tests: remove quote magic from the bison test wrapper.

[bison.git] / src / reader.c

diff --git a/src/reader.c b/src/reader.c
index 9e93cab66718bdb2e667c6aa30cb5c89eb8a5656..2e0aa5ee81449787635c1af365d39f63cb2472da 100644 (file)
--- a/src/reader.c
+++ b/src/reader.c
@@ -1,179 +1,159 @@
-/* Input parser for bison
-   Copyright (C) 1984, 1986, 1989, 1992, 1998, 2000, 2001, 2002
-   Free Software Foundation, Inc.
+/* Input parser for Bison
+
+   Copyright (C) 1984, 1986, 1989, 1992, 1998, 2000-2003, 2005-2007,
+   2009-2012 Free Software Foundation, Inc.

 
    This file is part of Bison, the GNU Compiler Compiler.
 
 
    This file is part of Bison, the GNU Compiler Compiler.
 

-   Bison is free software; you can redistribute it and/or modify
+   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
    it under the terms of the GNU General Public License as published by

-   the Free Software Foundation; either version 2, or (at your option)
-   any later version.
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.

 
 

-   Bison is distributed in the hope that it will be useful,
+   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
    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 Bison; see the file COPYING.  If not, write to
-   the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
-   Boston, MA 02111-1307, USA.  */
-
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */

 
 

+#include <config.h>

 #include "system.h"
 #include "system.h"

-#include "quotearg.h"
-#include "quote.h"
-#include "getargs.h"
+
+#include <quote.h>
+
+#include "complain.h"
+#include "conflicts.h"

 #include "files.h"
 #include "files.h"

-#include "symtab.h"
-#include "options.h"
+#include "getargs.h"

 #include "gram.h"
 #include "gram.h"

-#include "complain.h"
-#include "output.h"
+#include "muscle-tab.h"

 #include "reader.h"
 #include "reader.h"

-#include "conflicts.h"
-#include "muscle_tab.h"
+#include "symlist.h"
+#include "symtab.h"
+#include "scan-gram.h"
+#include "scan-code.h"
+
+static void prepare_percent_define_front_end_variables (void);
+static void check_and_convert_grammar (void);

 
 

-int lineno;

 static symbol_list *grammar = NULL;
 static symbol_list *grammar = NULL;

-static int start_flag = 0;
+static bool start_flag = false;
+merger_list *merge_functions;

 
 

-/* Nonzero if %union has been seen.  */
-int typed = 0;
+/* Was %union seen?  */
+bool union_seen = false;

 
 

-static symbol_list *
-symbol_list_new (symbol_t *sym)
-{
-  symbol_list *res = XMALLOC (symbol_list, 1);
-  res->next = NULL;
-  res->sym = sym;
-  res->line = lineno;
-  res->action = NULL;
-  res->action_line = 0;
-  res->ruleprec = NULL;
-  return res;
-}
+/* Was a tag seen?  */
+bool tag_seen = false;

 
 

+/* Should rules have a default precedence?  */
+bool default_prec = true;

 \f
 \f

-/*--------------------------------------------------------------.
-| Get the data type (alternative in the union) of the value for |
-| symbol N in rule RULE.                                        |
-`--------------------------------------------------------------*/
-
-char *
-get_type_name (int n, symbol_list *rule)
-{
-  int i;
-  symbol_list *rp;
-
-  if (n < 0)
-    {
-      complain (_("invalid $ value"));
-      return NULL;
-    }
-
-  rp = rule;
-  i = 0;
-
-  while (i < n)
-    {
-      rp = rp->next;
-      if (rp == NULL || rp->sym == NULL)
-       {
-         complain (_("invalid $ value"));
-         return NULL;
-       }
-      ++i;
-    }
-
-  return rp->sym->type_name;
-}
-
-

 /*-----------------------.
 | Set the start symbol.  |
 `-----------------------*/
 
 void
 /*-----------------------.
 | Set the start symbol.  |
 `-----------------------*/
 
 void

-grammar_start_symbol_set (symbol_t *s)
+grammar_start_symbol_set (symbol *sym, location loc)

 {
   if (start_flag)
 {
   if (start_flag)

-    complain (_("multiple %s declarations"), "%start");
+    complain_at (loc, _("multiple %s declarations"), "%start");

   else
     {
   else
     {

-      start_flag = 1;
-      startsymbol = s;
+      start_flag = true;
+      startsymbol = sym;
+      startsymbol_location = loc;

     }
 }
 
     }
 }
 

+\f

 
 

-/*----------------------------------------------------------------.
-| There are two prologues: one before %union, one after.  Augment |
-| the current one.                                                |
-`----------------------------------------------------------------*/
+/*------------------------------------------------------------------------.
+| Return the merger index for a merging function named NAME.  Records the |
+| function, if new, in MERGER_LIST.                                       |
+`------------------------------------------------------------------------*/

 
 

-void
-prologue_augment (const char *prologue, int location)
+static int
+get_merge_function (uniqstr name)

 {
 {

-  struct obstack *oout =
-    !typed ? &pre_prologue_obstack : &post_prologue_obstack;
-
-  if (!no_lines_flag)
+  merger_list *syms;
+  merger_list head;
+  int n;
+
+  if (! glr_parser)
+    return 0;
+
+  head.next = merge_functions;
+  for (syms = &head, n = 1; syms->next; syms = syms->next, n += 1)
+    if (UNIQSTR_EQ (name, syms->next->name))
+      break;
+  if (syms->next == NULL)

     {
     {

-      obstack_fgrow2 (oout, muscle_find ("linef"),
-                     location, quotearg_style (c_quoting_style,
-                                               muscle_find ("filename")));
+      syms->next = xmalloc (sizeof syms->next[0]);
+      syms->next->name = uniqstr_new (name);
+      /* After all symbol type declarations have been parsed, packgram invokes
+         record_merge_function_type to set the type.  */
+      syms->next->type = NULL;
+      syms->next->next = NULL;
+      merge_functions = head.next;

     }
     }

-  obstack_sgrow (oout, prologue);
+  return n;

 }
 
 }
 

+/*-------------------------------------------------------------------------.
+| For the existing merging function with index MERGER, record the result   |
+| type as TYPE as required by the lhs of the rule whose %merge declaration |
+| is at DECLARATION_LOC.                                                   |
+`-------------------------------------------------------------------------*/

 
 

+static void
+record_merge_function_type (int merger, uniqstr type, location declaration_loc)
+{
+  int merger_find;
+  merger_list *merge_function;

 
 

+  if (merger <= 0)
+    return;

 
 

-/*----------------------.
-| Handle the epilogue.  |
-`----------------------*/
+  if (type == NULL)
+    type = uniqstr_new ("");

 
 

-void
-epilogue_set (const char *epilogue, int location)
-{
-  struct obstack el_obstack;
-  obstack_init (&el_obstack);
-
-  if (!no_lines_flag)
+  merger_find = 1;
+  for (merge_function = merge_functions;
+       merge_function != NULL && merger_find != merger;
+       merge_function = merge_function->next)
+    merger_find += 1;
+  aver (merge_function != NULL && merger_find == merger);
+  if (merge_function->type != NULL && !UNIQSTR_EQ (merge_function->type, type))

     {
     {

-      obstack_fgrow2 (&el_obstack, muscle_find ("linef"),
-                     location, quotearg_style (c_quoting_style,
-                                               muscle_find ("filename")));
+      complain_at (declaration_loc,
+                   _("result type clash on merge function %s: <%s> != <%s>"),
+                   quote (merge_function->name), type, merge_function->type);
+      complain_at (merge_function->type_declaration_location,
+                   _("previous declaration"));

     }
     }

-  obstack_sgrow (&el_obstack, epilogue);
-  obstack_1grow (&el_obstack, 0);
-  muscle_insert ("epilogue", obstack_finish (&el_obstack));
+  merge_function->type = uniqstr_new (type);
+  merge_function->type_declaration_location = declaration_loc;

 }
 
 }
 

+/*--------------------------------------.
+| Free all merge-function definitions.  |
+`--------------------------------------*/

 
 

-\f
-
-/*-------------------------------------------------------------------.
-| Generate a dummy symbol, a nonterminal, whose name cannot conflict |
-| with the user's names.                                             |
-`-------------------------------------------------------------------*/
-
-static symbol_t *
-gensym (void)
+void
+free_merger_functions (void)

 {
 {

-  /* Incremented for each generated symbol */
-  static int gensym_count = 0;
-  static char buf[256];
-
-  symbol_t *sym;
-
-  sprintf (buf, "@%d", ++gensym_count);
-  sym = getsym (buf);
-  sym->class = nterm_sym;
-  sym->number = nvars++;
-  return sym;
+  merger_list *L0 = merge_functions;
+  while (L0)
+    {
+      merger_list *L1 = L0->next;
+      free (L0);
+      L0 = L1;
+    }

 }
 }

+

 \f
 /*-------------------------------------------------------------------.
 | Parse the input grammar into a one symbol_list structure.  Each    |
 \f
 /*-------------------------------------------------------------------.
 | Parse the input grammar into a one symbol_list structure.  Each    |


@@ -184,23 +164,16 @@ gensym (void)
 |                                                                    |
 | All actions are copied out, labelled by the rule number they apply |
 | to.                                                                |
 |                                                                    |
 | All actions are copied out, labelled by the rule number they apply |
 | to.                                                                |

-|                                                                    |
-| Bison used to allow some %directives in the rules sections, but    |
-| this is no longer consider appropriate: (i) the documented grammar |
-| doesn't claim it, (ii), it would promote bad style, (iii), error   |
-| recovery for %directives consists in skipping the junk until a `%' |
-| is seen and helrp synchronizing.  This scheme is definitely wrong  |
-| in the rules section.                                              |

 `-------------------------------------------------------------------*/
 
 /* The (currently) last symbol of GRAMMAR. */
 `-------------------------------------------------------------------*/
 
 /* The (currently) last symbol of GRAMMAR. */

-symbol_list *grammar_end = NULL;
+static symbol_list *grammar_end = NULL;

 
 

-/* Append S to the GRAMMAR. */
-void
-grammar_symbol_append (symbol_t *s)
+/* Append SYM to the grammar.  */
+static symbol_list *
+grammar_symbol_append (symbol *sym, location loc)

 {
 {

-  symbol_list *p = symbol_list_new (s);
+  symbol_list *p = symbol_list_sym_new (sym, loc);

 
   if (grammar_end)
     grammar_end->next = p;
 
   if (grammar_end)
     grammar_end->next = p;


@@ -208,35 +181,60 @@ grammar_symbol_append (symbol_t *s)
     grammar = p;
 
   grammar_end = p;
     grammar = p;
 
   grammar_end = p;

+
+  /* A null SYM stands for an end of rule; it is not an actual
+     part of it.  */
+  if (sym)
+    ++nritems;
+
+  return p;

 }
 
 }
 

-/* The rule currently being defined, and the previous rule.  Point to
-   the first symbol of each list: their lhs.  */
-symbol_list *current_rule = NULL;
-symbol_list *previous_rule = NULL;
+static void
+assign_named_ref (symbol_list *p, named_ref *name)
+{
+  symbol *sym = p->content.sym;

 
 

+  if (name->id == sym->tag)
+    {
+      warn_at (name->loc,
+               _("duplicated symbol name for %s ignored"),
+               quote (sym->tag));
+      named_ref_free (name);
+    }
+  else
+    p->named_ref = name;
+}
+
+
+/* The rule currently being defined, and the previous rule.
+   CURRENT_RULE points to the first LHS of the current rule, while
+   PREVIOUS_RULE_END points to the *end* of the previous rule (NULL).  */
+static symbol_list *current_rule = NULL;
+static symbol_list *previous_rule_end = NULL;

 
 

-/* Create a new rule for LHS in to the GRAMMAR. */
+
+/*----------------------------------------------.
+| Create a new rule for LHS in to the GRAMMAR.  |
+`----------------------------------------------*/

 
 void
 
 void

-grammar_rule_begin (symbol_t *lhs)
+grammar_current_rule_begin (symbol *lhs, location loc,
+                            named_ref *lhs_name)

 {
 {

-  if (!start_flag)
-    {
-      startsymbol = lhs;
-      start_flag = 1;
-    }
+  symbol_list* p;

 
   /* Start a new rule and record its lhs.  */
   ++nrules;
 
   /* Start a new rule and record its lhs.  */
   ++nrules;

-  ++nritems;
+  previous_rule_end = grammar_end;
+
+  p = grammar_symbol_append (lhs, loc);
+  if (lhs_name)
+    assign_named_ref (p, named_ref_copy (lhs_name));

 
 

-  previous_rule = grammar_end;
-  grammar_symbol_append (lhs);

   current_rule = grammar_end;
 
   /* Mark the rule's lhs as a nonterminal if not already so.  */
   current_rule = grammar_end;
 
   /* Mark the rule's lhs as a nonterminal if not already so.  */

-

   if (lhs->class == unknown_sym)
     {
       lhs->class = nterm_sym;
   if (lhs->class == unknown_sym)
     {
       lhs->class = nterm_sym;


@@ -244,57 +242,122 @@ grammar_rule_begin (symbol_t *lhs)
       ++nvars;
     }
   else if (lhs->class == token_sym)
       ++nvars;
     }
   else if (lhs->class == token_sym)

-    complain (_("rule given for %s, which is a token"), lhs->tag);
+    complain_at (loc, _("rule given for %s, which is a token"), lhs->tag);

 }
 
 }
 

-/* Check that the last rule (CURRENT_RULE) is properly defined.  For
-   instance, there should be no type clash on the default action.  */
+
+/*----------------------------------------------------------------------.
+| A symbol should be used if either:                                    |
+|   1. It has a destructor.                                             |
+|   2. The symbol is a mid-rule symbol (i.e., the generated LHS         |
+|      replacing a mid-rule action) that was assigned to or used, as in |
+|      "exp: { $$ = 1; } { $$ = $1; }".                                 |
+`----------------------------------------------------------------------*/
+
+static bool
+symbol_should_be_used (symbol_list const *s, bool *midrule_warning)
+{
+  if (symbol_destructor_get (s->content.sym)->code)
+    return true;
+  if ((s->midrule && s->midrule->action_props.is_value_used)
+      || (s->midrule_parent_rule
+          && symbol_list_n_get (s->midrule_parent_rule,
+                                s->midrule_parent_rhs_index)
+               ->action_props.is_value_used))
+    {
+      *midrule_warning = true;
+      return true;
+    }
+  return false;
+}
+
+/*----------------------------------------------------------------.
+| Check that the rule R is properly defined.  For instance, there |
+| should be no type clash on the default action.                  |
+`----------------------------------------------------------------*/

 
 static void
 
 static void

-grammar_current_rule_check (void)
+grammar_rule_check (const symbol_list *r)

 {
 {

-  symbol_t *lhs = current_rule->sym;
-  symbol_t *first_rhs = current_rule->next->sym;
+  /* Type check.

 
 

-  /* If there is an action, then there is nothing we can do: the user
-     is allowed to shoot in her foot.  */
-  if (current_rule->action)
-    return;
+     If there is an action, then there is nothing we can do: the user
+     is allowed to shoot herself in the foot.

 
 

-  /* If $$ is being set in default way, report if any type mismatch.
-     */
-  if (first_rhs)
+     Don't worry about the default action if $$ is untyped, since $$'s
+     value can't be used.  */
+  if (!r->action_props.code && r->content.sym->type_name)

     {
     {

-      const char *lhs_type = lhs->type_name       ? lhs->type_name       : "";
-      const char *rhs_type = first_rhs->type_name ? first_rhs->type_name : "";
-      if (strcmp (lhs_type, rhs_type))
-       complain (_("type clash (`%s' `%s') on default action"),
-                 lhs_type, rhs_type);
-    }
-  /* Warn if there is no default for $$ but we need one.  */
-  else
-    {
-      if (lhs->type_name)
-       complain (_("empty rule for typed nonterminal, and no action"));
+      symbol *first_rhs = r->next->content.sym;
+      /* If $$ is being set in default way, report if any type mismatch.  */
+      if (first_rhs)
+        {
+          char const *lhs_type = r->content.sym->type_name;
+          const char *rhs_type =
+            first_rhs->type_name ? first_rhs->type_name : "";
+          if (!UNIQSTR_EQ (lhs_type, rhs_type))
+            warn_at (r->location,
+                     _("type clash on default action: <%s> != <%s>"),
+                     lhs_type, rhs_type);
+        }
+      /* Warn if there is no default for $$ but we need one.  */
+      else
+        warn_at (r->location,
+                 _("empty rule for typed nonterminal, and no action"));

     }
     }

+
+  /* Check that symbol values that should be used are in fact used.  */
+  {
+    symbol_list const *l = r;
+    int n = 0;
+    for (; l && l->content.sym; l = l->next, ++n)
+      {
+        bool midrule_warning = false;
+        if (!l->action_props.is_value_used
+            && symbol_should_be_used (l, &midrule_warning)
+            /* The default action, $$ = $1, `uses' both.  */
+            && (r->action_props.code || (n != 0 && n != 1)))
+          {
+            void (*warn_at_ptr)(location, char const*, ...) =
+              midrule_warning ? midrule_value_at : warn_at;
+            if (n)
+              warn_at_ptr (r->location, _("unused value: $%d"), n);
+            else
+              warn_at_ptr (r->location, _("unset value: $$"));
+          }
+      }
+  }
+
+  /* See comments in grammar_current_rule_prec_set for how POSIX
+     mandates this complaint.  It's only for identifiers, so skip
+     it for char literals and strings, which are always tokens.  */
+  if (r->ruleprec
+      && r->ruleprec->tag[0] != '\'' && r->ruleprec->tag[0] != '"'
+      && !r->ruleprec->declared && !r->ruleprec->prec)
+    warn_at (r->location, _("token for %%prec is not defined: %s"),
+             r->ruleprec->tag);

 }
 
 
 }
 
 

-/* End the currently being grown rule. */
+/*-------------------------------------.
+| End the currently being grown rule.  |
+`-------------------------------------*/

 
 void
 
 void

-grammar_rule_end (void)
+grammar_current_rule_end (location loc)

 {
   /* Put an empty link in the list to mark the end of this rule  */
 {
   /* Put an empty link in the list to mark the end of this rule  */

-  grammar_symbol_append (NULL);
-  grammar_current_rule_check ();
+  grammar_symbol_append (NULL, grammar_end->location);
+  current_rule->location = loc;

 }
 
 
 }
 
 

-/* The previous action turns out the be a mid-rule action.  Attach it
-   to the current rule, i.e., create a dummy symbol, attach it this
-   mid-rule action, and append this dummy nonterminal to the current
-   rule.  */
+/*-------------------------------------------------------------------.
+| The previous action turns out the be a mid-rule action.  Attach it |
+| to the current rule, i.e., create a dummy symbol, attach it this   |
+| mid-rule action, and append this dummy nonterminal to the current  |
+| rule.                                                              |
+`-------------------------------------------------------------------*/

 
 void
 grammar_midrule_action (void)
 
 void
 grammar_midrule_action (void)


@@ -303,70 +366,124 @@ grammar_midrule_action (void)
      give the new rule this number by inserting the new rule before
      it.  */
 
      give the new rule this number by inserting the new rule before
      it.  */
 

-  /* Make a dummy nonterminal, a gensym.  */
-  symbol_t *sdummy = gensym ();
-  symbol_list *midrule_action = symbol_list_new (sdummy);
+  /* Make a DUMMY nonterminal, whose location is that of the midrule
+     action.  Create the MIDRULE.  */
+  location dummy_location = current_rule->action_props.location;
+  symbol *dummy = dummy_symbol_get (dummy_location);
+  symbol_list *midrule = symbol_list_sym_new (dummy, dummy_location);
+
+  /* Remember named_ref of previous action. */
+  named_ref *action_name = current_rule->action_props.named_ref;

 
   /* Make a new rule, whose body is empty, before the current one, so
      that the action just read can belong to it.  */
   ++nrules;
   ++nritems;
 
   /* Make a new rule, whose body is empty, before the current one, so
      that the action just read can belong to it.  */
   ++nrules;
   ++nritems;

-  /* Attach its lineno to that of the host rule.  */
-  midrule_action->line = current_rule->line;
-  /* Move the action from the host rule to this one.  */
-  midrule_action->action = current_rule->action;
-  midrule_action->action_line = current_rule->action_line;
-  current_rule->action = NULL;
-
-  if (previous_rule)
-    previous_rule->next = midrule_action;
+  /* Attach its location and actions to that of the DUMMY.  */
+  midrule->location = dummy_location;
+  code_props_rule_action_init (&midrule->action_props,
+                               current_rule->action_props.code,
+                               current_rule->action_props.location,
+                               midrule, 0,
+                               current_rule->action_props.is_predicate);
+  code_props_none_init (&current_rule->action_props);
+
+  if (previous_rule_end)
+    previous_rule_end->next = midrule;

   else
   else

-    grammar = midrule_action;
+    grammar = midrule;

 
 

-  /* End of the rule. */
-  previous_rule = symbol_list_new (NULL);
-  previous_rule->next = current_rule;
+  /* End the dummy's rule.  */
+  midrule->next = symbol_list_sym_new (NULL, dummy_location);
+  midrule->next->next = current_rule;

 
 

-  midrule_action->next = previous_rule;
+  previous_rule_end = midrule->next;

 
 

-  /* Insert the dummy generated by that rule into this rule.  */
-  ++nritems;
-  grammar_symbol_append (sdummy);
+  /* Insert the dummy nonterminal replacing the midrule action into
+     the current rule.  Bind it to its dedicated rule.  */
+  grammar_current_rule_symbol_append (dummy, dummy_location,
+                                      action_name);
+  grammar_end->midrule = midrule;
+  midrule->midrule_parent_rule = current_rule;
+  midrule->midrule_parent_rhs_index = symbol_list_length (current_rule->next);

 }
 
 /* Set the precedence symbol of the current rule to PRECSYM. */
 
 void
 }
 
 /* Set the precedence symbol of the current rule to PRECSYM. */
 
 void

-grammar_current_rule_prec_set (symbol_t *precsym)
+grammar_current_rule_prec_set (symbol *precsym, location loc)

 {
 {

+  /* POSIX says that any identifier is a nonterminal if it does not
+     appear on the LHS of a grammar rule and is not defined by %token
+     or by one of the directives that assigns precedence to a token.  We
+     ignore this here because the only kind of identifier that POSIX
+     allows to follow a %prec is a token and because assuming it's a
+     token now can produce more logical error messages.  Nevertheless,
+     grammar_rule_check does obey what we believe is the real intent of
+     POSIX here: that an error be reported for any identifier that
+     appears after %prec but that is not defined separately as a
+     token.  */
+  symbol_class_set (precsym, token_sym, loc, false);

   if (current_rule->ruleprec)
   if (current_rule->ruleprec)

-    complain (_("two @prec's in a row"));
+    complain_at (loc, _("only one %s allowed per rule"), "%prec");

   current_rule->ruleprec = precsym;
 }
 
   current_rule->ruleprec = precsym;
 }
 

-/* Attach a SYMBOL to the current rule.  If needed, move the previous
-   action as a mid-rule action.  */
+/* Attach dynamic precedence DPREC to the current rule. */

 
 void
 
 void

-grammar_current_rule_symbol_append (symbol_t *symbol)
+grammar_current_rule_dprec_set (int dprec, location loc)

 {
 {

-  if (current_rule->action)
-    grammar_midrule_action ();
-  ++nritems;
-  grammar_symbol_append (symbol);
+  if (! glr_parser)
+    warn_at (loc, _("%s affects only GLR parsers"), "%dprec");
+  if (dprec <= 0)
+    complain_at (loc, _("%s must be followed by positive number"), "%dprec");
+  else if (current_rule->dprec != 0)
+    complain_at (loc, _("only one %s allowed per rule"), "%dprec");
+  current_rule->dprec = dprec;

 }
 
 }
 

+/* Attach a merge function NAME with argument type TYPE to current
+   rule. */
+
+void
+grammar_current_rule_merge_set (uniqstr name, location loc)
+{
+  if (! glr_parser)
+    warn_at (loc, _("%s affects only GLR parsers"), "%merge");
+  if (current_rule->merger != 0)
+    complain_at (loc, _("only one %s allowed per rule"), "%merge");
+  current_rule->merger = get_merge_function (name);
+  current_rule->merger_declaration_location = loc;
+}

 
 

-/* Attach an ACTION to the current rule.  If needed, move the previous
+/* Attach SYM to the current rule.  If needed, move the previous

    action as a mid-rule action.  */
 
 void
    action as a mid-rule action.  */
 
 void

-grammar_current_rule_action_append (const char *action, int action_line)
+grammar_current_rule_symbol_append (symbol *sym, location loc,
+                                    named_ref *name)
+{
+  symbol_list *p;
+  if (current_rule->action_props.code)
+    grammar_midrule_action ();
+  p = grammar_symbol_append (sym, loc);
+  if (name)
+    assign_named_ref(p, name);
+}
+
+/* Attach an ACTION to the current rule.  */
+
+void
+grammar_current_rule_action_append (const char *action, location loc,
+                                    named_ref *name, bool is_predicate)

 {
 {

-  if (current_rule->action)
+  if (current_rule->action_props.code)

     grammar_midrule_action ();
     grammar_midrule_action ();

-  current_rule->action = action;
-  current_rule->action_line = action_line;
+  /* After all symbol declarations have been parsed, packgram invokes
+     code_props_translate_code.  */
+  code_props_rule_action_init (&current_rule->action_props, action, loc,
+                               current_rule, name, is_predicate);

 }
 
 \f
 }
 
 \f


@@ -378,60 +495,92 @@ grammar_current_rule_action_append (const char *action, int action_line)
 static void
 packgram (void)
 {
 static void
 packgram (void)
 {

-  unsigned int itemno;
-  int ruleno;
-  symbol_list *p;
+  unsigned int itemno = 0;
+  rule_number ruleno = 0;
+  symbol_list *p = grammar;

 
 

-  ritem = XCALLOC (item_number_t, nritems);
-  rules = XCALLOC (rule_t, nrules) - 1;
+  ritem = xnmalloc (nritems + 1, sizeof *ritem);

 
 

-  itemno = 0;
-  ruleno = 1;
+  /* This sentinel is used by build_relations in gram.c.  */
+  *ritem++ = 0;
+
+  rules = xnmalloc (nrules, sizeof *rules);

 
 

-  p = grammar;

   while (p)
     {
   while (p)
     {

-      symbol_t *ruleprec = p->ruleprec;
+      int rule_length = 0;
+      symbol *ruleprec = p->ruleprec;
+      record_merge_function_type (p->merger, p->content.sym->type_name,
+                                  p->merger_declaration_location);

       rules[ruleno].user_number = ruleno;
       rules[ruleno].number = ruleno;
       rules[ruleno].user_number = ruleno;
       rules[ruleno].number = ruleno;

-      rules[ruleno].lhs = p->sym;
+      rules[ruleno].lhs = p->content.sym;

       rules[ruleno].rhs = ritem + itemno;
       rules[ruleno].rhs = ritem + itemno;

-      rules[ruleno].line = p->line;
-      rules[ruleno].useful = TRUE;
-      rules[ruleno].action = p->action;
-      rules[ruleno].action_line = p->action_line;
-
-      p = p->next;
-      while (p && p->sym)
-       {
-         /* item_number_t = symbol_number_t.
-            But the former needs to contain more: negative rule numbers. */
-         ritem[itemno++] = symbol_number_as_item_number (p->sym->number);
-         /* A rule gets by default the precedence and associativity
-            of the last token in it.  */
-         if (p->sym->class == token_sym)
-           rules[ruleno].prec = p->sym;
-         if (p)
-           p = p->next;
-       }
+      rules[ruleno].prec = NULL;
+      rules[ruleno].dprec = p->dprec;
+      rules[ruleno].merger = p->merger;
+      rules[ruleno].precsym = NULL;
+      rules[ruleno].location = p->location;
+      rules[ruleno].useful = true;
+      rules[ruleno].action = p->action_props.code;
+      rules[ruleno].action_location = p->action_props.location;
+      rules[ruleno].is_predicate = p->action_props.is_predicate;
+
+      /* If the midrule's $$ is set or its $n is used, remove the `$' from the
+         symbol name so that it's a user-defined symbol so that the default
+         %destructor and %printer apply.  */
+      if (p->midrule_parent_rule
+          && (p->action_props.is_value_used
+              || symbol_list_n_get (p->midrule_parent_rule,
+                                    p->midrule_parent_rhs_index)
+                   ->action_props.is_value_used))
+        p->content.sym->tag += 1;
+
+      /* Don't check the generated rule 0.  It has no action, so some rhs
+         symbols may appear unused, but the parsing algorithm ensures that
+         %destructor's are invoked appropriately.  */
+      if (p != grammar)
+        grammar_rule_check (p);
+
+      for (p = p->next; p && p->content.sym; p = p->next)
+        {
+          ++rule_length;
+
+          /* Don't allow rule_length == INT_MAX, since that might
+             cause confusion with strtol if INT_MAX == LONG_MAX.  */
+          if (rule_length == INT_MAX)
+              fatal_at (rules[ruleno].location, _("rule is too long"));
+
+          /* item_number = symbol_number.
+             But the former needs to contain more: negative rule numbers. */
+          ritem[itemno++] =
+            symbol_number_as_item_number (p->content.sym->number);
+          /* A rule gets by default the precedence and associativity
+             of its last token.  */
+          if (p->content.sym->class == token_sym && default_prec)
+            rules[ruleno].prec = p->content.sym;
+        }

 
       /* If this rule has a %prec,
          the specified symbol's precedence replaces the default.  */
       if (ruleprec)
 
       /* If this rule has a %prec,
          the specified symbol's precedence replaces the default.  */
       if (ruleprec)

-       {
-         rules[ruleno].precsym = ruleprec;
-         rules[ruleno].prec = ruleprec;
-       }
-      ritem[itemno++] = -ruleno;
+        {
+          rules[ruleno].precsym = ruleprec;
+          rules[ruleno].prec = ruleprec;
+        }
+      /* An item ends by the rule number (negated).  */
+      ritem[itemno++] = rule_number_as_item_number (ruleno);
+      aver (itemno < ITEM_NUMBER_MAX);

       ++ruleno;
       ++ruleno;

+      aver (ruleno < RULE_NUMBER_MAX);

 
       if (p)
 
       if (p)

-       p = p->next;
+        p = p->next;

     }
 
     }
 

-  assert (itemno == nritems);
+  aver (itemno == nritems);

 
 

-  if (trace_flag)
+  if (trace_flag & trace_sets)

     ritem_print (stderr);
 }
 \f
     ritem_print (stderr);
 }
 \f


@@ -445,92 +594,154 @@ packgram (void)
 void
 reader (void)
 {
 void
 reader (void)
 {

-  gram_control_t gram_control;
-  lineno = 1;
-
-  /* Initialize the muscle obstack.  */
-  obstack_init (&muscle_obstack);
-

   /* Initialize the symbol table.  */
   symbols_new ();
 
   /* Initialize the symbol table.  */
   symbols_new ();
 

-  /* Construct the axiom symbol. */
-  axiom = getsym ("$axiom");
-  axiom->class = nterm_sym;
-  axiom->number = nvars++;
+  /* Construct the accept symbol. */
+  accept = symbol_get ("$accept", empty_location);
+  accept->class = nterm_sym;
+  accept->number = nvars++;

 
   /* Construct the error token */
 
   /* Construct the error token */

-  errtoken = getsym ("error");
+  errtoken = symbol_get ("error", empty_location);

   errtoken->class = token_sym;
   errtoken->number = ntokens++;
 
   /* Construct a token that represents all undefined literal tokens.
      It is always token number 2.  */
   errtoken->class = token_sym;
   errtoken->number = ntokens++;
 
   /* Construct a token that represents all undefined literal tokens.
      It is always token number 2.  */

-  undeftoken = getsym ("$undefined.");
+  undeftoken = symbol_get ("$undefined", empty_location);

   undeftoken->class = token_sym;
   undeftoken->number = ntokens++;
 
   undeftoken->class = token_sym;
   undeftoken->number = ntokens++;
 

-  /* Initialize the obstacks. */
-  obstack_init (&action_obstack);
-  obstack_init (&output_obstack);
-  obstack_init (&pre_prologue_obstack);
-  obstack_init (&post_prologue_obstack);
+  gram_in = xfopen (grammar_file, "r");
+
+  gram__flex_debug = trace_flag & trace_scan;
+  gram_debug = trace_flag & trace_parse;
+  gram_scanner_initialize ();
+  gram_parse ();
+  prepare_percent_define_front_end_variables ();
+
+  if (! complaint_issued)
+    check_and_convert_grammar ();
+
+  xfclose (gram_in);
+}
+
+static void
+prepare_percent_define_front_end_variables (void)
+{
+  /* Set %define front-end variable defaults.  */
+  muscle_percent_define_default ("lr.keep-unreachable-states", "false");
+  {
+    char *lr_type;
+    /* IELR would be a better default, but LALR is historically the
+       default.  */
+    muscle_percent_define_default ("lr.type", "lalr");
+    lr_type = muscle_percent_define_get ("lr.type");
+    if (STRNEQ (lr_type, "canonical-lr"))
+      muscle_percent_define_default ("lr.default-reductions", "most");
+    else
+      muscle_percent_define_default ("lr.default-reductions", "accepting");
+    free (lr_type);
+  }
+
+  /* Check %define front-end variables.  */
+  {
+    static char const * const values[] = {
+      "lr.type", "lalr", "ielr", "canonical-lr", NULL,
+      "lr.default-reductions", "most", "consistent", "accepting", NULL,
+      NULL
+    };
+    muscle_percent_define_check_values (values);
+  }
+}

 
 

-  finput = xfopen (infile, "r");
-  gram_in = finput;

 
 

-  gram_debug = !!getenv ("parse");
-  gram__flex_debug = !!getenv ("scan");
-  gram_parse (&gram_control);
+/*-------------------------------------------------------------.
+| Check the grammar that has just been read, and convert it to |
+| internal form.                                               |
+`-------------------------------------------------------------*/

 
 

-  /* Grammar has been read.  Do some checking */
+static void
+check_and_convert_grammar (void)
+{
+  /* Grammar has been read.  Do some checking.  */

   if (nrules == 0)
     fatal (_("no rules in the input grammar"));
 
   if (nrules == 0)
     fatal (_("no rules in the input grammar"));
 

+  /* If the user did not define her ENDTOKEN, do it now. */
+  if (!endtoken)
+    {
+      endtoken = symbol_get ("$end", empty_location);
+      endtoken->class = token_sym;
+      endtoken->number = 0;
+      /* Value specified by POSIX.  */
+      endtoken->user_token_number = 0;
+    }
+

   /* Report any undefined symbols and consider them nonterminals.  */
   symbols_check_defined ();
 
   /* Report any undefined symbols and consider them nonterminals.  */
   symbols_check_defined ();
 

-  /* If the user did not define her EOFTOKEN, do it now. */
-  if (!eoftoken)
+  /* Find the start symbol if no %start.  */
+  if (!start_flag)

     {
     {

-      eoftoken = getsym ("$");
-      eoftoken->class = token_sym;
-      eoftoken->number = 0;
-      /* Value specified by POSIX.  */
-      eoftoken->user_token_number = 0;
+      symbol_list *node;
+      for (node = grammar;
+           node != NULL && symbol_is_dummy (node->content.sym);
+           node = node->next)
+        {
+          for (node = node->next;
+               node != NULL && node->content.sym != NULL;
+               node = node->next)
+            ;
+        }
+      aver (node != NULL);
+      grammar_start_symbol_set (node->content.sym,
+                                node->content.sym->location);

     }
 
     }
 

-  /* Insert the initial rule, which line is that of the first rule
+  /* Insert the initial rule, whose line is that of the first rule

      (not that of the start symbol):
 
      (not that of the start symbol):
 

-     axiom: %start EOF.  */
+     accept: %start EOF.  */

   {
   {

-    symbol_list *p = symbol_list_new (axiom);
-    p->line = grammar->line;
-    p->next = symbol_list_new (startsymbol);
-    p->next->next = symbol_list_new (eoftoken);
-    p->next->next->next = symbol_list_new (NULL);
+    symbol_list *p = symbol_list_sym_new (accept, empty_location);
+    p->location = grammar->location;
+    p->next = symbol_list_sym_new (startsymbol, empty_location);
+    p->next->next = symbol_list_sym_new (endtoken, empty_location);
+    p->next->next->next = symbol_list_sym_new (NULL, empty_location);

     p->next->next->next->next = grammar;
     nrules += 1;
     nritems += 3;
     grammar = p;
   }
 
     p->next->next->next->next = grammar;
     nrules += 1;
     nritems += 3;
     grammar = p;
   }
 

-  if (nsyms > SHRT_MAX)
-    fatal (_("too many symbols (tokens plus nonterminals); maximum %d"),
-          SHRT_MAX);
-
-  assert (nsyms == ntokens + nvars);
-
-  xfclose (finput);
+  aver (nsyms <= SYMBOL_NUMBER_MAXIMUM && nsyms == ntokens + nvars);

 
   /* Assign the symbols their symbol numbers.  Write #defines for the
      token symbols into FDEFINES if requested.  */
   symbols_pack ();
 
 
   /* Assign the symbols their symbol numbers.  Write #defines for the
      token symbols into FDEFINES if requested.  */
   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.
+
+     Before invoking grammar_rule_check (in packgram below) on any rule, make
+     sure all actions have already been scanned in order to set `used' flags.
+     Otherwise, checking that a midrule's $$ should be set will not always work
+     properly because the check must forward-reference the midrule's parent
+     rule.  For the same reason, all the `used' flags must be set before
+     checking whether to remove `$' from any midrule symbol name (also in
+     packgram).  */
+  {
+    symbol_list *sym;
+    for (sym = grammar; sym; sym = sym->next)
+      code_props_translate_code (&sym->action_props);
+  }
+

   /* Convert the grammar into the format described in gram.h.  */
   packgram ();
 
   /* The grammar as a symbol_list is no longer needed. */
   /* Convert the grammar into the format described in gram.h.  */
   packgram ();
 
   /* The grammar as a symbol_list is no longer needed. */

-  LIST_FREE (symbol_list, grammar);
+  symbol_list_free (grammar);

 }
 }