#include <config.h>
-#include "lexer_t.h"
+#include "diagnostic.h"
+#include "lexer.h"
#include "token_t.h"
#include "symbol_table_t.h"
#include "adt/error.h"
+#include "adt/strset.h"
+#include "adt/util.h"
+#include "types.h"
+#include "type_t.h"
+#include "target_architecture.h"
+#include "parser.h"
+#include "warning.h"
#include <assert.h>
#include <errno.h>
#include <string.h>
+#include <stdbool.h>
#include <ctype.h>
//#define DEBUG_CHARS
-#define MAX_PUTBACK 2
+#define MAX_PUTBACK 3
-static
-void error_prefix_at(lexer_t *this, const char *input_name, unsigned linenr)
+#ifdef _WIN32
+/* No strtold on windows and no replacement yet */
+#define strtold(s, e) strtod(s, e)
+#endif
+
+#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 FILE *input;
+static char buf[1024 + MAX_PUTBACK];
+static const char *bufend;
+static const char *bufpos;
+static strset_t stringset;
+
+static void error_prefix_at(const char *input_name, unsigned linenr)
{
- (void) this;
- fprintf(stderr, "%s:%d: Error: ", input_name, linenr);
+ fprintf(stderr, "%s:%u: Error: ", input_name, linenr);
}
-static
-void error_prefix(lexer_t *this)
+static void error_prefix(void)
{
- error_prefix_at(this, this->source_position.input_name,
- this->source_position.linenr);
+ error_prefix_at(lexer_token.source_position.input_name,
+ lexer_token.source_position.linenr);
}
-static
-void parse_error(lexer_t *this, const char *msg)
+static void parse_error(const char *msg)
{
- error_prefix(this);
+ error_prefix();
fprintf(stderr, "%s\n", msg);
}
-static inline
-void next_char(lexer_t *this)
+static inline void next_real_char(void)
{
- this->bufpos++;
- if(this->bufpos >= this->bufend) {
- size_t s = fread(this->buf + MAX_PUTBACK, 1,
- sizeof(this->buf) - MAX_PUTBACK, this->input);
+ assert(bufpos <= bufend);
+ if (bufpos >= bufend) {
+ size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK,
+ input);
if(s == 0) {
- this->c = EOF;
+ c = EOF;
return;
}
- this->bufpos = this->buf + MAX_PUTBACK;
- this->bufend = this->buf + MAX_PUTBACK + s;
+ bufpos = buf + MAX_PUTBACK;
+ bufend = buf + MAX_PUTBACK + s;
}
- this->c = *(this->bufpos);
-#ifdef DEBUG_CHARS
- printf("nchar '%c'\n", this->c);
-#endif
+ c = *bufpos++;
}
-static inline
-void put_back(lexer_t *this, int c)
+static inline void put_back(int pc)
{
- char *p = (char*) this->bufpos - 1;
- this->bufpos--;
- assert(p >= this->buf);
- *p = c;
+ assert(bufpos > buf);
+ *(--bufpos - buf + buf) = (char) pc;
#ifdef DEBUG_CHARS
- printf("putback '%c'\n", c);
+ printf("putback '%c'\n", pc);
#endif
}
-static
-int replace_trigraph(lexer_t *this)
-{
-#define MATCH_TRIGRAPH(ch,replacement) \
- case ch: \
- this->c = replacement; \
- return 1;
-
- switch(this->c) {
- MATCH_TRIGRAPH('=', '#')
- MATCH_TRIGRAPH('(', '[')
- MATCH_TRIGRAPH('/', '\\')
- MATCH_TRIGRAPH(')', ']')
- MATCH_TRIGRAPH('\'', '^')
- MATCH_TRIGRAPH('<', '{')
- MATCH_TRIGRAPH('!', '|')
- MATCH_TRIGRAPH('>', '}')
- MATCH_TRIGRAPH('-', '~')
+static inline void next_char(void);
+
+#define MATCH_NEWLINE(code) \
+ case '\r': \
+ next_char(); \
+ if(c == '\n') { \
+ next_char(); \
+ } \
+ lexer_token.source_position.linenr++; \
+ code \
+ case '\n': \
+ next_char(); \
+ lexer_token.source_position.linenr++; \
+ code
+
+#define eat(c_type) do { assert(c == c_type); next_char(); } while(0)
+
+static void maybe_concat_lines(void)
+{
+ eat('\\');
+
+ switch(c) {
+ MATCH_NEWLINE(return;)
+
default:
break;
}
- return 0;
-}
-
-#define SKIP_TRIGRAPHS(no_trigraph_code) \
- case '?': \
- next_char(this); \
- if(this->c != '?') { \
- put_back(this, this->c); \
- this->c = '?'; \
- no_trigraph_code; \
- } \
- next_char(this); \
- if(replace_trigraph(this)) \
- break; \
- put_back(this, '?'); \
- put_back(this, this->c); \
- this->c = '?'; \
- no_trigraph_code;
-
-#define EAT_NEWLINE(newline_code) \
- if(this->c == '\r') { \
- next_char(this); \
- if(this->c == '\n') \
- next_char(this); \
- this->source_position.linenr++; \
- newline_code; \
- } else if(this->c == '\n') { \
- next_char(this); \
- this->source_position.linenr++; \
- newline_code; \
- }
-
-static
-void parse_symbol(lexer_t *this, token_t *token)
+ put_back(c);
+ c = '\\';
+}
+
+static inline void next_char(void)
+{
+ next_real_char();
+
+ /* filter trigraphs */
+ if(UNLIKELY(c == '\\')) {
+ maybe_concat_lines();
+ goto end_of_next_char;
+ }
+
+ if(LIKELY(c != '?'))
+ goto end_of_next_char;
+
+ next_real_char();
+ if(LIKELY(c != '?')) {
+ put_back(c);
+ c = '?';
+ goto end_of_next_char;
+ }
+
+ next_real_char();
+ switch(c) {
+ case '=': c = '#'; break;
+ case '(': c = '['; break;
+ case '/': c = '\\'; maybe_concat_lines(); break;
+ case ')': c = ']'; break;
+ case '\'': c = '^'; break;
+ case '<': c = '{'; break;
+ case '!': c = '|'; break;
+ case '>': c = '}'; break;
+ case '-': c = '~'; break;
+ default:
+ put_back(c);
+ put_back('?');
+ c = '?';
+ break;
+ }
+
+end_of_next_char:;
+#ifdef DEBUG_CHARS
+ printf("nchar '%c'\n", c);
+#endif
+}
+
+#define SYMBOL_CHARS \
+ 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 'Z': \
+ case '_':
+
+#define DIGITS \
+ case '0': \
+ case '1': \
+ case '2': \
+ case '3': \
+ case '4': \
+ case '5': \
+ case '6': \
+ case '7': \
+ case '8': \
+ case '9':
+
+static void parse_symbol(void)
{
symbol_t *symbol;
char *string;
- obstack_1grow(&symbol_obstack, this->c);
- next_char(this);
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
while(1) {
- switch(this->c) {
- case '\\':
- next_char(this);
- EAT_NEWLINE(break;)
- goto end_symbol;
-
- case 'A' ... 'Z':
- case 'a' ... 'z':
- case '_':
- obstack_1grow(&symbol_obstack, this->c);
- next_char(this);
+ switch(c) {
+ DIGITS
+ SYMBOL_CHARS
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
break;
- case '?':
- next_char(this);
- if(this->c != '?') {
- put_back(this, this->c);
- this->c = '?';
- goto end_symbol;
- }
- next_char(this);
- if(replace_trigraph(this))
- break;
- put_back(this, '?');
- put_back(this, this->c);
- this->c = '?';
- goto end_symbol;
-
default:
goto end_symbol;
}
}
+
end_symbol:
obstack_1grow(&symbol_obstack, '\0');
string = obstack_finish(&symbol_obstack);
symbol = symbol_table_insert(string);
- if(symbol->ID > 0) {
- token->type = symbol->ID;
- } else {
- token->type = T_IDENTIFIER;
- }
- token->v.symbol = symbol;
+ lexer_token.type = symbol->ID;
+ lexer_token.v.symbol = symbol;
if(symbol->string != string) {
obstack_free(&symbol_obstack, string);
}
}
-#if 0
-static
-preprocessor_token_type_t parse_pp_symbol(lexer_t *this)
+static void parse_integer_suffix(bool is_oct_hex)
{
- do {
- obstack_1grow(&symbol_obstack, this->c);
- next_char(this);
- } while(is_ident_char(this->c));
- obstack_1grow(&symbol_obstack, '\0');
+ bool is_unsigned = false;
+ bool min_long = false;
+ bool min_longlong = false;
+
+ if(c == 'U' || c == 'u') {
+ is_unsigned = true;
+ next_char();
+ if(c == 'L' || c == 'l') {
+ min_long = true;
+ next_char();
+ if(c == 'L' || c == 'l') {
+ min_longlong = true;
+ next_char();
+ }
+ }
+ } else if(c == 'l' || c == 'L') {
+ min_long = true;
+ next_char();
+ if(c == 'l' || c == 'L') {
+ min_longlong = true;
+ next_char();
+ if(c == 'u' || c == 'U') {
+ is_unsigned = true;
+ next_char();
+ }
+ } else if(c == 'u' || c == 'U') {
+ is_unsigned = true;
+ next_char();
+ lexer_token.datatype = type_unsigned_long;
+ }
+ }
- char *string = obstack_finish(&symbol_obstack);
- symbol_t *symbol = preprocessor_symbol_table_find(string);
- obstack_free(&symbol_obstack, string);
+ if(!is_unsigned) {
+ long long v = lexer_token.v.intvalue;
+ if(!min_long) {
+ if(v >= TARGET_INT_MIN && v <= TARGET_INT_MAX) {
+ lexer_token.datatype = type_int;
+ return;
+ } else if(is_oct_hex && v >= 0 && v <= TARGET_UINT_MAX) {
+ lexer_token.datatype = type_unsigned_int;
+ return;
+ }
+ }
+ if(!min_longlong) {
+ if(v >= TARGET_LONG_MIN && v <= TARGET_LONG_MAX) {
+ lexer_token.datatype = type_long;
+ return;
+ } else if(is_oct_hex && v >= 0 && v <= TARGET_ULONG_MAX) {
+ lexer_token.datatype = type_unsigned_long;
+ return;
+ }
+ }
+ unsigned long long uv = (unsigned long long) v;
+ if(is_oct_hex && uv > (unsigned long long) TARGET_LONGLONG_MAX) {
+ lexer_token.datatype = type_unsigned_long_long;
+ return;
+ }
- if(symbol == 0)
- return TP_ERROR;
+ lexer_token.datatype = type_long_long;
+ } else {
+ unsigned long long v = (unsigned long long) lexer_token.v.intvalue;
+ if(!min_long && v <= TARGET_UINT_MAX) {
+ lexer_token.datatype = type_unsigned_int;
+ return;
+ }
+ if(!min_longlong && v <= TARGET_ULONG_MAX) {
+ lexer_token.datatype = type_unsigned_long;
+ return;
+ }
+ lexer_token.datatype = type_unsigned_long_long;
+ }
+}
- return symbol->ID;
+static void parse_floating_suffix(void)
+{
+ switch(c) {
+ /* 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_long_double;
+ break;
+ default:
+ lexer_token.datatype = type_double;
+ break;
+ }
}
-#endif
-static
-void parse_number_hex(lexer_t *this, token_t *token)
+/**
+ * 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(this->c == 'x' || this->c == 'X');
- next_char(this);
+ assert(c == 'x' || c == 'X');
+ next_char();
- if (!isdigit(this->c) &&
- !('A' <= this->c && this->c <= 'F') &&
- !('a' <= this->c && this->c <= 'f')) {
- parse_error(this, "premature end of hex number literal");
- token->type = T_ERROR;
- return;
+ while(isxdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
+ obstack_1grow(&symbol_obstack, '\0');
+ char *string = obstack_finish(&symbol_obstack);
- int value = 0;
- for(;;) {
- if (isdigit(this->c)) {
- value = 16 * value + this->c - '0';
- } else if ('A' <= this->c && this->c <= 'F') {
- value = 16 * value + this->c - 'A' + 10;
- } else if ('a' <= this->c && this->c <= 'f') {
- value = 16 * value + this->c - 'a' + 10;
- } else {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
- }
- next_char(this);
+ if(c == '.' || c == 'p' || c == 'P') {
+ next_char();
+ panic("Hex floating point numbers not implemented yet");
+ }
+ if(*string == '\0') {
+ parse_error("invalid hex number");
+ lexer_token.type = T_ERROR;
+ }
+
+ const char *endptr;
+ lexer_token.type = T_INTEGER;
+ 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(true);
}
-static
-void parse_number_oct(lexer_t *this, token_t *token)
+static inline bool is_octal_digit(int chr)
{
- assert(this->c == 'o' || this->c == 'O');
- next_char(this);
+ return '0' <= chr && chr <= '7';
+}
- int value = 0;
- for(;;) {
- if ('0' <= this->c && this->c <= '7') {
- value = 8 * value + this->c - '0';
- } else {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
- }
- next_char(this);
+static void parse_number_oct(void)
+{
+ while(is_octal_digit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+ obstack_1grow(&symbol_obstack, '\0');
+ char *string = obstack_finish(&symbol_obstack);
+
+ const char *endptr;
+ lexer_token.type = T_INTEGER;
+ lexer_token.v.intvalue = parse_int_string(string, &endptr, 8);
+ if(*endptr != '\0') {
+ parse_error("octal number literal too long");
}
+
+ obstack_free(&symbol_obstack, string);
+ parse_integer_suffix(true);
}
-static
-void parse_number_dec(lexer_t *this, token_t *token, int first_char)
+static void parse_number_dec(void)
{
- int value = 0;
- if(first_char > 0) {
- assert(first_char >= '0' && first_char <= '9');
- value = first_char - '0';
+ bool is_float = false;
+ while(isdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
- for(;;) {
- if (isdigit(this->c)) {
- value = 10 * value + this->c - '0';
- } else {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
+ if(c == '.') {
+ obstack_1grow(&symbol_obstack, '.');
+ next_char();
+
+ while(isdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+ is_float = true;
+ }
+ if(c == 'e' || c == 'E') {
+ obstack_1grow(&symbol_obstack, 'e');
+ next_char();
+
+ if(c == '-' || c == '+') {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+
+ while(isdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
- next_char(this);
+ is_float = true;
+ }
+
+ obstack_1grow(&symbol_obstack, '\0');
+ char *string = obstack_finish(&symbol_obstack);
+
+ if(is_float) {
+ char *endptr;
+ lexer_token.type = T_FLOATINGPOINT;
+ lexer_token.v.floatvalue = strtold(string, &endptr);
+
+ if(*endptr != '\0') {
+ parse_error("invalid number literal");
+ }
+
+ parse_floating_suffix();
+ } else {
+ const char *endptr;
+ lexer_token.type = T_INTEGER;
+ lexer_token.v.intvalue = parse_int_string(string, &endptr, 10);
+
+ if(*endptr != '\0') {
+ parse_error("invalid number literal");
+ }
+
+ parse_integer_suffix(false);
}
+ obstack_free(&symbol_obstack, string);
}
-static
-void parse_number(lexer_t *this, token_t *token)
+static void parse_number(void)
{
- // TODO check for overflow
- // TODO check for various invalid inputs sequences
-
- if (this->c == '0') {
- next_char(this);
- switch (this->c) {
+ if (c == '0') {
+ next_char();
+ switch (c) {
case 'X':
- case 'x': parse_number_hex(this, token); break;
- case 'o':
- case 'O': parse_number_oct(this, token); break;
- default: parse_number_dec(this, token, '0');
+ case 'x':
+ parse_number_hex();
+ break;
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ parse_number_oct();
+ break;
+ case '8':
+ case '9':
+ next_char();
+ parse_error("invalid octal number");
+ lexer_token.type = T_ERROR;
+ return;
+ case '.':
+ case 'e':
+ case 'E':
+ default:
+ obstack_1grow(&symbol_obstack, '0');
+ parse_number_dec();
+ return;
}
} else {
- parse_number_dec(this, token, 0);
+ parse_number_dec();
}
}
-static
-int parse_escape_sequence(lexer_t *this)
+static int parse_octal_sequence(const int first_digit)
{
+ assert(is_octal_digit(first_digit));
+ int value = first_digit - '0';
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + c - '0';
+ next_char();
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + c - '0';
+ next_char();
+ return (char_type)value;
+}
+
+static int parse_hex_sequence(void)
+{
+ int value = 0;
while(1) {
- int c = this->c;
- next_char(this);
+ if (c >= '0' && c <= '9') {
+ value = 16 * value + c - '0';
+ } else if ('A' <= c && c <= 'F') {
+ value = 16 * value + c - 'A' + 10;
+ } else if ('a' <= c && c <= 'f') {
+ value = 16 * value + c - 'a' + 10;
+ } else {
+ break;
+ }
+ next_char();
+ }
- switch(c) {
- case '"': return '"';
- case '\'': return'\'';
- case '\\':
- EAT_NEWLINE(break;)
- return '\\';
- case 'a': return '\a';
- case 'b': return '\b';
- case 'f': return '\f';
- case 'n': return '\n';
- case 'r': return '\r';
- case 't': return '\t';
- case 'v': return '\v';
- case 'x': /* TODO parse hex number ... */
- parse_error(this, "hex escape sequences not implemented yet");
- return EOF;
- case 0 ... 8: /* TODO parse octal number ... */
- parse_error(this, "octal escape sequences not implemented yet");
- return EOF;
- case '?':
- if(this->c != '?') {
- return '?';
- }
- /* might be a trigraph */
- next_char(this);
- if(replace_trigraph(this)) {
- break;
- }
- put_back(this, this->c);
- this->c = '?';
- return '?';
+ return (char_type)value;
+}
- case EOF:
- parse_error(this, "reached end of file while parsing escape sequence");
- return EOF;
- default:
- parse_error(this, "unknown escape sequence");
- return EOF;
- }
+static int parse_escape_sequence(void)
+{
+ eat('\\');
+
+ int ec = c;
+ next_char();
+
+ switch(ec) {
+ case '"': return '"';
+ case '\'': return '\'';
+ case '\\': return '\\';
+ case '?': return '\?';
+ case 'a': return '\a';
+ case 'b': return '\b';
+ case 'f': return '\f';
+ case 'n': return '\n';
+ case 'r': return '\r';
+ case 't': return '\t';
+ case 'v': return '\v';
+ case 'x':
+ return parse_hex_sequence();
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ return parse_octal_sequence(ec);
+ case EOF:
+ parse_error("reached end of file while parsing escape sequence");
+ return EOF;
+ default:
+ parse_error("unknown escape sequence");
+ return EOF;
}
}
-static
-void parse_string_literal(lexer_t *this, token_t *token)
+string_t concat_strings(const string_t *const s1, const string_t *const s2)
{
- unsigned start_linenr = this->source_position.linenr;
- char *string;
- const char *result;
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
- assert(this->c == '"');
- next_char(this);
+ char *const concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
+ memcpy(concat, s1->begin, len1);
+ memcpy(concat + len1, s2->begin, len2 + 1);
- while(1) {
- switch(this->c) {
- SKIP_TRIGRAPHS(
- obstack_1grow(&symbol_obstack, '?');
- next_char(this);
- break;
- )
+#if 0 /* TODO hash */
+ const char *result = strset_insert(&stringset, concat);
+ if(result != concat) {
+ obstack_free(&symbol_obstack, concat);
+ }
+
+ return result;
+#else
+ return (string_t){ concat, len1 + len2 + 1 };
+#endif
+}
+static void parse_string_literal(void)
+{
+ const unsigned start_linenr = lexer_token.source_position.linenr;
+
+ assert(c == '"');
+ next_char();
+
+ int tc;
+ while(1) {
+ switch(c) {
case '\\':
- next_char(this);
- EAT_NEWLINE(break;)
- int c = parse_escape_sequence(this);
- obstack_1grow(&symbol_obstack, c);
+ tc = parse_escape_sequence();
+ obstack_1grow(&symbol_obstack, (char) tc);
break;
case EOF:
- error_prefix_at(this, this->source_position.input_name,
+ error_prefix_at(lexer_token.source_position.input_name,
start_linenr);
fprintf(stderr, "string has no end\n");
- token->type = T_ERROR;
+ lexer_token.type = T_ERROR;
return;
case '"':
- next_char(this);
+ next_char();
goto end_of_string;
default:
- obstack_1grow(&symbol_obstack, this->c);
- next_char(this);
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
break;
}
}
/* add finishing 0 to the string */
obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack);
+ const char *const string = obstack_finish(&symbol_obstack);
+#if 0 /* TODO hash */
/* check if there is already a copy of the string */
- result = strset_insert(&this->stringset, string);
+ result = strset_insert(&stringset, string);
if(result != string) {
obstack_free(&symbol_obstack, string);
}
+#else
+ const char *const result = string;
+#endif
- token->type = T_STRING_LITERAL;
- token->v.string = result;
+ lexer_token.type = T_STRING_LITERAL;
+ lexer_token.v.string.begin = result;
+ lexer_token.v.string.size = size;
}
-#define MATCH_NEWLINE(code) \
- case '\r': \
- next_char(this); \
- if(this->c == '\n') { \
- next_char(this); \
- } \
- this->source_position.linenr++; \
- code; \
- case '\n': \
- next_char(this); \
- this->source_position.linenr++; \
- code;
-
-static
-void parse_character_constant(lexer_t *this, token_t *token)
+static void parse_wide_character_constant(void)
{
- assert(this->c == '\'');
- next_char(this);
+ eat('\'');
int found_char = 0;
while(1) {
- switch(this->c) {
- SKIP_TRIGRAPHS(
- found_char = '?';
- break;
- )
-
+ switch(c) {
case '\\':
- next_char(this);
- EAT_NEWLINE(break;)
- found_char = '\\';
+ found_char = parse_escape_sequence();
break;
MATCH_NEWLINE(
- parse_error(this, "newline while parsing character constant");
+ parse_error("newline while parsing character constant");
break;
)
case '\'':
- next_char(this);
- goto end_of_char_constant;
+ next_char();
+ goto end_of_wide_char_constant;
case EOF:
- parse_error(this, "EOF while parsing character constant");
- token->type = T_ERROR;
+ parse_error("EOF while parsing character constant");
+ lexer_token.type = T_ERROR;
return;
default:
if(found_char != 0) {
- parse_error(this, "more than 1 characters in character "
+ parse_error("more than 1 characters in character "
"constant");
- goto end_of_char_constant;
+ goto end_of_wide_char_constant;
} else {
- found_char = this->c;
- next_char(this);
+ found_char = c;
+ next_char();
}
break;
}
}
-end_of_char_constant:
- token->type = T_INTEGER;
- token->v.intvalue = found_char;
+end_of_wide_char_constant:
+ lexer_token.type = T_INTEGER;
+ lexer_token.v.intvalue = found_char;
+ lexer_token.datatype = type_wchar_t;
}
-static
-void skip_multiline_comment(lexer_t *this)
+static void parse_wide_string_literal(void)
{
- unsigned start_linenr = this->source_position.linenr;
- int had_star = 0;
+ const unsigned start_linenr = lexer_token.source_position.linenr;
+
+ assert(c == '"');
+ next_char();
while(1) {
- switch(this->c) {
- case '*':
- next_char(this);
- had_star = 1;
- break;
+ switch(c) {
+ case '\\': {
+ wchar_rep_t tc = parse_escape_sequence();
+ obstack_grow(&symbol_obstack, &tc, sizeof(tc));
+ break;
+ }
- case '/':
- next_char(this);
- if(had_star) {
+ case EOF:
+ error_prefix_at(lexer_token.source_position.input_name,
+ start_linenr);
+ fprintf(stderr, "string has no end\n");
+ lexer_token.type = T_ERROR;
return;
+
+ case '"':
+ next_char();
+ goto end_of_string;
+
+ default: {
+ wchar_rep_t tc = c;
+ obstack_grow(&symbol_obstack, &tc, sizeof(tc));
+ next_char();
+ break;
}
- had_star = 0;
- break;
+ }
+ }
+
+end_of_string:;
+ /* TODO: concatenate multiple strings separated by whitespace... */
+
+ /* add finishing 0 to the string */
+ wchar_rep_t nul = L'\0';
+ obstack_grow(&symbol_obstack, &nul, sizeof(nul));
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack) / sizeof(wchar_rep_t);
+ const wchar_rep_t *const string = obstack_finish(&symbol_obstack);
+
+#if 0 /* TODO hash */
+ /* check if there is already a copy of the string */
+ const wchar_rep_t *const result = strset_insert(&stringset, string);
+ if(result != string) {
+ obstack_free(&symbol_obstack, string);
+ }
+#else
+ const wchar_rep_t *const result = string;
+#endif
+
+ lexer_token.type = T_WIDE_STRING_LITERAL;
+ lexer_token.v.wide_string.begin = result;
+ lexer_token.v.wide_string.size = size;
+}
+
+static void parse_character_constant(void)
+{
+ eat('\'');
+
+ int found_char = 0;
+ while(1) {
+ switch(c) {
case '\\':
- next_char(this);
- EAT_NEWLINE(break;)
- had_star = 0;
+ found_char = parse_escape_sequence();
break;
- case '?':
- next_char(this);
- if(this->c != '?') {
- had_star = 0;
- break;
- }
- next_char(this);
- if(replace_trigraph(this))
- break;
- put_back(this, this->c);
- this->c = '?';
- had_star = 0;
- /* we don't put back the 2nd ? as the comment text is discarded
- * anyway */
+ MATCH_NEWLINE(
+ parse_error("newline while parsing character constant");
break;
+ )
- MATCH_NEWLINE(had_star = 0; break;)
+ case '\'':
+ next_char();
+ goto end_of_char_constant;
case EOF:
- error_prefix_at(this, this->source_position.input_name,
- start_linenr);
- fprintf(stderr, "at end of file while looking for comment end\n");
+ parse_error("EOF while parsing character constant");
+ lexer_token.type = T_ERROR;
return;
+
default:
- had_star = 0;
- next_char(this);
+ if(found_char != 0) {
+ parse_error("more than 1 characters in character "
+ "constant");
+ goto end_of_char_constant;
+ } else {
+ found_char = c;
+ next_char();
+ }
break;
}
}
+
+end_of_char_constant:
+ lexer_token.type = T_INTEGER;
+ lexer_token.v.intvalue = found_char;
+ lexer_token.datatype = type_int;
}
-static
-void skip_line_comment(lexer_t *this)
+static void skip_multiline_comment(void)
{
+ unsigned start_linenr = lexer_token.source_position.linenr;
+
while(1) {
- switch(this->c) {
- case '?':
- next_char(this);
- if(this->c != '?')
- break;
- next_char(this);
- if(replace_trigraph(this))
- break;
- put_back(this, '?');
- /* we don't put back the 2nd ? as the comment text is discarded
- * anyway */
+ switch(c) {
+ case '*':
+ next_char();
+ if(c == '/') {
+ next_char();
+ return;
+ }
break;
- case '\\':
- next_char(this);
- if(this->c == '\n') {
- next_char(this);
- this->source_position.linenr++;
- }
+ MATCH_NEWLINE(break;)
+
+ case EOF:
+ error_prefix_at(lexer_token.source_position.input_name,
+ start_linenr);
+ fprintf(stderr, "at end of file while looking for comment end\n");
+ return;
+
+ default:
+ next_char();
break;
+ }
+ }
+}
+static void skip_line_comment(void)
+{
+ while(1) {
+ switch(c) {
case EOF:
- case '\r':
+ return;
+
case '\n':
+ case '\r':
return;
default:
- next_char(this);
+ next_char();
break;
}
}
}
-static
-void parse_preprocessor_directive(lexer_t *this, token_t *result_token)
+static token_t pp_token;
+
+static inline void next_pp_token(void)
{
- (void) result_token;
- /* skip whitespaces */
- while(this->c == ' ' || this->c == '\t' || this->c == '\r') {
- next_char(this);
- }
+ lexer_next_preprocessing_token();
+ pp_token = lexer_token;
}
-void preprocessor_next_token(lexer_t *this, token_t *token)
+static void eat_until_newline(void)
{
- /* skip whitespaces */
- while(this->c == ' ' || this->c == '\t' || this->c == '\r') {
- next_char(this);
+ while(pp_token.type != '\n' && pp_token.type != T_EOF) {
+ next_pp_token();
}
+}
- switch(this->c) {
- case 'A' ... 'Z':
- case 'a' ... 'z':
- case '_':
- parse_symbol(this, token);
+static void error_directive(void)
+{
+ error_prefix();
+ fprintf(stderr, "#error directive: \n");
+
+ /* parse pp-tokens until new-line */
+}
+
+static void define_directive(void)
+{
+ lexer_next_preprocessing_token();
+ if(lexer_token.type != T_IDENTIFIER) {
+ parse_error("expected identifier after #define\n");
+ eat_until_newline();
}
}
-void lexer_next_token(lexer_t *this, token_t *token)
+static void ifdef_directive(int is_ifndef)
{
- int line_begin = 0;
+ (void) is_ifndef;
+ lexer_next_preprocessing_token();
+ //expect_identifier();
+ //extect_newline();
+}
- while(1) {
- switch(this->c) {
- case ' ':
- case '\t':
- next_char(this);
- break;
+static void endif_directive(void)
+{
+ //expect_newline();
+}
- MATCH_NEWLINE(break;)
+static void parse_line_directive(void)
+{
+ if(pp_token.type != T_INTEGER) {
+ parse_error("expected integer");
+ } else {
+ lexer_token.source_position.linenr = (unsigned int)(pp_token.v.intvalue - 1);
+ next_pp_token();
+ }
+ if(pp_token.type == T_STRING_LITERAL) {
+ lexer_token.source_position.input_name = pp_token.v.string.begin;
+ next_pp_token();
+ }
- case 'A' ... 'Z':
- case 'a' ... 'z':
- case '_':
- parse_symbol(this, token);
- return;
+ eat_until_newline();
+}
- case '0' ... '9':
- parse_number(this, token);
- return;
+static void parse_preprocessor_identifier(void)
+{
+ assert(pp_token.type == T_IDENTIFIER);
+ symbol_t *symbol = pp_token.v.symbol;
- case '"':
- parse_string_literal(this, token);
- return;
+ switch(symbol->pp_ID) {
+ case TP_include:
+ printf("include - enable header name parsing!\n");
+ break;
+ case TP_define:
+ define_directive();
+ break;
+ case TP_ifdef:
+ ifdef_directive(0);
+ break;
+ case TP_ifndef:
+ ifdef_directive(1);
+ break;
+ case TP_endif:
+ endif_directive();
+ break;
+ case TP_line:
+ next_pp_token();
+ parse_line_directive();
+ break;
+ case TP_if:
+ case TP_else:
+ case TP_elif:
+ case TP_undef:
+ case TP_error:
+ error_directive();
+ break;
+ case TP_pragma:
+ if (warning.unknown_pragmas) {
+ warningf(lexer_token.source_position, "encountered unknown #pragma");
+ }
+ eat_until_newline();
+ break;
+ }
+}
- case '\'':
- parse_character_constant(this, token);
- return;
+static void parse_preprocessor_directive(void)
+{
+ next_pp_token();
- case '\\':
- next_char(this);
- if(this->c == '\n') {
- next_char(this);
- this->source_position.linenr++;
- break;
- } else {
- parse_error(this, "unexpected '\\' found");
- token->type = T_ERROR;
- }
- return;
+ switch(pp_token.type) {
+ case T_IDENTIFIER:
+ parse_preprocessor_identifier();
+ break;
+ case T_INTEGER:
+ parse_line_directive();
+ break;
+ default:
+ parse_error("invalid preprocessor directive");
+ eat_until_newline();
+ break;
+ }
+}
#define MAYBE_PROLOG \
- next_char(this); \
+ next_char(); \
while(1) { \
- switch(this->c) {
+ switch(c) {
#define MAYBE(ch, set_type) \
case ch: \
- next_char(this); \
- token->type = set_type; \
+ next_char(); \
+ lexer_token.type = set_type; \
return;
#define ELSE_CODE(code) \
- SKIP_TRIGRAPHS( \
- code; \
- ) \
- \
- case '\\': \
- next_char(this); \
- if(this->c == '\n') { \
- next_char(this); \
- this->source_position.linenr++; \
- break; \
- } \
- /* fallthrough */ \
default: \
code; \
} \
#define ELSE(set_type) \
ELSE_CODE( \
- token->type = set_type; \
+ lexer_token.type = set_type; \
return; \
)
+void lexer_next_preprocessing_token(void)
+{
+ while(1) {
+ switch(c) {
+ case ' ':
+ case '\t':
+ next_char();
+ break;
+
+ MATCH_NEWLINE(
+ lexer_token.type = '\n';
+ return;
+ )
+
+ SYMBOL_CHARS
+ parse_symbol();
+ /* might be a wide string ( L"string" ) */
+ if(lexer_token.type == T_IDENTIFIER &&
+ lexer_token.v.symbol == symbol_L) {
+ if(c == '"') {
+ parse_wide_string_literal();
+ } else if(c == '\'') {
+ parse_wide_character_constant();
+ }
+ }
+ return;
+
+ DIGITS
+ parse_number();
+ return;
+
+ case '"':
+ parse_string_literal();
+ return;
+
+ case '\'':
+ parse_character_constant();
+ return;
+
case '.':
MAYBE_PROLOG
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+ put_back(c);
+ c = '.';
+ parse_number_dec();
+ return;
+
case '.':
MAYBE_PROLOG
MAYBE('.', T_DOTDOTDOT)
ELSE_CODE(
- put_back(this, this->c);
- this->c = '.';
- token->type = '.';
+ put_back(c);
+ c = '.';
+ lexer_token.type = '.';
return;
)
ELSE('.')
ELSE('+')
case '-':
MAYBE_PROLOG
+ MAYBE('>', T_MINUSGREATER)
MAYBE('-', T_MINUSMINUS)
MAYBE('=', T_MINUSEQUAL)
ELSE('-')
MAYBE_PROLOG
MAYBE('=', T_SLASHEQUAL)
case '*':
- next_char(this);
- skip_multiline_comment(this);
- lexer_next_token(this, token);
+ next_char();
+ skip_multiline_comment();
+ lexer_next_preprocessing_token();
return;
case '/':
- next_char(this);
- skip_line_comment(this);
- lexer_next_token(this, token);
+ next_char();
+ skip_line_comment();
+ lexer_next_preprocessing_token();
return;
ELSE('/')
case '%':
MAYBE_PROLOG
MAYBE(':', T_PERCENTCOLONPERCENTCOLON)
ELSE_CODE(
- put_back(this, this->c);
- this->c = '%';
- token->type = T_PERCENTCOLON;
+ put_back(c);
+ c = '%';
+ lexer_token.type = T_PERCENTCOLON;
return;
)
ELSE(T_PERCENTCOLON)
MAYBE_PROLOG
MAYBE(':', T_LESSCOLON)
MAYBE('%', T_LESSPERCENT)
+ MAYBE('=', T_LESSEQUAL)
case '<':
MAYBE_PROLOG
MAYBE('=', T_LESSLESSEQUAL)
ELSE('<')
case '>':
MAYBE_PROLOG
+ MAYBE('=', T_GREATEREQUAL)
case '>':
MAYBE_PROLOG
MAYBE('=', T_GREATERGREATEREQUAL)
case '#':
MAYBE_PROLOG
MAYBE('#', T_HASHHASH)
- ELSE_CODE(
- if(line_begin) {
- parse_preprocessor_directive(this, token);
- return;
- } else {
- token->type = '#';
- return;
- }
- )
+ ELSE('#')
case '?':
- next_char(this);
- /* just a simple ? */
- if(this->c != '?') {
- token->type = '?';
- return;
- }
- /* might be a trigraph */
- next_char(this);
- if(replace_trigraph(this)) {
- break;
- }
- put_back(this, this->c);
- this->c = '?';
- token->type = '?';
- return;
-
case '[':
case ']':
case '(':
case '~':
case ';':
case ',':
- token->type = this->c;
- next_char(this);
+ case '\\':
+ lexer_token.type = c;
+ next_char();
return;
case EOF:
- token->type = T_EOF;
+ lexer_token.type = T_EOF;
return;
default:
- next_char(this);
- error_prefix(this);
- fprintf(stderr, "unknown character '%c' found\n", this->c);
- token->type = T_ERROR;
+ next_char();
+ error_prefix();
+ fprintf(stderr, "unknown character '%c' found\n", c);
+ lexer_token.type = T_ERROR;
return;
}
}
}
-void lexer_init(lexer_t *this, FILE *stream, const char *input_name)
+void lexer_next_token(void)
{
- memset(this, 0, sizeof(this[0]));
+ lexer_next_preprocessing_token();
+ if(lexer_token.type != '\n')
+ return;
+
+newline_found:
+ do {
+ lexer_next_preprocessing_token();
+ } while(lexer_token.type == '\n');
- this->input = stream;
+ if(lexer_token.type == '#') {
+ parse_preprocessor_directive();
+ goto newline_found;
+ }
+}
+
+void init_lexer(void)
+{
+ strset_init(&stringset);
+}
+
+void lexer_open_stream(FILE *stream, const char *input_name)
+{
+ input = stream;
+ lexer_token.source_position.linenr = 0;
+ lexer_token.source_position.input_name = input_name;
- this->source_position.linenr = 0;
- this->source_position.input_name = input_name;
- strset_init(&this->stringset);
+ symbol_L = symbol_table_insert("L");
+ bufpos = NULL;
+ bufend = NULL;
- /* we place a virtual '\n' at the beginning so the lexer knows we're at the
- * beginning of a line */
- this->c = '\n';
+ /* place a virtual \n at the beginning so the lexer knows that we're
+ * at the beginning of a line */
+ c = '\n';
}
-void lexer_destroy(lexer_t *this)
+void exit_lexer(void)
{
- (void) this;
+ strset_destroy(&stringset);
}
static __attribute__((unused))
void dbg_pos(const source_position_t source_position)
{
- fprintf(stdout, "%s:%d\n", source_position.input_name, source_position.linenr);
+ fprintf(stdout, "%s:%u\n", source_position.input_name,
+ source_position.linenr);
fflush(stdout);
}