(String s) {
+    System.err.println(s);
+  }
+
+  Integer yylval;
+
+  public Object getLVal() {
+    return yylval;
+  }
+
+  public int yylex() throws IOException {
+    int ttype = st.nextToken();
+    switch (ttype) {
+    case StreamTokenizer.TT_EOF:
+      return YYEOF;
+    case StreamTokenizer.TT_EOL:
+      return (int) '\n';
+    case StreamTokenizer.TT_WORD:
+      yylval = new Integer(st.sval);
+      return NUM;
+    default:
+      return ttype;
+    }
+  }
+}

examples/java/simple/Makefile

@@ -0,0 +1,26 @@
+# This Makefile is designed to be simple and readable.  It does not
+# aim at portability.  It requires GNU Make.
+
+BISON = bison
+JAVAC = javac
+JAVA = java
+XSLTPROC = xsltproc
+
+all: Calc.class
+
+%.java %.xml %.gv: %.y
+	$(BISON) $(BISONFLAGS) --xml --graph=$*.gv -o $*.java $<
+
+%.class: %.java
+	$(JAVAC) $(JAVACFLAGS) $<
+
+run: Calc.class
+	@echo "Type arithmetic expressions.  Quit with ctrl-d."
+	$(JAVA) $(JAVAFLAGS) Calc
+
+html: Calc.html
+%.html: %.xml
+	$(XSLTPROC) $(XSLTPROCFLAGS) -o $@ $$($(BISON) --print-datadir)/xslt/xml2xhtml.xsl $<
+
+clean:
+	rm -f *.class Calc.java Calc.html Calc.xml Calc.gv

examples/java/simple/local.mk

@@ -0,0 +1,36 @@
+## Copyright (C) 2018-2020 Free Software Foundation, Inc.
+##
+## This program is free software: you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation, either version 3 of the License, or
+## (at your option) any later version.
+##
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+## GNU General Public License for more details.
+##
+## You should have received a copy of the GNU General Public License
+## along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+java_simpledir = $(docdir)/%D%
+
+## ------ ##
+## Calc.  ##
+## ------ ##
+
+if ENABLE_JAVA
+  check_SCRIPTS += %D%/Calc.class
+  TESTS += %D%/Calc.test
+endif
+EXTRA_DIST += %D%/Calc.test
+
+%D%/Calc.java: %D%/Calc.y $(dependencies)
+	$(AM_V_GEN)$(MKDIR_P) %D%
+	$(AM_V_at)$(BISON) $(srcdir)/%D%/Calc.y -o $@
+
+%D%/Calc.class: %D%/Calc.java
+	$(AM_V_GEN) $(SHELL) $(top_builddir)/javacomp.sh %D%/Calc.java
+
+dist_java_simple_DATA = %D%/Calc.y %D%/Makefile
+CLEANFILES += %D%/*.class %D%/Calc.java

examples/test

@@ -30,19 +30,23 @@ exit=true
 # top_builddir.
 cwd=$(pwd)
 
+# Whether to strip '> ...' lines from the expected output.
+# See bistromathic.test.
+strip_prompt=false
+
 # The exercised program.
-for p in "$cwd/examples/$medir/$me"
-do
-    if test -x "$p"; then
-        prog=$p
-        break
-    elif test -f "$p.class"; then
-        pwd
-        prog="$SHELL $cwd/javaexec.sh -cp $(dirname $p) $(basename $p)"
-        break
-    fi
-done
-if test x"$prog" = x; then
+abs_medir=$cwd/examples/$medir
+if test -x "$abs_medir/$me"; then
+    prog ()
+    {
+        "$abs_medir/$me" "$@"
+    }
+elif test -f "$abs_medir/$me.class"; then
+    prog ()
+    {
+        "$SHELL" "$cwd/javaexec.sh" -cp "$abs_medir" "$me" "$@"
+    }
+else
     echo "$me: ERROR: cannot find program to exercise in:"
     echo "$me: ERROR:   $cwd/examples/$medir/$me"
     exit 1
@@ -55,7 +59,7 @@ cleanup ()
 {
   status=$?
   if test -z "$DEBUG"; then
-     cd $cwd
+     cd "$cwd"
      rm -rf $$.dir
   fi
   exit $status
@@ -64,33 +68,65 @@ trap cleanup 0 1 2 13 15
 mkdir $$.dir
 cd $$.dir
 
-# run [-noerr] EXPECTED-EXIT-STATUS EXPECTED-OUTPUT [PARSER-OPTIONS]
-# ------------------------------------------------------------------
+# run [-noerr, -n] EXPECTED-EXIT-STATUS EXPECTED-OUTPUT [PARSER-OPTIONS]
+# ----------------------------------------------------------------------
 # -noerr: ignore stderr, otherwise merge it into effective output.
+# -n: not final end-of-line in expected-output
 run ()
 {
   noerr=false
-  case $1 in
-    (-noerr) noerr=true; shift;;
-  esac
+  echo=echo
+  while true; do
+    case $1 in
+      (-noerr) noerr=true; shift;;
+      (-n)     echo=printf; shift;;
+      (*)      break;;
+    esac
+  done
 
   # Expected exit status.
   sta_exp=$1
   shift
+
   # Expected output.
-  out_exp=$1
+  $echo "$1" |
+    sed -e 's/Reducing stack by rule .* (line .*):/Reducing stack by rule XX (line XXX):/g' |
+    if $strip_prompt; then
+      # An extra EOL added in bistromathic's main.  Keep that empty line.
+      sed -e '/^> ./d;s/^> $//g'
+    else
+      cat
+    fi >exp
   shift
+
   # Effective exit status.
   sta_eff=0
-  $prog "$@" - <input >out_eff 2>err_eff || sta_eff=$?
+
+  prog "$@" - <input >out_eff 2>err_eff || sta_eff=$?
+
   # Combine effective output and error streams.
-  out_eff=$(cat out_eff && $noerr || sed -e 's/^/err: /g' err_eff)
+  {
+    cat out_eff
+    if ! $noerr; then
+       sed -e 's/^/err: /g' \
+           -e 's/Reducing stack by rule .* (line .*):/Reducing stack by rule XX (line XXX):/g' \
+           err_eff
+    fi
+  } >eff
+
   if test $sta_eff -eq $sta_exp; then
-    if test "$out_eff" = "$out_exp"; then
+    if cmp eff exp 2>/dev/null; then
       echo "$me: PASS: $number"
     else
-      echo "$me: FAIL: $number (expected output: $out_exp, effective: $out_eff)"
-      cat err_eff
+      echo "$me: FAIL: $number"
+      echo "$me: input:"
+      sed -e 's/^/  /' input
+      echo "$me: expected output:"
+      sed -e 's/^/  /' exp
+      echo "$me: effective output:"
+      sed -e 's/^/  /' eff
+      echo "$me: diff:"
+      diff -u exp eff | sed -e 's/^/  /'
       exit=false
     fi
   else

gnulib-po/.gitignore

@@ -1,3 +1,4 @@
+/remove-potcdate.sin~
 /*.gmo
 /*.po
 /LINGUAS

lib/.gitignore

@@ -58,7 +58,6 @@
 /dirname-lgpl.c
 /dirname.c
 /dirname.h
-/dosname.h
 /dup-safer-flag.c
 /dup-safer.c
 /dup2.c
@@ -96,9 +95,11 @@
 /fstrcmp.c
 /fstrcmp.h
 /fsync.c
+/getdelim.c
 /getdtablesize.c
 /gethrxtime.c
 /gethrxtime.h
+/getline.c
 /getopt-cdefs.h
 /getopt-cdefs.in.h
 /getopt-core.h
@@ -138,6 +139,8 @@
 /isnanl-nolibm.h
 /isnanl.c
 /iswblank.c
+/iswdigit.c
+/iswxdigit.c
 /itold.c
 /lc-charset-dispatch.c
 /lc-charset-dispatch.h
@@ -147,6 +150,7 @@
 /limits.in.h
 /localcharset.c
 /localcharset.h
+/locale.h
 /locale.in.h
 /localtime-buffer.c
 /localtime-buffer.h
@@ -203,6 +207,8 @@
 /raise.c
 /rawmemchr.c
 /rawmemchr.valgrind
+/readline.c
+/readline.h
 /readlink.c
 /realloc.c
 /ref-add.sed

m4/.gitignore

@@ -1,3 +1,4 @@
+/extern-inline.m4~
 /*~
 /00gnulib.m4
 /__inline.m4
@@ -38,8 +39,10 @@
 /fseterr.m4
 /fstat.m4
 /fsync.m4
+/getdelim.m4
 /getdtablesize.m4
 /gethrxtime.m4
+/getline.m4
 /getopt.m4
 /getprogname.m4
 /getrusage.m4
@@ -69,6 +72,8 @@
 /isnanf.m4
 /isnanl.m4
 /iswblank.m4
+/iswdigit.m4
+/iswxdigit.m4
 /javacomp.m4
 /javaexec.m4
 /jm-winsz1.m4
@@ -134,6 +139,7 @@
 /quotearg.m4
 /raise.m4
 /rawmemchr.m4
+/readline.m4
 /readlink.m4
 /realloc.m4
 /relocatable-lib.m4
@@ -210,3 +216,4 @@
 /xalloc.m4
 /xsize.m4
 /xstrndup.m4
+/zzgnulib.m4

po/POTFILES.in

@@ -1,3 +1,5 @@
+examples/c/bistromathic/parse.y
+examples/java/calc/Calc.y
 src/complain.c
 src/conflicts.c
 src/files.c
@@ -5,6 +7,7 @@ src/fixits.c
 src/getargs.c
 src/gram.c
 src/graphviz.c
+src/i18n-strings.c
 src/location.c
 src/main.c
 src/muscle-tab.c

src/complain.c

@@ -23,6 +23,7 @@
 #include "system.h"
 
 #include <argmatch.h>
+#include <ctype.h>
 #include <progname.h>
 #include <stdarg.h>
 #include <sys/stat.h>
@@ -41,9 +42,9 @@ bool warnings_are_errors = false;
 /** Whether -Werror/-Wno-error was applied to a warning.  */
 typedef enum
   {
-    errority_unset = 0,     /** No explict status.  */
-    errority_disabled = 1,  /** Explictly disabled with -Wno-error=foo.  */
-    errority_enabled = 2    /** Explictly enabled with -Werror=foo. */
+    errority_unset = 0,     /** No explicit status.  */
+    errority_disabled = 1,  /** Explicitly disabled with -Wno-error=foo.  */
+    errority_enabled = 2    /** Explicitly enabled with -Werror=foo. */
   } errority;
 
 /** For each warning type, its errority.  */
@@ -263,6 +264,20 @@ severity_prefix (severity s)
 }
 
 
+static void
+severity_print (severity s, FILE *out)
+{
+  if (s != severity_disabled)
+    {
+      const char* style = severity_style (s);
+      begin_use_class (style, out);
+      fprintf (out, "%s:", severity_prefix (s));
+      end_use_class (style, out);
+      fputc (' ', out);
+    }
+}
+
+
 /*-----------.
 | complain.  |
 `-----------*/
@@ -409,7 +424,6 @@ warnings_print_categories (warnings warn_flags, FILE *out)
  *
  * \param loc     the location, defaulting to the current file,
  *                or the program name.
- * \param indent  optional indentation for the error message.
  * \param flags   the category for this message.
  * \param sever   to decide the prefix to put before the message
  *                (e.g., "warning").
@@ -421,33 +435,21 @@ warnings_print_categories (warnings warn_flags, FILE *out)
  */
 static
 void
-error_message (const location *loc, int *indent, warnings flags,
+error_message (const location *loc, warnings flags,
                severity sever, const char *message, va_list args)
 {
-  int pos = 0;
+  const char* style = flags & note ? "note" : severity_style (sever);
 
   if (loc)
-    pos += location_print (*loc, stderr);
+    location_print (*loc, stderr);
   else
-    pos += fprintf (stderr, "%s", grammar_file ? grammar_file : program_name);
-  pos += fprintf (stderr, ": ");
-
-  if (indent)
-    {
-      if (*indent)
-        sever = severity_disabled;
-      if (!*indent)
-        *indent = pos;
-      else if (*indent > pos)
-        fprintf (stderr, "%*s", *indent - pos, "");
-    }
-
-  const char* style = severity_style (sever);
+    fprintf (stderr, "%s", grammar_file ? grammar_file : program_name);
+  fprintf (stderr, ": ");
 
   if (sever != severity_disabled)
     {
       begin_use_class (style, stderr);
-      fprintf (stderr, "%s:", severity_prefix (sever));
+      fprintf (stderr, "%s:", flags & note ? _("note") : severity_prefix (sever));
       end_use_class (style, stderr);
       fputc (' ', stderr);
     }
@@ -458,34 +460,32 @@ error_message (const location *loc, int *indent, warnings flags,
   if (! (flags & silent) && sever != severity_disabled)
     warnings_print_categories (flags, stderr);
 
-  {
-    size_t l = strlen (message);
-    if (l < 2 || message[l - 2] != ':' || message[l - 1] != ' ')
-      {
-        putc ('\n', stderr);
-        flush (stderr);
-        if (loc && feature_flag & feature_caret && !(flags & no_caret))
-          location_caret (*loc, style, stderr);
-      }
-  }
+  size_t l = strlen (message);
+  if (l < 2 || message[l - 2] != ':' || message[l - 1] != ' ')
+    {
+      putc ('\n', stderr);
+      flush (stderr);
+      if (loc && !(flags & no_caret))
+        location_caret (*loc, style, stderr);
+    }
   flush (stderr);
 }
 
 /** Raise a complaint (fatal error, error or just warning).  */
 
 static void
-complains (const location *loc, int *indent, warnings flags,
+complains (const location *loc, warnings flags,
            const char *message, va_list args)
 {
-  severity s = warning_severity (flags);
   if ((flags & complaint) && complaint_status < status_complaint)
     complaint_status = status_complaint;
 
+  severity s = warning_severity (flags);
   if (severity_warning <= s)
     {
       if (severity_error <= s && ! complaint_status)
         complaint_status = status_warning_as_error;
-      error_message (loc, indent, flags, s, message, args);
+      error_message (loc, flags, s, message, args);
     }
 
   if (flags & fatal)
@@ -497,41 +497,39 @@ complain (location const *loc, warnings flags, const char *message, ...)
 {
   va_list args;
   va_start (args, message);
-  complains (loc, NULL, flags, message, args);
+  complains (loc, flags, message, args);
   va_end (args);
 }
 
 void
-complain_indent (location const *loc, warnings flags, int *indent,
-                 const char *message, ...)
+subcomplain (location const *loc, warnings flags, const char *message, ...)
 {
   va_list args;
   va_start (args, message);
-  complains (loc, indent, flags, message, args);
+  complains (loc, flags | note | silent, message, args);
   va_end (args);
 }
 
 void
-complain_args (location const *loc, warnings w, int *indent,
+complain_args (location const *loc, warnings w,
                int argc, char *argv[])
 {
   switch (argc)
   {
   case 1:
-    complain_indent (loc, w, indent, "%s", _(argv[0]));
+    complain (loc, w, "%s", _(argv[0]));
     break;
   case 2:
-    complain_indent (loc, w, indent, _(argv[0]), argv[1]);
+    complain (loc, w, _(argv[0]), argv[1]);
     break;
   case 3:
-    complain_indent (loc, w, indent, _(argv[0]), argv[1], argv[2]);
+    complain (loc, w, _(argv[0]), argv[1], argv[2]);
     break;
   case 4:
-    complain_indent (loc, w, indent, _(argv[0]), argv[1], argv[2], argv[3]);
+    complain (loc, w, _(argv[0]), argv[1], argv[2], argv[3]);
     break;
   case 5:
-    complain_indent (loc, w, indent, _(argv[0]), argv[1], argv[2], argv[3],
-                     argv[4]);
+    complain (loc, w, _(argv[0]), argv[1], argv[2], argv[3], argv[4]);
     break;
   default:
     complain (loc, fatal, "too many arguments for complains");
@@ -555,8 +553,7 @@ deprecated_directive (location const *loc, char const *old, char const *upd)
       complain (loc, Wdeprecated,
                 _("deprecated directive: %s, use %s"),
                 quote (old), quote_n (1, upd));
-      if (feature_flag & feature_caret)
-        location_caret_suggestion (*loc, upd, stderr);
+      location_caret_suggestion (*loc, upd, stderr);
       /* Register updates only if -Wdeprecated is enabled.  */
       fixits_register (loc, upd);
     }
@@ -566,13 +563,11 @@ void
 duplicate_directive (char const *directive,
                      location first, location second)
 {
-  int i = 0;
   if (feature_flag & feature_caret)
-    complain_indent (&second, Wother, &i, _("duplicate directive"));
+    complain (&second, Wother, _("duplicate directive"));
   else
-    complain_indent (&second, Wother, &i, _("duplicate directive: %s"), quote (directive));
-  i += SUB_INDENT;
-  complain_indent (&first, Wother, &i, _("previous declaration"));
+    complain (&second, Wother, _("duplicate directive: %s"), quote (directive));
+  subcomplain (&first, Wother, _("previous declaration"));
   fixits_register (&second, "");
 }
 
@@ -580,11 +575,57 @@ void
 duplicate_rule_directive (char const *directive,
                           location first, location second)
 {
-  int i = 0;
-  complain_indent (&second, complaint, &i,
-                   _("only one %s allowed per rule"), directive);
-  i += SUB_INDENT;
-  complain_indent (&first, complaint, &i,
-                   _("previous declaration"));
+  complain (&second, complaint, _("only one %s allowed per rule"), directive);
+  subcomplain (&first, complaint, _("previous declaration"));
   fixits_register (&second, "");
 }
+
+void
+syntax_error (location loc,
+              int argc, const char* argv[])
+{
+  if (complaint_status < status_complaint)
+    complaint_status = status_complaint;
+  assert (argc <= 5);
+  const char *format = NULL;
+  switch (argc)
+    {
+#define CASE(N, S)                          \
+      case N:                               \
+        format = S;                         \
+        break
+    default: /* Avoid compiler warnings. */
+      CASE (0, _("syntax error"));
+      CASE (1, _("unexpected %0$s"));
+      CASE (2, _("expected %1$s before %0$s"));
+      CASE (3, _("expected %1$s or %2$s before %0$s"));
+      CASE (4, _("expected %1$s or %2$s or %3$s before %0$s"));
+      CASE (5, _("expected %1$s or %2$s or %3$s or %4$s before %0$s"));
+#undef CASE
+    }
+  location_print (loc, stderr);
+  fputs (": ", stderr);
+  severity_print (severity_error, stderr);
+
+  while (*format)
+    if (format[0] == '%'
+        && isdigit (format[1])
+        && format[2] == '$'
+        && format[3] == 's'
+        && (format[1] - '0') < argc)
+      {
+        int i = format[1] - '0';
+        const char *style = i == 0 ? "unexpected" : "expected";
+        begin_use_class (style, stderr);
+        fputs (argv[i], stderr);
+        end_use_class (style, stderr);
+        format += 4;
+      }
+    else
+      {
+        fputc (*format, stderr);
+        ++format;
+      }
+  fputc ('\n', stderr);
+  location_caret (loc, "error", stderr);
+}

src/complain.h

@@ -21,9 +21,6 @@
 
 # include "location.h"
 
-/* Sub-messages indent. */
-# define SUB_INDENT (4)
-
 /*---------------.
 | Error stream.  |
 `---------------*/
@@ -119,6 +116,7 @@ typedef enum
     fatal             = 1 << 12, /**< All fatal errors.  */
     silent            = 1 << 13, /**< Do not display the warning type.  */
     no_caret          = 1 << 14, /**< Do not display caret location.  */
+    note              = 1 << 15, /**< Display as a note.  */
 
     /**< All above warnings.  */
     Weverything       = ~complaint & ~fatal & ~silent,
@@ -137,13 +135,13 @@ void complain (location const *loc, warnings flags, char const *message, ...)
   __attribute__ ((__format__ (__printf__, 3, 4)));
 
 /** Likewise, but with an \a argc/argv interface.  */
-void complain_args (location const *loc, warnings w, int *indent,
+void complain_args (location const *loc, warnings w,
                     int argc, char *arg[]);
 
-/** Make a complaint with location and some indentation.  */
-void complain_indent (location const *loc, warnings flags, int *indent,
-                      char const *message, ...)
-  __attribute__ ((__format__ (__printf__, 4, 5)));
+/** Make a subcomplain with location and note.  */
+void subcomplain (location const *loc, warnings flags,
+                  char const *message, ...)
+  __attribute__ ((__format__ (__printf__, 3, 4)));
 
 
 /** GNU Bison extension not valid with POSIX Yacc.  */
@@ -161,6 +159,11 @@ void duplicate_directive (char const *directive,
 void duplicate_rule_directive (char const *directive,
                                location first, location second);
 
+/** Report a syntax error, where argv[0] is the unexpected
+    token, and argv[1...argc] are the expected ones.  */
+void syntax_error (location loc,
+                   int argc, const char* argv[]);
+
 /** Warnings treated as errors shouldn't stop the execution as regular
     errors should (because due to their nature, it is safe to go
     on). Thus, there are three possible execution statuses.  */

src/gram.h

@@ -42,9 +42,9 @@
 
    Internally, we cannot use the number 0 for a rule because for
    instance RITEM stores both symbol (the RHS) and rule numbers: the
-   symbols are shorts >= 0, and rule number are stored negative.
+   symbols are integers >= 0, and rule numbers are stored negative.
    Therefore 0 cannot be used, since it would be both the rule number
-   0, and the token $end).
+   0, and the token $end.
 
    Actions are accessed via the rule number.
 
@@ -55,8 +55,7 @@
 
    RULES[R].lhs -- the symbol of the left hand side of rule R.
 
-   RULES[R].rhs -- the index in RITEM of the beginning of the portion
-   for rule R.
+   RULES[R].rhs -- the beginning of the portion of RITEM for rule R.
 
    RULES[R].prec -- the symbol providing the precedence level of R.
 
@@ -75,16 +74,16 @@
 
    RULES[R].line -- the line where R was defined.
 
-   RULES[R].useful -- whether the rule is used (i.e., false if thrown
-   away by reduce).
+   RULES[R].useful -- whether the rule is used.  False if thrown away
+   by reduce().
 
    The right hand side is stored as symbol numbers in a portion of
    RITEM.
 
    The length of the portion is one greater than the number of symbols
    in the rule's right hand side.  The last element in the portion
-   contains minus R, which identifies it as the end of a portion and
-   says which rule it is for.
+   contains -R, which identifies it as the end of a portion and says
+   which rule it is for.
 
    The portions of RITEM come in order of increasing rule number.
    NRITEMS is the total length of RITEM.  Each element of RITEM is

src/i18n-strings.c

@@ -0,0 +1,37 @@
+/* This file is not compiled in, it is used only to expose more
+   strings to gettextize.
+
+   Why is this needed?
+
+   Bison emits strings to translate in the generated code, for builtin
+   tokens.  So they appear only in generated parsers, which are not
+   shipped, so they are not in the src tree, so we cannot use them in
+   our POTFILE.
+
+   Except src/parse-gram.c, which is in the source tree.  And even in
+   the git repo.  But to avoid useless diffs in the repo, we do not
+   keep the #line directives in the src tree.  Yet, for the user, we
+   ship a src/parse-gram.c _with_ the #lines.  This is done in a
+   dist-hook which regenerates src/parse-gram.c when we run "make
+   dist".
+
+   Unfortunately, then, update-po traverses the whole tree and sees
+   that the location of the strings to translate in src/parse-gram.c
+   have changed, so the bison.pot is to be updated.  And that is not
+   possible in the "make dist" which is run within "make distcheck"
+   (not the one preparing the dist for distcheck, the one run by
+   distcheck to check that a distributed tarball can build a tarball)
+   because then the src tree is read-only.
+
+   So let's not put src/parse-gram.c in the POTFILE, and expose these
+   strings to gettextize by hand.
+  */
+
+// Please syntax-check.
+#include <config.h>
+
+static const char *const msgid[] =
+  {
+    N_("end of file"),
+    N_("invalid token")
+  }

src/local.mk

@@ -130,6 +130,7 @@ src_bison_LDADD =                               \
   lib/libbison.a
 
 
+EXTRA_DIST += %D%/i18n-strings.c
 
 
 ## ------ ##

src/location.c

@@ -175,6 +175,8 @@ location_print (location loc, FILE *out)
     }
   else
     {
+      aver (loc.start.file);
+      aver (loc.end.file);
       int end_col = 0 != loc.end.column ? loc.end.column - 1 : 0;
       res += fprintf (out, "%s",
                       quotearg_n_style (3, escape_quoting_style, loc.start.file));
@@ -317,7 +319,7 @@ caret_getc_internal (mbchar_t *res)
 
 /* Move CARET_INFO (which has a valid FILE) to the line number LINE.
    Compute and cache that line's length in CARET_INFO.LINE_LEN.
-   Return whether succesful.*/
+   Return whether successful.  */
 static bool
 caret_set_line (int line)
 {
@@ -400,6 +402,8 @@ caret_set_column (int col)
 void
 location_caret (location loc, const char *style, FILE *out)
 {
+  if (!(feature_flag & feature_caret))
+    return;
   if (!loc.start.line)
     return;
   if (!caret_set_file (loc.start.file))
@@ -421,12 +425,14 @@ location_caret (location loc, const char *style, FILE *out)
       {
         /* The last column to highlight.  Only the first line of
            multiline locations are quoted, in which case the ending
-           column is the end of line.  Single point locations (with
-           equal boundaries) denote the character that they
-           follow.  */
-        int col_end
+           column is the end of line.
+
+           We used to work with byte offsets, and that was much
+           easier.  However, we went back to using (visual) columns to
+           support truncating of long lines.  */
+        const int col_end
           = loc.start.line == loc.end.line
-          ? loc.end.column + (loc.start.column == loc.end.column)
+          ? loc.end.column
           : caret_info.line_len;
         /* Quote the file (at most the first line in the case of
            multiline locations).  */
@@ -436,24 +442,28 @@ location_caret (location loc, const char *style, FILE *out)
              expected (maybe the file was changed since the scanner
              ran), we might reach the end before we actually saw the
              opening column.  */
-          bool opened = false;
+          enum { before, inside, after } state = before;
           while (!mb_iseof (c) && !mb_iseq (c, '\n'))
             {
-              if (caret_info.pos.column == loc.start.column)
+              // We might have already opened (and even closed!) the
+              // style and yet have the equality of the columns if we
+              // just saw zero-width characters.
+              if (state == before
+                  && caret_info.pos.column == loc.start.column)
                 {
                   begin_use_class (style, out);
-                  opened = true;
+                  state = inside;
                 }
               if (skip < caret_info.pos.column)
                 mb_putc (c, out);
               boundary_compute (&caret_info.pos, mb_ptr (c), mb_len (c));
               caret_getc (c);
-              if (opened
+              if (state == inside
                   && (caret_info.pos.column == col_end
                       || width < caret_info.pos.column - skip))
                 {
                   end_use_class (style, out);
-                  opened = false;
+                  state = after;
                 }
               if (width < caret_info.pos.column - skip)
                 {
@@ -461,6 +471,12 @@ location_caret (location loc, const char *style, FILE *out)
                   break;
                 }
             }
+          if (state == inside)
+            {
+              // The line is shorter than expected.
+              end_use_class (style, out);
+              state = after;
+            }
           putc ('\n', out);
         }
 
@@ -486,6 +502,8 @@ location_caret (location loc, const char *style, FILE *out)
 void
 location_caret_suggestion (location loc, const char *s, FILE *out)
 {
+  if (!(feature_flag & feature_caret))
+    return;
   const char *style = "fixit-insert";
   fprintf (out, "      | %*s",
            loc.start.column - 1 - caret_info.skip

src/location.h

@@ -42,16 +42,14 @@ typedef struct
 
   /* If positive, the column (starting at 1) just after the boundary.
      This is neither a byte count, nor a character count; it is a
-     column count.  If this is INT_MAX, the column number has
+     (visual) column count.  If this is INT_MAX, the column number has
      overflowed.
 
-     Meaningless and not displayed if nonpositive.
-  */
+     Meaningless and not displayed if nonpositive.  */
   int column;
 
-  /* If nonnegative, the byte number (starting at 0) in the current line.
-     Never displayed, used when printing error messages with colors to
-     know where colors start and end.  */
+  /* If nonnegative, the byte number (starting at 0) in the current
+     line.  Not displayed (unless --trace=location).  */
   int byte;
 
 } boundary;
@@ -126,12 +124,12 @@ void caret_init (void);
    left-over by the usage of location_caret.  */
 void caret_free (void);
 
-/* Quote the line containing LOC onto OUT.  Highlight the part of LOC
-   with the color STYLE.  */
+/* If -fcaret is enabled, quote the line containing LOC onto OUT.
+   Highlight the part of LOC with the color STYLE.  */
 void location_caret (location loc, const char* style, FILE *out);
 
-/* Display a suggestion of replacement for LOC with S.  To call after
-   location_caret.  */
+/* If -fcaret is enabled, display a suggestion of replacement for LOC
+   with S.  To call after location_caret.  */
 void location_caret_suggestion (location loc, const char *s, FILE *out);
 
 /* Return -1, 0, 1, depending whether a is before, equal, or

src/lr0.c

@@ -58,10 +58,11 @@ core_print (size_t core_size, item_number *core, FILE *out)
     }
 }
 
-/*------------------------------------------------------------------.
-| A state was just discovered from another state.  Queue it for     |
-| later examination, in order to find its transitions.  Return it.  |
-`------------------------------------------------------------------*/
+/*-----------------------------------------------------------------.
+| A state was just discovered by transitioning on SYM from another |
+| state.  Queue this state for later examination, in order to find |
+| its outgoing transitions.  Return it.                            |
+`-----------------------------------------------------------------*/
 
 static state *
 state_list_append (symbol_number sym, size_t core_size, item_number *core)
@@ -98,7 +99,7 @@ static state **shiftset;
 
 
 /* KERNEL_BASE[symbol-number] -> list of item numbers (offsets inside
-   RITEM) of lenngth KERNEL_SIZE[symbol-number]. */
+   RITEM) of length KERNEL_SIZE[symbol-number]. */
 static item_number **kernel_base;
 static int *kernel_size;
 

src/lr0.h

@@ -21,8 +21,6 @@
 #ifndef LR0_H_
 # define LR0_H_
 
-# include "state.h"
-
 void generate_states (void);
 
 #endif /* !LR0_H_ */

src/muscle-tab.c

@@ -292,7 +292,6 @@ muscle_location_grow (char const *key, location loc)
 
 #define COMMON_DECODE(Value)                                    \
   case '$':                                                     \
-    ++(Value); aver (*(Value) == '[');                          \
     ++(Value); aver (*(Value) == ']');                          \
     ++(Value); aver (*(Value) == '[');                          \
     obstack_sgrow (&muscle_obstack, "$");                       \
@@ -525,15 +524,13 @@ muscle_percent_define_insert (char const *var, location variable_loc,
             = atoi (muscle_find_const (how_name));
           if (how_old == MUSCLE_PERCENT_DEFINE_F)
             goto end;
-          int i = 0;
           /* If assigning the same value, make it a warning.  */
           warnings warn = STREQ (value, current_value) ? Wother : complaint;
-          complain_indent (&variable_loc, warn, &i,
-                           _("%%define variable %s redefined"),
-                           quote (variable));
-          i += SUB_INDENT;
+          complain (&variable_loc, warn,
+                    _("%%define variable %s redefined"),
+                    quote (variable));
           location loc = muscle_percent_define_get_loc (variable);
-          complain_indent (&loc, warn, &i, _("previous definition"));
+          subcomplain (&loc, warn, _("previous definition"));
           fixits_register (&variable_loc, "");
           warned = true;
         }
@@ -739,14 +736,12 @@ muscle_percent_define_check_values (char const * const *values)
           if (!*values)
             {
               location loc = muscle_percent_define_get_loc (*variablep);
-              int i = 0;
-              complain_indent (&loc, complaint, &i,
-                               _("invalid value for %%define variable %s: %s"),
-                               quote (*variablep), quote_n (1, value));
-              i += SUB_INDENT;
+              complain (&loc, complaint,
+                        _("invalid value for %%define variable %s: %s"),
+                        quote (*variablep), quote_n (1, value));
               for (values = variablep + 1; *values; ++values)
-                complain_indent (&loc, complaint | no_caret | silent, &i,
-                                 _("accepted value: %s"), quote (*values));
+                subcomplain (&loc, complaint | no_caret | silent,
+                             _("accepted value: %s"), quote (*values));
             }
           else
             while (*values)

src/output.c

@@ -50,6 +50,10 @@ static struct obstack format_obstack;
 | result of formatting the FIRST and then TABLE_DATA[BEGIN..END[ (of |
 | TYPE), and to the muscle NAME_max, the max value of the            |
 | TABLE_DATA.                                                        |
+|                                                                    |
+| For the typical case of outputting a complete table from 0, pass   |
+| TABLE[0] as FIRST, and 1 as BEGIN.  For instance                   |
+| muscle_insert_base_table ("pact", base, base[0], 1, nstates);      |
 `-------------------------------------------------------------------*/
 
 
@@ -132,42 +136,130 @@ string_output (FILE *out, char const *string)
 }
 
 
+/* Store in BUFFER a copy of SRC where trigraphs are escaped, return
+   the size of the result (including the final NUL).  If called with
+   BUFFERSIZE = 0, returns the needed size for BUFFER.  */
+static ptrdiff_t
+escape_trigraphs (char *buffer, ptrdiff_t buffersize, const char *src)
+{
+#define STORE(c)                                \
+  do                                            \
+    {                                           \
+      if (res < buffersize)                     \
+        buffer[res] = (c);                      \
+      ++res;                                    \
+    }                                           \
+  while (0)
+  ptrdiff_t res = 0;
+  for (ptrdiff_t i = 0, len = strlen (src); i < len; ++i)
+    {
+      if (i + 2 < len
+          && src[i] == '?' && src[i+1] == '?')
+        {
+          switch (src[i+2])
+            {
+            case '!': case '\'':
+            case '(': case ')': case '-': case '/':
+            case '<': case '=': case '>':
+              i += 1;
+              STORE ('?');
+              STORE ('"');
+              STORE ('"');
+              STORE ('?');
+              continue;
+            }
+        }
+      STORE (src[i]);
+    }
+  STORE ('\0');
+#undef STORE
+  return res;
+}
+
+/* Same as xstrdup, except that trigraphs are escaped.  */
+static char *
+xescape_trigraphs (const char *src)
+{
+  ptrdiff_t bufsize = escape_trigraphs (NULL, 0, src);
+  char *buf = xcharalloc (bufsize);
+  escape_trigraphs (buf, bufsize, src);
+  return buf;
+}
+
+/* The tag to show in the generated parsers.  Use "end of file" rather
+   than "$end".  But keep "$end" in the reports, it's shorter and more
+   consistent.  Support i18n if the user already uses it.  */
+static const char *
+symbol_tag (const symbol *sym)
+{
+  const bool eof_is_user_defined
+    = !endtoken->alias || STRNEQ (endtoken->alias->tag, "$end");
+
+  if (!eof_is_user_defined && sym->content == endtoken->content)
+    return "\"end of file\"";
+  else if (sym->content == undeftoken->content)
+    return "\"invalid token\"";
+  else
+    return sym->tag;
+}
+
 /* Generate the b4_<MUSCLE_NAME> (e.g., b4_tname) table with the
    symbol names (aka tags). */
 
 static void
 prepare_symbol_names (char const *muscle_name)
 {
+  // Whether to add a pair of quotes around the name.
+  const bool quote = STREQ (muscle_name, "tname");
+  bool has_translations = false;
+
   /* We assume that the table will be output starting at column 2. */
-  int j = 2;
+  int col = 2;
   struct quoting_options *qo = clone_quoting_options (0);
   set_quoting_style (qo, c_quoting_style);
   set_quoting_flags (qo, QA_SPLIT_TRIGRAPHS);
   for (int i = 0; i < nsyms; i++)
     {
-      char *cp = quotearg_alloc (symbols[i]->tag, -1, qo);
+      const char *tag = symbol_tag (symbols[i]);
+      bool translatable = !quote && symbols[i]->translatable;
+      if (translatable)
+        has_translations = true;
+
+      char *cp
+        = tag[0] == '"' && !quote
+        ? xescape_trigraphs (tag)
+        : quotearg_alloc (tag, -1, qo);
       /* Width of the next token, including the two quotes, the
          comma and the space.  */
-      int width = strlen (cp) + 2;
+      int width
+        = strlen (cp) + 2
+        + (translatable ? strlen ("N_()") : 0);
 
-      if (j + width > 75)
+      if (col + width > 75)
         {
           obstack_sgrow (&format_obstack, "\n ");
-          j = 1;
+          col = 1;
         }
 
       if (i)
         obstack_1grow (&format_obstack, ' ');
+      if (translatable)
+        obstack_sgrow (&format_obstack, "]b4_symbol_translate([");
       obstack_escape (&format_obstack, cp);
+      if (translatable)
+        obstack_sgrow (&format_obstack, "])[");
       free (cp);
       obstack_1grow (&format_obstack, ',');
-      j += width;
+      col += width;
     }
   free (qo);
   obstack_sgrow (&format_obstack, " ]b4_null[");
 
   /* Finish table and store. */
   muscle_insert (muscle_name, obstack_finish0 (&format_obstack));
+
+  /* Announce whether translation support is needed.  */
+  MUSCLE_INSERT_BOOL ("has_translations_flag", has_translations);
 }
 
 
@@ -182,7 +274,6 @@ prepare_symbols (void)
   MUSCLE_INSERT_INT ("tokens_number", ntokens);
   MUSCLE_INSERT_INT ("nterms_number", nvars);
   MUSCLE_INSERT_INT ("symbols_number", nsyms);
-  MUSCLE_INSERT_INT ("undef_token_number", undeftoken->content->number);
   MUSCLE_INSERT_INT ("user_token_number_max", max_user_token_number);
 
   muscle_insert_symbol_number_table ("translate",
@@ -192,6 +283,27 @@ prepare_symbols (void)
 
   /* tname -- token names.  */
   prepare_symbol_names ("tname");
+  prepare_symbol_names ("symbol_names");
+
+  /* translatable -- whether a token is translatable. */
+  {
+    bool translatable = false;
+    for (int i = 0; i < ntokens; ++i)
+      if (symbols[i]->translatable)
+        {
+          translatable = true;
+          break;
+        }
+    if (translatable)
+      {
+        int *values = xnmalloc (nsyms, sizeof *values);
+        for (int i = 0; i < ntokens; ++i)
+          values[i] = symbols[i]->translatable;
+        muscle_insert_int_table ("translatable", values,
+                                 values[0], 1, ntokens);
+        free (values);
+      }
+  }
 
   /* Output YYTOKNUM. */
   {
@@ -451,14 +563,13 @@ prepare_symbol_definitions (void)
 
       /* Its tag.  Typically for documentation purpose.  */
       SET_KEY ("tag");
-      MUSCLE_INSERT_STRING (key, sym->tag);
+      MUSCLE_INSERT_STRING (key, symbol_tag (sym));
 
       SET_KEY ("user_number");
       MUSCLE_INSERT_INT (key, sym->content->user_token_number);
 
       SET_KEY ("is_token");
-      MUSCLE_INSERT_INT (key,
-                         i < ntokens && sym != errtoken && sym != undeftoken);
+      MUSCLE_INSERT_INT (key, i < ntokens);
 
       SET_KEY ("number");
       MUSCLE_INSERT_INT (key, sym->content->number);
@@ -504,9 +615,7 @@ prepare_symbol_definitions (void)
 static void
 prepare_actions (void)
 {
-  /* Figure out the actions for the specified state, indexed by
-     lookahead token type.  */
-
+  /* Figure out the actions for the specified state.  */
   muscle_insert_rule_number_table ("defact", yydefact,
                                    yydefact[0], 1, nstates);
 

src/parse-gram.h

@@ -1,4 +1,4 @@
-/* A Bison parser, made by GNU Bison 3.5.22-948cd-dirty.  */
+/* A Bison parser, made by GNU Bison 3.5.90.  */
 
 /* Bison interface for Yacc-like parsers in C
 
@@ -31,8 +31,9 @@
    This special exception was added by the Free Software Foundation in
    version 2.2 of Bison.  */
 
-/* Undocumented macros, especially those whose name start with YY_,
-   are private implementation details.  Do not rely on them.  */
+/* DO NOT RELY ON FEATURES THAT ARE NOT DOCUMENTED in the manual,
+   especially those whose name start with YY_ or yy_.  They are
+   private implementation details that can be changed or removed.  */
 
 #ifndef YY_GRAM_SRC_PARSE_GRAM_H_INCLUDED
 # define YY_GRAM_SRC_PARSE_GRAM_H_INCLUDED
@@ -72,159 +73,124 @@ extern int gram_debug;
   } value_type;
 
 
-/* Token type.  */
+/* Token kinds.  */
 #ifndef GRAM_TOKENTYPE
 # define GRAM_TOKENTYPE
   enum gram_tokentype
   {
-    GRAM_EOF = 0,
-    STRING = 3,
-    PERCENT_TOKEN = 4,
-    PERCENT_NTERM = 5,
-    PERCENT_TYPE = 6,
-    PERCENT_DESTRUCTOR = 7,
-    PERCENT_PRINTER = 8,
-    PERCENT_LEFT = 9,
-    PERCENT_RIGHT = 10,
-    PERCENT_NONASSOC = 11,
-    PERCENT_PRECEDENCE = 12,
-    PERCENT_PREC = 13,
-    PERCENT_DPREC = 14,
-    PERCENT_MERGE = 15,
-    PERCENT_CODE = 16,
-    PERCENT_DEFAULT_PREC = 17,
-    PERCENT_DEFINE = 18,
-    PERCENT_DEFINES = 19,
-    PERCENT_ERROR_VERBOSE = 20,
-    PERCENT_EXPECT = 21,
-    PERCENT_EXPECT_RR = 22,
-    PERCENT_FLAG = 23,
-    PERCENT_FILE_PREFIX = 24,
-    PERCENT_GLR_PARSER = 25,
-    PERCENT_INITIAL_ACTION = 26,
-    PERCENT_LANGUAGE = 27,
-    PERCENT_NAME_PREFIX = 28,
-    PERCENT_NO_DEFAULT_PREC = 29,
-    PERCENT_NO_LINES = 30,
-    PERCENT_NONDETERMINISTIC_PARSER = 31,
-    PERCENT_OUTPUT = 32,
-    PERCENT_PURE_PARSER = 33,
-    PERCENT_REQUIRE = 34,
-    PERCENT_SKELETON = 35,
-    PERCENT_START = 36,
-    PERCENT_TOKEN_TABLE = 37,
-    PERCENT_VERBOSE = 38,
-    PERCENT_YACC = 39,
-    BRACED_CODE = 40,
-    BRACED_PREDICATE = 41,
-    BRACKETED_ID = 42,
-    CHAR = 43,
-    COLON = 44,
-    EPILOGUE = 45,
-    EQUAL = 46,
-    ID = 47,
-    ID_COLON = 48,
-    PERCENT_PERCENT = 49,
-    PIPE = 50,
-    PROLOGUE = 51,
-    SEMICOLON = 52,
-    TAG = 53,
-    TAG_ANY = 54,
-    TAG_NONE = 55,
-    INT = 56,
-    PERCENT_PARAM = 57,
-    PERCENT_UNION = 58,
-    PERCENT_EMPTY = 59
+    GRAM_EMPTY = -2,
+    GRAM_EOF = 0,                  /* "end of file"  */
+    GRAM_error = 1,                /* error  */
+    GRAM_UNDEF = 2,                /* "invalid token"  */
+    STRING = 3,                    /* "string"  */
+    TSTRING = 4,                   /* "translatable string"  */
+    PERCENT_TOKEN = 5,             /* "%token"  */
+    PERCENT_NTERM = 6,             /* "%nterm"  */
+    PERCENT_TYPE = 7,              /* "%type"  */
+    PERCENT_DESTRUCTOR = 8,        /* "%destructor"  */
+    PERCENT_PRINTER = 9,           /* "%printer"  */
+    PERCENT_LEFT = 10,             /* "%left"  */
+    PERCENT_RIGHT = 11,            /* "%right"  */
+    PERCENT_NONASSOC = 12,         /* "%nonassoc"  */
+    PERCENT_PRECEDENCE = 13,       /* "%precedence"  */
+    PERCENT_PREC = 14,             /* "%prec"  */
+    PERCENT_DPREC = 15,            /* "%dprec"  */
+    PERCENT_MERGE = 16,            /* "%merge"  */
+    PERCENT_CODE = 17,             /* "%code"  */
+    PERCENT_DEFAULT_PREC = 18,     /* "%default-prec"  */
+    PERCENT_DEFINE = 19,           /* "%define"  */
+    PERCENT_DEFINES = 20,          /* "%defines"  */
+    PERCENT_ERROR_VERBOSE = 21,    /* "%error-verbose"  */
+    PERCENT_EXPECT = 22,           /* "%expect"  */
+    PERCENT_EXPECT_RR = 23,        /* "%expect-rr"  */
+    PERCENT_FLAG = 24,             /* "%<flag>"  */
+    PERCENT_FILE_PREFIX = 25,      /* "%file-prefix"  */
+    PERCENT_GLR_PARSER = 26,       /* "%glr-parser"  */
+    PERCENT_INITIAL_ACTION = 27,   /* "%initial-action"  */
+    PERCENT_LANGUAGE = 28,         /* "%language"  */
+    PERCENT_NAME_PREFIX = 29,      /* "%name-prefix"  */
+    PERCENT_NO_DEFAULT_PREC = 30,  /* "%no-default-prec"  */
+    PERCENT_NO_LINES = 31,         /* "%no-lines"  */
+    PERCENT_NONDETERMINISTIC_PARSER = 32, /* "%nondeterministic-parser"  */
+    PERCENT_OUTPUT = 33,           /* "%output"  */
+    PERCENT_PURE_PARSER = 34,      /* "%pure-parser"  */
+    PERCENT_REQUIRE = 35,          /* "%require"  */
+    PERCENT_SKELETON = 36,         /* "%skeleton"  */
+    PERCENT_START = 37,            /* "%start"  */
+    PERCENT_TOKEN_TABLE = 38,      /* "%token-table"  */
+    PERCENT_VERBOSE = 39,          /* "%verbose"  */
+    PERCENT_YACC = 40,             /* "%yacc"  */
+    BRACED_CODE = 41,              /* "{...}"  */
+    BRACED_PREDICATE = 42,         /* "%?{...}"  */
+    BRACKETED_ID = 43,             /* "[identifier]"  */
+    CHAR = 44,                     /* "character literal"  */
+    COLON = 45,                    /* ":"  */
+    EPILOGUE = 46,                 /* "epilogue"  */
+    EQUAL = 47,                    /* "="  */
+    ID = 48,                       /* "identifier"  */
+    ID_COLON = 49,                 /* "identifier:"  */
+    PERCENT_PERCENT = 50,          /* "%%"  */
+    PIPE = 51,                     /* "|"  */
+    PROLOGUE = 52,                 /* "%{...%}"  */
+    SEMICOLON = 53,                /* ";"  */
+    TAG = 54,                      /* "<tag>"  */
+    TAG_ANY = 55,                  /* "<*>"  */
+    TAG_NONE = 56,                 /* "<>"  */
+    INT = 57,                      /* "integer literal"  */
+    PERCENT_PARAM = 58,            /* "%param"  */
+    PERCENT_UNION = 59,            /* "%union"  */
+    PERCENT_EMPTY = 60             /* "%empty"  */
   };
+  typedef enum gram_tokentype gram_token_kind_t;
 #endif
 
 /* Value type.  */
 #if ! defined GRAM_STYPE && ! defined GRAM_STYPE_IS_DECLARED
 union GRAM_STYPE
 {
+  assoc precedence_declarator;             /* precedence_declarator  */
+  char* STRING;                            /* "string"  */
+  char* TSTRING;                           /* "translatable string"  */
+  char* BRACED_CODE;                       /* "{...}"  */
+  char* BRACED_PREDICATE;                  /* "%?{...}"  */
+  char* EPILOGUE;                          /* "epilogue"  */
+  char* PROLOGUE;                          /* "%{...%}"  */
+  code_props_type code_props_type;         /* code_props_type  */
+  int INT;                                 /* "integer literal"  */
+  int yykind_82;                           /* int.opt  */
+  named_ref* yykind_95;                    /* named_ref.opt  */
+  param_type PERCENT_PARAM;                /* "%param"  */
+  symbol* token_decl;                      /* token_decl  */
+  symbol* alias;                           /* alias  */
+  symbol* token_decl_for_prec;             /* token_decl_for_prec  */
+  symbol* id;                              /* id  */
+  symbol* id_colon;                        /* id_colon  */
+  symbol* symbol;                          /* symbol  */
+  symbol* string_as_id;                    /* string_as_id  */
+  symbol_list* generic_symlist;            /* generic_symlist  */
+  symbol_list* generic_symlist_item;       /* generic_symlist_item  */
+  symbol_list* nterm_decls;                /* nterm_decls  */
+  symbol_list* token_decls;                /* token_decls  */
+  symbol_list* yykind_80;                  /* token_decl.1  */
+  symbol_list* token_decls_for_prec;       /* token_decls_for_prec  */
+  symbol_list* yykind_85;                  /* token_decl_for_prec.1  */
+  symbol_list* symbol_decls;               /* symbol_decls  */
+  symbol_list* yykind_88;                  /* symbol_decl.1  */
+  uniqstr PERCENT_ERROR_VERBOSE;           /* "%error-verbose"  */
+  uniqstr PERCENT_FLAG;                    /* "%<flag>"  */
+  uniqstr PERCENT_FILE_PREFIX;             /* "%file-prefix"  */
+  uniqstr PERCENT_NAME_PREFIX;             /* "%name-prefix"  */
+  uniqstr PERCENT_PURE_PARSER;             /* "%pure-parser"  */
+  uniqstr BRACKETED_ID;                    /* "[identifier]"  */
+  uniqstr ID;                              /* "identifier"  */
+  uniqstr ID_COLON;                        /* "identifier:"  */
+  uniqstr TAG;                             /* "<tag>"  */
+  uniqstr yykind_74;                       /* tag.opt  */
+  uniqstr tag;                             /* tag  */
+  uniqstr variable;                        /* variable  */
+  unsigned char CHAR;                      /* "character literal"  */
+  value_type value;                        /* value  */
 
-  /* precedence_declarator  */
-  assoc precedence_declarator;
-  /* "string"  */
-  char* STRING;
-  /* "{...}"  */
-  char* BRACED_CODE;
-  /* "%?{...}"  */
-  char* BRACED_PREDICATE;
-  /* "epilogue"  */
-  char* EPILOGUE;
-  /* "%{...%}"  */
-  char* PROLOGUE;
-  /* code_props_type  */
-  code_props_type code_props_type;
-  /* "integer"  */
-  int INT;
-  /* int.opt  */
-  int yytype_81;
-  /* named_ref.opt  */
-  named_ref* yytype_93;
-  /* "%param"  */
-  param_type PERCENT_PARAM;
-  /* token_decl  */
-  symbol* token_decl;
-  /* token_decl_for_prec  */
-  symbol* token_decl_for_prec;
-  /* id  */
-  symbol* id;
-  /* id_colon  */
-  symbol* id_colon;
-  /* symbol  */
-  symbol* symbol;
-  /* string_as_id  */
-  symbol* string_as_id;
-  /* string_as_id.opt  */
-  symbol* yytype_100;
-  /* generic_symlist  */
-  symbol_list* generic_symlist;
-  /* generic_symlist_item  */
-  symbol_list* generic_symlist_item;
-  /* nterm_decls  */
-  symbol_list* nterm_decls;
-  /* token_decls  */
-  symbol_list* token_decls;
-  /* token_decl.1  */
-  symbol_list* yytype_79;
-  /* token_decls_for_prec  */
-  symbol_list* token_decls_for_prec;
-  /* token_decl_for_prec.1  */
-  symbol_list* yytype_83;
-  /* symbol_decls  */
-  symbol_list* symbol_decls;
-  /* symbol_decl.1  */
-  symbol_list* yytype_86;
-  /* "%error-verbose"  */
-  uniqstr PERCENT_ERROR_VERBOSE;
-  /* "%<flag>"  */
-  uniqstr PERCENT_FLAG;
-  /* "%file-prefix"  */
-  uniqstr PERCENT_FILE_PREFIX;
-  /* "%name-prefix"  */
-  uniqstr PERCENT_NAME_PREFIX;
-  /* "%pure-parser"  */
-  uniqstr PERCENT_PURE_PARSER;
-  /* "[identifier]"  */
-  uniqstr BRACKETED_ID;
-  /* "identifier"  */
-  uniqstr ID;
-  /* "identifier:"  */
-  uniqstr ID_COLON;
-  /* "<tag>"  */
-  uniqstr TAG;
-  /* tag.opt  */
-  uniqstr yytype_73;
-  /* tag  */
-  uniqstr tag;
-  /* variable  */
-  uniqstr variable;
-  /* "character literal"  */
-  unsigned char CHAR;
-  /* value  */
-  value_type value;
 
 };
 typedef union GRAM_STYPE GRAM_STYPE;

src/parse-gram.y

@@ -125,7 +125,7 @@
 %define api.token.raw
 %define api.value.type union
 %define locations
-%define parse.error verbose
+%define parse.error custom
 %define parse.lac full
 %define parse.trace
 %defines
@@ -140,31 +140,26 @@
   boundary_set (&@$.end, grammar_file, 1, 1, 1);
 }
 
-/* Define the tokens together with their human representation.  */
-%token GRAM_EOF 0 "end of file"
-%token STRING     "string"
-
-%token PERCENT_TOKEN       "%token"
-%token PERCENT_NTERM       "%nterm"
-
-%token PERCENT_TYPE        "%type"
-%token PERCENT_DESTRUCTOR  "%destructor"
-%token PERCENT_PRINTER     "%printer"
-
-%token PERCENT_LEFT        "%left"
-%token PERCENT_RIGHT       "%right"
-%token PERCENT_NONASSOC    "%nonassoc"
-%token PERCENT_PRECEDENCE  "%precedence"
-
-%token PERCENT_PREC          "%prec"
-%token PERCENT_DPREC         "%dprec"
-%token PERCENT_MERGE         "%merge"
-
-/*----------------------.
-| Global Declarations.  |
-`----------------------*/
-
 %token
+  STRING              _("string")
+  TSTRING             _("translatable string")
+
+  PERCENT_TOKEN       "%token"
+  PERCENT_NTERM       "%nterm"
+
+  PERCENT_TYPE        "%type"
+  PERCENT_DESTRUCTOR  "%destructor"
+  PERCENT_PRINTER     "%printer"
+
+  PERCENT_LEFT        "%left"
+  PERCENT_RIGHT       "%right"
+  PERCENT_NONASSOC    "%nonassoc"
+  PERCENT_PRECEDENCE  "%precedence"
+
+  PERCENT_PREC        "%prec"
+  PERCENT_DPREC       "%dprec"
+  PERCENT_MERGE       "%merge"
+
   PERCENT_CODE            "%code"
   PERCENT_DEFAULT_PREC    "%default-prec"
   PERCENT_DEFINE          "%define"
@@ -190,24 +185,23 @@
   PERCENT_TOKEN_TABLE     "%token-table"
   PERCENT_VERBOSE         "%verbose"
   PERCENT_YACC            "%yacc"
-;
 
-%token BRACED_CODE     "{...}"
-%token BRACED_PREDICATE "%?{...}"
-%token BRACKETED_ID    "[identifier]"
-%token CHAR            "character literal"
-%token COLON           ":"
-%token EPILOGUE        "epilogue"
-%token EQUAL           "="
-%token ID              "identifier"
-%token ID_COLON        "identifier:"
-%token PERCENT_PERCENT "%%"
-%token PIPE            "|"
-%token PROLOGUE        "%{...%}"
-%token SEMICOLON       ";"
-%token TAG             "<tag>"
-%token TAG_ANY         "<*>"
-%token TAG_NONE        "<>"
+  BRACED_CODE       "{...}"
+  BRACED_PREDICATE  "%?{...}"
+  BRACKETED_ID      _("[identifier]")
+  CHAR              _("character literal")
+  COLON             ":"
+  EPILOGUE          _("epilogue")
+  EQUAL             "="
+  ID                _("identifier")
+  ID_COLON          _("identifier:")
+  PERCENT_PERCENT   "%%"
+  PIPE              "|"
+  PROLOGUE          "%{...%}"
+  SEMICOLON         ";"
+  TAG               _("<tag>")
+  TAG_ANY           "<*>"
+  TAG_NONE          "<>"
 
  /* Experimental feature, don't rely on it.  */
 %code pre-printer  {tron (yyo);}
@@ -216,8 +210,8 @@
 %type <unsigned char> CHAR
 %printer { fputs (char_name ($$), yyo); } <unsigned char>
 
-%type <char*> "{...}" "%?{...}" "%{...%}" EPILOGUE STRING
-%printer { fputs ($$, yyo); }  <char*>
+%type <char*> "{...}" "%?{...}" "%{...%}" EPILOGUE STRING TSTRING
+%printer { fputs ($$, yyo); } <char*>
 
 %type <uniqstr>
   BRACKETED_ID ID ID_COLON
@@ -230,7 +224,7 @@
 %printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG
 %printer { fprintf (yyo, "<%s>", $$); } TAG tag
 
-%token <int> INT "integer"
+%token <int> INT _("integer literal")
 %printer { fprintf (yyo, "%d", $$); } <int>
 
 %type <symbol*> id id_colon string_as_id symbol token_decl token_decl_for_prec
@@ -532,7 +526,7 @@ token_decl.1:
 
 // One symbol declaration for %token or %nterm.
 token_decl:
-  id int.opt[num] string_as_id.opt[alias]
+  id int.opt[num] alias
     {
       $$ = $id;
       symbol_class_set ($id, current_class, @id, true);
@@ -549,6 +543,19 @@ int.opt:
 | INT
 ;
 
+%type <symbol*> alias;
+alias:
+  %empty         { $$ = NULL; }
+| string_as_id   { $$ = $1; }
+| TSTRING
+    {
+      $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
+      symbol_class_set ($$, token_sym, @1, false);
+      $$->translatable = true;
+    }
+;
+
+
 /*-------------------------------------.
 | token_decls_for_prec (%left, etc.).  |
 `-------------------------------------*/
@@ -748,14 +755,11 @@ id:
         }
       if (muscle_percent_define_ifdef (var))
         {
-          int indent = 0;
-          complain_indent (&@1, complaint, &indent,
-                           _("character literals cannot be used together"
-                             " with %s"), var);
-          indent += SUB_INDENT;
+          complain (&@1, complaint,
+                    _("character literals cannot be used together"
+                    " with %s"), var);
           location loc = muscle_percent_define_get_loc (var);
-          complain_indent (&loc, complaint, &indent,
-                           _("definition of %s"), var);
+          subcomplain (&loc, complaint, _("definition of %s"), var);
         }
       $$ = symbol_get (char_name ($1), @1);
       symbol_class_set ($$, token_sym, @1, false);
@@ -782,12 +786,6 @@ string_as_id:
     }
 ;
 
-%type <symbol*> string_as_id.opt;
-string_as_id.opt:
-  %empty             { $$ = NULL; }
-| string_as_id
-;
-
 epilogue.opt:
   %empty
 | "%%" EPILOGUE
@@ -799,6 +797,32 @@ epilogue.opt:
 
 %%
 
+int
+yyreport_syntax_error (const yypcontext_t *ctx)
+{
+  int res = 0;
+  /* Arguments of format: reported tokens (one for the "unexpected",
+     one per "expected"). */
+  enum { ARGS_MAX = 5 };
+  const char *argv[ARGS_MAX];
+  int argc = 0;
+  yysymbol_kind_t unexpected = yypcontext_token (ctx);
+  if (unexpected != YYSYMBOL_YYEMPTY)
+    {
+      argv[argc++] = yysymbol_name (unexpected);
+      yysymbol_kind_t expected[ARGS_MAX - 1];
+      int nexpected = yypcontext_expected_tokens (ctx, expected, ARGS_MAX - 1);
+      if (nexpected < 0)
+        res = nexpected;
+      else
+        for (int i = 0; i < nexpected; ++i)
+          argv[argc++] = yysymbol_name (expected[i]);
+    }
+  syntax_error (*yypcontext_location (ctx), argc, argv);
+  return res;
+}
+
+
 /* Return the location of the left-hand side of a rule whose
    right-hand side is RHS[1] ... RHS[N].  Ignore empty nonterminals in
    the right-hand side, and return an empty location equal to the end
@@ -1021,9 +1045,9 @@ handle_require (location const *loc, char const *version)
       return;
     }
 
-  /* Pretend to be at least 3.5, to check features published in that
-     version while developping it.  */
-  const char* api_version = "3.5";
+  /* Pretend to be at least that version, to check features published
+     in that version while developping it.  */
+  const char* api_version = "3.6";
   const char* package_version =
     0 < strverscmp (api_version, PACKAGE_VERSION)
     ? api_version : PACKAGE_VERSION;

src/print-xml.c

@@ -329,7 +329,7 @@ print_reductions (FILE *out, int level, state *s)
 
 /*--------------------------------------------------------------.
 | Report on OUT all the actions (shifts, gotos, reductions, and |
-| explicit erros from %nonassoc) of S.                          |
+| explicit errors from %nonassoc) of S.                         |
 `--------------------------------------------------------------*/
 
 static void

src/print.c

@@ -319,7 +319,7 @@ print_reductions (FILE *out, state *s)
 
 /*--------------------------------------------------------------.
 | Report on OUT all the actions (shifts, gotos, reductions, and |
-| explicit erros from %nonassoc) of S.                          |
+| explicit errors from %nonassoc) of S.                         |
 `--------------------------------------------------------------*/
 
 static void

src/reader.c

@@ -124,16 +124,13 @@ record_merge_function_type (int merger, uniqstr type, location declaration_loc)
   aver (merge_function != NULL && merger_find == merger);
   if (merge_function->type != NULL && !UNIQSTR_EQ (merge_function->type, type))
     {
-      int indent = 0;
-      complain_indent (&declaration_loc, complaint, &indent,
-                       _("result type clash on merge function %s: "
-                         "<%s> != <%s>"),
-                       quote (merge_function->name), type,
-                       merge_function->type);
-      indent += SUB_INDENT;
-      complain_indent (&merge_function->type_declaration_loc, complaint,
-                       &indent,
-                       _("previous declaration"));
+      complain (&declaration_loc, complaint,
+                _("result type clash on merge function %s: "
+                "<%s> != <%s>"),
+                quote (merge_function->name), type,
+                merge_function->type);
+      subcomplain (&merge_function->type_declaration_loc, complaint,
+                   _("previous declaration"));
     }
   merge_function->type = uniqstr_new (type);
   merge_function->type_declaration_loc = declaration_loc;
@@ -610,7 +607,7 @@ packgram (void)
 {
   int itemno = 0;
   ritem = xnmalloc (nritems + 1, sizeof *ritem);
-  /* This sentinel is used by build_relations in gram.c.  */
+  /* This sentinel is used by build_relations() in lalr.c.  */
   *ritem++ = 0;
 
   rule_number ruleno = 0;
@@ -781,11 +778,16 @@ check_and_convert_grammar (void)
   /* If the user did not define her ENDTOKEN, do it now. */
   if (!endtoken)
     {
-      endtoken = symbol_get ("$end", empty_loc);
+      endtoken = symbol_get ("YYEOF", empty_loc);
       endtoken->content->class = token_sym;
       endtoken->content->number = 0;
       /* Value specified by POSIX.  */
       endtoken->content->user_token_number = 0;
+      {
+        symbol *alias = symbol_get ("$end", empty_loc);
+        symbol_class_set (alias, token_sym, empty_loc, false);
+        symbol_make_alias (endtoken, alias, empty_loc);
+      }
     }
 
   /* Report any undefined symbols and consider them nonterminals.  */
@@ -820,9 +822,9 @@ check_and_convert_grammar (void)
   /* Assign the symbols their symbol numbers.  */
   symbols_pack ();
 
-  /* Scan rule actions after invoking symbol_check_alias_consistency (in
-     symbols_pack above) so that token types are set correctly before the rule
-     action type checking.
+  /* Scan rule actions after invoking symbol_check_alias_consistency
+     (in symbols_pack above) so that token types are set correctly
+     before the rule action type checking.
 
      Before invoking grammar_rule_check_and_complete (in packgram
      below) on any rule, make sure all actions have already been

src/reader.h

@@ -35,12 +35,6 @@ typedef struct merger_list
   location type_declaration_loc;
 } merger_list;
 
-/* From the parser.  */
-extern int gram_debug;
-int gram_parse (void);
-
-
-/* From reader.c. */
 void grammar_start_symbol_set (symbol *sym, location loc);
 void grammar_current_rule_begin (symbol *lhs, location loc,
                                  named_ref *lhs_named_ref);

src/scan-code.l

@@ -81,7 +81,7 @@ static bool untyped_var_seen;
    historically almost any character is allowed in a tag.  We disallow
    NUL and newline, as this simplifies our implementation.  We allow
    "->" as a means to dereference a pointer.  */
-tag      ([^\0\n>]|->)+
+tag      ([^\0\n>]|->)*[^-]
 
 /* Zero or more instances of backslash-newline.  Following GCC, allow
    white space between the backslash and the newline.  */
@@ -321,14 +321,14 @@ static void
 show_sub_message (warnings warning,
                   const char* cp, bool explicit_bracketing,
                   int midrule_rhs_index, char dollar_or_at,
-                  int indent, const variant *var)
+                  const variant *var)
 {
   const char *at_spec = get_at_spec (var->symbol_index);
 
   if (var->err == 0)
-    complain_indent (&var->loc, warning, &indent,
-                     _("refers to: %c%s at %s"), dollar_or_at,
-                     var->id, at_spec);
+    subcomplain (&var->loc, warning,
+                 _("refers to: %c%s at %s"), dollar_or_at,
+                 var->id, at_spec);
   else
     {
       const char *id;
@@ -375,8 +375,8 @@ show_sub_message (warnings warning,
                         _(", cannot be accessed from midrule action at $%d"),
                         midrule_rhs_index);
 
-      complain_indent (&id_loc, warning, &indent, "%s",
-                        obstack_finish0 (&msg_buf));
+      subcomplain (&id_loc, warning, "%s",
+                   obstack_finish0 (&msg_buf));
       obstack_free (&msg_buf, 0);
     }
 }
@@ -384,14 +384,13 @@ show_sub_message (warnings warning,
 static void
 show_sub_messages (warnings warning,
                    const char* cp, bool explicit_bracketing,
-                   int midrule_rhs_index, char dollar_or_at,
-                   int indent)
+                   int midrule_rhs_index, char dollar_or_at)
 {
   for (int i = 0; i < variant_count; ++i)
     show_sub_message (warning | silent,
                       cp, explicit_bracketing,
                       midrule_rhs_index, dollar_or_at,
-                      indent, &variant_table[i]);
+                      &variant_table[i]);
 }
 
 /* Returned from "parse_ref" when the reference
@@ -492,47 +491,44 @@ parse_ref (char *cp, symbol_list *rule, int rule_length,
       {
         int len = (explicit_bracketing || !ref_tail_fields) ?
           cp_end - cp : ref_tail_fields - cp;
-        int indent = 0;
 
-        complain_indent (text_loc, complaint, &indent,
-                         _("invalid reference: %s"), quote (text));
-        indent += SUB_INDENT;
+        complain (text_loc, complaint,
+                     _("invalid reference: %s"), quote (text));
         if (len == 0)
           {
             location sym_loc = *text_loc;
             sym_loc.start.column += 1;
             sym_loc.end = sym_loc.start;
-            complain_indent (&sym_loc, complaint, &indent,
-                             _("syntax error after '%c', expecting integer, "
-                               "letter, '_', '[', or '$'"),
-                             dollar_or_at);
+            subcomplain (&sym_loc, complaint,
+                         _("syntax error after '%c', expecting integer, "
+                         "letter, '_', '[', or '$'"),
+                         dollar_or_at);
           }
         else if (midrule_rhs_index)
-          complain_indent (&rule->rhs_loc, complaint, &indent,
-                           _("symbol not found in production before $%d: "
-                             "%.*s"),
-                           midrule_rhs_index, len, cp);
+          subcomplain (&rule->rhs_loc, complaint,
+                       _("symbol not found in production before $%d: "
+                       "%.*s"),
+                       midrule_rhs_index, len, cp);
         else
-          complain_indent (&rule->rhs_loc, complaint, &indent,
-                           _("symbol not found in production: %.*s"),
-                           len, cp);
+          subcomplain (&rule->rhs_loc, complaint,
+                       _("symbol not found in production: %.*s"),
+                       len, cp);
 
         if (variant_count > 0)
           show_sub_messages (complaint,
                              cp, explicit_bracketing, midrule_rhs_index,
-                             dollar_or_at, indent);
+                             dollar_or_at);
         return INVALID_REF;
       }
     case 1:
       {
-        int indent = 0;
         if (variant_count > 1)
           {
-            complain_indent (text_loc, Wother, &indent,
-                             _("misleading reference: %s"), quote (text));
+            complain (text_loc, Wother,
+                         _("misleading reference: %s"), quote (text));
             show_sub_messages (Wother,
                                cp, explicit_bracketing, midrule_rhs_index,
-                               dollar_or_at, indent + SUB_INDENT);
+                               dollar_or_at);
           }
         {
           int symbol_index =
@@ -543,12 +539,11 @@ parse_ref (char *cp, symbol_list *rule, int rule_length,
     case 2:
     default:
       {
-        int indent = 0;
-        complain_indent (text_loc, complaint, &indent,
-                         _("ambiguous reference: %s"), quote (text));
+        complain (text_loc, complaint,
+                  _("ambiguous reference: %s"), quote (text));
         show_sub_messages (complaint,
                            cp, explicit_bracketing, midrule_rhs_index,
-                           dollar_or_at, indent + SUB_INDENT);
+                           dollar_or_at);
         return INVALID_REF;
       }
     }
@@ -617,7 +612,7 @@ handle_action_dollar (symbol_list *rule, char *text, const location *dollar_loc)
   char *cp = fetch_type_name (text + 1, &type_name, dollar_loc);
   int n = parse_ref (cp, effective_rule, effective_rule_length,
                      rule->midrule_parent_rhs_index, text, dollar_loc, '$');
-  /* End type_name.  Don't do it ealier: parse_ref depends on TEXT.  */
+  /* End type_name.  Don't do it earlier: parse_ref depends on TEXT.  */
   if (type_name)
     cp[-1] = '\0';
 

src/scan-gram.l

@@ -51,7 +51,7 @@ static boundary scanner_cursor;
 
 #define YY_USER_ACTION  location_compute (loc, &scanner_cursor, yytext, yyleng);
 
-/* Report that yytext is an extension, and evaluate to its token type.  */
+/* Report that yytext is an extension, and evaluate to its token kind.  */
 #define BISON_DIRECTIVE(Directive)                              \
   (bison_directive (loc, yytext), PERCENT_ ## Directive)
 
@@ -83,6 +83,23 @@ static boundary scanner_cursor;
       unput (Msg[i - 1]);                                       \
   } while (0)
 
+
+#define STRING_GROW_ESCAPE(Char)                                \
+  do {                                                          \
+    long c = Char;                                              \
+    if (0 < c && c <= UCHAR_MAX)                                \
+      obstack_1grow (&obstack_for_string, c);                   \
+    else                                                        \
+      {                                                         \
+        complain (loc, complaint,                               \
+                  _("invalid number after \\-escape: %s"),      \
+                  yytext + 1);                                  \
+        /* Avoid additional errors about empty char literal. */ \
+        obstack_1grow (&obstack_for_string, '?');               \
+      }                                                         \
+  } while (0)
+
+
 /* The current file name.  Might change with #line.  */
 static uniqstr current_file = NULL;
 
@@ -110,8 +127,8 @@ static void unexpected_newline (boundary, char const *);
 %}
  /* A C-like comment in directives/rules. */
 %x SC_YACC_COMMENT
- /* Strings and characters in directives/rules. */
-%x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
+ /* Characters and strings in directives/rules. */
+%x SC_ESCAPED_CHARACTER SC_ESCAPED_STRING SC_ESCAPED_TSTRING
  /* A identifier was just read in directives/rules.  Special state
     to capture the sequence 'identifier :'. */
 %x SC_AFTER_IDENTIFIER
@@ -290,6 +307,7 @@ eqopt    ({sp}=)?
 
   "%"{id} {
     complain (loc, complaint, _("invalid directive: %s"), quote (yytext));
+    return GRAM_error;
   }
 
   ":"                     return COLON;
@@ -311,6 +329,7 @@ eqopt    ({sp}=)?
      accept "1FOO" as "1 FOO".  */
   {int}{id} {
     complain (loc, complaint, _("invalid identifier: %s"), quote (yytext));
+    return GRAM_error;
   }
 
   /* Characters.  */
@@ -318,6 +337,7 @@ eqopt    ({sp}=)?
 
   /* Strings. */
   "\""        token_start = loc->start; BEGIN SC_ESCAPED_STRING;
+  "_(\""      token_start = loc->start; BEGIN SC_ESCAPED_TSTRING;
 
   /* Prologue. */
   "%{"        code_start = loc->start; BEGIN SC_PROLOGUE;
@@ -360,10 +380,11 @@ eqopt    ({sp}=)?
     BEGIN SC_BRACKETED_ID;
   }
 
-  [^\[%A-Za-z0-9_<>{}\"\'*;|=/, \f\r\n\t\v]+|. {
+  [^\[%A-Za-z0-9_<>{}""''*;|=/, \f\r\n\t\v]+|. {
     complain (loc, complaint, "%s: %s",
               ngettext ("invalid character", "invalid characters", yyleng),
               quote_mem (yytext, yyleng));
+    return GRAM_error;
   }
 
   <<EOF>> {
@@ -378,9 +399,13 @@ eqopt    ({sp}=)?
   | added value.                                                  |
   `--------------------------------------------------------------*/
 
-<SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING,SC_TAG>
+<SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_TSTRING,SC_TAG>
 {
-  \0        complain (loc, complaint, _("invalid null character"));
+  \0         {
+    complain (loc, complaint, _("invalid null character"));
+    STRING_FREE;
+    return GRAM_error;
+  }
 }
 
 
@@ -436,6 +461,7 @@ eqopt    ({sp}=)?
         complain (loc, complaint,
                   _("unexpected identifier in bracketed name: %s"),
                   quote (yytext));
+        return GRAM_error;
       }
     else
       {
@@ -456,7 +482,10 @@ eqopt    ({sp}=)?
           }
       }
     else
-      complain (loc, complaint, _("an identifier expected"));
+      {
+        complain (loc, complaint, _("an identifier expected"));
+        return GRAM_error;
+      }
   }
 
   [^\].A-Za-z0-9_/ \f\r\n\t\v]+|. {
@@ -464,6 +493,7 @@ eqopt    ({sp}=)?
               ngettext ("invalid character in bracketed name",
                         "invalid characters in bracketed name", yyleng),
               quote_mem (yytext, yyleng));
+    return GRAM_error;
   }
 
   <<EOF>> {
@@ -539,6 +569,20 @@ eqopt    ({sp}=)?
   {eol}     unexpected_newline (token_start, "\"");
 }
 
+<SC_ESCAPED_TSTRING>
+{
+  "\")" {
+    STRING_FINISH;
+    BEGIN INITIAL;
+    loc->start = token_start;
+    complain (loc, Wyacc,
+              _("POSIX Yacc does not support string literals"));
+    RETURN_VALUE (TSTRING, last_string);
+  }
+  <<EOF>>   unexpected_eof (token_start, "\"");
+  "\n"      unexpected_newline (token_start, "\"");
+}
+
   /*----------------------------------------------------------.
   | Scanning a Bison character literal, decoding its escapes. |
   | The initial quote is already eaten.                       |
@@ -548,22 +592,27 @@ eqopt    ({sp}=)?
 {
   "'" {
     STRING_FINISH;
+    BEGIN INITIAL;
     loc->start = token_start;
     val->CHAR = last_string[0];
 
-    /* FIXME: Eventually, make these errors.  */
     if (last_string[0] == '\0')
-    {
-      complain (loc, Wother, _("empty character literal"));
-      /* '\0' seems dangerous even if we are about to complain.  */
-      val->CHAR = '\'';
-    }
+      {
+        complain (loc, complaint, _("empty character literal"));
+        STRING_FREE;
+        return GRAM_error;
+      }
     else if (last_string[1] != '\0')
-      complain (loc, Wother,
-                _("extra characters in character literal"));
-    STRING_FREE;
-    BEGIN INITIAL;
-    return CHAR;
+      {
+        complain (loc, complaint, _("extra characters in character literal"));
+        STRING_FREE;
+        return GRAM_error;
+      }
+    else
+      {
+        STRING_FREE;
+        return CHAR;
+      }
   }
   {eol}     unexpected_newline (token_start, "'");
   <<EOF>>   unexpected_eof (token_start, "'");
@@ -601,26 +650,16 @@ eqopt    ({sp}=)?
   | Decode escaped characters.  |
   `----------------------------*/
 
-<SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
+<SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_TSTRING>
 {
   \\[0-7]{1,3} {
     verify (UCHAR_MAX < ULONG_MAX);
-    long c = strtol (yytext + 1, NULL, 8);
-    if (0 < c && c <= UCHAR_MAX)
-      obstack_1grow (&obstack_for_string, c);
-    else
-      complain (loc, complaint, _("invalid number after \\-escape: %s"),
-                yytext+1);
+    STRING_GROW_ESCAPE (strtol (yytext + 1, NULL, 8));
   }
 
   \\x[0-9abcdefABCDEF]+ {
     verify (UCHAR_MAX < ULONG_MAX);
-    long c = strtol (yytext + 2, NULL, 16);
-    if (0 < c && c <= UCHAR_MAX)
-      obstack_1grow (&obstack_for_string, c);
-    else
-      complain (loc, complaint, _("invalid number after \\-escape: %s"),
-                yytext+1);
+    STRING_GROW_ESCAPE (strtol (yytext + 2, NULL, 16));
   }
 
   \\a   obstack_1grow (&obstack_for_string, '\a');
@@ -635,13 +674,9 @@ eqopt    ({sp}=)?
   \\("\""|"'"|"?"|"\\")  obstack_1grow (&obstack_for_string, yytext[1]);
 
   \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} {
-    int c = convert_ucn_to_byte (yytext);
-    if (c <= 0)
-      complain (loc, complaint, _("invalid number after \\-escape: %s"),
-                yytext+1);
-    else
-      obstack_1grow (&obstack_for_string, c);
+    STRING_GROW_ESCAPE (convert_ucn_to_byte (yytext));
   }
+
   \\(.|{eol})      {
     char const *p = yytext + 1;
     /* Quote only if escaping won't make the character visible.  */
@@ -651,6 +686,7 @@ eqopt    ({sp}=)?
       p = quotearg_style_mem (escape_quoting_style, p, 1);
     complain (loc, complaint, _("invalid character after \\-escape: %s"),
               p);
+    obstack_1grow (&obstack_for_string, '?');
   }
 }
 
@@ -797,7 +833,7 @@ eqopt    ({sp}=)?
   | By default, grow the string obstack with the input.  |
   `-----------------------------------------------------*/
 
-<SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PREDICATE,SC_PROLOGUE,SC_EPILOGUE,SC_STRING,SC_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER>
+<SC_COMMENT,SC_LINE_COMMENT,SC_BRACED_CODE,SC_PREDICATE,SC_PROLOGUE,SC_EPILOGUE,SC_STRING,SC_CHARACTER,SC_ESCAPED_CHARACTER,SC_ESCAPED_STRING,SC_ESCAPED_TSTRING>
 {
   /* Accept multibyte characters in one block instead of byte after
      byte, so that add_column_width and mbsnwidth can compute correct

src/scan-skel.l

@@ -84,7 +84,7 @@ static void fail_for_invalid_at (char const *at);
 "@output("      at_init (&argc, argv, &at_ptr, &at_output);
 
   /* This pattern must not match more than the previous @ patterns. */
-@[^@{}\'(\n]*   fail_for_invalid_at (yytext);
+@[^@{}''(\n]*   fail_for_invalid_at (yytext);
 \n              out_lineno++; ECHO;
 [^@\n]+         ECHO;
 
@@ -189,7 +189,7 @@ flag (const char *arg)
   else if (STREQ (arg, "fatal"))
     return fatal;
   else if (STREQ (arg, "note"))
-    return silent | complaint | no_caret;
+    return silent | complaint | no_caret | note;
   else if (STREQ (arg, "warn"))
     return Wother;
   else
@@ -225,14 +225,7 @@ at_complain (int argc, char *argv[], char **out_namep, int *out_linenop)
       boundary_set_from_string (&loc.end, argv[3]);
       locp = &loc;
     }
-  static int indent;
-  if (w & silent)
-    indent += SUB_INDENT;
-  else
-    indent = 0;
-  complain_args (locp, w, &indent, argc - 4, argv + 4);
-  if (w & silent)
-    indent -= SUB_INDENT;
+  complain_args (locp, w, argc - 4, argv + 4);
 }
 
 static void

src/symtab.c

@@ -65,6 +65,20 @@ static bool *used_assoc = NULL;
 bool tag_seen = false;
 
 
+/* Whether SYM was defined by the user.  */
+
+static bool
+symbol_is_user_defined (symbol *sym)
+{
+  const bool eof_is_user_defined
+    = !endtoken->alias || STRNEQ (endtoken->alias->tag, "$end");
+  return sym->tag[0] != '$'
+    && (eof_is_user_defined || (sym != endtoken && sym->alias != errtoken))
+    && sym != errtoken && sym->alias != errtoken
+    && sym != undeftoken && sym->alias != undeftoken;
+}
+
+
 /*--------------------------.
 | Create a new sym_content. |
 `--------------------------*/
@@ -77,10 +91,12 @@ sym_content_new (symbol *s)
   res->symbol = s;
 
   res->type_name = NULL;
+  res->type_loc = empty_loc;
   for (int i = 0; i < CODE_PROPS_SIZE; ++i)
     code_props_none_init (&res->props[i]);
 
   res->number = NUMBER_UNDEFINED;
+  res->prec_loc = empty_loc;
   res->prec = 0;
   res->assoc = undef_assoc;
   res->user_token_number = USER_NUMBER_UNDEFINED;
@@ -109,6 +125,7 @@ symbol_new (uniqstr tag, location loc)
 
   res->tag = tag;
   res->location = loc;
+  res->translatable = false;
   res->location_of_lhs = false;
   res->alias = NULL;
   res->content = sym_content_new (res);
@@ -149,9 +166,9 @@ symbol_free (void *ptr)
 /* If needed, swap first and second so that first has the earliest
    location (according to location_cmp).
 
-   Many symbol features (e.g., user token numbers) are not assigned
-   during the parsing, but in a second step, via a traversal of the
-   symbol table sorted on tag.
+   Many symbol features (e.g., token codes) are not assigned during
+   parsing, but in a second step, via a traversal of the symbol table
+   sorted on tag.
 
    However, error messages make more sense if we keep the first
    declaration first.
@@ -276,6 +293,7 @@ is_identifier (uniqstr s)
 /*-----------------------------------------------.
 | Get the identifier associated to this symbol.  |
 `-----------------------------------------------*/
+
 uniqstr
 symbol_id_get (symbol const *sym)
 {
@@ -294,42 +312,29 @@ static void
 complain_symbol_redeclared (symbol *s, const char *what, location first,
                             location second)
 {
-  int i = 0;
   locations_sort (&first, &second);
-  complain_indent (&second, complaint, &i,
-                   _("%s redeclaration for %s"), what, s->tag);
-  i += SUB_INDENT;
-  complain_indent (&first, complaint, &i,
-                   _("previous declaration"));
+  complain (&second, complaint, _("%s redeclaration for %s"), what, s->tag);
+  subcomplain (&first, complaint, _("previous declaration"));
 }
 
 static void
 complain_semantic_type_redeclared (semantic_type *s, const char *what, location first,
                                    location second)
 {
-  int i = 0;
   locations_sort (&first, &second);
-  complain_indent (&second, complaint, &i,
-                   _("%s redeclaration for <%s>"), what, s->tag);
-  i += SUB_INDENT;
-  complain_indent (&first, complaint, &i,
-                   _("previous declaration"));
+  complain (&second, complaint, _("%s redeclaration for <%s>"), what, s->tag);
+  subcomplain (&first, complaint, _("previous declaration"));
 }
 
 static void
 complain_class_redeclared (symbol *sym, symbol_class class, location second)
 {
-  int i = 0;
-  complain_indent (&second, complaint, &i,
-                   class == token_sym
-                   ? _("symbol %s redeclared as a token")
-                   : _("symbol %s redeclared as a nonterminal"), sym->tag);
+  complain (&second, complaint,
+            class == token_sym
+            ? _("symbol %s redeclared as a token")
+            : _("symbol %s redeclared as a nonterminal"), sym->tag);
   if (!location_empty (sym->location))
-    {
-      i += SUB_INDENT;
-      complain_indent (&sym->location, complaint, &i,
-                       _("previous definition"));
-    }
+    subcomplain (&sym->location, complaint, _("previous definition"));
 }
 
 static const symbol *
@@ -429,6 +434,7 @@ symbol_code_props_set (symbol *sym, code_props_type kind,
     sym->content->props[kind] = *code;
 }
 
+
 /*-----------------------------------------------------.
 | Set the DESTRUCTOR or PRINTER associated with TYPE.  |
 `-----------------------------------------------------*/
@@ -467,7 +473,7 @@ symbol_code_props_get (symbol *sym, code_props_type kind)
     }
 
   /* Apply default code props's only to user-defined symbols.  */
-  if (sym->tag[0] != '$' && sym != errtoken)
+  if (symbol_is_user_defined (sym))
     {
       code_props *code = &semantic_type_get (sym->content->type_name ? "*" : "",
                                              NULL)->props[kind];
@@ -545,15 +551,16 @@ symbol_class_set (symbol *sym, symbol_class class, location loc, bool declaring)
         {
           if (s->status == declared)
             {
-              int i = 0;
-              complain_indent (&loc, Wother, &i,
-                               _("symbol %s redeclared"), sym->tag);
-              i += SUB_INDENT;
-              complain_indent (&sym->location, Wother, &i,
-                               _("previous declaration"));
+              complain (&loc, Wother,
+                        _("symbol %s redeclared"), sym->tag);
+              subcomplain (&sym->location, Wother,
+                           _("previous declaration"));
             }
           else
-            s->status = declared;
+            {
+              sym->location = loc;
+              s->status = declared;
+            }
         }
     }
 }
@@ -569,13 +576,13 @@ symbol_user_token_number_set (symbol *sym, int user_token_number, location loc)
   int *user_token_numberp = &sym->content->user_token_number;
   if (sym->content->class != token_sym)
     complain (&loc, complaint,
-              _("nonterminals cannot be given an explicit number"));
+              _("nonterminals cannot be given a token code"));
   else if (*user_token_numberp != USER_NUMBER_UNDEFINED
            && *user_token_numberp != user_token_number)
-    complain (&loc, complaint, _("redefining user token number of %s"),
+    complain (&loc, complaint, _("redefining code of token %s"),
               sym->tag);
   else if (user_token_number == INT_MAX)
-    complain (&loc, complaint, _("user token number of %s too large"),
+    complain (&loc, complaint, _("code of token %s too large"),
               sym->tag);
   else
     {
@@ -718,49 +725,47 @@ symbol_make_alias (symbol *sym, symbol *str, location loc)
 `-------------------------------------------------------------------*/
 
 static void
-symbol_pack (symbol *this)
+symbol_pack (symbol *sym)
 {
-  aver (this->content->number != NUMBER_UNDEFINED);
-  if (this->content->class == nterm_sym)
-    this->content->number += ntokens;
+  aver (sym->content->number != NUMBER_UNDEFINED);
+  if (sym->content->class == nterm_sym)
+    sym->content->number += ntokens;
 
-  symbols[this->content->number] = this->content->symbol;
+  symbols[sym->content->number] = sym->content->symbol;
 }
 
 static void
 complain_user_token_number_redeclared (int num, symbol *first, symbol *second)
 {
-  int i = 0;
   symbols_sort (&first, &second);
-  complain_indent (&second->location, complaint, &i,
-                   _("user token number %d redeclaration for %s"),
-                   num, second->tag);
-  i += SUB_INDENT;
-  complain_indent (&first->location, complaint, &i,
-                   _("previous declaration for %s"),
-                   first->tag);
+  complain (&second->location, complaint,
+            _("code %d reassigned to token %s"),
+            num, second->tag);
+  subcomplain (&first->location, complaint,
+               _("previous declaration for %s"),
+               first->tag);
 }
 
-/*--------------------------------------------------.
-| Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
-`--------------------------------------------------*/
+/*-------------------------------------------------.
+| Put SYM in TOKEN_TRANSLATIONS if it is a token.  |
+`-------------------------------------------------*/
 
 static void
-symbol_translation (symbol *this)
+symbol_translation (symbol *sym)
 {
   /* Nonterminal? */
-  if (this->content->class == token_sym
-      && !this->is_alias)
+  if (sym->content->class == token_sym
+      && !sym->is_alias)
     {
       /* A token which translation has already been set?*/
-      if (token_translations[this->content->user_token_number]
+      if (token_translations[sym->content->user_token_number]
           != undeftoken->content->number)
         complain_user_token_number_redeclared
-          (this->content->user_token_number,
-           symbols[token_translations[this->content->user_token_number]], this);
+          (sym->content->user_token_number,
+           symbols[token_translations[sym->content->user_token_number]], sym);
       else
-        token_translations[this->content->user_token_number]
-          = this->content->number;
+        token_translations[sym->content->user_token_number]
+          = sym->content->number;
     }
 }
 
@@ -842,16 +847,26 @@ symbols_new (void)
   accept->content->class = nterm_sym;
   accept->content->number = nvars++;
 
-  /* Construct the error token */
-  errtoken = symbol_get ("error", empty_loc);
+  /* Construct the YYerror/"error" token */
+  errtoken = symbol_get ("YYerror", empty_loc);
   errtoken->content->class = token_sym;
   errtoken->content->number = ntokens++;
+  {
+    symbol *alias = symbol_get ("error", empty_loc);
+    symbol_class_set (alias, token_sym, empty_loc, false);
+    symbol_make_alias (errtoken, alias, empty_loc);
+  }
 
-  /* Construct a token that represents all undefined literal tokens.
-     It is always token number 2.  */
-  undeftoken = symbol_get ("$undefined", empty_loc);
+  /* Construct the YYUNDEF/"$undefined" token that represents all
+     undefined literal tokens.  It is always symbol number 2.  */
+  undeftoken = symbol_get ("YYUNDEF", empty_loc);
   undeftoken->content->class = token_sym;
   undeftoken->content->number = ntokens++;
+  {
+    symbol *alias = symbol_get ("$undefined", empty_loc);
+    symbol_class_set (alias, token_sym, empty_loc, false);
+    symbol_make_alias (undeftoken, alias, empty_loc);
+  }
 
   semantic_type_table = hash_xinitialize (HT_INITIAL_CAPACITY,
                                           NULL,
@@ -954,7 +969,7 @@ dummy_symbol_get (location loc)
 }
 
 bool
-symbol_is_dummy (const symbol *sym)
+symbol_is_dummy (symbol const *sym)
 {
   return sym->tag[0] == '@' || (sym->tag[0] == '$' && sym->tag[1] == '@');
 }
@@ -1029,17 +1044,17 @@ symbols_token_translations_init (void)
 {
   bool num_256_available_p = true;
 
-  /* Find the highest user token number, and whether 256, the POSIX
-     preferred user token number for the error token, is used.  */
+  /* Find the highest token code, and whether 256, the POSIX preferred
+     token code for the error token, is used.  */
   max_user_token_number = 0;
   for (int i = 0; i < ntokens; ++i)
     {
-      sym_content *this = symbols[i]->content;
-      if (this->user_token_number != USER_NUMBER_UNDEFINED)
+      sym_content *sym = symbols[i]->content;
+      if (sym->user_token_number != USER_NUMBER_UNDEFINED)
         {
-          if (this->user_token_number > max_user_token_number)
-            max_user_token_number = this->user_token_number;
-          if (this->user_token_number == 256)
+          if (sym->user_token_number > max_user_token_number)
+            max_user_token_number = sym->user_token_number;
+          if (sym->user_token_number == 256)
             num_256_available_p = false;
         }
     }
@@ -1055,17 +1070,17 @@ symbols_token_translations_init (void)
 
   for (int i = 0; i < ntokens; ++i)
     {
-      sym_content *this = symbols[i]->content;
-      if (this->user_token_number == USER_NUMBER_UNDEFINED)
+      sym_content *sym = symbols[i]->content;
+      if (sym->user_token_number == USER_NUMBER_UNDEFINED)
         {
           IGNORE_TYPE_LIMITS_BEGIN
           if (INT_ADD_WRAPV (max_user_token_number, 1, &max_user_token_number))
             complain (NULL, fatal, _("token number too large"));
           IGNORE_TYPE_LIMITS_END
-          this->user_token_number = max_user_token_number;
+          sym->user_token_number = max_user_token_number;
         }
-      if (this->user_token_number > max_user_token_number)
-        max_user_token_number = this->user_token_number;
+      if (sym->user_token_number > max_user_token_number)
+        max_user_token_number = sym->user_token_number;
     }
 
   token_translations = xnmalloc (max_user_token_number + 1,
@@ -1080,6 +1095,21 @@ symbols_token_translations_init (void)
 }
 
 
+/* Whether some symbol requires internationalization.  */
+static bool
+has_translations (void)
+{
+  for (const void *entry = hash_get_first (symbol_table);
+       entry;
+       entry = hash_get_next (symbol_table, entry))
+    {
+      const symbol *sym = (const symbol *) entry;
+      if (sym->translatable)
+        return true;
+    }
+  return false;
+}
+
 /*----------------------------------------------------------------.
 | Assign symbol numbers, and write definition of token names into |
 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
@@ -1122,6 +1152,18 @@ symbols_pack (void)
     complain (&startsymbol_loc, fatal,
               _("the start symbol %s is a token"),
               startsymbol->tag);
+
+  // If some user tokens are internationalized, the internal ones
+  // should be too.
+  if (has_translations ())
+    {
+      const bool eof_is_user_defined
+        = !endtoken->alias || STRNEQ (endtoken->alias->tag, "$end");
+      if (!eof_is_user_defined)
+        endtoken->alias->translatable = true;
+      undeftoken->alias->translatable = true;
+      errtoken->alias->translatable = true;
+    }
 }
 
 /*---------------------------------.

src/symtab.h

@@ -97,6 +97,9 @@ struct symbol
   /** The "defining" location.  */
   location location;
 
+  /** Whether this symbol is translatable. */
+  bool translatable;
+
   /** Whether \a location is about the first uses as left-hand side
       symbol of a rule (true), or simply the first occurrence (e.g.,
       in a %type, or as a rhs symbol of a rule).  The former type of
@@ -117,6 +120,8 @@ struct symbol
 
 struct sym_content
 {
+  /** The main symbol that denotes this content (it contains the
+      possible alias). */
   symbol *symbol;
 
   /** Its \c \%type.
@@ -130,7 +135,7 @@ struct sym_content
   /** Its \c \%type's location.  */
   location type_loc;
 
-  /** Any \c \%destructor (resp. \%printer) declared specificially for this
+  /** Any \c \%destructor (resp. \%printer) declared specifically for this
       symbol.
 
       Access this field only through <tt>symbol</tt>'s interface functions. For
@@ -179,7 +184,7 @@ symbol *dummy_symbol_get (location loc);
 void symbol_print (symbol const *s, FILE *f);
 
 /** Is this a dummy nonterminal?  */
-bool symbol_is_dummy (const symbol *sym);
+bool symbol_is_dummy (symbol const *sym);
 
 /** The name of the code_props type: "\%destructor" or "\%printer".  */
 char const *code_props_type_string (code_props_type kind);
@@ -225,7 +230,8 @@ void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 /** Set the \c class associated with \c sym.
 
     Whether \c declaring means whether this class definition comes
-    from %nterm or %token (but not %type, prec/assoc, etc.).  */
+    from %nterm or %token (but not %type, prec/assoc, etc.).  A symbol
+    can have "declaring" set only at most once.  */
 void symbol_class_set (symbol *sym, symbol_class class, location loc,
                        bool declaring);
 

src/system.h

@@ -73,9 +73,14 @@ typedef size_t uintptr_t;
 # include <verify.h>
 # include <xalloc.h>
 
+// Clang and ICC like to pretend they are GCC.
+# if defined __GNUC__ && !defined __clang__ && !defined __ICC
+#  define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
+# endif
 
-/* See https://lists.gnu.org/archive/html/bug-bison/2019-10/msg00061.html. */
-# if defined __GNUC__ && ! defined __clang__ && ! defined __ICC && __GNUC__ < 5
+// See https://lists.gnu.org/archive/html/bug-bison/2019-10/msg00061.html
+// and https://trac.macports.org/ticket/59927.
+# if defined GCC_VERSION && 405 <= GCC_VERSION
 #  define IGNORE_TYPE_LIMITS_BEGIN \
      _Pragma ("GCC diagnostic push") \
      _Pragma ("GCC diagnostic ignored \"-Wtype-limits\"")
@@ -209,10 +214,10 @@ typedef size_t uintptr_t;
 
 /* Output Str both quoted for M4 (i.e., embed in [[...]]), and escaped
    for our postprocessing (i.e., escape M4 special characters).  If
-   Str is empty (or NULL), output "[]" instead of "[[]]" as it make M4
-   programming easier (m4_ifval can be used).
+   Str is empty (or NULL), output "[]" instead of "[[]]" as it makes
+   M4 programming easier (m4_ifval can be used).
 
-   For instance "[foo]" -> "[[@{foo@}]]", "$$" -> "[[$][$][]]". */
+   For instance "[foo]" -> "[[@{foo@}]]", "$$" -> "[[$][$][]]".  */
 
 # define obstack_quote(Obs, Str)                \
   do {                                          \