X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/d2729d44ab05f89e96d199ad6b10e127c1cb0a38..8c7d6a3de82ca7fcbdfa6ad31c3bf70f63c6a069:/src/lex.c?ds=sidebyside

diff --git a/src/lex.c b/src/lex.c
index f0fbf485..4bd0a6f2 100644
--- a/src/lex.c
+++ b/src/lex.c
@@ -1,75 +1,37 @@
 /* Token-reader for Bison's input parser,
-   Copyright (C) 1984, 1986, 1989, 1992 Free Software Foundation, Inc.
+   Copyright 1984, 1986, 1989, 1992, 2000, 2001 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
-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.
+   Bison 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 2, or (at your option)
+   any later version.
 
-Bison 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.
+   Bison 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 Bison; see the file COPYING.  If not, write to
-the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.  */
+   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.  */
 
-
-/*
-   lex is the entry point.  It is called from reader.c.
-   It returns one of the token-type codes defined in lex.h.
-   When an identifier is seen, the code IDENTIFIER is returned
-   and the name is looked up in the symbol table using symtab.c;
-   symval is set to a pointer to the entry found.  */
-
-#include <stdio.h>
 #include "system.h"
+#include "getargs.h"
 #include "files.h"
 #include "getopt.h"		/* for optarg */
 #include "symtab.h"
 #include "lex.h"
-#include "alloc.h"
-
-/* flags set by % directives */
-extern int definesflag;    	/* for -d */
-extern int toknumflag;   	/* for -k */
-extern int noparserflag;	/* for -n */
-extern int fixed_outfiles;  	/* for -y */
-extern int nolinesflag;    	/* for -l */
-extern int rawtoknumflag;	/* for -r */
-extern int verboseflag;	/* for -v */
-extern int debugflag;  		/* for -t */
-extern char *spec_name_prefix; 	/* for -p */
-extern char *spec_file_prefix;	/* for -b */
-/*spec_outfile is declared in files.h, for -o */
-
-extern int lineno;
-extern int translations;
-
-void init_lex PARAMS((void));
-char *grow_token_buffer PARAMS((char *));
-int skip_white_space PARAMS((void));
-int safegetc PARAMS((FILE *));
-int literalchar PARAMS((char **, int *, char));
-void unlex PARAMS((int));
-int lex PARAMS((void));
-int parse_percent_token PARAMS((void));
-
-/* functions from main.c */
-extern char *printable_version PARAMS((int));
-extern void fatal PARAMS((char *));
-extern void warn PARAMS((char *));
-extern void warni PARAMS((char *, int));
-extern void warns PARAMS((char *, char *));
+#include "xalloc.h"
+#include "complain.h"
+#include "gram.h"
+#include "quote.h"
 
 /* Buffer for storing the current token.  */
-char *token_buffer;
-
-/* Allocated size of token_buffer, not including space for terminator.  */
-int maxtoken;
+struct obstack token_obstack;
+const char *token_buffer = NULL;
 
 bucket *symval;
 int numval;
@@ -81,29 +43,18 @@ static bucket *unlexed_symval;	/* by the next call to lex */
 void
 init_lex (void)
 {
-  maxtoken = 100;
-  token_buffer = NEW2 (maxtoken + 1, char);
+  obstack_init (&token_obstack);
   unlexed = -1;
 }
 
 
-char *
-grow_token_buffer (char *p)
-{
-  int offset = p - token_buffer;
-  maxtoken *= 2;
-  token_buffer = (char *) xrealloc(token_buffer, maxtoken + 1);
-  return token_buffer + offset;
-}
-
-
 int
 skip_white_space (void)
 {
-  register int c;
-  register int inside;
+  int c;
+  int inside;
 
-  c = getc(finput);
+  c = getc (finput);
 
   for (;;)
     {
@@ -112,15 +63,16 @@ skip_white_space (void)
       switch (c)
 	{
 	case '/':
-	  c = getc(finput);
+	  /* FIXME: Should probably be merged with copy_comment.  */
+	  c = getc (finput);
 	  if (c != '*' && c != '/')
 	    {
-	      warn(_("unexpected `/' found and ignored"));
+	      complain (_("unexpected `/' found and ignored"));
 	      break;
 	    }
 	  cplus_comment = (c == '/');
 
-	  c = getc(finput);
+	  c = getc (finput);
 
 	  inside = 1;
 	  while (inside)
@@ -128,12 +80,12 @@ skip_white_space (void)
 	      if (!cplus_comment && c == '*')
 		{
 		  while (c == '*')
-		    c = getc(finput);
+		    c = getc (finput);
 
 		  if (c == '/')
 		    {
 		      inside = 0;
-		      c = getc(finput);
+		      c = getc (finput);
 		    }
 		}
 	      else if (c == '\n')
@@ -141,12 +93,12 @@ skip_white_space (void)
 		  lineno++;
 		  if (cplus_comment)
 		    inside = 0;
-		  c = getc(finput);
+		  c = getc (finput);
 		}
 	      else if (c == EOF)
-		fatal(_("unterminated comment"));
+		fatal (_("unterminated comment"));
 	      else
-		c = getc(finput);
+		c = getc (finput);
 	    }
 
 	  break;
@@ -157,44 +109,56 @@ skip_white_space (void)
 	case ' ':
 	case '\t':
 	case '\f':
-	  c = getc(finput);
+	  c = getc (finput);
 	  break;
 
 	default:
-	  return (c);
+	  return c;
 	}
     }
 }
 
-/* do a getc, but give error message if EOF encountered */
-int
-safegetc (FILE *f)
+
+/*-----------------------------------------------------.
+| Do a getc, but give error message if EOF encountered |
+`-----------------------------------------------------*/
+
+static int
+xgetc (FILE *f)
 {
-  register int c = getc(f);
+  int c = getc (f);
   if (c == EOF)
-    fatal(_("Unexpected end of file"));
+    fatal (_("unexpected end of file"));
   return c;
 }
 
-/* read one literal character from finput.  process \ escapes.
-   append the normalized string version of the char to *pp.
-   assign the character code to *pcode
-   return 1 unless the character is an unescaped `term' or \n
-	report error for \n
-*/
-int
-literalchar (char **pp, int *pcode, char term)
+
+/*------------------------------------------------------------------.
+| Read one literal character from finput.  Process \ escapes.       |
+| Append the normalized string version of the char to OUT.  Assign  |
+| the character code to *PCODE. Return 1 unless the character is an |
+| unescaped `term' or \n report error for \n.                       |
+`------------------------------------------------------------------*/
+
+/* FIXME: We could directly work in the obstack, but that would make
+   it more difficult to move to quotearg some day.  So for the time
+   being, I prefer have literalchar behave like quotearg, and change
+   my mind later if I was wrong.  */
+
+static int
+literalchar (struct obstack *out, int *pcode, char term)
 {
-  register int c;
-  register char *p;
-  register int code;
+  int c;
+  char buf[4096];
+  char *cp;
+  int code;
   int wasquote = 0;
 
-  c = safegetc(finput);
+  c = xgetc (finput);
   if (c == '\n')
     {
-      warn(_("unescaped newline in constant"));
-      ungetc(c, finput);
+      complain (_("unescaped newline in constant"));
+      ungetc (c, finput);
       code = '?';
       wasquote = 1;
     }
@@ -206,17 +170,27 @@ literalchar (char **pp, int *pcode, char term)
     }
   else
     {
-      c = safegetc(finput);
-      if (c == 't')  code = '\t';
-      else if (c == 'n')  code = '\n';
-      else if (c == 'a')  code = '\007';
-      else if (c == 'r')  code = '\r';
-      else if (c == 'f')  code = '\f';
-      else if (c == 'b')  code = '\b';
-      else if (c == 'v')  code = '\013';
-      else if (c == '\\')  code = '\\';
-      else if (c == '\'')  code = '\'';
-      else if (c == '\"')  code = '\"';
+      c = xgetc (finput);
+      if (c == 't')
+	code = '\t';
+      else if (c == 'n')
+	code = '\n';
+      else if (c == 'a')
+	code = '\007';
+      else if (c == 'r')
+	code = '\r';
+      else if (c == 'f')
+	code = '\f';
+      else if (c == 'b')
+	code = '\b';
+      else if (c == 'v')
+	code = '\013';
+      else if (c == '\\')
+	code = '\\';
+      else if (c == '\'')
+	code = '\'';
+      else if (c == '\"')
+	code = '\"';
       else if (c <= '7' && c >= '0')
 	{
 	  code = 0;
@@ -225,72 +199,116 @@ literalchar (char **pp, int *pcode, char term)
 	      code = (code * 8) + (c - '0');
 	      if (code >= 256 || code < 0)
 		{
-		  warni(_("octal value outside range 0...255: `\\%o'"), code);
+		  complain (_("octal value outside range 0...255: `\\%o'"),
+			    code);
 		  code &= 0xFF;
 		  break;
 		}
-	      c = safegetc(finput);
+	      c = xgetc (finput);
 	    }
-	  ungetc(c, finput);
+	  ungetc (c, finput);
 	}
       else if (c == 'x')
 	{
-	  c = safegetc(finput);
+	  c = xgetc (finput);
 	  code = 0;
 	  while (1)
 	    {
 	      if (c >= '0' && c <= '9')
-		code *= 16,  code += c - '0';
+		code *= 16, code += c - '0';
 	      else if (c >= 'a' && c <= 'f')
-		code *= 16,  code += c - 'a' + 10;
+		code *= 16, code += c - 'a' + 10;
 	      else if (c >= 'A' && c <= 'F')
-		code *= 16,  code += c - 'A' + 10;
+		code *= 16, code += c - 'A' + 10;
 	      else
 		break;
-	      if (code >= 256 || code<0)
+	      if (code >= 256 || code < 0)
 		{
-		  warni(_("hexadecimal value above 255: `\\x%x'"), code);
+		  complain (_("hexadecimal value above 255: `\\x%x'"), code);
 		  code &= 0xFF;
 		  break;
 		}
-	      c = safegetc(finput);
+	      c = xgetc (finput);
 	    }
-	  ungetc(c, finput);
+	  ungetc (c, finput);
 	}
       else
 	{
-	  warns (_("unknown escape sequence: `\\' followed by `%s'"),
-		 printable_version(c));
+	  char badchar [] = "c";
+	  badchar[0] = c;
+	  complain (_("unknown escape sequence: `\\' followed by `%s'"),
+		    quote (badchar));
 	  code = '?';
 	}
-    } /* has \ */
+    }				/* has \ */
 
-  /* now fill token_buffer with the canonical name for this character
-     as a literal token.  Do not use what the user typed,
-     so that `\012' and `\n' can be interchangeable.  */
+  /* now fill BUF with the canonical name for this character as a
+     literal token.  Do not use what the user typed, so that `\012'
+     and `\n' can be interchangeable.  */
 
-  p = *pp;
-  if (code == '\\')  {*p++ = '\\'; *p++ = '\\';}
-  else if (code == '\'')  {*p++ = '\\'; *p++ = '\'';}
-  else if (code == '\"')  {*p++ = '\\'; *p++ = '\"';}
+  cp = buf;
+  if (code == term && wasquote)
+    *cp++ = code;
+  else if (code == '\\')
+    {
+      *cp++ = '\\';
+      *cp++ = '\\';
+    }
+  else if (code == '\'')
+    {
+      *cp++ = '\\';
+      *cp++ = '\'';
+    }
+  else if (code == '\"')
+    {
+      *cp++ = '\\';
+      *cp++ = '\"';
+    }
   else if (code >= 040 && code < 0177)
-    *p++ = code;
-  else if (code == '\t')  {*p++ = '\\'; *p++ = 't';}
-  else if (code == '\n')  {*p++ = '\\'; *p++ = 'n';}
-  else if (code == '\r')  {*p++ = '\\'; *p++ = 'r';}
-  else if (code == '\v')  {*p++ = '\\'; *p++ = 'v';}
-  else if (code == '\b')  {*p++ = '\\'; *p++ = 'b';}
-  else if (code == '\f')  {*p++ = '\\'; *p++ = 'f';}
+    *cp++ = code;
+  else if (code == '\t')
+    {
+      *cp++ = '\\';
+      *cp++ = 't';
+    }
+  else if (code == '\n')
+    {
+      *cp++ = '\\';
+      *cp++ = 'n';
+    }
+  else if (code == '\r')
+    {
+      *cp++ = '\\';
+      *cp++ = 'r';
+    }
+  else if (code == '\v')
+    {
+      *cp++ = '\\';
+      *cp++ = 'v';
+    }
+  else if (code == '\b')
+    {
+      *cp++ = '\\';
+      *cp++ = 'b';
+    }
+  else if (code == '\f')
+    {
+      *cp++ = '\\';
+      *cp++ = 'f';
+    }
   else
     {
-      *p++ = '\\';
-      *p++ = code / 0100 + '0';
-      *p++ = ((code / 010) & 07) + '0';
-      *p++ = (code & 07) + '0';
+      *cp++ = '\\';
+      *cp++ = code / 0100 + '0';
+      *cp++ = ((code / 010) & 07) + '0';
+      *cp++ = (code & 07) + '0';
     }
-  *pp = p;
+  *cp = '\0';
+
+  if (out)
+    obstack_sgrow (out, buf);
   *pcode = code;
-  return  ! wasquote;
+  return !wasquote;
 }
 
 
@@ -301,253 +319,247 @@ unlex (int token)
   unlexed_symval = symval;
 }
 
+/*-----------------------------------------------------------------.
+| We just read `<' from FIN.  Store in TOKEN_BUFFER, the type name |
+| specified between the `<...>'.                                   |
+`-----------------------------------------------------------------*/
 
-int
+void
+read_type_name (FILE *fin)
+{
+  int c = getc (fin);
+
+  while (c != '>')
+    {
+      if (c == EOF)
+	fatal (_("unterminated type name at end of file"));
+      if (c == '\n')
+	{
+	  complain (_("unterminated type name"));
+	  ungetc (c, fin);
+	  break;
+	}
+
+      obstack_1grow (&token_obstack, c);
+      c = getc (fin);
+    }
+  obstack_1grow (&token_obstack, '\0');
+  token_buffer = obstack_finish (&token_obstack);
+}
+
+
+token_t
 lex (void)
 {
-  register int c;
-  char *p;
+  int c;
+
+  /* Just to make sure. */
+  token_buffer = NULL;
 
   if (unlexed >= 0)
     {
       symval = unlexed_symval;
       c = unlexed;
       unlexed = -1;
-      return (c);
+      return c;
     }
 
-  c = skip_white_space();
-  *token_buffer = c;	/* for error messages (token buffer always valid) */
-  token_buffer[1] = 0;
+  c = skip_white_space ();
 
   switch (c)
     {
     case EOF:
-      strcpy(token_buffer, "EOF");
-      return (ENDFILE);
-
-    case 'A':  case 'B':  case 'C':  case 'D':  case 'E':
-    case 'F':  case 'G':  case 'H':  case 'I':  case 'J':
-    case 'K':  case 'L':  case 'M':  case 'N':  case 'O':
-    case 'P':  case 'Q':  case 'R':  case 'S':  case 'T':
-    case 'U':  case 'V':  case 'W':  case 'X':  case 'Y':
+      token_buffer = "EOF";
+      return tok_eof;
+
+    case 'A':    case 'B':    case 'C':    case 'D':    case 'E':
+    case 'F':    case 'G':    case 'H':    case 'I':    case 'J':
+    case 'K':    case 'L':    case 'M':    case 'N':    case 'O':
+    case 'P':    case 'Q':    case 'R':    case 'S':    case 'T':
+    case 'U':    case 'V':    case 'W':    case 'X':    case 'Y':
     case 'Z':
-    case 'a':  case 'b':  case 'c':  case 'd':  case 'e':
-    case 'f':  case 'g':  case 'h':  case 'i':  case 'j':
-    case 'k':  case 'l':  case 'm':  case 'n':  case 'o':
-    case 'p':  case 'q':  case 'r':  case 's':  case 't':
-    case 'u':  case 'v':  case 'w':  case 'x':  case 'y':
+    case 'a':    case 'b':    case 'c':    case 'd':    case 'e':
+    case 'f':    case 'g':    case 'h':    case 'i':    case 'j':
+    case 'k':    case 'l':    case 'm':    case 'n':    case 'o':
+    case 'p':    case 'q':    case 'r':    case 's':    case 't':
+    case 'u':    case 'v':    case 'w':    case 'x':    case 'y':
     case 'z':
-    case '.':  case '_':
-      p = token_buffer;
-      while (isalnum(c) || c == '_' || c == '.')
-	{
-	  if (p == token_buffer + maxtoken)
-	    p = grow_token_buffer(p);
+    case '.':    case '_':
 
-	  *p++ = c;
-	  c = getc(finput);
+      while (isalnum (c) || c == '_' || c == '.')
+	{
+	  obstack_1grow (&token_obstack, c);
+	  c = getc (finput);
 	}
-
-      *p = 0;
-      ungetc(c, finput);
-      symval = getsym(token_buffer);
-      return (IDENTIFIER);
-
-    case '0':  case '1':  case '2':  case '3':  case '4':
-    case '5':  case '6':  case '7':  case '8':  case '9':
+      obstack_1grow (&token_obstack, '\0');
+      token_buffer = obstack_finish (&token_obstack);
+      ungetc (c, finput);
+      symval = getsym (token_buffer);
+      return tok_identifier;
+
+    case '0':    case '1':    case '2':    case '3':    case '4':
+    case '5':    case '6':    case '7':    case '8':    case '9':
       {
 	numval = 0;
 
-	p = token_buffer;
-	while (isdigit(c))
+	while (isdigit (c))
 	  {
-	    if (p == token_buffer + maxtoken)
-	      p = grow_token_buffer(p);
-
-	    *p++ = c;
-	    numval = numval*10 + c - '0';
-	    c = getc(finput);
+	    obstack_1grow (&token_obstack, c);
+	    numval = numval * 10 + c - '0';
+	    c = getc (finput);
 	  }
-	*p = 0;
-	ungetc(c, finput);
-	return (NUMBER);
+	obstack_1grow (&token_obstack, '\0');
+	token_buffer = obstack_finish (&token_obstack);
+	ungetc (c, finput);
+	return tok_number;
       }
 
     case '\'':
-
       /* parse the literal token and compute character code in  code  */
 
       translations = -1;
       {
 	int code, discode;
-	char discard[10], *dp;
 
-	p = token_buffer;
-	*p++ = '\'';
-	literalchar(&p, &code, '\'');
+	obstack_1grow (&token_obstack, '\'');
+	literalchar (&token_obstack, &code, '\'');
 
-	c = getc(finput);
+	c = getc (finput);
 	if (c != '\'')
 	  {
-	    warn(_("use \"...\" for multi-character literal tokens"));
+	    complain (_("use \"...\" for multi-character literal tokens"));
 	    while (1)
-	      {
-		dp = discard;
-		if (! literalchar(&dp, &discode, '\''))
-		  break;
-	      }
+	      if (!literalchar (0, &discode, '\''))
+		break;
 	  }
-	*p++ = '\'';
-	*p = 0;
-	symval = getsym(token_buffer);
-	symval->class = STOKEN;
-	if (! symval->user_token_number)
+	obstack_1grow (&token_obstack, '\'');
+	obstack_1grow (&token_obstack, '\0');
+	token_buffer = obstack_finish (&token_obstack);
+	symval = getsym (token_buffer);
+	symval->class = token_sym;
+	if (!symval->user_token_number)
 	  symval->user_token_number = code;
-	return (IDENTIFIER);
+	return tok_identifier;
       }
 
     case '\"':
-
       /* parse the literal string token and treat as an identifier */
 
       translations = -1;
       {
-	int code;	/* ignored here */
-	p = token_buffer;
-	*p++ = '\"';
-	while (literalchar(&p, &code, '\"'))  /* read up to and including " */
-	  {
-	    if (p >= token_buffer + maxtoken - 4)
-	      p = grow_token_buffer(p);
-	  }
-	*p = 0;
+	int code;		/* ignored here */
+
+	obstack_1grow (&token_obstack, '\"');
+	/* Read up to and including ".  */
+	while (literalchar (&token_obstack, &code, '\"'))
+	  /* nothing */;
+	obstack_1grow (&token_obstack, '\0');
+	token_buffer = obstack_finish (&token_obstack);
 
-	symval = getsym(token_buffer);
-	symval->class = STOKEN;
+	symval = getsym (token_buffer);
+	symval->class = token_sym;
 
-	return (IDENTIFIER);
+	return tok_identifier;
       }
 
     case ',':
-      return (COMMA);
+      return tok_comma;
 
     case ':':
-      return (COLON);
+      return tok_colon;
 
     case ';':
-      return (SEMICOLON);
+      return tok_semicolon;
 
     case '|':
-      return (BAR);
+      return tok_bar;
 
     case '{':
-      return (LEFT_CURLY);
+      return tok_left_curly;
 
     case '=':
       do
 	{
-	  c = getc(finput);
-	  if (c == '\n') lineno++;
+	  c = getc (finput);
+	  if (c == '\n')
+	    lineno++;
 	}
-      while(c==' ' || c=='\n' || c=='\t');
+      while (c == ' ' || c == '\n' || c == '\t');
 
       if (c == '{')
 	{
-	  strcpy(token_buffer, "={");
-	  return(LEFT_CURLY);
+	  token_buffer = "={";
+	  return tok_left_curly;
 	}
       else
 	{
-	  ungetc(c, finput);
-	  return(ILLEGAL);
+	  ungetc (c, finput);
+	  return tok_illegal;
 	}
 
     case '<':
-      p = token_buffer;
-      c = getc(finput);
-      while (c != '>')
-	{
-	  if (c == EOF)
-	    fatal(_("unterminated type name at end of file"));
-	  if (c == '\n')
-	    {
-	      warn(_("unterminated type name"));
-	      ungetc(c, finput);
-	      break;
-	    }
-
-	  if (p == token_buffer + maxtoken)
-	    p = grow_token_buffer(p);
-
-	  *p++ = c;
-	  c = getc(finput);
-	}
-      *p = 0;
-      return (TYPENAME);
-
+      read_type_name (finput);
+      return tok_typename;
 
     case '%':
-      return (parse_percent_token());
+      return parse_percent_token ();
 
     default:
-      return (ILLEGAL);
+      return tok_illegal;
     }
 }
 
-/* the following table dictates the action taken for the various
-	% directives.  A setflag value causes the named flag to be
-	set.  A retval action returns the code.
-*/
-struct percent_table_struct {
-	char *name;
-	void *setflag;
-	int retval;
-} percent_table[] =
+/* the following table dictates the action taken for the various %
+   directives.  A set_flag value causes the named flag to be set.  A
+   retval action returns the code.  */
+struct percent_table_struct
 {
-  {"token", NULL, TOKEN},
-  {"term", NULL, TOKEN},
-  {"nterm", NULL, NTERM},
-  {"type", NULL, TYPE},
-  {"guard", NULL, GUARD},
-  {"union", NULL, UNION},
-  {"expect", NULL, EXPECT},
-  {"thong", NULL, THONG},
-  {"start", NULL, START},
-  {"left", NULL, LEFT},
-  {"right", NULL, RIGHT},
-  {"nonassoc", NULL, NONASSOC},
-  {"binary", NULL, NONASSOC},
-  {"semantic_parser", NULL, SEMANTIC_PARSER},
-  {"pure_parser", NULL, PURE_PARSER},
-  {"prec", NULL, PREC},
-
-  {"no_lines", &nolinesflag, NOOP}, /* -l */
-  {"raw", &rawtoknumflag, NOOP}, /* -r */
-  {"token_table", &toknumflag, NOOP}, /* -k */
+  const char *name;
+  void *set_flag;
+  int retval;
+};
 
+struct percent_table_struct percent_table[] =
+{
+  { "token",		NULL,			tok_token },
+  { "term",		NULL,			tok_token },
+  { "nterm",		NULL,			tok_nterm },
+  { "type",		NULL,			tok_type },
+  { "guard",		NULL,			tok_guard },
+  { "union",		NULL,			tok_union },
+  { "expect",		NULL,			tok_expect },
+  { "thong",		NULL,			tok_thong },
+  { "start",		NULL,			tok_start },
+  { "left",		NULL,			tok_left },
+  { "right",		NULL,			tok_right },
+  { "nonassoc",		NULL,			tok_nonassoc },
+  { "binary",		NULL,			tok_nonassoc },
+  { "prec",		NULL,			tok_prec },
+  { "locations",	&locations_flag,	tok_noop },	/* -l */
+  { "no_lines",		&no_lines_flag,		tok_noop },	/* -l */
+  { "raw",		NULL,			tok_obsolete },	/* -r */
+  { "token_table",	&token_table_flag,	tok_noop },	/* -k */
+  { "yacc",		&yacc_flag,		tok_noop },	/* -y */
+  { "fixed_output_files",&yacc_flag,		tok_noop },	/* -y */
+  { "defines",		&defines_flag,		tok_noop },	/* -d */
+  { "no_parser",	&no_parser_flag,	tok_noop },	/* -n */
 #if 0
-  /* These can be utilized after main is reoganized so
-     open_files() is deferred 'til after read_declarations().
-     But %{ and %union both put information into files
-     that have to be opened before read_declarations().
-     */
-  {"yacc", &fixed_outfiles, NOOP}, /* -y */
-  {"fixed_output_files", &fixed_outfiles, NOOP}, /* -y */
-  {"defines", &definesflag, NOOP}, /* -d */
-  {"no_parser", &noparserflag, NOOP}, /* -n */
-  {"output_file", &spec_outfile, SETOPT}, /* -o */
-  {"file_prefix", &spec_file_prefix, SETOPT}, /* -b */
-  {"name_prefix", &spec_name_prefix, SETOPT}, /* -p */
-
-  /* These would be acceptable, but they do not affect processing */
-  {"verbose", &verboseflag, NOOP}, /* -v */
-  {"debug", &debugflag, NOOP},	/* -t */
-  /*	{"help", <print usage stmt>, NOOP},*/	/* -h */
-  /*	{"version", <print version number> ,  NOOP},*/	/* -V */
+  /* For the time being, this is not enabled yet, while it's possible
+     though, since we use obstacks.  The only risk is with semantic
+     parsers which will output an `include' of an output file: be sure
+     that the naem included is indeed the name of the output file.  */
+  { "output_file",	&spec_outfile,		tok_setopt },	/* -o */
+  { "file_prefix",	&spec_file_prefix,	tok_setopt },	/* -b */
+  { "name_prefix",	&spec_name_prefix,	tok_setopt },	/* -p */
 #endif
-
-  {NULL, NULL, ILLEGAL}
+  { "header_extension",	NULL,			tok_hdrext},
+  { "source_extension",	NULL,			tok_srcext},
+  { "verbose",		&verbose_flag,		tok_noop },	/* -v */
+  { "debug",		&debug_flag,		tok_noop },	/* -t */
+  { "semantic_parser",	&semantic_parser,	tok_noop },
+  { "pure_parser",	&pure_parser,		tok_noop },
+
+  { NULL, NULL, tok_illegal}
 };
 
 /* Parse a token which starts with %.
@@ -556,71 +568,73 @@ struct percent_table_struct {
 int
 parse_percent_token (void)
 {
-  register int c;
-  register char *p;
-  register struct percent_table_struct *tx;
+  int c;
+  struct percent_table_struct *tx;
 
-  p = token_buffer;
-  c = getc(finput);
-  *p++ = '%';
-  *p++ = c;	/* for error msg */
-  *p = 0;
+  c = getc (finput);
 
   switch (c)
     {
     case '%':
-      return (TWO_PERCENTS);
+      return tok_two_percents;
 
     case '{':
-      return (PERCENT_LEFT_CURLY);
+      return tok_percent_left_curly;
 
     case '<':
-      return (LEFT);
+      return tok_left;
 
     case '>':
-      return (RIGHT);
+      return tok_right;
 
     case '2':
-      return (NONASSOC);
+      return tok_nonassoc;
 
     case '0':
-      return (TOKEN);
+      return tok_token;
 
     case '=':
-      return (PREC);
+      return tok_prec;
     }
-  if (!isalpha(c))
-    return (ILLEGAL);
 
-  p = token_buffer;
-  *p++ = '%';
-  while (isalpha(c) || c == '_' || c == '-')
-    {
-      if (p == token_buffer + maxtoken)
-	p = grow_token_buffer(p);
+  if (!isalpha (c))
+    return tok_illegal;
 
-      if (c == '-') c = '_';
-      *p++ = c;
-      c = getc(finput);
+  obstack_1grow (&token_obstack, '%');
+  while (isalpha (c) || c == '_' || c == '-')
+    {
+      if (c == '-')
+	c = '_';
+      obstack_1grow (&token_obstack, c);
+      c = getc (finput);
     }
 
-  ungetc(c, finput);
-
-  *p = 0;
+  ungetc (c, finput);
+  obstack_1grow (&token_obstack, '\0');
+  token_buffer = obstack_finish (&token_obstack);
 
   /* table lookup % directive */
   for (tx = percent_table; tx->name; tx++)
-    if (strcmp(token_buffer+1, tx->name) == 0)
+    if (strcmp (token_buffer + 1, tx->name) == 0)
       break;
-  if (tx->retval == SETOPT)
+
+  if (tx->set_flag)
     {
-      *((char **)(tx->setflag)) = optarg;
-      return NOOP;
+      *((int *) (tx->set_flag)) = 1;
+      return tok_noop;
     }
-  if (tx->setflag)
+
+  switch (tx->retval)
     {
-      *((int *)(tx->setflag)) = 1;
-      return NOOP;
+    case tok_setopt:
+      *((char **) (tx->set_flag)) = optarg;
+      return tok_noop;
+      break;
+
+    case tok_obsolete:
+      fatal (_("`%s' is no longer supported"), token_buffer);
+      break;
     }
+
   return tx->retval;
 }