version 3.7.4

* NEWS: Record release date.

.prev-version

@@ -1 +1 @@
-3.7
+3.7.3

.travis.yml

@@ -50,6 +50,7 @@ jobs:
         - make -j2 dist-xz
         # Can help understanding why we get "dirty" tarballs.
         - git status
+        - git diff
         - dist=$(echo bison*.xz)
 
         # Unfortunately we cannot deterministically know the name of the tarball without the full
@@ -129,7 +130,7 @@ jobs:
         - CC=icc
         - CXX=icpc
       install:
-        - source /opt/intel/inteloneapi/compiler/latest/env/vars.sh
+        - source /opt/intel/oneapi/compiler/latest/env/vars.sh
       addons:
         apt:
           sources:

NEWS

@@ -1,5 +1,66 @@
 GNU Bison NEWS
 
+* Noteworthy changes in release 3.7.4 (2020-11-14) [stable]
+
+** Bug fixes
+
+*** Bug fixes in yacc.c
+
+  In Yacc mode, all the tokens are defined twice: once as an enum, and then
+  as a macro.  YYEMPTY was missing its macro.
+
+*** Bug fixes in lalr1.cc
+
+  The lalr1.cc skeleton used to emit internal assertions (using YY_ASSERT)
+  even when the `parse.assert` %define variable is not enabled.  It no
+  longer does.
+
+  The private internal macro YY_ASSERT now obeys the `api.prefix` %define
+  variable.
+
+  When there is a very large number of tokens, some assertions could be long
+  enough to hit arbitrary limits in Visual C++.  They have been rewritten to
+  work around this limitation.
+
+** Changes
+
+  The YYBISON macro in generated "regular C parsers" (from the "yacc.c"
+  skeleton) used to be defined to 1.  It is now defined to the version of
+  Bison as an integer (e.g., 30704 for version 3.7.4).
+
+* Noteworthy changes in release 3.7.3 (2020-10-13) [stable]
+
+** Bug fixes
+
+  Fix concurrent build issues.
+
+  The bison executable is no longer linked uselessly against libreadline.
+
+  Fix incorrect use of yytname in glr.cc.
+
+* Noteworthy changes in release 3.7.2 (2020-09-05) [stable]
+
+  This release of Bison fixes all known bugs reported for Bison in MITRE's
+  Common Vulnerabilities and Exposures (CVE) system.  These vulnerabilities
+  are only about bison-the-program itself, not the generated code.
+
+  Although these bugs are typically irrelevant to how Bison is used, they
+  are worth fixing if only to give users peace of mind.
+
+  There is no known vulnerability in the generated parsers.
+
+** Bug fixes
+
+  Fix concurrent build issues (introduced in Bison 3.5).
+
+  Push parsers always use YYMALLOC/YYFREE (no direct calls to malloc/free).
+
+  Fix portability issues of the test suite, and of bison itself.
+
+  Some unlikely crashes found by fuzzing have been fixed.  This is only
+  about bison itself, not the generated parsers.
+
+
 * Noteworthy changes in release 3.7.1 (2020-08-02) [stable]
 
 ** Bug fixes
@@ -554,7 +615,8 @@ GNU Bison NEWS
   \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.
+  Some unlikely crashes found by fuzzing have been fixed.  This is only
+  about bison itself, not the generated parsers.
 
 
 * Noteworthy changes in release 3.5.2 (2020-02-13) [stable]

README

@@ -27,7 +27,7 @@ Bison from the git repo.  Roughly, run:
 then proceed with the usual `configure && make` steps.
 
 ## Build from tarball
-See the [INSTALL file](INSTALL] for generic compilation and installation
+See the [INSTALL file](INSTALL) for generic compilation and installation
 instructions.
 
 Bison requires GNU m4 1.4.6 or later.  See

TODO

@@ -1,4 +1,12 @@
-* Bison 3.7
+* Soon
+** gnulib
+Bruno notes:
+
+> I haven't looked deeply, but it strikes me that gnulib/lib/bitset/array.c
+> does not make use of the 'ffsl' function, nor or the 'integer_length_l'
+> function. Maybe because in Bison, all bitsets are so dense that it does
+> not give a performance advantage?
+
 ** Cex
 *** Improve gnulib
 Don't do this (counterexample.c):

cfg.mk

@@ -126,7 +126,7 @@ _sed_rm_comments_q = $(subst ','\'',$(_sed_remove_comments))
 
 _space_before_paren_exempt =? \\n\\$$
 _space_before_paren_exempt = \
-  (^ *\#|(LA)?LR\([01]\)|percent_(code|define)|b4_syncline|m4_(define|init)|symbol)
+  (^ *\#|(LA)?LR\([01]\)|percent_(code|define)|b4_syncline|m4_(define|init))
 # Ensure that there is a space before each open parenthesis in C code.
 sc_space_before_open_paren:
 	@if $(VC_LIST_EXCEPT) | grep -l '\.[ch]$$' > /dev/null; then	\
@@ -156,6 +156,7 @@ exclude = \
 $(call exclude,                                                                 \
   bindtextdomain=^lib/main.c$$                                                  \
   cast_of_argument_to_free=^src/muscle-tab.c$$                                  \
+  error_message_uppercase=etc/bench.pl.in$$                                     \
   po_check=^tests|(^po/POTFILES.in|.md)$$                                       \
   preprocessor_indentation=^data/|^lib/|^src/parse-gram.[ch]$$                  \
   program_name=^lib/main.c$$                                                    \

data/skeletons/bison.m4

@@ -49,7 +49,7 @@ m4_define([m4_shift4], [m4_shift(m4_shift(m4_shift(m4_shift($@))))])
 # b4_generated_by
 # ---------------
 m4_define([b4_generated_by],
-[b4_comment([A Bison parser, made by GNU Bison b4_version.])
+[b4_comment([A Bison parser, made by GNU Bison b4_version_string.])
 ])
 
 # b4_copyright(TITLE, [YEARS])
@@ -633,11 +633,11 @@ m4_define([_b4_type_action],
 
 ])])
 
-# b4_type_foreach(MACRO)
-# ----------------------
+# b4_type_foreach(MACRO, [SEP])
+# -----------------------------
 # Invoke MACRO(SYMBOL-NUMS) for each set of SYMBOL-NUMS for each type set.
 m4_define([b4_type_foreach],
-          [m4_map([$1], m4_defn([b4_type_names]))])
+          [m4_map_sep([$1], [$2], m4_defn([b4_type_names]))])
 
 
 

data/skeletons/c++.m4

@@ -321,8 +321,9 @@ m4_define([b4_symbol_type_define],
       /// Copy constructor.
       basic_symbol (const basic_symbol& that);]b4_variant_if([[
 
-      /// Constructor for valueless symbols, and symbols from each type.
-]b4_type_foreach([b4_basic_symbol_constructor_define])], [[
+      /// Constructors for typed symbols.
+]b4_type_foreach([b4_basic_symbol_constructor_define], [
+])], [[
       /// Constructor for valueless symbols.
       basic_symbol (typename Base::kind_type t]b4_locations_if([,
                     YY_MOVE_REF (location_type) l])[);

data/skeletons/c.m4

@@ -58,11 +58,11 @@ m4_define([b4_cpp_guard_close],
 # b4_pull_flag if they use the values of the %define variables api.pure or
 # api.push-pull.
 m4_define([b4_identification],
-[[/* Identify Bison output.  */
-#define YYBISON 1
+[[/* Identify Bison output, and Bison version.  */
+#define YYBISON ]b4_version[
 
-/* Bison version.  */
-#define YYBISON_VERSION "]b4_version["
+/* Bison version string.  */
+#define YYBISON_VERSION "]b4_version_string["
 
 /* Skeleton name.  */
 #define YYSKELETON_NAME ]b4_skeleton[]m4_ifdef([b4_pure_flag], [[
@@ -509,10 +509,11 @@ m4_define([b4_token_define],
 # ----------------
 # Output the definition of the tokens.
 m4_define([b4_token_defines],
-[b4_any_token_visible_if([/* Token kinds.  */
-m4_join([
+[[/* Token kinds.  */
+#define ]b4_symbol([-2], [id])[ -2
+]m4_join([
 ], b4_symbol_map([b4_token_define]))
-])])
+])
 
 
 # b4_token_enum(TOKEN-NUM)

data/skeletons/d.m4

@@ -103,12 +103,12 @@ m4_define([b4_location_type_if],
 # b4_identification
 # -----------------
 m4_define([b4_identification],
-[/** Version number for the Bison executable that generated this parser.  */
-  public static immutable string yy_bison_version = "b4_version";
+[[/** Version number for the Bison executable that generated this parser.  */
+  public static immutable string yy_bison_version = "]b4_version_string[";
 
   /** Name of the skeleton that generated this parser.  */
-  public static immutable string yy_bison_skeleton = b4_skeleton;
-])
+  public static immutable string yy_bison_skeleton = ]b4_skeleton[;
+]])
 
 
 ## ------------ ##

data/skeletons/glr.cc

@@ -172,7 +172,7 @@ m4_pushdef([b4_parse_param], m4_defn([b4_parse_param_orig]))dnl
                            const location_type* yylocationp]])[) const
   {
     *yycdebug_ << (yykind < YYNTOKENS ? "token" : "nterm")
-               << ' ' << yytname[yykind] << " ("]b4_locations_if([[
+               << ' ' << yysymbol_name (yykind) << " ("]b4_locations_if([[
                << *yylocationp << ": "]])[;
     yy_symbol_value_print_ (yykind, yyvaluep]b4_locations_if([[, yylocationp]])[);
     *yycdebug_ << ')';

data/skeletons/java.m4

@@ -71,12 +71,12 @@ m4_define([b4_lexer_if],
 # b4_identification
 # -----------------
 m4_define([b4_identification],
-[  /** Version number for the Bison executable that generated this parser.  */
-  public static final String bisonVersion = "b4_version";
+[[  /** Version number for the Bison executable that generated this parser.  */
+  public static final String bisonVersion = "]b4_version_string[";
 
   /** Name of the skeleton that generated this parser.  */
-  public static final String bisonSkeleton = b4_skeleton;
-])
+  public static final String bisonSkeleton = ]b4_skeleton[;
+]])
 
 
 ## ------------ ##

data/skeletons/location.cc

@@ -22,7 +22,7 @@ m4_pushdef([b4_copyright_years],
 # b4_position_file
 # ----------------
 # Name of the file containing the position class, if we want this file.
-b4_defines_if([b4_required_version_if([302], [],
+b4_defines_if([b4_required_version_if([30200], [],
                                       [m4_define([b4_position_file], [position.hh])])])])
 
 

data/skeletons/stack.hh

@@ -19,7 +19,7 @@
 # b4_stack_file
 # -------------
 # Name of the file containing the stack class, if we want this file.
-b4_defines_if([b4_required_version_if([302], [],
+b4_defines_if([b4_required_version_if([30200], [],
                                       [m4_define([b4_stack_file], [stack.hh])])])
 
 

data/skeletons/variant.hh

@@ -20,6 +20,13 @@
 ## variant.  ##
 ## --------- ##
 
+# b4_assert
+# ---------
+# The name of YY_ASSERT.
+m4_define([b4_assert],
+          [b4_api_PREFIX[]_ASSERT])
+
+
 # b4_symbol_variant(YYTYPE, YYVAL, ACTION, [ARGS])
 # ------------------------------------------------
 # Run some ACTION ("build", or "destroy") on YYVAL of symbol type
@@ -71,12 +78,12 @@ m4_map([      b4_symbol_tag_comment], [$@])dnl
 # -------------------
 # The needed includes for variants support.
 m4_define([b4_variant_includes],
-[b4_parse_assert_if([[#include <typeinfo>]])[
-#ifndef YY_ASSERT
+[b4_parse_assert_if([[#include <typeinfo>
+#ifndef ]b4_assert[
 # include <cassert>
-# define YY_ASSERT assert
+# define ]b4_assert[ assert
 #endif
-]])
+]])])
 
 
 
@@ -110,8 +117,8 @@ m4_define([b4_value_type_declare],
     template <typename T>
     semantic_type (YY_RVREF (T) t)]b4_parse_assert_if([
       : yytypeid_ (&typeid (T))])[
-    {
-      YY_ASSERT (sizeof (T) <= size);
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (sizeof (T) <= size);]])[
       new (yyas_<T> ()) T (YY_MOVE (t));
     }
 
@@ -125,7 +132,7 @@ m4_define([b4_value_type_declare],
     /// Destruction, allowed only if empty.
     ~semantic_type () YY_NOEXCEPT
     {]b4_parse_assert_if([
-      YY_ASSERT (!yytypeid_);
+      ]b4_assert[ (!yytypeid_);
     ])[}
 
 # if 201103L <= YY_CPLUSPLUS
@@ -133,10 +140,10 @@ m4_define([b4_value_type_declare],
     template <typename T, typename... U>
     T&
     emplace (U&&... u)
-    {]b4_parse_assert_if([
-      YY_ASSERT (!yytypeid_);
-      YY_ASSERT (sizeof (T) <= size);
-      yytypeid_ = & typeid (T);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (!yytypeid_);
+      ]b4_assert[ (sizeof (T) <= size);
+      yytypeid_ = & typeid (T);]])[
       return *new (yyas_<T> ()) T (std::forward <U>(u)...);
     }
 # else
@@ -144,10 +151,10 @@ m4_define([b4_value_type_declare],
     template <typename T>
     T&
     emplace ()
-    {]b4_parse_assert_if([
-      YY_ASSERT (!yytypeid_);
-      YY_ASSERT (sizeof (T) <= size);
-      yytypeid_ = & typeid (T);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (!yytypeid_);
+      ]b4_assert[ (sizeof (T) <= size);
+      yytypeid_ = & typeid (T);]])[
       return *new (yyas_<T> ()) T ();
     }
 
@@ -155,10 +162,10 @@ m4_define([b4_value_type_declare],
     template <typename T>
     T&
     emplace (const T& t)
-    {]b4_parse_assert_if([
-      YY_ASSERT (!yytypeid_);
-      YY_ASSERT (sizeof (T) <= size);
-      yytypeid_ = & typeid (T);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (!yytypeid_);
+      ]b4_assert[ (sizeof (T) <= size);
+      yytypeid_ = & typeid (T);]])[
       return *new (yyas_<T> ()) T (t);
     }
 # endif
@@ -185,10 +192,10 @@ m4_define([b4_value_type_declare],
     template <typename T>
     T&
     as () YY_NOEXCEPT
-    {]b4_parse_assert_if([
-      YY_ASSERT (yytypeid_);
-      YY_ASSERT (*yytypeid_ == typeid (T));
-      YY_ASSERT (sizeof (T) <= size);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (yytypeid_);
+      ]b4_assert[ (*yytypeid_ == typeid (T));
+      ]b4_assert[ (sizeof (T) <= size);]])[
       return *yyas_<T> ();
     }
 
@@ -196,10 +203,10 @@ m4_define([b4_value_type_declare],
     template <typename T>
     const T&
     as () const YY_NOEXCEPT
-    {]b4_parse_assert_if([
-      YY_ASSERT (yytypeid_);
-      YY_ASSERT (*yytypeid_ == typeid (T));
-      YY_ASSERT (sizeof (T) <= size);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (yytypeid_);
+      ]b4_assert[ (*yytypeid_ == typeid (T));
+      ]b4_assert[ (sizeof (T) <= size);]])[
       return *yyas_<T> ();
     }
 
@@ -214,9 +221,9 @@ m4_define([b4_value_type_declare],
     template <typename T>
     void
     swap (self_type& that) YY_NOEXCEPT
-    {]b4_parse_assert_if([
-      YY_ASSERT (yytypeid_);
-      YY_ASSERT (*yytypeid_ == *that.yytypeid_);])[
+    {]b4_parse_assert_if([[
+      ]b4_assert[ (yytypeid_);
+      ]b4_assert[ (*yytypeid_ == *that.yytypeid_);]])[
       std::swap (as<T> (), that.as<T> ());
     }
 
@@ -388,11 +395,67 @@ m4_define([_b4_token_maker_define],
 ])])
 
 
-m4_define([_b4_type_clause],
-[b4_symbol_if([$1], [is_token],
-              [b4_symbol_if([$1], [has_id],
-                            [tok == token::b4_symbol([$1], [id])],
-                            [tok == b4_symbol([$1], [code])])])])
+# b4_token_kind(SYMBOL-NUM)
+# -------------------------
+# Some tokens don't have an ID.
+m4_define([b4_token_kind],
+[b4_symbol_if([$1], [has_id],
+              [token::b4_symbol([$1], [id])],
+              [b4_symbol([$1], [code])])])
+
+
+# _b4_tok_in(SYMBOL-NUM, ...)
+# ---------------------------
+# See b4_tok_in below.  The SYMBOL-NUMs... are tokens only.
+#
+# We iterate over the tokens to group them by "range" of token numbers (not
+# symbols numbers!).
+#
+# b4_fst is the start of that range.
+# b4_prev is the previous value.
+# b4_val is the current value.
+# If b4_val is the successor of b4_prev in token numbers, update the latter,
+#   otherwise emit the code for range b4_fst .. b4_prev.
+# $1 is also used as a terminator in the foreach, but it will not be printed.
+#
+m4_define([_b4_tok_in],
+[m4_pushdef([b4_prev], [$1])dnl
+m4_pushdef([b4_fst], [$1])dnl
+m4_pushdef([b4_sep], [])dnl
+m4_foreach([b4_val], m4_dquote(m4_shift($@, $1)),
+           [m4_if(b4_symbol(b4_val, [code]), m4_eval(b4_symbol(b4_prev, [code]) + 1), [],
+                  [b4_sep[]m4_if(b4_fst, b4_prev,
+                         [tok == b4_token_kind(b4_fst)],
+                         [(b4_token_kind(b4_fst) <= tok && tok <= b4_token_kind(b4_prev))])[]dnl
+m4_define([b4_fst], b4_val)dnl
+m4_define([b4_sep], [
+                   || ])])dnl
+m4_define([b4_prev], b4_val)])dnl
+m4_popdef([b4_sep])dnl
+m4_popdef([b4_fst])dnl
+m4_popdef([b4_prev])dnl
+])
+
+
+# _b4_filter_tokens(SYMBOL-NUM, ...)
+# ----------------------------------
+# Expand as the list of tokens amongst SYMBOL-NUM.
+m4_define([_b4_filter_tokens],
+[m4_pushdef([b4_sep])dnl
+m4_foreach([b4_val], [$@],
+           [b4_symbol_if(b4_val, [is_token], [b4_sep[]b4_val[]m4_define([b4_sep], [,])])])dnl
+m4_popdef([b4_sep])dnl
+])
+
+
+# b4_tok_in(SYMBOL-NUM, ...)
+# ---------------------------
+# A C++ conditional that checks that `tok` is a member of this list of symbol
+# numbers.
+m4_define([b4_tok_in],
+          [_$0(_b4_filter_tokens($@))])
+
+
 
 
 # _b4_token_constructor_define(SYMBOL-NUM...)
@@ -410,9 +473,6 @@ m4_define([_b4_token_constructor_define],
         : super_type(]b4_join([token_type (tok)],
                               b4_symbol_if([$1], [has_type], [std::move (v)]),
                               b4_locations_if([std::move (l)]))[)
-      {
-        YY_ASSERT (]m4_join([ || ], m4_map_sep([_b4_type_clause], [, ], [$@]))[);
-      }
 #else
       symbol_type (]b4_join(
           [int tok],
@@ -422,10 +482,10 @@ m4_define([_b4_token_constructor_define],
         : super_type(]b4_join([token_type (tok)],
                               b4_symbol_if([$1], [has_type], [v]),
                               b4_locations_if([l]))[)
-      {
-        YY_ASSERT (]m4_join([ || ], m4_map_sep([_b4_type_clause], [, ], [$@]))[);
-      }
 #endif
+      {]b4_parse_assert_if([[
+        ]b4_assert[ (]b4_tok_in($@)[);
+      ]])[}
 ]])])
 
 

data/skeletons/yacc.c

@@ -1486,7 +1486,7 @@ yypstate_new (void)
   yypstate *yyps;]b4_pure_if([], [[
   if (yypstate_allocated)
     return YY_NULLPTR;]])[
-  yyps = YY_CAST (yypstate *, malloc (sizeof *yyps));
+  yyps = YY_CAST (yypstate *, YYMALLOC (sizeof *yyps));
   if (!yyps)
     return YY_NULLPTR;]b4_pure_if([], [[
   yypstate_allocated = 1;]])[
@@ -1515,7 +1515,7 @@ yypstate_delete (yypstate *yyps)
 #endif]b4_lac_if([[
       if (yyes != yyesa)
         YYSTACK_FREE (yyes);]])[
-      free (yyps);]b4_pure_if([], [[
+      YYFREE (yyps);]b4_pure_if([], [[
       yypstate_allocated = 0;]])[
     }
 }

doc/bison.texi

@@ -303,7 +303,7 @@ Grammar Rules for @code{rpcalc}
 
 * Rpcalc Input::            Explanation of the @code{input} nonterminal
 * Rpcalc Line::             Explanation of the @code{line} nonterminal
-* Rpcalc Expr::             Explanation of the @code{expr} nonterminal
+* Rpcalc Exp::              Explanation of the @code{exp} nonterminal
 
 Location Tracking Calculator: @code{ltcalc}
 
@@ -1788,7 +1788,7 @@ rule are referred to as @code{$1}, @code{$2}, and so on.
 @menu
 * Rpcalc Input::            Explanation of the @code{input} nonterminal
 * Rpcalc Line::             Explanation of the @code{line} nonterminal
-* Rpcalc Expr::             Explanation of the @code{expr} nonterminal
+* Rpcalc Exp::              Explanation of the @code{exp} nonterminal
 @end menu
 
 @node Rpcalc Input
@@ -1853,8 +1853,8 @@ uninitialized (its value will be unpredictable).  This would be a bug if
 that value were ever used, but we don't use it: once rpcalc has printed the
 value of the user's input line, that value is no longer needed.
 
-@node Rpcalc Expr
-@subsubsection Explanation of @code{expr}
+@node Rpcalc Exp
+@subsubsection Explanation of @code{exp}
 
 The @code{exp} grouping has several rules, one for each kind of expression.
 The first rule handles the simplest expressions: those that are just
@@ -6296,7 +6296,10 @@ Introduced in Bison 3.3 to replace @code{parser_class_name}.
 
 @item Default Value: @code{YY} for Java, @code{yy} otherwise.
 
-@item History: introduced in Bison 2.6
+@item History:
+introduced in Bison 2.6, with its argument in double quotes.  Uses braces
+since Bison 3.0 (double quotes are still supported for backward
+compatibility).
 @end itemize
 @end deffn
 
@@ -9942,10 +9945,11 @@ very documentation.  To solve a conflict, one must understand it: when does
 it occur?  Is it because of a flaw in the grammar?  Is it rather because
 LR(1) cannot cope with this grammar?
 
-On difficulty is that conflicts occur in the @emph{automaton}, and it can be
-tricky to related them to issues in the @emph{grammar} itself.  With
-experience and patience, analysis the detailed description of the automaton
-(@pxref{Understanding}) allows to find example strings that reach these conflicts.
+One difficulty is that conflicts occur in the @emph{automaton}, and it can
+be tricky to relate them to issues in the @emph{grammar} itself.  With
+experience and patience, analysis of the detailed description of the
+automaton (@pxref{Understanding}) allows one to find example strings that
+reach these conflicts.
 
 That task is made much easier thanks to the generation of counterexamples,
 initially developed by Chinawat Isradisaikul and Andrew Myers
@@ -10098,7 +10102,7 @@ sequence.y:8.3-45: @dwarning{warning}: rule useless in parser due to conflicts [
 
 Each of these three conflicts, again, prove that the grammar is ambiguous.
 For instance, the second conflict (the reduce/reduce one) shows that the
-grammar accept the empty input in two different ways.
+grammar accepts the empty input in two different ways.
 
 @sp 1
 
@@ -15430,6 +15434,12 @@ Macro to discard a value from the parser stack and fake a lookahead
 token.  @xref{Action Features}.
 @end deffn
 
+@deffn {Macro} YYBISON
+The version of Bison as an integer, for instance 30704 for version 3.7.4.
+Defined in @file{yacc.c} only.  Before version 3.7.4, @code{YYBISON} was
+defined to 1.
+@end deffn
+
 @deffn {Variable} yychar
 External integer variable that contains the integer value of the
 lookahead token.  (In a pure parser, it is a local variable within

doc/local.mk

@@ -57,7 +57,7 @@ MAINTAINERCLEANFILES = $(CROSS_OPTIONS_TEXI)
 
 # Fix Info's @code in @deftype
 # https://lists.gnu.org/archive/html/help-texinfo/2019-11/msg00004.html
-all: $(srcdir)/$(%C%_bison).info.bak
+all-local: $(srcdir)/$(%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)*)}'	\

etc/bench.pl.in

@@ -185,13 +185,13 @@ my $verbose = 1;
 
 =over 4
 
-=item C<verbose($level, $message)>
+=item C<verbose ($level, $message)>
 
 Report the C<$message> is C<$level> E<lt>= C<$verbose>.
 
 =cut
 
-sub verbose($$)
+sub verbose ($$)
 {
   my ($level, $message) = @_;
   print STDERR $message
@@ -201,13 +201,13 @@ sub verbose($$)
 
 ######################################################################
 
-=item C<directives($bench, @directive)>
+=item C<directives ($bench, @directive)>
 
 Format the list of directives for Bison for bench named C<$bench>.
 
 =cut
 
-sub directives($@)
+sub directives ($@)
 {
   my ($bench, @directive) = @_;
   my $res = "/* Directives for bench '$bench'. */\n";
@@ -218,6 +218,27 @@ sub directives($@)
 
 ######################################################################
 
+=item C<is_pure (@directive)>
+
+Whether api.pure is set.
+
+=cut
+
+sub is_pure (@)
+{
+  my (@directive) = @_;
+  for my $dir (@directive)
+    {
+      if ($dir =~ /\A%define api.pure/)
+        {
+          return 1;
+        }
+    }
+  return 0;
+}
+
+######################################################################
+
 =item C<generate_grammar_triangular ($base, $max, @directive)>
 
 Create a large triangular grammar which looks like :
@@ -389,18 +410,14 @@ sub generate_grammar_calc ($$@)
 %define api.value.type union
 $directives
 
-%{
+%code provides {
 static int power (int base, int exponent);
 /* yyerror receives the location if:
    - %location & %pure & %glr
    - %location & %pure & %yacc & %parse-param. */
 static void yyerror (const char *s);
-#if YYPURE
-static int yylex (YYSTYPE* yylvalp);
-#else
-static int yylex (void);
-#endif
-%}
+static int yylex (@{[is_pure (@directive) ? "YYSTYPE *yylvalp" : "void"]});
+}
 
 /* Bison Declarations */
 %token
@@ -467,12 +484,7 @@ yyerror (const char *s)
 }
 
 static int
-#if YYPURE
-# define yylval (*yylvalp)
-yylex (YYSTYPE* yylvalp)
-#else
-yylex (void)
-#endif
+yylex (@{[is_pure (@directive) ? "YYSTYPE *yylvalp" : "void"]})
 {
   int c;
 
@@ -498,7 +510,7 @@ yylex (void)
     case '5': case '6': case '7': case '8': case '9':
       {
         int nchars = 0;
-        int n = sscanf (input - 1, "%d%n", &yylval.NUM, &nchars);
+        int n = sscanf (input - 1, "%d%n", &@{[is_pure (@directive) ? "yylvalp->" : "yylval."]}NUM, &nchars);
         assert (n == 1);
         input += nchars - 1;
         return NUM;
@@ -506,7 +518,7 @@ yylex (void)
 
     default:
       yyerror ("error: invalid character");
-      return yylex ();
+      return yylex (@{[is_pure (@directive) ? "yylvalp" : ""]});
     }
 }
 EOF
@@ -592,10 +604,10 @@ $directives
   // Prototype of the yylex function providing subsequent tokens.
   static
 #if USE_TOKEN_CTOR
-  yy::parser::symbol_type yylex();
+  yy::parser::symbol_type yylex ();
 #else
-  yy::parser::token_type yylex(yy::parser::semantic_type* yylvalp,
-                               yy::parser::location_type* yyllocp);
+  yy::parser::token_type yylex (yy::parser::semantic_type *yylvalp,
+                                yy::parser::location_type *yyllocp);
 #endif
 
   // Conversion to string.
@@ -618,8 +630,8 @@ EOF
       print $out <<'EOF';
 %token <std::string> TEXT
 %token <int> NUMBER
-%printer { std::cerr << "Number: " << $$; } <int>
-%printer { std::cerr << "Text: " << $$; } <std::string>
+%printer { yyo << "Number: " << $$; } <int>
+%printer { yyo << "Text: " << $$; } <std::string>
 %type <std::string> text result
 
 %%
@@ -641,8 +653,8 @@ EOF
 %union {int ival; std::string* sval;}
 %token <sval> TEXT
 %token <ival> NUMBER
-%printer { std::cerr << "Number: " << $$; } <ival>
-%printer { std::cerr << "Text: " << *$$; } <sval>
+%printer { yyo << "Number: " << $$; } <ival>
+%printer { yyo << "Text: " << *$$; } <sval>
 %type <sval> text result
 
 %%
@@ -664,10 +676,10 @@ EOF
 
 static
 #if USE_TOKEN_CTOR
-yy::parser::symbol_type yylex()
+yy::parser::symbol_type yylex ()
 #else
-yy::parser::token_type yylex(yy::parser::semantic_type* yylvalp,
-                             yy::parser::location_type* yyllocp)
+yy::parser::token_type yylex (yy::parser::semantic_type *yylvalp,
+                              yy::parser::location_type *yyllocp)
 #endif
 {
   typedef yy::parser::location_type location_type;

examples/c/bistromathic/bistromathic.test

@@ -251,29 +251,29 @@ 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: Reducing stack by rule XX (line XXX):
 err:    $1 = token ( (1.1: )
 err:    $2 = token error (1.2-3: )
 err:    $3 = token ) (1.4: )
 err: -> $$ = nterm exp (1.1-4: 666)
-err: Entering state 7
-err: Stack now 0 7
+err: Entering state 8
+err: Stack now 0 8
 err: Return for a new token:
 err: Reading a token
 err: Now at end of input.
 err: LAC: initial context established for end of file
-err: LAC: checking lookahead end of file: R2 G8 S19
-err: Reducing stack by rule 2 (line 126):
+err: LAC: checking lookahead end of file: R2 G7 S14
+err: Reducing stack by rule XX (line XXX):
 err:    $1 = nterm exp (1.1-4: 666)
 err: -> $$ = nterm input (1.1-4: )
-err: Entering state 8
-err: Stack now 0 8
+err: Entering state 7
+err: Stack now 0 7
 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: Entering state 14
+err: Stack now 0 7 14
+err: Stack now 0 7 14
 err: Cleanup: popping token end of file (1.5: )
 err: Cleanup: popping nterm input (1.1-4: )' -p
 

examples/c/lexcalc/local.mk

@@ -31,6 +31,13 @@ endif FLEX_WORKS
 
 %D%/parse.c: $(dependencies)
 
+# Tell Make scan.o depends on parse.h, except that Make sees only
+# parse.c, not parse.h.  We can't use BUILT_SOURCES to this end, since
+# we use the built bison.
+%D%/lexcalc$(DASH)scan.o: %D%/parse.c
+# Likewise, but for Automake before 1.16.
+%D%/examples_c_lexcalc_lexcalc$(DASH)scan.o: %D%/parse.c
+
 EXTRA_DIST += %D%/lexcalc.test
 dist_lexcalc_DATA = %D%/parse.y %D%/scan.l %D%/Makefile %D%/README.md
 CLEANFILES += %D%/parse.[ch] %D%/scan.c %D%/parse.output

examples/d/local.mk

@@ -27,7 +27,7 @@ EXTRA_DIST += %D%/calc.test
 
 %D%/calc.d: %D%/calc.y $(dependencies)
 	$(AM_V_GEN)$(MKDIR_P) %D%
-	$(AM_V_at)$(BISON) $(srcdir)/%D%/calc.y -o $@
+	$(AM_V_at)$(BISON) -o $@ $(srcdir)/%D%/calc.y
 
 %D%/calc: %D%/calc.d
 	$(AM_V_GEN) $(DC) $(DCFLAGS) -of$@ %D%/calc.d

examples/java/calc/local.mk

@@ -27,7 +27,7 @@ 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 $@
+	$(AM_V_at)$(BISON) -o $@ $(srcdir)/%D%/Calc.y
 
 %D%/Calc.class: %D%/Calc.java
 	$(AM_V_GEN) $(SHELL) $(top_builddir)/javacomp.sh %D%/Calc.java

examples/java/simple/local.mk

@@ -27,7 +27,7 @@ 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 $@
+	$(AM_V_at)$(BISON) -o $@ $(srcdir)/%D%/Calc.y
 
 %D%/Calc.class: %D%/Calc.java
 	$(AM_V_GEN) $(SHELL) $(top_builddir)/javacomp.sh %D%/Calc.java

lib/.gitignore

@@ -336,9 +336,11 @@
 /sys_types.in.h
 /sys_wait.in.h
 /sysexits.in.h
+/termios.h
 /termios.in.h
 /textstyle.h
 /textstyle.in.h
+/thread-optim.h
 /time.h
 /time.in.h
 /timespec.c

m4/.gitignore

@@ -67,6 +67,7 @@
 /intlmacosx.m4
 /intmax.m4
 /intmax_t.m4
+/inttypes-pri.m4
 /inttypes.m4
 /inttypes_h.m4
 /isnan.m4
@@ -127,6 +128,7 @@
 /open.m4
 /pathmax.m4
 /perror.m4
+/pid_t.m4
 /pipe2.m4
 /po.m4
 /posix_spawn.m4

src/counterexample.c

@@ -712,10 +712,10 @@ ssb_equals (const search_state_bundle *s1, const search_state_bundle *s2)
 typedef gl_list_t ssb_list;
 
 static size_t
-visited_hasher (const search_state *ss, size_t maximum)
+visited_hasher (const search_state *ss, size_t max)
 {
-  return (parse_state_hasher (ss->states[0], maximum)
-          + parse_state_hasher (ss->states[1], maximum)) % maximum;
+  return (parse_state_hasher (ss->states[0], max)
+          + parse_state_hasher (ss->states[1], max)) % max;
 }
 
 static bool

src/local.mk

@@ -103,6 +103,8 @@ src_bison_SOURCES =                             \
   src/state.h                                   \
   src/state-item.c                              \
   src/state-item.h                              \
+  src/strversion.c                              \
+  src/strversion.h                              \
   src/symlist.c                                 \
   src/symlist.h                                 \
   src/symtab.c                                  \
@@ -143,7 +145,6 @@ src_bison_LDADD =                               \
   $(LIB_SETLOCALE_NULL)                         \
   $(LIBICONV)                                   \
   $(LIBINTL)                                    \
-  $(LIBREADLINE)                                \
   $(LIBTEXTSTYLE)
 
 

src/location.c

@@ -155,7 +155,9 @@ int
 location_print (location loc, FILE *out)
 {
   int res = 0;
-  if (trace_flag & trace_locations)
+  if (location_empty (loc))
+    res += fprintf (out, "(empty location)");
+  else if (trace_flag & trace_locations)
     {
       res += boundary_print (&loc.start, out);
       res += fprintf (out, "-");

src/muscle-tab.c

@@ -127,9 +127,6 @@ muscle_init (void)
 
   muscle_table = hash_xinitialize (HT_INITIAL_CAPACITY, NULL, hash_muscle,
                                    hash_compare_muscles, muscle_entry_free);
-
-  /* Version and input file.  */
-  MUSCLE_INSERT_STRING ("version", VERSION);
 }
 
 

src/output.c

@@ -42,6 +42,7 @@
 #include "scan-skel.h"
 #include "symtab.h"
 #include "tables.h"
+#include "strversion.h"
 
 static struct obstack format_obstack;
 
@@ -249,7 +250,7 @@ prepare_symbol_names (char const *muscle_name)
       if (i)
         obstack_1grow (&format_obstack, ' ');
       if (translatable)
-        obstack_sgrow (&format_obstack, "]b4_symbol_translate([");
+        obstack_sgrow (&format_obstack, "]b4_symbol_translate""([");
       obstack_escape (&format_obstack, cp);
       if (translatable)
         obstack_sgrow (&format_obstack, "])[");
@@ -554,7 +555,7 @@ prepare_symbol_definitions (void)
   /* Map "orig NUM" to new numbers.  See data/README.  */
   for (symbol_number i = ntokens; i < nsyms + nuseless_nonterminals; ++i)
     {
-      obstack_printf (&format_obstack, "symbol(orig %d, number)", i);
+      obstack_printf (&format_obstack, "symbol""(orig %d, number)", i);
       const char *key = obstack_finish0 (&format_obstack);
       MUSCLE_INSERT_INT (key, nterm_map ? nterm_map[i - ntokens] : i);
     }
@@ -565,12 +566,12 @@ prepare_symbol_definitions (void)
       const char *key;
 
 #define SET_KEY(Entry)                                          \
-      obstack_printf (&format_obstack, "symbol(%d, %s)",        \
+      obstack_printf (&format_obstack, "symbol""(%d, %s)",        \
                       i, Entry);                                \
       key = obstack_finish0 (&format_obstack);
 
 #define SET_KEY2(Entry, Suffix)                                 \
-      obstack_printf (&format_obstack, "symbol(%d, %s_%s)",     \
+      obstack_printf (&format_obstack, "symbol""(%d, %s_%s)",     \
                       i, Entry, Suffix);                        \
       key = obstack_finish0 (&format_obstack);
 
@@ -807,6 +808,9 @@ prepare (void)
   char const *cp = getenv ("BISON_USE_PUSH_FOR_PULL");
   bool use_push_for_pull_flag = cp && *cp && strtol (cp, 0, 10);
 
+  /* Versions.  */
+  MUSCLE_INSERT_STRING ("version_string", VERSION);
+  MUSCLE_INSERT_INT ("version", strversion_to_int (VERSION));
   MUSCLE_INSERT_INT ("required_version", required_version);
 
   /* Flags. */

src/parse-gram.c

@@ -1,4 +1,4 @@
-/* A Bison parser, made by GNU Bison 3.7.14-eb26-dirty.  */
+/* A Bison parser, made by GNU Bison 3.7.3.7-d831b.  */
 
 /* Bison implementation for Yacc-like parsers in C
 
@@ -46,10 +46,10 @@
    USER NAME SPACE" below.  */
 
 /* Identify Bison output.  */
-#define YYBISON 1
+#define YYBISON 30703
 
 /* Bison version.  */
-#define YYBISON_VERSION "3.7.14-eb26-dirty"
+#define YYBISON_VERSION "3.7.3.7-d831b"
 
 /* Skeleton name.  */
 #define YYSKELETON_NAME "yacc.c"
@@ -218,8 +218,6 @@ typedef enum yysymbol_kind_t yysymbol_kind_t;
   #include "system.h"
 
   #include <c-ctype.h>
-  #include <errno.h>
-  #include <intprops.h>
   #include <quotearg.h>
   #include <vasnprintf.h>
   #include <xmemdup0.h>
@@ -233,6 +231,7 @@ typedef enum yysymbol_kind_t yysymbol_kind_t;
   #include "reader.h"
   #include "scan-code.h"
   #include "scan-gram.h"
+  #include "strversion.h"
 
   /* Pretend to be at least that version, to check features published
      in that version while developping it.  */
@@ -639,19 +638,19 @@ union yyalloc
   /* YYRLINE[YYN] -- Source line where rule number YYN was defined.  */
 static const yytype_int16 yyrline[] =
 {
-       0,   312,   312,   321,   322,   326,   327,   333,   337,   342,
-     343,   344,   345,   346,   347,   348,   353,   358,   359,   360,
-     361,   362,   363,   363,   364,   365,   366,   367,   368,   369,
-     370,   371,   375,   376,   385,   386,   390,   401,   405,   409,
-     417,   427,   428,   438,   439,   445,   458,   458,   463,   463,
-     468,   472,   482,   483,   484,   485,   489,   490,   495,   496,
-     500,   501,   505,   506,   507,   520,   529,   533,   537,   545,
-     546,   550,   563,   564,   569,   570,   571,   589,   593,   597,
-     605,   607,   612,   619,   629,   633,   637,   645,   650,   662,
-     663,   669,   670,   671,   678,   678,   686,   687,   688,   693,
-     696,   698,   700,   702,   704,   706,   708,   710,   712,   717,
-     718,   727,   751,   752,   753,   754,   766,   768,   792,   797,
-     798,   803,   811,   812
+       0,   311,   311,   320,   321,   325,   326,   332,   336,   341,
+     342,   343,   344,   345,   346,   347,   352,   357,   358,   359,
+     360,   361,   362,   362,   363,   364,   365,   366,   367,   368,
+     369,   370,   374,   375,   384,   385,   389,   400,   404,   408,
+     416,   426,   427,   437,   438,   444,   457,   457,   462,   462,
+     467,   471,   481,   482,   483,   484,   488,   489,   494,   495,
+     499,   500,   504,   505,   506,   519,   528,   532,   536,   544,
+     545,   549,   562,   563,   568,   569,   570,   588,   592,   596,
+     604,   606,   611,   618,   628,   632,   636,   644,   649,   661,
+     662,   668,   669,   670,   677,   677,   685,   686,   687,   692,
+     695,   697,   699,   701,   703,   705,   707,   709,   711,   716,
+     717,   726,   750,   751,   752,   753,   765,   767,   791,   796,
+     797,   802,   810,   811
 };
 #endif
 
@@ -3032,41 +3031,11 @@ handle_pure_parser (location const *loc, char const *directive)
 }
 
 
-/* Convert VERSION into an int (MAJOR * 100 + MINOR).  Return -1 on
-   errors.
-
-   Changes of behavior are only on minor version changes, so "3.0.5"
-   is the same as "3.0": 300. */
-static int
-str_to_version (char const *version)
-{
-  IGNORE_TYPE_LIMITS_BEGIN
-  int res = 0;
-  errno = 0;
-  char *cp = NULL;
-  long major = strtol (version, &cp, 10);
-  if (errno || cp == version || *cp != '.' || major < 0
-      || INT_MULTIPLY_WRAPV (major, 100, &res))
-    return -1;
-
-  ++cp;
-  char *cp1 = NULL;
-  long minor = strtol (cp, &cp1, 10);
-  if (errno || cp1 == cp || (*cp1 != '\0' && *cp1 != '.')
-      || ! (0 <= minor && minor < 100)
-      || INT_ADD_WRAPV (minor, res, &res))
-    return -1;
-
-  IGNORE_TYPE_LIMITS_END
-  return res;
-}
-
-
 static void
 handle_require (location const *loc, char const *version_quoted)
 {
   char *version = unquote (version_quoted);
-  required_version = str_to_version (version);
+  required_version = strversion_to_int (version);
   if (required_version == -1)
     {
       complain (loc, complaint, _("invalid version requirement: %s"),
@@ -3147,8 +3116,7 @@ char_name (char c)
     }
 }
 
-static
-void
+static void
 current_lhs (symbol *sym, location loc, named_ref *ref)
 {
   current_lhs_symbol = sym;

src/parse-gram.h

@@ -1,4 +1,4 @@
-/* A Bison parser, made by GNU Bison 3.7.14-eb26-dirty.  */
+/* A Bison parser, made by GNU Bison 3.7.3.7-d831b.  */
 
 /* Bison interface for Yacc-like parsers in C
 

src/parse-gram.y

@@ -42,8 +42,6 @@
   #include "system.h"
 
   #include <c-ctype.h>
-  #include <errno.h>
-  #include <intprops.h>
   #include <quotearg.h>
   #include <vasnprintf.h>
   #include <xmemdup0.h>
@@ -57,6 +55,7 @@
   #include "reader.h"
   #include "scan-code.h"
   #include "scan-gram.h"
+  #include "strversion.h"
 
   /* Pretend to be at least that version, to check features published
      in that version while developping it.  */
@@ -1043,41 +1042,11 @@ handle_pure_parser (location const *loc, char const *directive)
 }
 
 
-/* Convert VERSION into an int (MAJOR * 100 + MINOR).  Return -1 on
-   errors.
-
-   Changes of behavior are only on minor version changes, so "3.0.5"
-   is the same as "3.0": 300. */
-static int
-str_to_version (char const *version)
-{
-  IGNORE_TYPE_LIMITS_BEGIN
-  int res = 0;
-  errno = 0;
-  char *cp = NULL;
-  long major = strtol (version, &cp, 10);
-  if (errno || cp == version || *cp != '.' || major < 0
-      || INT_MULTIPLY_WRAPV (major, 100, &res))
-    return -1;
-
-  ++cp;
-  char *cp1 = NULL;
-  long minor = strtol (cp, &cp1, 10);
-  if (errno || cp1 == cp || (*cp1 != '\0' && *cp1 != '.')
-      || ! (0 <= minor && minor < 100)
-      || INT_ADD_WRAPV (minor, res, &res))
-    return -1;
-
-  IGNORE_TYPE_LIMITS_END
-  return res;
-}
-
-
 static void
 handle_require (location const *loc, char const *version_quoted)
 {
   char *version = unquote (version_quoted);
-  required_version = str_to_version (version);
+  required_version = strversion_to_int (version);
   if (required_version == -1)
     {
       complain (loc, complaint, _("invalid version requirement: %s"),
@@ -1158,8 +1127,7 @@ char_name (char c)
     }
 }
 
-static
-void
+static void
 current_lhs (symbol *sym, location loc, named_ref *ref)
 {
   current_lhs_symbol = sym;

src/parse-simulation.c

@@ -28,7 +28,7 @@
 #include "lssi.h"
 #include "nullable.h"
 
-typedef struct parse_state
+struct parse_state
 {
   // Path of state-items the parser has traversed.
   struct si_chunk
@@ -58,7 +58,7 @@ typedef struct parse_state
   // Causes chunk contents to be freed when the reference count is
   // one. Used when only the chunk metadata will be needed.
   bool free_contents_early;
-} parse_state;
+};
 
 
 static void

src/reader.c

@@ -406,8 +406,8 @@ grammar_midrule_action (void)
      action.  Create the MIDRULE.  */
   location dummy_loc = current_rule->action_props.location;
   symbol *dummy = dummy_symbol_get (dummy_loc);
-  symbol_type_set(dummy,
-                  current_rule->action_props.type, current_rule->action_props.location);
+  symbol_type_set (dummy,
+                   current_rule->action_props.type, current_rule->action_props.location);
   symbol_list *midrule = symbol_list_sym_new (dummy, dummy_loc);
 
   /* Remember named_ref of previous action. */
@@ -815,8 +815,11 @@ check_and_convert_grammar (void)
     grammar = p;
   }
 
-  aver (nsyms <= SYMBOL_NUMBER_MAXIMUM);
-  aver (nsyms == ntokens + nnterms);
+  if (SYMBOL_NUMBER_MAXIMUM - nnterms < ntokens)
+    complain (NULL, fatal, "too many symbols in input grammar (limit is %d)",
+              SYMBOL_NUMBER_MAXIMUM);
+
+  nsyms = ntokens + nnterms;
 
   /* Assign the symbols their symbol numbers.  */
   symbols_pack ();

src/scan-gram.l

@@ -567,6 +567,8 @@ eqopt    ({sp}=)?
               _("POSIX Yacc does not support string literals"));
     RETURN_VALUE (STRING, last_string);
   }
+  <<EOF>>   unexpected_eof (token_start, "\"");
+  "\n"      unexpected_newline (token_start, "\"");
 }
 
 <SC_ESCAPED_TSTRING>
@@ -580,13 +582,10 @@ eqopt    ({sp}=)?
               _("POSIX Yacc does not support string literals"));
     RETURN_VALUE (TSTRING, last_string);
   }
+  <<EOF>>   unexpected_eof (token_start, "\")");
+  "\n"      unexpected_newline (token_start, "\")");
 }
 
-<SC_ESCAPED_STRING,SC_ESCAPED_TSTRING>
-{
-  <<EOF>>   unexpected_eof (token_start, "\"");
-  "\n"      unexpected_newline (token_start, "\"");
-}
 
 
   /*----------------------------------------------------------.
@@ -692,6 +691,15 @@ eqopt    ({sp}=)?
               p);
     STRING_1GROW ('?');
   }
+
+  "\\"             {
+    // None of the other rules matched: the last character of this
+    // file is "\".  But Flex does not support "\\<<EOF>>".
+    unexpected_eof (token_start,
+                    YY_START == SC_ESCAPED_CHARACTER ? "?'"
+                    : YY_START == SC_ESCAPED_STRING ? "?\""
+                    : "?\")");
+  }
 }
 
   /*--------------------------------------------.

src/strversion.c

@@ -0,0 +1,67 @@
+/* Convert version string to int.
+
+   Copyright (C) 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/>.  */
+
+#include <config.h>
+#include "system.h"
+
+#include "strversion.h"
+
+#include <errno.h>
+#include <intprops.h>
+
+int
+strversion_to_int (char const *version)
+{
+  IGNORE_TYPE_LIMITS_BEGIN
+  int res = 0;
+  errno = 0;
+  char *cp = NULL;
+
+  {
+    long major = strtol (version, &cp, 10);
+    if (errno || cp == version || *cp != '.' || major < 0
+        || INT_MULTIPLY_WRAPV (major, 10000, &res))
+      return -1;
+  }
+
+  {
+    ++cp;
+    char *prev = cp;
+    long minor = strtol (cp, &cp, 10);
+    if (errno || cp == prev || (*cp != '\0' && *cp != '.')
+        || ! (0 <= minor && minor < 100)
+        || INT_MULTIPLY_WRAPV (minor, 100, &minor)
+        || INT_ADD_WRAPV (minor, res, &res))
+      return -1;
+  }
+
+  if (*cp == '.')
+    {
+      ++cp;
+      char *prev = cp;
+      long micro = strtol (cp, &cp, 10);
+      if (errno || cp == prev || (*cp != '\0' && *cp != '.')
+          || ! (0 <= micro && micro < 100)
+          || INT_ADD_WRAPV (micro, res, &res))
+        return -1;
+    }
+
+  IGNORE_TYPE_LIMITS_END
+  return res;
+}

src/strversion.h

@@ -0,0 +1,28 @@
+/* Convert version string to int.
+
+   Copyright (C) 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/>.  */
+
+#ifndef STRVERSION_H_
+# define STRVERSION_H_
+
+/* Convert VERSION into an int (MAJOR * 10000 + MINOR * 100 + MICRO).
+   E.g., "3.7.4" => 30704, "3.8" => 30800.
+   Return -1 on errors. */
+int strversion_to_int (char const *version);
+
+#endif

src/symtab.c

@@ -137,11 +137,6 @@ symbol_new (uniqstr tag, location loc)
   res->alias = NULL;
   res->content = sym_content_new (res);
   res->is_alias = false;
-
-  if (nsyms == SYMBOL_NUMBER_MAXIMUM)
-    complain (NULL, fatal, _("too many symbols in input grammar (limit is %d)"),
-              SYMBOL_NUMBER_MAXIMUM);
-  nsyms++;
   return res;
 }
 
@@ -182,11 +177,11 @@ symbol_free (void *ptr)
 */
 
 static void
-symbols_sort (symbol **first, symbol **second)
+symbols_sort (const symbol **first, const symbol **second)
 {
   if (0 < location_cmp ((*first)->location, (*second)->location))
     {
-      symbol* tmp = *first;
+      const symbol* tmp = *first;
       *first = *second;
       *second = tmp;
     }
@@ -243,7 +238,11 @@ semantic_type_new (uniqstr tag, const location *loc)
 | Print a symbol.  |
 `-----------------*/
 
-#define SYMBOL_ATTR_PRINT(Attr)                         \
+#define SYMBOL_INT_ATTR_PRINT(Attr)                     \
+  if (s->content)                                       \
+    fprintf (f, " %s = %d", #Attr, s->content->Attr)
+
+#define SYMBOL_STR_ATTR_PRINT(Attr)                     \
   if (s->content && s->content->Attr)                   \
     fprintf (f, " %s { %s }", #Attr, s->content->Attr)
 
@@ -264,7 +263,11 @@ symbol_print (symbol const *s, FILE *f)
                : c == nterm_sym    ? "nterm"
                : NULL, /* abort.  */
                s->tag);
-      SYMBOL_ATTR_PRINT (type_name);
+      putc (' ', f);
+      location_print (s->location, f);
+      SYMBOL_INT_ATTR_PRINT (code);
+      SYMBOL_INT_ATTR_PRINT (number);
+      SYMBOL_STR_ATTR_PRINT (type_name);
       SYMBOL_CODE_PRINT (destructor);
       SYMBOL_CODE_PRINT (printer);
     }
@@ -371,7 +374,7 @@ symbol_from_uniqstr_fuzzy (const uniqstr key)
 }
 
 static void
-complain_symbol_undeclared (symbol *sym)
+complain_symbol_undeclared (const symbol *sym)
 {
   assert (sym->content->status != declared);
   const symbol *best = symbol_from_uniqstr_fuzzy (sym->tag);
@@ -398,7 +401,10 @@ void
 symbol_location_as_lhs_set (symbol *sym, location loc)
 {
   if (!sym->location_of_lhs)
-    sym->location = loc;
+    {
+      sym->location = loc;
+      sym->location_of_lhs = true;
+    }
 }
 
 
@@ -548,10 +554,6 @@ symbol_class_set (symbol *sym, symbol_class class, location loc, bool declaring)
       if (class == token_sym && s->class == pct_type_sym)
         complain_pct_type_on_token (&sym->location);
 
-      if (class == nterm_sym && s->class != nterm_sym)
-        s->number = nnterms++;
-      else if (class == token_sym && s->number == NUMBER_UNDEFINED)
-        s->number = ntokens++;
       s->class = class;
 
       if (declaring)
@@ -573,9 +575,9 @@ symbol_class_set (symbol *sym, symbol_class class, location loc, bool declaring)
 }
 
 
-/*------------------------------------------------.
-| Set the USER_TOKEN_NUMBER associated with SYM.  |
-`------------------------------------------------*/
+/*----------------------------.
+| Set the token code of SYM.  |
+`----------------------------*/
 
 void
 symbol_code_set (symbol *sym, int code, location loc)
@@ -598,10 +600,6 @@ symbol_code_set (symbol *sym, int code, location loc)
       if (code == 0 && !eoftoken)
         {
           eoftoken = sym->content->symbol;
-          /* It is always mapped to 0, so it was already counted in
-             NTOKENS.  */
-          if (eoftoken->content->number != NUMBER_UNDEFINED)
-            --ntokens;
           eoftoken->content->number = 0;
         }
     }
@@ -621,9 +619,11 @@ symbol_check_defined (symbol *sym)
     {
       complain_symbol_undeclared (sym);
       s->class = nterm_sym;
-      s->number = nnterms++;
     }
 
+  if (s->number == NUMBER_UNDEFINED)
+    s->number = s->class == token_sym ? ntokens++ : nnterms++;
+
   if (s->class == token_sym
       && sym->tag[0] == '"'
       && !sym->is_alias)
@@ -742,7 +742,7 @@ symbol_pack (symbol *sym)
 }
 
 static void
-complain_code_redeclared (int num, symbol *first, symbol *second)
+complain_code_redeclared (int num, const symbol *first, const symbol *second)
 {
   symbols_sort (&first, &second);
   complain (&second->location, complaint,
@@ -758,13 +758,11 @@ complain_code_redeclared (int num, symbol *first, symbol *second)
 `-------------------------------------------------*/
 
 static void
-symbol_translation (symbol *sym)
+symbol_translation (const symbol *sym)
 {
-  /* Nonterminal? */
-  if (sym->content->class == token_sym
-      && !sym->is_alias)
+  if (sym->content->class == token_sym && !sym->is_alias)
     {
-      /* A token which translation has already been set?*/
+      /* A token whose translation has already been set? */
       if (token_translations[sym->content->code]
           != undeftoken->content->number)
         complain_code_redeclared
@@ -969,7 +967,6 @@ dummy_symbol_get (location loc)
   assure (len < sizeof buf);
   symbol *sym = symbol_get (buf, loc);
   sym->content->class = nterm_sym;
-  sym->content->number = nnterms++;
   return sym;
 }
 
@@ -1002,7 +999,7 @@ symbol_cmp (void const *a, void const *b)
 }
 
 /* Store in *SORTED an array of pointers to the symbols contained in
-   TABLE, sorted (alphabetically) by tag. */
+   TABLE, sorted by order of appearance (i.e., by location). */
 
 static void
 table_sort (struct hash_table *table, symbol ***sorted)

src/symtab.h

@@ -227,7 +227,7 @@ void symbol_precedence_set (symbol *sym, int prec, assoc a, location loc);
 void symbol_class_set (symbol *sym, symbol_class class, location loc,
                        bool declaring);
 
-/** Set the \c code associated with \c sym.  */
+/** Set the token \c code of \c sym, specified by the user at \c loc.  */
 void symbol_code_set (symbol *sym, int code, location loc);
 
 

tests/calc.at

@@ -915,6 +915,12 @@ AT_BISON_OPTION_PUSHDEFS([$1])
 
 AT_DATA_CALC_Y([$1])
 AT_FULL_COMPILE(AT_JAVA_IF([[Calc]], [[calc]]), AT_DEFINES_IF([[lex], [main]], [[], []]), [$2], [-Wno-deprecated])
+
+AT_YACC_IF(
+  [# No direct calls to malloc/free.
+  AT_CHECK([[$EGREP '(malloc|free) *\(' calc.[ch] | $EGREP -v 'INFRINGES ON USER NAME SPACE']],
+           [1])])
+
 AT_PUSH_IF([AT_JAVA_IF(
  [# Verify that this is a push parser.
   AT_CHECK_JAVA_GREP([[Calc.java]],

tests/headers.at

@@ -204,17 +204,20 @@ AT_SETUP([Several parsers])
 # Generate and compile to *.o.  Make sure there is no (allowed) YY*
 # nor yy* identifiers in the header after applying api.prefix.  Check
 # that headers can be compiled by a C++ compiler.
+#
+# They should all use parse.assert to make sure that we don't even
+# conflict of YY_ASSERT.
 m4_pushdef([AT_TEST],
-[AT_BISON_OPTION_PUSHDEFS([%define api.prefix {$1_} $2])
+[AT_BISON_OPTION_PUSHDEFS([%define api.prefix {$1_} %define parse.assert $2])
 AT_DATA_GRAMMAR([$1.y],
 [[%define api.prefix {$1_}
+%define parse.assert
 $2
 %define parse.error verbose
-%union
-{
-  int integer;
-}
-%{
+]AT_VARIANT_IF([],
+[%union {int integer;}])[
+
+%code {
 #include <stdio.h> /* printf. */
 ]AT_PUSH_IF([[
 #if defined __GNUC__ && (7 == __GNUC__ || 9 == __GNUC__)
@@ -223,8 +226,10 @@ $2
 ]])[
   ]AT_YYERROR_DECLARE[
   ]AT_YYLEX_DECLARE[
-%}
+}
+
 %%
+
 exp:
   'x' '1' { printf ("x1\n"); }
 | 'x' '2' { printf ("x2\n"); }
@@ -235,9 +240,12 @@ exp:
 | 'x' '7' { printf ("x7\n"); }
 | 'x' '8' { printf ("x8\n"); }
 | 'x' '9' { printf ("x9\n"); }
+| 'x' 'a' { printf ("xa\n"); }
+| 'x' 'b' { printf ("xb\n"); }
 ;
 
 %%
+
 ]AT_YYERROR_DEFINE[
 ]AT_YYLEX_DEFINE(["$1"])[
 ]])
@@ -270,6 +278,8 @@ extern "C"
 #endif
 #include "x5.hh"
 #include "x9.hh"
+#include "xa.hh"
+#include "xb.hh"
 
 #define RUN(S)                                  \
   do {                                          \
@@ -292,6 +302,10 @@ main (void)
   RUN(x8_parse());
   x9_::parser p9;
   RUN(p9.parse());
+  xa_::parser pa;
+  RUN(pa.parse());
+  xb_::parser pb;
+  RUN(pb.parse());
   return 0;
 }
 ]])# main.cc
@@ -304,7 +318,9 @@ AT_TEST([x5], [%locations %debug %language "c++"])
 AT_TEST([x6], [%define api.pure])
 AT_TEST([x7], [%define api.push-pull both])
 AT_TEST([x8], [%define api.pure %define api.push-pull both])
-AT_TEST([x9], [%locations %code requires {#include "location.hh"} %define api.location.type {x5_::location} %debug %language "c++"])
+AT_TEST([x9], [%locations %code requires {#include "location.hh"} %define api.location.type {::x5_::location} %debug %language "c++"])
+AT_TEST([xa], [%locations %code requires {#include "location.hh"} %define api.location.type {::x5_::location} %language "c++" %define api.value.type variant])
+AT_TEST([xb], [%locations %define api.location.file none %language "c++" %define api.value.type variant])
 #AT_TEST([x5], [%locations %language "c++" %glr-parser])
 
 # Check that api.prefix works properly:
@@ -340,6 +356,8 @@ AT_PERL_CHECK([[-n -0777 -e '
       |YY_NULLPTR
       |YY_RVREF
       |YY_\w+_INCLUDED
+      |FILE\ \*yyo   # Function argument.
+      |const\ yylocp # Function argument.
       )\b}{}gx;
   while (/^(.*YY.*)$/gm)
   {
@@ -357,7 +375,7 @@ AT_PERL_CHECK([[-n -0777 -e '
 # Do this late, so that other checks have been performed.
 AT_SKIP_IF_CANNOT_LINK_C_AND_CXX
 
-AT_COMPILE_CXX([parser], [[x[1-9].o -DCC_IS_CXX=$CC_IS_CXX main.cc]])
+AT_COMPILE_CXX([parser], [[x[1-9a-b].o -DCC_IS_CXX=$CC_IS_CXX main.cc]])
 AT_PARSER_CHECK([parser], [0], [[expout]])
 
 m4_popdef([AT_TEST])

tests/input.at

@@ -135,6 +135,10 @@ input.y:9.1-10.0: error: missing '%}' at end of file
 AT_CLEANUP
 
 
+## ------------------------ ##
+## Invalid inputs with {}.  ##
+## ------------------------ ##
+
 AT_SETUP([Invalid inputs with {}])
 
 # We used to SEGV here.  See
@@ -816,6 +820,33 @@ input.y:3.8-10: note: previous declaration
 AT_CLEANUP
 
 
+## ---------------- ##
+## EOF redeclared.  ##
+## ---------------- ##
+
+AT_SETUP([EOF redeclared])
+
+# We used to crash when redefining a token after having defined EOF.
+# See https://lists.gnu.org/r/bug-bison/2020-08/msg00008.html.
+
+AT_DATA([[input.y]],
+[[%token FOO BAR FOO 0
+%%
+input: %empty
+]])
+
+AT_BISON_CHECK([-fcaret input.y], [0], [],
+[[input.y:1.16-18: warning: symbol FOO redeclared [-Wother]
+    1 | %token FOO BAR FOO 0
+      |                ^~~
+input.y:1.8-10: note: previous declaration
+    1 | %token FOO BAR FOO 0
+      |        ^~~
+]])
+
+AT_CLEANUP
+
+
 ## --------------------------- ##
 ## Symbol class redefinition.  ##
 ## --------------------------- ##
@@ -1364,11 +1395,6 @@ AT_CLEANUP
 AT_SETUP([Torturing the Scanner])
 
 AT_BISON_OPTION_PUSHDEFS
-AT_DATA([input.y], [])
-AT_BISON_CHECK([input.y], [1], [],
-[[input.y:1.1: error: unexpected end of file
-]])
-
 
 AT_DATA([input.y],
 [{}
@@ -2475,6 +2501,99 @@ input.y:5.19: error: invalid character after \-escape: \001
 AT_CLEANUP
 
 
+## ------------------------ ##
+## Unexpected end of file.  ##
+## ------------------------ ##
+
+AT_SETUP([[Unexpected end of file]])
+
+
+AT_DATA([input.y], [])
+AT_BISON_CHECK([-fcaret input.y], [1], [],
+[[input.y:1.1: error: unexpected end of file
+]])
+
+
+AT_DATA_NO_FINAL_EOL([char.y],
+[[%token FOO ']])
+
+AT_BISON_CHECK([-fcaret char.y], [1], [],
+[[char.y:1.12: error: missing "'" at end of file
+    1 | %token FOO '
+      |            ^
+char.y:1.12: error: empty character literal
+    1 | %token FOO '
+      |            ^
+]])
+
+
+AT_DATA_NO_FINAL_EOL([escape-in-char.y],
+[[%token FOO '\]])
+
+AT_BISON_CHECK([-fcaret escape-in-char.y], [1], [],
+[[escape-in-char.y:1.12-13: error: missing '?\'' at end of file
+    1 | %token FOO '\
+      |            ^~
+escape-in-char.y:1.14: error: unexpected end of file
+    1 | %token FOO '\
+      |              ^
+]])
+
+
+AT_DATA_NO_FINAL_EOL([string.y],
+[[%token FOO "]])
+
+AT_BISON_CHECK([-fcaret string.y], [1], [],
+[[string.y:1.12: error: missing '"' at end of file
+    1 | %token FOO "
+      |            ^
+string.y:1.13: error: unexpected end of file
+    1 | %token FOO "
+      |             ^
+]])
+
+
+AT_DATA_NO_FINAL_EOL([escape-in-string.y],
+[[%token FOO "\]])
+
+AT_BISON_CHECK([-fcaret escape-in-string.y], [1], [],
+[[escape-in-string.y:1.12-13: error: missing '?"' at end of file
+    1 | %token FOO "\
+      |            ^~
+escape-in-string.y:1.14: error: unexpected end of file
+    1 | %token FOO "\
+      |              ^
+]])
+
+
+AT_DATA_NO_FINAL_EOL([tstring.y],
+[[%token FOO _("]])
+
+AT_BISON_CHECK([-fcaret tstring.y], [1], [],
+[[tstring.y:1.12-14: error: missing '")' at end of file
+    1 | %token FOO _("
+      |            ^~~
+tstring.y:1.15: error: unexpected end of file
+    1 | %token FOO _("
+      |               ^
+]])
+
+
+AT_DATA_NO_FINAL_EOL([escape-in-tstring.y],
+[[%token FOO _("\]])
+
+AT_BISON_CHECK([-fcaret escape-in-tstring.y], [1], [],
+[[escape-in-tstring.y:1.12-15: error: missing '?")' at end of file
+    1 | %token FOO _("\
+      |            ^~~~
+escape-in-tstring.y:1.16: error: unexpected end of file
+    1 | %token FOO _("\
+      |                ^
+]])
+
+AT_CLEANUP
+
+
 ## ------------------------- ##
 ## LAC: Errors for %define.  ##
 ## ------------------------- ##

tests/local.at

@@ -351,7 +351,7 @@ AT_TOKEN_CTOR_IF(
 [m4_pushdef([AT_LOC], [[(]AT_NAME_PREFIX[lloc)]])
  m4_pushdef([AT_VAL], [[(]AT_NAME_PREFIX[lval)]])
  m4_pushdef([AT_YYLEX_FORMALS],     [])
- m4_pushdef([AT_YYLEX_RETURN],      [yy::parser::symbol_type])
+ m4_pushdef([AT_YYLEX_RETURN],      [AT_NAMESPACE::parser::symbol_type])
  m4_pushdef([AT_YYLEX_ARGS],        [])
  m4_pushdef([AT_USE_LEX_ARGS],      [])
  m4_pushdef([AT_YYLEX_PRE_FORMALS], [])
@@ -390,6 +390,7 @@ AT_LOCATION_TYPE_SPAN_IF(
 
 
 AT_GLR_IF([AT_KEYWORDS([glr])])
+AT_PUSH_IF([AT_KEYWORDS([push])])
 ])# _AT_BISON_OPTION_PUSHDEFS
 
 

tests/output.at

@@ -759,7 +759,7 @@ AT_TEST([x1],
          ])
 
 # Check the CPP guard and Doxyen comments.
-AT_CHECK([sed -ne 's/#line [0-9]\+ "/#line "/p;/INCLUDED/p;/\\file/{p;n;p;}' out/include/ast/loc.hh], [],
+AT_CHECK([[sed -ne 's/#line [0-9][0-9]* "/#line "/p;/INCLUDED/p;/\\file/{p;n;p;}' out/include/ast/loc.hh]], [],
 [[ ** \file bar/include/ast/loc.hh
  ** Define the x1::location class.
 #ifndef YY_YY_BAR_INCLUDE_AST_LOC_HH_INCLUDED
@@ -771,7 +771,7 @@ AT_CHECK([sed -ne 's/#line [0-9]\+ "/#line "/p;/INCLUDED/p;/\\file/{p;n;p;}' out
 #endif // !YY_YY_BAR_INCLUDE_AST_LOC_HH_INCLUDED
 ]])
 
-AT_CHECK([sed -ne 's/^#line [0-9]\+ "/#line "/p;/INCLUDED/p;/\\file/{p;n;p;}' out/x1.hh], [],
+AT_CHECK([[sed -ne 's/^#line [0-9][0-9]* "/#line "/p;/INCLUDED/p;/\\file/{p;n;p;}' out/x1.hh]], [],
 [[ ** \file bar/x1.hh
  ** Define the x1::parser class.
 #ifndef YY_YY_BAR_X1_HH_INCLUDED