#include "adt/error.h"
#include "adt/strset.h"
#include "adt/util.h"
+#include "type_t.h"
#include <assert.h>
#include <errno.h>
//#define DEBUG_CHARS
#define MAX_PUTBACK 3
+#ifdef _WIN32
+/* No strtold on windows and no replacement yet */
+#define strtold(s, e) strtod(s, e)
+#endif
+
+#define HAS_SIGNED_CHAR
+//#define HAS_UNSIGNED_CHAR
+
+#if defined HAS_SIGNED_CHAR
+typedef signed char char_type;
+#elif defined HAS_UNSIGNED_CHAR
+typedef unsigned char char_type;
+#else
+# error signedness of char not determined
+#endif
+
static int c;
token_t lexer_token;
symbol_t *symbol_L;
static const char *bufpos;
static strset_t stringset;
+static type_t *type_int = NULL;
+static type_t *type_uint = NULL;
+static type_t *type_long = NULL;
+static type_t *type_ulong = NULL;
+static type_t *type_longlong = NULL;
+static type_t *type_ulonglong = NULL;
+static type_t *type_float = NULL;
+static type_t *type_double = NULL;
+static type_t *type_longdouble = NULL;
+
static void error_prefix_at(const char *input_name, unsigned linenr)
{
fprintf(stderr, "%s:%u: Error: ", input_name, linenr);
static void parse_integer_suffix(void)
{
- if(c == 'U' || c == 'U') {
- /* TODO do something with the suffixes... */
+ if(c == 'U' || c == 'u') {
next_char();
if(c == 'L' || c == 'l') {
next_char();
if(c == 'L' || c == 'l') {
next_char();
+ lexer_token.datatype = type_ulonglong;
+ } else {
+ lexer_token.datatype = type_ulong;
}
+ } else {
+ lexer_token.datatype = type_uint;
}
} else if(c == 'l' || c == 'L') {
next_char();
next_char();
if(c == 'u' || c == 'U') {
next_char();
+ lexer_token.datatype = type_ulonglong;
+ } else {
+ lexer_token.datatype = type_longlong;
}
} else if(c == 'u' || c == 'U') {
next_char();
+ lexer_token.datatype = type_ulong;
+ } else {
+ lexer_token.datatype = type_int;
}
+ } else {
+ lexer_token.datatype = type_int;
}
}
/* TODO: do something usefull with the suffixes... */
case 'f':
case 'F':
+ next_char();
+ lexer_token.datatype = type_float;
+ break;
case 'l':
case 'L':
next_char();
+ lexer_token.datatype = type_longdouble;
break;
default:
+ lexer_token.datatype = type_double;
break;
}
}
-static inline bool is_hex_digit(int c)
-{
- return (c >= '0' && c <= '9')
- || (c >= 'a' && c <= 'z')
- || (c >= 'A' && c <= 'Z');
+/**
+ * A replacement for strtoull. Only those parts needed for
+ * our parser are implemented.
+ */
+static unsigned long long parse_int_string(const char *s, const char **endptr, int base) {
+ unsigned long long v = 0;
+
+ switch (base) {
+ case 16:
+ for (;; ++s) {
+ /* check for overrun */
+ if (v >= 0x1000000000000000ULL)
+ break;
+ switch (tolower(*s)) {
+ case '0': v <<= 4; break;
+ case '1': v <<= 4; v |= 0x1; break;
+ case '2': v <<= 4; v |= 0x2; break;
+ case '3': v <<= 4; v |= 0x3; break;
+ case '4': v <<= 4; v |= 0x4; break;
+ case '5': v <<= 4; v |= 0x5; break;
+ case '6': v <<= 4; v |= 0x6; break;
+ case '7': v <<= 4; v |= 0x7; break;
+ case '8': v <<= 4; v |= 0x8; break;
+ case '9': v <<= 4; v |= 0x9; break;
+ case 'a': v <<= 4; v |= 0xa; break;
+ case 'b': v <<= 4; v |= 0xb; break;
+ case 'c': v <<= 4; v |= 0xc; break;
+ case 'd': v <<= 4; v |= 0xd; break;
+ case 'e': v <<= 4; v |= 0xe; break;
+ case 'f': v <<= 4; v |= 0xf; break;
+ default:
+ goto end;
+ }
+ }
+ break;
+ case 8:
+ for (;; ++s) {
+ /* check for overrun */
+ if (v >= 0x2000000000000000ULL)
+ break;
+ switch (tolower(*s)) {
+ case '0': v <<= 3; break;
+ case '1': v <<= 3; v |= 1; break;
+ case '2': v <<= 3; v |= 2; break;
+ case '3': v <<= 3; v |= 3; break;
+ case '4': v <<= 3; v |= 4; break;
+ case '5': v <<= 3; v |= 5; break;
+ case '6': v <<= 3; v |= 6; break;
+ case '7': v <<= 3; v |= 7; break;
+ default:
+ goto end;
+ }
+ }
+ break;
+ case 10:
+ for (;; ++s) {
+ /* check for overrun */
+ if (v > 0x1999999999999999ULL)
+ break;
+ switch (tolower(*s)) {
+ case '0': v *= 10; break;
+ case '1': v *= 10; v += 1; break;
+ case '2': v *= 10; v += 2; break;
+ case '3': v *= 10; v += 3; break;
+ case '4': v *= 10; v += 4; break;
+ case '5': v *= 10; v += 5; break;
+ case '6': v *= 10; v += 6; break;
+ case '7': v *= 10; v += 7; break;
+ case '8': v *= 10; v += 8; break;
+ case '9': v *= 10; v += 9; break;
+ default:
+ goto end;
+ }
+ }
+ break;
+ default:
+ assert(0);
+ break;
+ }
+end:
+ *endptr = s;
+ return v;
}
static void parse_number_hex(void)
assert(c == 'x' || c == 'X');
next_char();
- while(is_hex_digit(c)) {
+ while(isxdigit(c)) {
obstack_1grow(&symbol_obstack, c);
next_char();
}
lexer_token.type = T_ERROR;
}
- char *endptr;
+ const char *endptr;
lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = strtoll(string, &endptr, 16);
+ lexer_token.v.intvalue = parse_int_string(string, &endptr, 16);
if(*endptr != '\0') {
parse_error("hex number literal too long");
}
obstack_free(&symbol_obstack, string);
+ parse_integer_suffix();
}
static inline bool is_octal_digit(int chr)
obstack_1grow(&symbol_obstack, '\0');
char *string = obstack_finish(&symbol_obstack);
- char *endptr;
+ const char *endptr;
lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = strtoll(string, &endptr, 8);
+ lexer_token.v.intvalue = parse_int_string(string, &endptr, 8);
if(*endptr != '\0') {
parse_error("octal number literal too long");
}
obstack_1grow(&symbol_obstack, '\0');
char *string = obstack_finish(&symbol_obstack);
- char *endptr;
if(is_float) {
+ char *endptr;
lexer_token.type = T_FLOATINGPOINT;
lexer_token.v.floatvalue = strtold(string, &endptr);
parse_floating_suffix();
} else {
+ const char *endptr;
lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = strtoll(string, &endptr, 10);
+ lexer_token.v.intvalue = parse_int_string(string, &endptr, 10);
if(*endptr != '\0') {
parse_error("invalid number literal");
if (!is_octal_digit(c)) return value;
value = 8 * value + c - '0';
next_char();
- return value;
+ return (char_type)value;
}
static int parse_hex_sequence(void)
next_char();
}
- return value;
+ return (char_type)value;
}
static int parse_escape_sequence(void)
if(pp_token.type != T_INTEGER) {
parse_error("expected integer");
} else {
- lexer_token.source_position.linenr = pp_token.v.intvalue - 1;
+ lexer_token.source_position.linenr = (unsigned int)(pp_token.v.intvalue - 1);
next_pp_token();
}
if(pp_token.type == T_STRING_LITERAL) {
void init_lexer(void)
{
strset_init(&stringset);
+
+ type_int = make_atomic_type(ATOMIC_TYPE_INT, TYPE_QUALIFIER_CONST);
+ type_uint = make_atomic_type(ATOMIC_TYPE_UINT, TYPE_QUALIFIER_CONST);
+ type_long = make_atomic_type(ATOMIC_TYPE_LONG, TYPE_QUALIFIER_CONST);
+ type_ulong = make_atomic_type(ATOMIC_TYPE_ULONG, TYPE_QUALIFIER_CONST);
+ type_longlong = make_atomic_type(ATOMIC_TYPE_LONGLONG,
+ TYPE_QUALIFIER_CONST);
+ type_ulonglong = make_atomic_type(ATOMIC_TYPE_ULONGLONG,
+ TYPE_QUALIFIER_CONST);
+
+ type_float = make_atomic_type(ATOMIC_TYPE_FLOAT, TYPE_QUALIFIER_CONST);
+ type_double = make_atomic_type(ATOMIC_TYPE_DOUBLE,
+ TYPE_QUALIFIER_CONST);
+ type_longdouble = make_atomic_type(ATOMIC_TYPE_LONG_DOUBLE,
+ TYPE_QUALIFIER_CONST);
}
void lexer_open_stream(FILE *stream, const char *input_name)