.

[bison.git] / src / lex.c

diff --git a/src/lex.c b/src/lex.c
index 408e30fcb8549fb664cf8cde73d953054aa26e7f..4bd0a6f26344d4fbc883b38f6c5543166255b27a 100644 (file)
--- a/src/lex.c
+++ b/src/lex.c
@@ -1,5 +1,5 @@
 /* Token-reader for Bison's input parser,
 /* Token-reader for Bison's input parser,

-   Copyright 1984, 1986, 1989, 1992, 2000 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.
 


@@ -30,10 +30,8 @@
 #include "quote.h"
 
 /* Buffer for storing the current token.  */
 #include "quote.h"
 
 /* Buffer for storing the current token.  */

-char *token_buffer;
-
-/* Allocated size of token_buffer, not including space for terminator.  */
-static int maxtoken;
+struct obstack token_obstack;
+const char *token_buffer = NULL;

 
 bucket *symval;
 int numval;
 
 bucket *symval;
 int numval;


@@ -45,22 +43,11 @@ static bucket *unlexed_symval;      /* by the next call to lex */
 void
 init_lex (void)
 {
 void
 init_lex (void)
 {

-  maxtoken = 100;
-  token_buffer = XCALLOC (char, maxtoken + 1);
+  obstack_init (&token_obstack);

   unlexed = -1;
 }
 
 
   unlexed = -1;
 }
 
 

-static char *
-grow_token_buffer (char *p)
-{
-  int offset = p - token_buffer;
-  maxtoken *= 2;
-  token_buffer = XREALLOC (token_buffer, char, maxtoken + 1);
-  return token_buffer + offset;
-}
-
-

 int
 skip_white_space (void)
 {
 int
 skip_white_space (void)
 {


@@ -148,16 +135,22 @@ xgetc (FILE *f)
 
 /*------------------------------------------------------------------.
 | Read one literal character from finput.  Process \ escapes.       |
 
 /*------------------------------------------------------------------.
 | Read one literal character from finput.  Process \ escapes.       |

-| Append the normalized string version of the char to *PP.  Assign  |
+| Append the normalized string version of the char to OUT.  Assign  |

 | the character code to *PCODE. Return 1 unless the character is an |
 | the character code to *PCODE. Return 1 unless the character is an |

-| unescaped `term' or \n report error for \n                        |
+| 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
 static int

-literalchar (char **pp, int *pcode, char term)
+literalchar (struct obstack *out, int *pcode, char term)

 {
   int c;
 {
   int c;

-  char *p;
+  char buf[4096];
+  char *cp;

   int code;
   int wasquote = 0;
 
   int code;
   int wasquote = 0;
 


@@ -241,76 +234,79 @@ literalchar (char **pp, int *pcode, char term)
        }
       else
        {
        }
       else
        {

-         char buf [] = "c";
-         buf[0] = c;
+         char badchar [] = "c";
+         badchar[0] = c;

          complain (_("unknown escape sequence: `\\' followed by `%s'"),
          complain (_("unknown escape sequence: `\\' followed by `%s'"),

-                   quote (buf));
+                   quote (badchar));

          code = '?';
        }
     }                          /* has \ */
 
          code = '?';
        }
     }                          /* 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;
+  cp = buf;

   if (code == term && wasquote)
   if (code == term && wasquote)

-    *p++ = code;
+    *cp++ = code;

   else if (code == '\\')
     {
   else if (code == '\\')
     {

-      *p++ = '\\';
-      *p++ = '\\';
+      *cp++ = '\\';
+      *cp++ = '\\';

     }
   else if (code == '\'')
     {
     }
   else if (code == '\'')
     {

-      *p++ = '\\';
-      *p++ = '\'';
+      *cp++ = '\\';
+      *cp++ = '\'';

     }
   else if (code == '\"')
     {
     }
   else if (code == '\"')
     {

-      *p++ = '\\';
-      *p++ = '\"';
+      *cp++ = '\\';
+      *cp++ = '\"';

     }
   else if (code >= 040 && code < 0177)
     }
   else if (code >= 040 && code < 0177)

-    *p++ = code;
+    *cp++ = code;

   else if (code == '\t')
     {
   else if (code == '\t')
     {

-      *p++ = '\\';
-      *p++ = 't';
+      *cp++ = '\\';
+      *cp++ = 't';

     }
   else if (code == '\n')
     {
     }
   else if (code == '\n')
     {

-      *p++ = '\\';
-      *p++ = 'n';
+      *cp++ = '\\';
+      *cp++ = 'n';

     }
   else if (code == '\r')
     {
     }
   else if (code == '\r')
     {

-      *p++ = '\\';
-      *p++ = 'r';
+      *cp++ = '\\';
+      *cp++ = 'r';

     }
   else if (code == '\v')
     {
     }
   else if (code == '\v')
     {

-      *p++ = '\\';
-      *p++ = 'v';
+      *cp++ = '\\';
+      *cp++ = 'v';

     }
   else if (code == '\b')
     {
     }
   else if (code == '\b')
     {

-      *p++ = '\\';
-      *p++ = 'b';
+      *cp++ = '\\';
+      *cp++ = 'b';

     }
   else if (code == '\f')
     {
     }
   else if (code == '\f')
     {

-      *p++ = '\\';
-      *p++ = 'f';
+      *cp++ = '\\';
+      *cp++ = 'f';

     }
   else
     {
     }
   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;
 }
   *pcode = code;
   return !wasquote;
 }


@@ -331,7 +327,6 @@ unlex (int token)
 void
 read_type_name (FILE *fin)
 {
 void
 read_type_name (FILE *fin)
 {

-  char *p = token_buffer;

   int c = getc (fin);
 
   while (c != '>')
   int c = getc (fin);
 
   while (c != '>')


@@ -345,13 +340,11 @@ read_type_name (FILE *fin)
          break;
        }
 
          break;
        }
 

-      if (p == token_buffer + maxtoken)
-       p = grow_token_buffer (p);
-
-      *p++ = c;
+      obstack_1grow (&token_obstack, c);

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

-  *p = 0;
+  obstack_1grow (&token_obstack, '\0');
+  token_buffer = obstack_finish (&token_obstack);

 }
 
 
 }
 
 


@@ -359,7 +352,9 @@ token_t
 lex (void)
 {
   int c;
 lex (void)
 {
   int c;

-  char *p;
+
+  /* Just to make sure. */
+  token_buffer = NULL;

 
   if (unlexed >= 0)
     {
 
   if (unlexed >= 0)
     {


@@ -370,14 +365,11 @@ lex (void)
     }
 
   c = skip_white_space ();
     }
 
   c = skip_white_space ();

-  /* for error messages (token buffer always valid) */
-  *token_buffer = c;
-  token_buffer[1] = 0;

 
   switch (c)
     {
     case EOF:
 
   switch (c)
     {
     case EOF:

-      strcpy (token_buffer, "EOF");
+      token_buffer = "EOF";

       return tok_eof;
 
     case 'A':    case 'B':    case 'C':    case 'D':    case 'E':
       return tok_eof;
 
     case 'A':    case 'B':    case 'C':    case 'D':    case 'E':


@@ -394,17 +386,13 @@ lex (void)
     case 'z':
     case '.':    case '_':
 
     case 'z':
     case '.':    case '_':
 

-      p = token_buffer;

       while (isalnum (c) || c == '_' || c == '.')
        {
       while (isalnum (c) || c == '_' || c == '.')
        {

-         if (p == token_buffer + maxtoken)
-           p = grow_token_buffer (p);
-
-         *p++ = c;
+         obstack_1grow (&token_obstack, c);

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

-
-      *p = 0;
+      obstack_1grow (&token_obstack, '\0');
+      token_buffer = obstack_finish (&token_obstack);

       ungetc (c, finput);
       symval = getsym (token_buffer);
       return tok_identifier;
       ungetc (c, finput);
       symval = getsym (token_buffer);
       return tok_identifier;


@@ -414,17 +402,14 @@ lex (void)
       {
        numval = 0;
 
       {
        numval = 0;
 

-       p = token_buffer;

        while (isdigit (c))
          {
        while (isdigit (c))
          {

-           if (p == token_buffer + maxtoken)
-             p = grow_token_buffer (p);
-
-           *p++ = c;
+           obstack_1grow (&token_obstack, c);

            numval = numval * 10 + c - '0';
            c = getc (finput);
          }
            numval = numval * 10 + c - '0';
            c = getc (finput);
          }

-       *p = 0;
+       obstack_1grow (&token_obstack, '\0');
+       token_buffer = obstack_finish (&token_obstack);

        ungetc (c, finput);
        return tok_number;
       }
        ungetc (c, finput);
        return tok_number;
       }


@@ -435,25 +420,21 @@ lex (void)
       translations = -1;
       {
        int code, discode;
       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);
        if (c != '\'')
          {
            complain (_("use \"...\" for multi-character literal tokens"));
            while (1)
 
        c = getc (finput);
        if (c != '\'')
          {
            complain (_("use \"...\" for multi-character literal tokens"));
            while (1)

-             {
-               dp = discard;
-               if (!literalchar (&dp, &discode, '\''))
-                 break;
-             }
+             if (!literalchar (0, &discode, '\''))
+               break;

          }
          }

-       *p++ = '\'';
-       *p = 0;
+       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 = getsym (token_buffer);
        symval->class = token_sym;
        if (!symval->user_token_number)


@@ -467,15 +448,13 @@ lex (void)
       translations = -1;
       {
        int code;               /* ignored here */
       translations = -1;
       {
        int code;               /* ignored here */

-       p = token_buffer;
-       *p++ = '\"';
+
+       obstack_1grow (&token_obstack, '\"');

        /* Read up to and including ".  */
        /* Read up to and including ".  */

-       while (literalchar (&p, &code, '\"'))
-         {
-           if (p >= token_buffer + maxtoken - 4)
-             p = grow_token_buffer (p);
-         }
-       *p = 0;
+       while (literalchar (&token_obstack, &code, '\"'))
+         /* nothing */;
+       obstack_1grow (&token_obstack, '\0');
+       token_buffer = obstack_finish (&token_obstack);

 
        symval = getsym (token_buffer);
        symval->class = token_sym;
 
        symval = getsym (token_buffer);
        symval->class = token_sym;


@@ -509,7 +488,7 @@ lex (void)
 
       if (c == '{')
        {
 
       if (c == '{')
        {

-         strcpy (token_buffer, "={");
+         token_buffer = "={";

          return tok_left_curly;
        }
       else
          return tok_left_curly;
        }
       else


@@ -556,14 +535,14 @@ struct percent_table_struct percent_table[] =
   { "nonassoc",                NULL,                   tok_nonassoc },
   { "binary",          NULL,                   tok_nonassoc },
   { "prec",            NULL,                   tok_prec },
   { "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",             &raw_flag,              tok_noop },             /* -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 */
+  { "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
   /* For the time being, this is not enabled yet, while it's possible
      though, since we use obstacks.  The only risk is with semantic
 #if 0
   /* For the time being, this is not enabled yet, while it's possible
      though, since we use obstacks.  The only risk is with semantic


@@ -573,12 +552,13 @@ struct percent_table_struct percent_table[] =
   { "file_prefix",     &spec_file_prefix,      tok_setopt },   /* -b */
   { "name_prefix",     &spec_name_prefix,      tok_setopt },   /* -p */
 #endif
   { "file_prefix",     &spec_file_prefix,      tok_setopt },   /* -b */
   { "name_prefix",     &spec_name_prefix,      tok_setopt },   /* -p */
 #endif

-  { "verbose",         &verbose_flag,          tok_noop },             /* -v */
-  { "debug",           &debug_flag,            tok_noop },             /* -t */
+  { "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 },
   { "semantic_parser", &semantic_parser,       tok_noop },
   { "pure_parser",     &pure_parser,           tok_noop },

-/*    {"help", <print usage stmt>, tok_noop}, *//* -h */
-/*    {"version", <print version number> ,  tok_noop}, *//* -V */
+

   { NULL, NULL, tok_illegal}
 };
 
   { NULL, NULL, tok_illegal}
 };
 


@@ -589,14 +569,9 @@ int
 parse_percent_token (void)
 {
   int c;
 parse_percent_token (void)
 {
   int c;

-  char *p;

   struct percent_table_struct *tx;
 
   struct percent_table_struct *tx;
 

-  p = token_buffer;

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

-  *p++ = '%';
-  *p++ = c;                    /* for error msg */
-  *p = 0;

 
   switch (c)
     {
 
   switch (c)
     {


@@ -621,40 +596,45 @@ parse_percent_token (void)
     case '=':
       return tok_prec;
     }
     case '=':
       return tok_prec;
     }

+

   if (!isalpha (c))
     return tok_illegal;
 
   if (!isalpha (c))
     return tok_illegal;
 

-  p = token_buffer;
-  *p++ = '%';
+  obstack_1grow (&token_obstack, '%');

   while (isalpha (c) || c == '_' || c == '-')
     {
   while (isalpha (c) || c == '_' || c == '-')
     {

-      if (p == token_buffer + maxtoken)
-       p = grow_token_buffer (p);
-

       if (c == '-')
        c = '_';
       if (c == '-')
        c = '_';

-      *p++ = c;
+      obstack_1grow (&token_obstack, c);

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

-
-  *p = 0;
+  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)
       break;
 
 
   /* table lookup % directive */
   for (tx = percent_table; tx->name; tx++)
     if (strcmp (token_buffer + 1, tx->name) == 0)
       break;
 

-  if (tx->retval == tok_setopt)
+  if (tx->set_flag)

     {
     {

-      *((char **) (tx->set_flag)) = optarg;
+      *((int *) (tx->set_flag)) = 1;

       return tok_noop;
     }
       return tok_noop;
     }

-  if (tx->set_flag)
+
+  switch (tx->retval)

     {
     {

-      *((int *) (tx->set_flag)) = 1;
+    case tok_setopt:
+      *((char **) (tx->set_flag)) = optarg;

       return tok_noop;
       return tok_noop;

+      break;
+
+    case tok_obsolete:
+      fatal (_("`%s' is no longer supported"), token_buffer);
+      break;

     }
     }

+

   return tx->retval;
 }
   return tx->retval;
 }