static struct obstack token_obstack;
const char *token_buffer = NULL;
-bucket *symval = NULL;
+symbol_t *symval = NULL;
int numval;
/* A token to be reread, see unlex and lex. */
static token_t unlexed = tok_undef;
-static bucket *unlexed_symval = NULL;
+static symbol_t *unlexed_symval = NULL;
static const char *unlexed_token_buffer = NULL;
void
| Do a getc, but give error message if EOF encountered |
`-----------------------------------------------------*/
-static int
+int
xgetc (FILE *f)
{
int c = getc (f);
}
-/*-----------------------------------------------------------------.
-| Read one literal character from FINPUT. Process \-escapes. |
-| Append the char to OUT and assign it *PCODE. Return 1 unless the |
-| character is an unescaped `term' or \n report error for \n. |
-`-----------------------------------------------------------------*/
+/*---------------------------------------------------------------.
+| Read one literal character from FINPUT, process \-escapes, and |
+| return the character. |
+`---------------------------------------------------------------*/
-int
-literalchar (struct obstack *out, int *pcode, char term)
+char
+literalchar (void)
{
int c;
- int code;
- int wasquote = 0;
+ int res;
c = xgetc (finput);
if (c == '\n')
{
complain (_("unescaped newline in constant"));
ungetc (c, finput);
- code = '?';
- wasquote = 1;
+ res = '?';
}
else if (c != '\\')
{
- code = c;
- if (c == term)
- wasquote = 1;
+ res = c;
}
else
{
c = xgetc (finput);
if (c == 't')
- code = '\t';
+ res = '\t';
else if (c == 'n')
- code = '\n';
+ res = '\n';
else if (c == 'a')
- code = '\007';
+ res = '\007';
else if (c == 'r')
- code = '\r';
+ res = '\r';
else if (c == 'f')
- code = '\f';
+ res = '\f';
else if (c == 'b')
- code = '\b';
+ res = '\b';
else if (c == 'v')
- code = '\013';
+ res = '\013';
else if (c == '\\')
- code = '\\';
+ res = '\\';
else if (c == '\'')
- code = '\'';
+ res = '\'';
else if (c == '\"')
- code = '\"';
+ res = '\"';
else if (c <= '7' && c >= '0')
{
- code = 0;
+ res = 0;
while (c <= '7' && c >= '0')
{
- code = (code * 8) + (c - '0');
- if (code >= 256 || code < 0)
+ res = (res * 8) + (c - '0');
+ if (res >= 256 || res < 0)
{
complain (_("octal value outside range 0...255: `\\%o'"),
- code);
- code &= 0xFF;
+ res);
+ res &= 0xFF;
break;
}
c = xgetc (finput);
else if (c == 'x')
{
c = xgetc (finput);
- code = 0;
+ res = 0;
while (1)
{
if (c >= '0' && c <= '9')
- code *= 16, code += c - '0';
+ res *= 16, res += c - '0';
else if (c >= 'a' && c <= 'f')
- code *= 16, code += c - 'a' + 10;
+ res *= 16, res += c - 'a' + 10;
else if (c >= 'A' && c <= 'F')
- code *= 16, code += c - 'A' + 10;
+ res *= 16, res += c - 'A' + 10;
else
break;
- if (code >= 256 || code < 0)
+ if (res >= 256 || res < 0)
{
- complain (_("hexadecimal value above 255: `\\x%x'"), code);
- code &= 0xFF;
+ complain (_("hexadecimal value above 255: `\\x%x'"), res);
+ res &= 0xFF;
break;
}
c = xgetc (finput);
badchar[0] = c;
complain (_("unknown escape sequence: `\\' followed by `%s'"),
quote (badchar));
- code = '?';
+ res = '?';
}
} /* has \ */
- if (out)
- obstack_1grow (out, code);
- *pcode = code;
- return !wasquote;
+ return res;
}
/* parse the literal token and compute character code in code */
{
- int code;
+ int code = literalchar ();
obstack_1grow (&token_obstack, '\'');
- literalchar (&token_obstack, &code, '\'');
+ obstack_1grow (&token_obstack, code);
c = getc (finput);
if (c != '\'')
{
- int discode;
complain (_("use \"...\" for multi-character literal tokens"));
- while (1)
- if (!literalchar (0, &discode, '\''))
- break;
+ while (literalchar () != '\'')
+ /* Skip. */;
}
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 == SUNDEF)
- symval->user_token_number = code;
+ if (symval->number == NUMBER_UNDEFINED)
+ {
+ symval->number = ntokens++;
+ symval->class = token_sym;
+ if (symval->user_token_number == SUNDEF)
+ symval->user_token_number = code;
+ }
return tok_identifier;
}
obstack_1grow (&token_obstack, '\"');
/* Read up to and including ". */
- while (literalchar (&token_obstack, &code, '\"'))
- /* nothing */;
+ do
+ {
+ code = literalchar ();
+ obstack_1grow (&token_obstack, code);
+ }
+ while (code != '\"');
obstack_1grow (&token_obstack, '\0');
token_buffer = obstack_finish (&token_obstack);
symval = getsym (token_buffer);
- symval->class = token_sym;
+ if (symval->number == NUMBER_UNDEFINED)
+ {
+ symval->number = ntokens++;
+ symval->class = token_sym;
+ }
return tok_identifier;
}
size_t arg_offset = 0;
int c = getc (finput);
+ obstack_1grow (&token_obstack, '%');
+ obstack_1grow (&token_obstack, c);
- switch (c)
+ if (!isalpha (c))
{
- case '%':
- return tok_two_percents;
+ obstack_1grow (&token_obstack, '\0');
+ token_buffer = obstack_finish (&token_obstack);
- case '{':
- return tok_percent_left_curly;
+ switch (c)
+ {
+ case '%':
+ return tok_two_percents;
- /* FIXME: Who the heck are those 5 guys!?! `%<' = `%left'!!!
- Let's ask for there removal. */
- case '<':
- return tok_left;
+ case '{':
+ return tok_percent_left_curly;
- case '>':
- return tok_right;
+ /* The following guys are here for backward compatibility with
+ very ancient Yacc versions. The paper of Johnson mentions
+ them (as ancient :). */
+ case '<':
+ return tok_left;
- case '2':
- return tok_nonassoc;
+ case '>':
+ return tok_right;
- case '0':
- return tok_token;
+ case '2':
+ return tok_nonassoc;
- case '=':
- return tok_prec;
- }
+ case '0':
+ return tok_token;
- if (!isalpha (c))
- return tok_illegal;
+ case '=':
+ return tok_prec;
- obstack_1grow (&token_obstack, '%');
- while (isalpha (c) || c == '_' || c == '-')
+ default:
+ return tok_illegal;
+ }
+ }
+
+ while (c = getc (finput), isalpha (c) || c == '_' || c == '-')
{
if (c == '_')
c = '-';
obstack_1grow (&token_obstack, c);
- c = getc (finput);
}
/* %DIRECTIVE="ARG". Separate into
arg_offset = obstack_object_size (&token_obstack);
/* Read up to and including `"'. Do not append the closing
`"' in the output: it's not part of the ARG. */
- while (literalchar (NULL, &code, '"'))
+ while ((code = literalchar ()) != '"')
obstack_1grow (&token_obstack, code);
}
/* else: should be an error. */