* src/scan-skel.l: Postprocess quadrigraphs.

[bison.git] / src / reader.c

diff --git a/src/reader.c b/src/reader.c
index db168b68edcaef7539823557814c331ea0cb7040..3375fa2a3f48d250d39401e6a436c7e777a0ff35 100644 (file)
--- a/src/reader.c
+++ b/src/reader.c
@@ -199,13 +199,13 @@ symbol_pack (symbol_t *this)
       /* This symbol and its alias are a single token defn.
         Allocate a tokno, and assign to both check agreement of
         prec and assoc fields and make both the same */
       /* This symbol and its alias are a single token defn.
         Allocate a tokno, and assign to both check agreement of
         prec and assoc fields and make both the same */

-      if (this->number == -1)
+      if (this->number == NUMBER_UNDEFINED)

        {
          if (this == eoftoken || this->alias == eoftoken)
            this->number = this->alias->number = 0;
          else
            {
        {
          if (this == eoftoken || this->alias == eoftoken)
            this->number = this->alias->number = 0;
          else
            {

-             assert (this->alias->number != -1);
+             assert (this->alias->number != NUMBER_UNDEFINED);

              this->number = this->alias->number;
            }
        }
              this->number = this->alias->number;
            }
        }


@@ -215,7 +215,7 @@ symbol_pack (symbol_t *this)
     }
   else /* this->class == token_sym */
     {
     }
   else /* this->class == token_sym */
     {

-      assert (this->number != -1);
+      assert (this->number != NUMBER_UNDEFINED);

     }
 
   symbols[this->number] = this;
     }
 
   symbols[this->number] = this;


@@ -237,7 +237,7 @@ symbol_translation (symbol_t *this)
       && this->user_token_number != SALIAS)
     {
       /* A token which translation has already been set? */
       && this->user_token_number != SALIAS)
     {
       /* A token which translation has already been set? */

-      if (token_translations[this->user_token_number] != 2)
+      if (token_translations[this->user_token_number] != undeftoken->number)

        complain (_("tokens %s and %s both assigned number %d"),
                  symbols[token_translations[this->user_token_number]]->tag,
                  this->tag, this->user_token_number);
        complain (_("tokens %s and %s both assigned number %d"),
                  symbols[token_translations[this->user_token_number]]->tag,
                  this->tag, this->user_token_number);


@@ -326,12 +326,34 @@ get_type_name (int n, symbol_list *rule)
          complain (_("invalid $ value"));
          return NULL;
        }
          complain (_("invalid $ value"));
          return NULL;
        }

-      i++;
+      ++i;

     }
 
   return rp->sym->type_name;
 }
 \f
     }
 
   return rp->sym->type_name;
 }
 \f

+/*------------------------------------------------------------------.
+| Copy the character C to OOUT, and insert quadigraphs when needed. |
+`------------------------------------------------------------------*/
+
+static inline void
+copy_character (struct obstack *oout, int c)
+{
+  switch (c)
+    {
+    case '[':
+      obstack_sgrow (oout, "@<:@");
+      break;
+
+    case ']':
+      obstack_sgrow (oout, "@:>@");
+      break;
+
+    default:
+      obstack_1grow (oout, c);
+    }
+}
+

 /*------------------------------------------------------------.
 | Dump the string from FIN to OOUT if non null.  MATCH is the |
 | delimiter of the string (either ' or ").                    |
 /*------------------------------------------------------------.
 | Dump the string from FIN to OOUT if non null.  MATCH is the |
 | delimiter of the string (either ' or ").                    |


@@ -359,17 +381,17 @@ copy_string2 (FILE *fin, struct obstack *oout, int match, int store)
          continue;
        }
 
          continue;
        }
 

-      obstack_1grow (oout, c);
+      copy_character (oout, c);

 
       if (c == '\\')
        {
          c = getc (fin);
          if (c == EOF)
            fatal (_("unterminated string at end of file"));
 
       if (c == '\\')
        {
          c = getc (fin);
          if (c == EOF)
            fatal (_("unterminated string at end of file"));

-         obstack_1grow (oout, c);
+         copy_character (oout, c);

 
          if (c == '\n')
 
          if (c == '\n')

-           lineno++;
+           ++lineno;

        }
 
       c = getc (fin);
        }
 
       c = getc (fin);


@@ -452,7 +474,7 @@ copy_comment (FILE *fin, struct obstack *oout)
        }
       else if (c == '\n')
        {
        }
       else if (c == '\n')
        {

-         lineno++;
+         ++lineno;

          obstack_1grow (oout, c);
          if (cplus_comment)
            ended = 1;
          obstack_1grow (oout, c);
          if (cplus_comment)
            ended = 1;


@@ -463,7 +485,7 @@ copy_comment (FILE *fin, struct obstack *oout)
        fatal (_("unterminated comment"));
       else
        {
        fatal (_("unterminated comment"));
       else
        {

-         obstack_1grow (oout, c);
+         copy_character (oout, c);

          c = getc (fin);
        }
     }
          c = getc (fin);
        }
     }


@@ -591,7 +613,7 @@ copy_dollar (FILE *fin, struct obstack *oout,
 `-------------------------------------------------------------------*/
 
 static void
 `-------------------------------------------------------------------*/
 
 static void

-copy_definition (void)
+copy_definition (struct obstack *oout)

 {
   int c;
   /* -1 while reading a character if prev char was %. */
 {
   int c;
   /* -1 while reading a character if prev char was %. */


@@ -599,7 +621,7 @@ copy_definition (void)
 
   if (!no_lines_flag)
     {
 
   if (!no_lines_flag)
     {

-      obstack_fgrow2 (&attrs_obstack, muscle_find ("linef"),
+      obstack_fgrow2 (oout, muscle_find ("linef"),

                      lineno, quotearg_style (c_quoting_style,
                                              muscle_find ("filename")));
     }
                      lineno, quotearg_style (c_quoting_style,
                                              muscle_find ("filename")));
     }


@@ -613,8 +635,8 @@ copy_definition (void)
       switch (c)
        {
        case '\n':
       switch (c)
        {
        case '\n':

-         obstack_1grow (&attrs_obstack, c);
-         lineno++;
+         obstack_1grow (oout, c);
+         ++lineno;

          break;
 
        case '%':
          break;
 
        case '%':


@@ -623,18 +645,18 @@ copy_definition (void)
 
        case '\'':
        case '"':
 
        case '\'':
        case '"':

-         copy_string (finput, &attrs_obstack, c);
+         copy_string (finput, oout, c);

          break;
 
        case '/':
          break;
 
        case '/':

-         copy_comment (finput, &attrs_obstack);
+         copy_comment (finput, oout);

          break;
 
        case EOF:
          fatal ("%s", _("unterminated `%{' definition"));
 
        default:
          break;
 
        case EOF:
          fatal ("%s", _("unterminated `%{' definition"));
 
        default:

-         obstack_1grow (&attrs_obstack, c);
+         copy_character (oout, c);

        }
 
       c = getc (finput);
        }
 
       c = getc (finput);


@@ -643,7 +665,7 @@ copy_definition (void)
        {
          if (c == '}')
            return;
        {
          if (c == '}')
            return;

-         obstack_1grow (&attrs_obstack, '%');
+         obstack_1grow (oout, '%');

        }
       after_percent = 0;
     }
        }
       after_percent = 0;
     }


@@ -705,7 +727,7 @@ parse_token_decl (symbol_class what_is, symbol_class what_is_not)
          symbol->class = what_is;
          if (what_is == nterm_sym && oldclass != nterm_sym)
            symbol->number = nvars++;
          symbol->class = what_is;
          if (what_is == nterm_sym && oldclass != nterm_sym)
            symbol->number = nvars++;

-         if (what_is == token_sym && symbol->number == -1)
+         if (what_is == token_sym && symbol->number == NUMBER_UNDEFINED)

            symbol->number = ntokens++;
 
          if (typename)
            symbol->number = ntokens++;
 
          if (typename)


@@ -826,7 +848,8 @@ parse_assoc_decl (associativity assoc)
   char *name = NULL;
   int prev = 0;
 
   char *name = NULL;
   int prev = 0;
 

-  lastprec++;                  /* Assign a new precedence level, never 0.  */
+  /* Assign a new precedence level, never 0.  */
+  ++lastprec;

 
   for (;;)
     {
 
   for (;;)
     {


@@ -856,7 +879,7 @@ parse_assoc_decl (associativity assoc)
          symval->assoc = assoc;
          if (symval->class == nterm_sym)
            complain (_("symbol %s redefined"), symval->tag);
          symval->assoc = assoc;
          if (symval->class == nterm_sym)
            complain (_("symbol %s redefined"), symval->tag);

-         if (symval->number == -1)
+         if (symval->number == NUMBER_UNDEFINED)

            {
              symval->number = ntokens++;
              symval->class = token_sym;
            {
              symval->number = ntokens++;
              symval->class = token_sym;


@@ -931,7 +954,7 @@ parse_union_decl (void)
       switch (c)
        {
        case '\n':
       switch (c)
        {
        case '\n':

-         lineno++;
+         ++lineno;

          break;
 
        case '/':
          break;
 
        case '/':


@@ -939,7 +962,7 @@ parse_union_decl (void)
          break;
 
        case '{':
          break;
 
        case '{':

-         count++;
+         ++count;

          break;
 
        case '}':
          break;
 
        case '}':


@@ -1153,7 +1176,7 @@ parse_skel_decl (void)
 /*----------------------------------------------------------------.
 | Read from finput until `%%' is seen.  Discard the `%%'.  Handle |
 | any `%' declarations, and copy the contents of any `%{ ... %}'  |
 /*----------------------------------------------------------------.
 | Read from finput until `%%' is seen.  Discard the `%%'.  Handle |
 | any `%' declarations, and copy the contents of any `%{ ... %}'  |

-| groups to ATTRS_OBSTACK.                                        |
+| groups to PRE_PROLOGUE_OBSTACK or POST_PROLOGUE_OBSTACK.        |

 `----------------------------------------------------------------*/
 
 static void
 `----------------------------------------------------------------*/
 
 static void


@@ -1173,7 +1196,10 @@ read_declarations (void)
              return;
 
            case tok_percent_left_curly:
              return;
 
            case tok_percent_left_curly:

-             copy_definition ();
+              if (!typed)
+               copy_definition (&pre_prologue_obstack);
+             else
+               copy_definition (&post_prologue_obstack);

              break;
 
            case tok_token:
              break;
 
            case tok_token:


@@ -1276,12 +1302,12 @@ parse_braces (symbol_list *rule, int stack_offset)
          {
          case '\n':
            obstack_1grow (&action_obstack, c);
          {
          case '\n':
            obstack_1grow (&action_obstack, c);

-           lineno++;
+           ++lineno;

            break;
 
          case '{':
            obstack_1grow (&action_obstack, c);
            break;
 
          case '{':
            obstack_1grow (&action_obstack, c);

-           count++;
+           ++count;

            break;
 
          case '\'':
            break;
 
          case '\'':


@@ -1431,8 +1457,8 @@ readgram (void)
          }
        /* start a new rule and record its lhs.  */
 
          }
        /* start a new rule and record its lhs.  */
 

-       nrules++;
-       nitems++;
+       ++nrules;
+       ++nritems;

 
        p = symbol_list_new (lhs);
 
 
        p = symbol_list_new (lhs);
 


@@ -1451,7 +1477,7 @@ readgram (void)
          {
            lhs->class = nterm_sym;
            lhs->number = nvars;
          {
            lhs->class = nterm_sym;
            lhs->number = nvars;

-           nvars++;
+           ++nvars;

          }
        else if (lhs->class == token_sym)
          complain (_("rule given for %s, which is a token"), lhs->tag);
          }
        else if (lhs->class == token_sym)
          complain (_("rule given for %s, which is a token"), lhs->tag);


@@ -1509,8 +1535,8 @@ readgram (void)
                /* Make a new rule, whose body is empty, before the
                   current one, so that the action just read can
                   belong to it.  */
                /* Make a new rule, whose body is empty, before the
                   current one, so that the action just read can
                   belong to it.  */

-               nrules++;
-               nitems++;
+               ++nrules;
+               ++nritems;

                p = symbol_list_new (sdummy);
                /* Attach its lineno to that of the host rule. */
                p->line = crule->line;
                p = symbol_list_new (sdummy);
                /* Attach its lineno to that of the host rule. */
                p->line = crule->line;


@@ -1531,7 +1557,7 @@ readgram (void)
 
                /* Insert the dummy generated by that rule into this
                   rule.  */
 
                /* Insert the dummy generated by that rule into this
                   rule.  */

-               nitems++;
+               ++nritems;

                p = symbol_list_new (sdummy);
                p1->next = p;
                p1 = p;
                p = symbol_list_new (sdummy);
                p1->next = p;
                p1 = p;


@@ -1541,7 +1567,7 @@ readgram (void)
 
            if (t == tok_identifier)
              {
 
            if (t == tok_identifier)
              {

-               nitems++;
+               ++nritems;

                p = symbol_list_new (symval);
                p1->next = p;
                p1 = p;
                p = symbol_list_new (symval);
                p1->next = p;
                p1 = p;


@@ -1550,9 +1576,9 @@ readgram (void)
              {
                parse_action (crule, rulelength);
                action_flag = 1;
              {
                parse_action (crule, rulelength);
                action_flag = 1;

-               xactions++;     /* JF */
+               ++xactions;     /* JF */

              }
              }

-           rulelength++;
+           ++rulelength;

          }                     /* end of  read rhs of rule */
 
        /* Put an empty link in the list to mark the end of this rule  */
          }                     /* end of  read rhs of rule */
 
        /* Put an empty link in the list to mark the end of this rule  */


@@ -1584,7 +1610,7 @@ readgram (void)
              complain (_("two actions at end of one rule"));
            parse_action (crule, rulelength);
            action_flag = 1;
              complain (_("two actions at end of one rule"));
            parse_action (crule, rulelength);
            action_flag = 1;

-           xactions++; /* -wjh */
+           ++xactions; /* -wjh */

            t = lex ();
          }
        /* If $$ is being set in default way, report if any type
            t = lex ();
          }
        /* If $$ is being set in default way, report if any type


@@ -1632,13 +1658,13 @@ readgram (void)
   p->next->next->next = symbol_list_new (NULL);
   p->next->next->next->next = grammar;
   nrules += 1;
   p->next->next->next = symbol_list_new (NULL);
   p->next->next->next->next = grammar;
   nrules += 1;

-  nitems += 3;
+  nritems += 3;

   grammar = p;
   startval = axiom;
 
   grammar = p;
   startval = axiom;
 

-  if (nsyms > MAXSHORT)
+  if (nsyms > SHRT_MAX)

     fatal (_("too many symbols (tokens plus nonterminals); maximum %d"),
     fatal (_("too many symbols (tokens plus nonterminals); maximum %d"),

-          MAXSHORT);
+          SHRT_MAX);

 
   assert (nsyms == ntokens + nvars);
 }
 
   assert (nsyms == ntokens + nvars);
 }


@@ -1662,7 +1688,7 @@ read_additionnal_code (void)
     }
 
   while ((c = getc (finput)) != EOF)
     }
 
   while ((c = getc (finput)) != EOF)

-    obstack_1grow (&el_obstack, c);
+    copy_character (&el_obstack, c);

 
   obstack_1grow (&el_obstack, 0);
   muscle_insert ("epilogue", obstack_finish (&el_obstack));
 
   obstack_1grow (&el_obstack, 0);
   muscle_insert ("epilogue", obstack_finish (&el_obstack));


@@ -1677,27 +1703,48 @@ read_additionnal_code (void)
 static void
 token_translations_init (void)
 {
 static void
 token_translations_init (void)
 {

-  int last_user_token_number = 256;
+  int num_256_available_p = TRUE;

   int i;
 
   int i;
 

-  /* Set the user numbers. */
+  /* Find the highest user token number, and whether 256, the POSIX
+     preferred user token number for the error token, is used.  */
+  max_user_token_number = 0;
+  for (i = 0; i < ntokens; ++i)
+    {
+      symbol_t *this = symbols[i];
+      if (this->user_token_number != SUNDEF)
+       {
+         if (this->user_token_number > max_user_token_number)
+           max_user_token_number = this->user_token_number;
+         if (this->user_token_number == 256)
+           num_256_available_p = FALSE;
+       }
+    }
+
+  /* If 256 is not used, assign it to error, to follow POSIX.  */
+  if (num_256_available_p && errtoken->user_token_number == SUNDEF)
+    errtoken->user_token_number = 256;
+
+  /* Set the missing user numbers. */
+  if (max_user_token_number < 256)
+    max_user_token_number = 256;
+

   for (i = 0; i < ntokens; ++i)
     {
       symbol_t *this = symbols[i];
       if (this->user_token_number == SUNDEF)
   for (i = 0; i < ntokens; ++i)
     {
       symbol_t *this = symbols[i];
       if (this->user_token_number == SUNDEF)

-       this->user_token_number = ++last_user_token_number;
+       this->user_token_number = ++max_user_token_number;

       if (this->user_token_number > max_user_token_number)
        max_user_token_number = this->user_token_number;
     }
 
       if (this->user_token_number > max_user_token_number)
        max_user_token_number = this->user_token_number;
     }
 

-  token_translations = XCALLOC (short, max_user_token_number + 1);
+  token_translations = XCALLOC (token_number_t, max_user_token_number + 1);

 
   /* Initialize all entries for literal tokens to 2, the internal
      token number for $undefined., which represents all invalid
      inputs.  */
   for (i = 0; i < max_user_token_number + 1; i++)
 
   /* Initialize all entries for literal tokens to 2, the internal
      token number for $undefined., which represents all invalid
      inputs.  */
   for (i = 0; i < max_user_token_number + 1; i++)

-    token_translations[i] = 2;
-
+    token_translations[i] = undeftoken->number;

   symbols_do (symbol_translation, NULL);
 }
 
   symbols_do (symbol_translation, NULL);
 }
 


@@ -1717,8 +1764,6 @@ packsymbols (void)
 
   token_translations_init ();
 
 
   token_translations_init ();
 

-  error_token_number = errtoken->number;
-

   if (startval->class == unknown_sym)
     fatal (_("the start symbol %s is undefined"), startval->tag);
   else if (startval->class == token_sym)
   if (startval->class == unknown_sym)
     fatal (_("the start symbol %s is undefined"), startval->tag);
   else if (startval->class == token_sym)


@@ -1740,11 +1785,7 @@ packgram (void)
   int ruleno;
   symbol_list *p;
 
   int ruleno;
   symbol_list *p;
 

-  /* We use short to index items.  */
-  if (nitems >= MAXSHORT)
-    fatal (_("too many items (max %d)"), MAXSHORT);
-
-  ritem = XCALLOC (short, nitems + 1);
+  ritem = XCALLOC (item_number_t, nritems + 1);

   rules = XCALLOC (rule_t, nrules) - 1;
 
   itemno = 0;
   rules = XCALLOC (rule_t, nrules) - 1;
 
   itemno = 0;


@@ -1768,7 +1809,9 @@ packgram (void)
       p = p->next;
       while (p && p->sym)
        {
       p = p->next;
       while (p && p->sym)
        {

-         ritem[itemno++] = p->sym->number;
+         /* item_number_t = token_number_t.
+            But the former needs to contain more: negative rule numbers. */
+         ritem[itemno++] = token_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)
          /* A rule gets by default the precedence and associativity
             of the last token in it.  */
          if (p->sym->class == token_sym)


@@ -1785,15 +1828,14 @@ packgram (void)
          rules[ruleno].prec = ruleprec;
        }
       ritem[itemno++] = -ruleno;
          rules[ruleno].prec = ruleprec;
        }
       ritem[itemno++] = -ruleno;

-      ruleno++;
+      ++ruleno;

 
       if (p)
        p = p->next;
     }
 
   ritem[itemno] = 0;
 
       if (p)
        p = p->next;
     }
 
   ritem[itemno] = 0;

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

 
   if (trace_flag)
     ritem_print (stderr);
 
   if (trace_flag)
     ritem_print (stderr);


@@ -1828,19 +1870,18 @@ reader (void)
   errtoken = getsym ("error");
   errtoken->class = token_sym;
   errtoken->number = ntokens++;
   errtoken = getsym ("error");
   errtoken->class = token_sym;
   errtoken->number = ntokens++;

-  errtoken->user_token_number = 256;   /* Value specified by POSIX.  */

 
   /* Construct a token that represents all undefined literal tokens.
      It is always token number 2.  */
   undeftoken = getsym ("$undefined.");
   undeftoken->class = token_sym;
   undeftoken->number = ntokens++;
 
   /* Construct a token that represents all undefined literal tokens.
      It is always token number 2.  */
   undeftoken = getsym ("$undefined.");
   undeftoken->class = token_sym;
   undeftoken->number = ntokens++;

-  undeftoken->user_token_number = 2;

 
   /* Initialize the obstacks. */
   obstack_init (&action_obstack);
 
   /* Initialize the obstacks. */
   obstack_init (&action_obstack);

-  obstack_init (&attrs_obstack);

   obstack_init (&output_obstack);
   obstack_init (&output_obstack);

+  obstack_init (&pre_prologue_obstack);
+  obstack_init (&post_prologue_obstack);

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