+/*
+ * This file is part of cparser.
+ * Copyright (C) 2007-2009 Matthias Braun <matze@braunis.de>
+ *
+ * This program 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
+ * of the License, or (at your option) any later version.
+ *
+ * This program 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 this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ * 02111-1307, USA.
+ */
#include <config.h>
+#include "adt/strutil.h"
+#include "input.h"
#include "diagnostic.h"
#include "lexer.h"
+#include "symbol_t.h"
#include "token_t.h"
#include "symbol_table_t.h"
#include "adt/error.h"
#include "type_t.h"
#include "target_architecture.h"
#include "parser.h"
+#include "warning.h"
+#include "lang_features.h"
#include <assert.h>
#include <errno.h>
#include <stdbool.h>
#include <ctype.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
-
-#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
+#ifndef _WIN32
+#include <strings.h>
#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;
+#define MAX_PUTBACK 16 // 3 would be enough, but 16 gives a nicer alignment
+#define BUF_SIZE 1024
-static void error_prefix_at(const char *input_name, unsigned linenr)
-{
- fprintf(stderr, "%s:%u: Error: ", input_name, linenr);
-}
+static input_t *input;
+static utf32 input_buf[BUF_SIZE + MAX_PUTBACK];
+static const utf32 *bufpos;
+static const utf32 *bufend;
+static utf32 c;
+static source_position_t lexer_pos;
+token_t lexer_token;
+static symbol_t *symbol_L;
+static strset_t stringset;
+bool allow_dollar_in_symbol = true;
-static void error_prefix(void)
+/**
+ * Prints a parse error message at the current token.
+ *
+ * @param msg the error message
+ */
+static void parse_error(const char *msg)
{
- error_prefix_at(lexer_token.source_position.input_name,
- lexer_token.source_position.linenr);
+ errorf(&lexer_pos, "%s", msg);
}
-static void parse_error(const char *msg)
+/**
+ * Prints an internal error message at the current token.
+ *
+ * @param msg the error message
+ */
+static NORETURN internal_error(const char *msg)
{
- error_prefix();
- fprintf(stderr, "%s\n", msg);
+ internal_errorf(&lexer_pos, "%s", msg);
}
static inline void next_real_char(void)
{
- bufpos++;
- if(bufpos >= bufend) {
- size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK,
- input);
- if(s == 0) {
+ assert(bufpos <= bufend);
+ if (bufpos >= bufend) {
+ size_t n = decode(input, input_buf+MAX_PUTBACK, BUF_SIZE);
+ if (n == 0) {
c = EOF;
return;
}
- bufpos = buf + MAX_PUTBACK;
- bufend = buf + MAX_PUTBACK + s;
+ bufpos = input_buf + MAX_PUTBACK;
+ bufend = bufpos + n;
}
- c = *(bufpos);
+ c = *bufpos++;
+ ++lexer_pos.colno;
}
-static inline void put_back(int pc)
+/**
+ * Put a character back into the buffer.
+ *
+ * @param pc the character to put back
+ */
+static inline void put_back(utf32 const pc)
{
- assert(bufpos >= buf);
- //assert(bufpos < buf+MAX_PUTBACK || *bufpos == pc);
-
- char *p = buf + (bufpos - buf);
- *p = (char) pc;
-
- /* going backwards in the buffer is legal as long as it's not more often
- * than MAX_PUTBACK */
- bufpos--;
-
-#ifdef DEBUG_CHARS
- printf("putback '%c'\n", pc);
-#endif
+ *(--bufpos - input_buf + input_buf) = pc;
+ --lexer_pos.colno;
}
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)
+#define MATCH_NEWLINE(code) \
+ case '\r': \
+ next_char(); \
+ if (c == '\n') { \
+ case '\n': \
+ next_char(); \
+ } \
+ lexer_pos.lineno++; \
+ lexer_pos.colno = 1; \
+ code
+
+#define eat(c_type) (assert(c == c_type), next_char())
static void maybe_concat_lines(void)
{
eat('\\');
- switch(c) {
+ switch (c) {
MATCH_NEWLINE(return;)
default:
c = '\\';
}
+/**
+ * Set c to the next input character, ie.
+ * after expanding trigraphs.
+ */
static inline void next_char(void)
{
next_real_char();
/* filter trigraphs */
- if(UNLIKELY(c == '\\')) {
+ if (UNLIKELY(c == '\\')) {
maybe_concat_lines();
- goto end_of_next_char;
+ return;
}
- if(LIKELY(c != '?'))
- goto end_of_next_char;
+ if (LIKELY(c != '?'))
+ return;
next_real_char();
- if(LIKELY(c != '?')) {
+ if (LIKELY(c != '?')) {
put_back(c);
c = '?';
- goto end_of_next_char;
+ return;
}
next_real_char();
- switch(c) {
+ switch (c) {
case '=': c = '#'; break;
case '(': c = '['; break;
case '/': c = '\\'; maybe_concat_lines(); break;
c = '?';
break;
}
-
-end_of_next_char:;
-#ifdef DEBUG_CHARS
- printf("nchar '%c'\n", c);
-#endif
}
#define SYMBOL_CHARS \
+ case '$': if (!allow_dollar_in_symbol) goto dollar_sign; \
case 'a': \
case 'b': \
case 'c': \
case '8': \
case '9':
+/**
+ * Read a symbol from the input and build
+ * the lexer_token.
+ */
static void parse_symbol(void)
{
- symbol_t *symbol;
- char *string;
-
obstack_1grow(&symbol_obstack, (char) c);
next_char();
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
DIGITS
SYMBOL_CHARS
obstack_1grow(&symbol_obstack, (char) c);
break;
default:
+dollar_sign:
goto end_symbol;
}
}
end_symbol:
obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
- symbol = symbol_table_insert(string);
+ char *string = obstack_finish(&symbol_obstack);
+ symbol_t *symbol = symbol_table_insert(string);
- lexer_token.type = symbol->ID;
- lexer_token.v.symbol = symbol;
+ lexer_token.kind = symbol->ID;
+ lexer_token.identifier.symbol = symbol;
- if(symbol->string != string) {
+ if (symbol->string != string) {
obstack_free(&symbol_obstack, string);
}
}
-static void parse_integer_suffix(bool is_oct_hex)
+static string_t identify_string(char *string, size_t len)
{
- bool is_unsigned = false;
- bool min_long = false;
- bool min_longlong = false;
+ /* TODO hash */
+#if 0
+ const char *result = strset_insert(&stringset, concat);
+ if (result != concat) {
+ obstack_free(&symbol_obstack, concat);
+ }
+#else
+ const char *result = string;
+#endif
+ return (string_t) {result, len};
+}
- 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;
+/**
+ * parse suffixes like 'LU' or 'f' after numbers
+ */
+static void parse_number_suffix(void)
+{
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ while (true) {
+ switch (c) {
+ SYMBOL_CHARS
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- lexer_token.datatype = type_unsigned_long;
+ break;
+ default:
+ dollar_sign:
+ goto finish_suffix;
}
}
-
- 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;
- }
-
- 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;
+finish_suffix:
+ if (obstack_object_size(&symbol_obstack) == 0) {
+ lexer_token.number.suffix.begin = NULL;
+ lexer_token.number.suffix.size = 0;
+ return;
}
-}
-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;
- }
+ obstack_1grow(&symbol_obstack, '\0');
+ size_t size = obstack_object_size(&symbol_obstack);
+ char *string = obstack_finish(&symbol_obstack);
+
+ lexer_token.number.suffix = identify_string(string, size);
}
/**
- * A replacement for strtoull. Only those parts needed for
- * our parser are implemented.
+ * Parses a hex number including hex floats and set the
+ * lexer_token.
*/
-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();
+ bool is_float = false;
+ bool has_digits = false;
- while(isxdigit(c)) {
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ while (isxdigit(c)) {
+ has_digits = true;
obstack_1grow(&symbol_obstack, (char) c);
next_char();
}
- obstack_1grow(&symbol_obstack, '\0');
- char *string = obstack_finish(&symbol_obstack);
- if(c == '.' || c == 'p' || c == 'P') {
+ if (c == '.') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- panic("Hex floating point numbers not implemented yet");
+
+ while (isxdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
}
- if(*string == '\0') {
- parse_error("invalid hex number");
- lexer_token.type = T_ERROR;
+ if (c == 'p' || c == 'P') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+
+ if (c == '-' || c == '+') {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+
+ while (isxdigit(c)) {
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
+ } else if (is_float) {
+ errorf(&lexer_token.base.source_position,
+ "hexadecimal floatingpoint constant requires an exponent");
}
+ obstack_1grow(&symbol_obstack, '\0');
+
+ size_t size = obstack_object_size(&symbol_obstack) - 1;
+ char *string = obstack_finish(&symbol_obstack);
+ lexer_token.number.number = identify_string(string, size);
+
+ lexer_token.kind =
+ is_float ? T_FLOATINGPOINT_HEXADECIMAL : T_INTEGER_HEXADECIMAL;
- 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");
+ if (!has_digits) {
+ errorf(&lexer_token.base.source_position,
+ "invalid number literal '0x%S'", &lexer_token.number.number);
+ lexer_token.number.number.begin = "0";
+ lexer_token.number.number.size = 1;
}
- obstack_free(&symbol_obstack, string);
- parse_integer_suffix(true);
+ parse_number_suffix();
}
-static inline bool is_octal_digit(int chr)
+/**
+ * Returns true if the given char is a octal digit.
+ *
+ * @param char the character to check
+ */
+static bool is_octal_digit(utf32 chr)
{
return '0' <= chr && chr <= '7';
}
-static void parse_number_oct(void)
+/**
+ * Parses a number and sets the lexer_token.
+ */
+static void parse_number(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);
+ bool is_float = false;
+ bool has_digits = false;
- 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");
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ if (c == '0') {
+ next_char();
+ if (c == 'x' || c == 'X') {
+ next_char();
+ parse_number_hex();
+ return;
+ } else {
+ has_digits = true;
+ }
+ obstack_1grow(&symbol_obstack, '0');
}
- obstack_free(&symbol_obstack, string);
- parse_integer_suffix(true);
-}
-
-static void parse_number_dec(void)
-{
- bool is_float = false;
- while(isdigit(c)) {
+ while (isdigit(c)) {
+ has_digits = true;
obstack_1grow(&symbol_obstack, (char) c);
next_char();
}
- if(c == '.') {
+ if (c == '.') {
+ is_float = true;
obstack_1grow(&symbol_obstack, '.');
next_char();
- while(isdigit(c)) {
+ while (isdigit(c)) {
+ has_digits = true;
obstack_1grow(&symbol_obstack, (char) c);
next_char();
}
- is_float = true;
}
- if(c == 'e' || c == 'E') {
+ if (c == 'e' || c == 'E') {
+ is_float = true;
obstack_1grow(&symbol_obstack, 'e');
next_char();
- if(c == '-' || c == '+') {
+ if (c == '-' || c == '+') {
obstack_1grow(&symbol_obstack, (char) c);
next_char();
}
- while(isdigit(c)) {
+ while (isdigit(c)) {
obstack_1grow(&symbol_obstack, (char) c);
next_char();
}
- 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");
+ size_t size = obstack_object_size(&symbol_obstack) - 1;
+ char *string = obstack_finish(&symbol_obstack);
+ lexer_token.number.number = identify_string(string, size);
+
+ /* is it an octal number? */
+ if (is_float) {
+ lexer_token.kind = T_FLOATINGPOINT;
+ } else if (string[0] == '0') {
+ lexer_token.kind = T_INTEGER_OCTAL;
+
+ /* check for invalid octal digits */
+ for (size_t i= 0; i < size; ++i) {
+ char t = string[i];
+ if (t >= '8')
+ errorf(&lexer_token.base.source_position,
+ "invalid digit '%c' in octal number", t);
}
-
- 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");
- }
+ lexer_token.kind = T_INTEGER;
+ }
- parse_integer_suffix(false);
+ if (!has_digits) {
+ errorf(&lexer_token.base.source_position, "invalid number literal '%S'",
+ &lexer_token.number.number);
}
- obstack_free(&symbol_obstack, string);
+
+ parse_number_suffix();
}
-static void parse_number(void)
+/**
+ * Returns the value of a digit.
+ * The only portable way to do it ...
+ */
+static int digit_value(utf32 const digit)
{
- if (c == '0') {
- next_char();
- switch (c) {
- case 'X':
- 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();
+ switch (digit) {
+ case '0': return 0;
+ case '1': return 1;
+ case '2': return 2;
+ case '3': return 3;
+ case '4': return 4;
+ case '5': return 5;
+ case '6': return 6;
+ case '7': return 7;
+ case '8': return 8;
+ case '9': return 9;
+ case 'a':
+ case 'A': return 10;
+ case 'b':
+ case 'B': return 11;
+ case 'c':
+ case 'C': return 12;
+ case 'd':
+ case 'D': return 13;
+ case 'e':
+ case 'E': return 14;
+ case 'f':
+ case 'F': return 15;
+ default:
+ internal_error("wrong character given");
}
}
-static int parse_octal_sequence(const int first_digit)
+/**
+ * Parses an octal character sequence.
+ *
+ * @param first_digit the already read first digit
+ */
+static utf32 parse_octal_sequence(utf32 const first_digit)
{
assert(is_octal_digit(first_digit));
- int value = first_digit - '0';
+ utf32 value = digit_value(first_digit);
if (!is_octal_digit(c)) return value;
- value = 8 * value + c - '0';
+ value = 8 * value + digit_value(c);
next_char();
if (!is_octal_digit(c)) return value;
- value = 8 * value + c - '0';
+ value = 8 * value + digit_value(c);
next_char();
- return (char_type)value;
+ return value;
}
-static int parse_hex_sequence(void)
+/**
+ * Parses a hex character sequence.
+ */
+static utf32 parse_hex_sequence(void)
{
- int value = 0;
- while(1) {
- 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;
- }
+ utf32 value = 0;
+ while (isxdigit(c)) {
+ value = 16 * value + digit_value(c);
next_char();
}
-
- return (char_type)value;
+ return value;
}
-static int parse_escape_sequence(void)
+/**
+ * Parse an escape sequence.
+ */
+static utf32 parse_escape_sequence(void)
{
eat('\\');
- int ec = c;
+ utf32 const ec = c;
next_char();
- switch(ec) {
+ switch (ec) {
case '"': return '"';
case '\'': return '\'';
case '\\': return '\\';
case EOF:
parse_error("reached end of file while parsing escape sequence");
return EOF;
- default:
- parse_error("unknown escape sequence");
+ /* \E is not documented, but handled, by GCC. It is acceptable according
+ * to ยง6.11.4, whereas \e is not. */
+ case 'E':
+ case 'e':
+ if (c_mode & _GNUC)
+ return 27; /* hopefully 27 is ALWAYS the code for ESCAPE */
+ break;
+ case 'u':
+ case 'U':
+ parse_error("universal character parsing not implemented yet");
return EOF;
+ default:
+ break;
}
+ /* ยง6.4.4.4:8 footnote 64 */
+ parse_error("unknown escape sequence");
+ return EOF;
}
-const char *concat_strings(const char *s1, const char *s2)
+/**
+ * Concatenate two strings.
+ */
+string_t concat_strings(const string_t *const s1, const string_t *const s2)
{
- size_t len1 = strlen(s1);
- size_t len2 = strlen(s2);
+ const size_t len1 = s1->size - 1;
+ const size_t len2 = s2->size - 1;
- char *concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
- memcpy(concat, s1, len1);
- memcpy(concat + len1, s2, len2 + 1);
+ char *const concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
+ memcpy(concat, s1->begin, len1);
+ memcpy(concat + len1, s2->begin, len2 + 1);
- const char *result = strset_insert(&stringset, concat);
- if(result != concat) {
- obstack_free(&symbol_obstack, concat);
- }
+ return identify_string(concat, len1 + len2 + 1);
+}
- return result;
+string_t make_string(const char *string)
+{
+ size_t len = strlen(string) + 1;
+ char *const space = obstack_alloc(&symbol_obstack, len);
+ memcpy(space, string, len);
+
+ return identify_string(space, len);
}
+/**
+ * Parse a string literal and set lexer_token.
+ */
static void parse_string_literal(void)
{
- unsigned start_linenr = lexer_token.source_position.linenr;
- char *string;
- const char *result;
-
- assert(c == '"');
- next_char();
+ eat('"');
- int tc;
- while(1) {
- switch(c) {
- case '\\':
- tc = parse_escape_sequence();
- obstack_1grow(&symbol_obstack, (char) tc);
+ while (true) {
+ switch (c) {
+ case '\\': {
+ utf32 const tc = parse_escape_sequence();
+ if (tc >= 0x100) {
+ warningf(WARN_OTHER, &lexer_pos, "escape sequence out of range");
+ }
+ obstack_1grow(&symbol_obstack, tc);
break;
+ }
- 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;
+ case EOF: {
+ errorf(&lexer_token.base.source_position, "string has no end");
+ lexer_token.kind = T_ERROR;
return;
+ }
case '"':
next_char();
goto end_of_string;
default:
- obstack_1grow(&symbol_obstack, (char) c);
+ obstack_grow_symbol(&symbol_obstack, c);
next_char();
break;
}
/* add finishing 0 to the string */
obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
-
- /* check if there is already a copy of the string */
- result = strset_insert(&stringset, string);
- if(result != string) {
- obstack_free(&symbol_obstack, string);
- }
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack);
+ char *string = obstack_finish(&symbol_obstack);
- lexer_token.type = T_STRING_LITERAL;
- lexer_token.v.string = result;
+ lexer_token.kind = T_STRING_LITERAL;
+ lexer_token.string.string = identify_string(string, size);
}
+/**
+ * Parse a wide character constant and set lexer_token.
+ */
static void parse_wide_character_constant(void)
{
eat('\'');
- int found_char = 0;
- while(1) {
- switch(c) {
- case '\\':
- found_char = parse_escape_sequence();
+ while (true) {
+ switch (c) {
+ case '\\': {
+ const utf32 tc = parse_escape_sequence();
+ obstack_grow_symbol(&symbol_obstack, tc);
break;
+ }
MATCH_NEWLINE(
parse_error("newline while parsing character constant");
next_char();
goto end_of_wide_char_constant;
- case EOF:
- parse_error("EOF while parsing character constant");
- lexer_token.type = T_ERROR;
+ case EOF: {
+ errorf(&lexer_token.base.source_position,
+ "EOF while parsing character constant");
+ lexer_token.kind = T_ERROR;
return;
+ }
default:
- if(found_char != 0) {
- parse_error("more than 1 characters in character "
- "constant");
- goto end_of_wide_char_constant;
- } else {
- found_char = c;
- next_char();
- }
+ obstack_grow_symbol(&symbol_obstack, c);
+ next_char();
break;
}
}
-end_of_wide_char_constant:
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = found_char;
- lexer_token.datatype = type_wchar_t;
-}
-
-static void parse_wide_string_literal(void)
-{
- const unsigned start_linenr = lexer_token.source_position.linenr;
-
- assert(c == '"');
- next_char();
-
- while(1) {
- switch(c) {
- case '\\': {
- wchar_rep_t tc = parse_escape_sequence();
- obstack_grow(&symbol_obstack, &tc, sizeof(tc));
- break;
- }
-
- 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;
+end_of_wide_char_constant:;
+ obstack_1grow(&symbol_obstack, '\0');
+ size_t size = (size_t) obstack_object_size(&symbol_obstack) - 1;
+ char *string = obstack_finish(&symbol_obstack);
- case '"':
- next_char();
- goto end_of_string;
+ lexer_token.kind = T_WIDE_CHARACTER_CONSTANT;
+ lexer_token.string.string = identify_string(string, size);
- default: {
- wchar_rep_t tc = c;
- obstack_grow(&symbol_obstack, &tc, sizeof(tc));
- next_char();
- break;
- }
- }
+ if (size == 0) {
+ errorf(&lexer_token.base.source_position, "empty character constant");
}
+}
-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;
+/**
+ * Parse a wide string literal and set lexer_token.
+ */
+static void parse_wide_string_literal(void)
+{
+ parse_string_literal();
+ if (lexer_token.kind == T_STRING_LITERAL)
+ lexer_token.kind = T_WIDE_STRING_LITERAL;
}
+/**
+ * Parse a character constant and set lexer_token.
+ */
static void parse_character_constant(void)
{
eat('\'');
- int found_char = 0;
- while(1) {
- switch(c) {
- case '\\':
- found_char = parse_escape_sequence();
+ while (true) {
+ switch (c) {
+ case '\\': {
+ utf32 const tc = parse_escape_sequence();
+ if (tc >= 0x100) {
+ warningf(WARN_OTHER, &lexer_pos, "escape sequence out of range");
+ }
+ obstack_1grow(&symbol_obstack, tc);
break;
+ }
MATCH_NEWLINE(
parse_error("newline while parsing character constant");
next_char();
goto end_of_char_constant;
- case EOF:
- parse_error("EOF while parsing character constant");
- lexer_token.type = T_ERROR;
+ case EOF: {
+ errorf(&lexer_token.base.source_position,
+ "EOF while parsing character constant");
+ lexer_token.kind = T_ERROR;
return;
+ }
default:
- if(found_char != 0) {
- parse_error("more than 1 characters in character "
- "constant");
- goto end_of_char_constant;
- } else {
- found_char = c;
- next_char();
- }
+ obstack_grow_symbol(&symbol_obstack, c);
+ next_char();
break;
+
}
}
-end_of_char_constant:
- lexer_token.type = T_INTEGER;
- lexer_token.v.intvalue = found_char;
- lexer_token.datatype = type_int;
+end_of_char_constant:;
+ obstack_1grow(&symbol_obstack, '\0');
+ const size_t size = (size_t)obstack_object_size(&symbol_obstack)-1;
+ char *const string = obstack_finish(&symbol_obstack);
+
+ lexer_token.kind = T_CHARACTER_CONSTANT;
+ lexer_token.string.string = identify_string(string, size);
+
+ if (size == 0) {
+ errorf(&lexer_token.base.source_position, "empty character constant");
+ }
}
+/**
+ * Skip a multiline comment.
+ */
static void skip_multiline_comment(void)
{
- unsigned start_linenr = lexer_token.source_position.linenr;
-
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
+ case '/':
+ next_char();
+ if (c == '*') {
+ /* nested comment, warn here */
+ warningf(WARN_COMMENT, &lexer_pos, "'/*' within comment");
+ }
+ break;
case '*':
next_char();
- if(c == '/') {
+ if (c == '/') {
next_char();
return;
}
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");
+ case EOF: {
+ errorf(&lexer_token.base.source_position,
+ "at end of file while looking for comment end");
return;
+ }
default:
next_char();
}
}
+/**
+ * Skip a single line comment.
+ */
static void skip_line_comment(void)
{
- while(1) {
- switch(c) {
+ while (true) {
+ switch (c) {
case EOF:
return;
case '\r':
return;
+ case '\\':
+ next_char();
+ if (c == '\n' || c == '\r') {
+ warningf(WARN_COMMENT, &lexer_pos, "multi-line comment");
+ return;
+ }
+ break;
+
default:
next_char();
break;
}
}
+/** The current preprocessor token. */
static token_t pp_token;
+/**
+ * Read the next preprocessor token.
+ */
static inline void next_pp_token(void)
{
lexer_next_preprocessing_token();
pp_token = lexer_token;
}
+/**
+ * Eat all preprocessor tokens until newline.
+ */
static void eat_until_newline(void)
{
- while(pp_token.type != '\n' && pp_token.type != T_EOF) {
+ while (pp_token.kind != '\n' && pp_token.kind != T_EOF) {
next_pp_token();
}
}
-static void error_directive(void)
+/**
+ * Parse the line directive.
+ */
+static void parse_line_directive(void)
{
- error_prefix();
- fprintf(stderr, "#error directive: \n");
+ if (pp_token.kind != T_INTEGER) {
+ parse_error("expected integer");
+ } else {
+ /* use offset -1 as this is about the next line */
+ lexer_pos.lineno = atoi(pp_token.number.number.begin) - 1;
+ next_pp_token();
+ }
+ if (pp_token.kind == T_STRING_LITERAL) {
+ lexer_pos.input_name = pp_token.string.string.begin;
+ lexer_pos.is_system_header = false;
+ next_pp_token();
+
+ /* attempt to parse numeric flags as outputted by gcc preprocessor */
+ while (pp_token.kind == T_INTEGER) {
+ /* flags:
+ * 1 - indicates start of a new file
+ * 2 - indicates return from a file
+ * 3 - indicates system header
+ * 4 - indicates implicit extern "C" in C++ mode
+ *
+ * currently we're only interested in "3"
+ */
+ if (streq(pp_token.number.number.begin, "3")) {
+ lexer_pos.is_system_header = true;
+ }
+ next_pp_token();
+ }
+ }
- /* parse pp-tokens until new-line */
+ eat_until_newline();
}
-static void define_directive(void)
+/**
+ * STDC pragmas.
+ */
+typedef enum stdc_pragma_kind_t {
+ STDC_UNKNOWN,
+ STDC_FP_CONTRACT,
+ STDC_FENV_ACCESS,
+ STDC_CX_LIMITED_RANGE
+} stdc_pragma_kind_t;
+
+/**
+ * STDC pragma values.
+ */
+typedef enum stdc_pragma_value_kind_t {
+ STDC_VALUE_UNKNOWN,
+ STDC_VALUE_ON,
+ STDC_VALUE_OFF,
+ STDC_VALUE_DEFAULT
+} stdc_pragma_value_kind_t;
+
+/**
+ * Parse a pragma directive.
+ */
+static void parse_pragma(void)
{
- lexer_next_preprocessing_token();
- if(lexer_token.type != T_IDENTIFIER) {
- parse_error("expected identifier after #define\n");
+ bool unknown_pragma = true;
+
+ next_pp_token();
+ if (pp_token.kind != T_IDENTIFIER) {
+ warningf(WARN_UNKNOWN_PRAGMAS, &pp_token.base.source_position,
+ "expected identifier after #pragma");
eat_until_newline();
+ return;
}
-}
-static void ifdef_directive(int is_ifndef)
-{
- (void) is_ifndef;
- lexer_next_preprocessing_token();
- //expect_identifier();
- //extect_newline();
-}
-
-static void endif_directive(void)
-{
- //expect_newline();
-}
+ symbol_t *symbol = pp_token.identifier.symbol;
+ if (symbol->pp_ID == TP_STDC) {
+ stdc_pragma_kind_t kind = STDC_UNKNOWN;
+ /* a STDC pragma */
+ if (c_mode & _C99) {
+ next_pp_token();
-static void parse_line_directive(void)
-{
- if(pp_token.type != T_INTEGER) {
- parse_error("expected integer");
+ switch (pp_token.identifier.symbol->pp_ID) {
+ case TP_FP_CONTRACT:
+ kind = STDC_FP_CONTRACT;
+ break;
+ case TP_FENV_ACCESS:
+ kind = STDC_FENV_ACCESS;
+ break;
+ case TP_CX_LIMITED_RANGE:
+ kind = STDC_CX_LIMITED_RANGE;
+ break;
+ default:
+ break;
+ }
+ if (kind != STDC_UNKNOWN) {
+ stdc_pragma_value_kind_t value = STDC_VALUE_UNKNOWN;
+ next_pp_token();
+ switch (pp_token.identifier.symbol->pp_ID) {
+ case TP_ON:
+ value = STDC_VALUE_ON;
+ break;
+ case TP_OFF:
+ value = STDC_VALUE_OFF;
+ break;
+ case TP_DEFAULT:
+ value = STDC_VALUE_DEFAULT;
+ break;
+ default:
+ break;
+ }
+ if (value != STDC_VALUE_UNKNOWN) {
+ unknown_pragma = false;
+ } else {
+ errorf(&pp_token.base.source_position,
+ "bad STDC pragma argument");
+ }
+ }
+ }
} 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;
- next_pp_token();
+ unknown_pragma = true;
}
-
eat_until_newline();
+ if (unknown_pragma) {
+ warningf(WARN_UNKNOWN_PRAGMAS, &pp_token.base.source_position,
+ "encountered unknown #pragma");
+ }
}
+/**
+ * Parse a preprocessor non-null directive.
+ */
static void parse_preprocessor_identifier(void)
{
- assert(pp_token.type == T_IDENTIFIER);
- symbol_t *symbol = pp_token.v.symbol;
+ assert(pp_token.kind == T_IDENTIFIER);
+ symbol_t *symbol = pp_token.identifier.symbol;
- 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;
+ switch (symbol->pp_ID) {
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:
- warningf(lexer_token.source_position, "encountered unknown #pragma");
- eat_until_newline();
+ parse_pragma();
+ break;
+ case TP_error:
+ /* TODO; output the rest of the line */
+ parse_error("#error directive");
break;
}
}
+/**
+ * Parse a preprocessor directive.
+ */
static void parse_preprocessor_directive(void)
{
next_pp_token();
- switch(pp_token.type) {
+ switch (pp_token.kind) {
case T_IDENTIFIER:
parse_preprocessor_identifier();
break;
case T_INTEGER:
parse_line_directive();
break;
+ case '\n':
+ /* NULL directive, see ยง6.10.7 */
+ break;
default:
parse_error("invalid preprocessor directive");
eat_until_newline();
#define MAYBE_PROLOG \
next_char(); \
- while(1) { \
- switch(c) {
+ while (true) { \
+ switch (c) {
#define MAYBE(ch, set_type) \
case ch: \
next_char(); \
- lexer_token.type = set_type; \
+ lexer_token.kind = set_type; \
return;
+/* must use this as last thing */
+#define MAYBE_MODE(ch, set_type, mode) \
+ case ch: \
+ if (c_mode & mode) { \
+ next_char(); \
+ lexer_token.kind = set_type; \
+ return; \
+ } \
+ /* fallthrough */
+
#define ELSE_CODE(code) \
default: \
- code; \
+ code \
+ return; \
} \
- } /* end of while(1) */ \
- break;
+ } /* end of while (true) */ \
#define ELSE(set_type) \
ELSE_CODE( \
- lexer_token.type = set_type; \
- return; \
+ lexer_token.kind = set_type; \
)
void lexer_next_preprocessing_token(void)
{
- while(1) {
- switch(c) {
+ while (true) {
+ lexer_token.base.source_position = lexer_pos;
+
+ switch (c) {
case ' ':
case '\t':
next_char();
break;
MATCH_NEWLINE(
- lexer_token.type = '\n';
+ lexer_token.kind = '\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();
+ if (lexer_token.identifier.symbol == symbol_L) {
+ switch (c) {
+ case '"': parse_wide_string_literal(); break;
+ case '\'': parse_wide_character_constant(); break;
}
}
return;
case '.':
MAYBE_PROLOG
- case '0':
- case '1':
- case '2':
- case '3':
- case '4':
- case '5':
- case '6':
- case '7':
- case '8':
- case '9':
+ DIGITS
put_back(c);
c = '.';
- parse_number_dec();
+ parse_number();
return;
case '.':
ELSE_CODE(
put_back(c);
c = '.';
- lexer_token.type = '.';
- return;
+ lexer_token.kind = '.';
)
ELSE('.')
case '&':
ELSE('/')
case '%':
MAYBE_PROLOG
- MAYBE('>', T_PERCENTGREATER)
+ MAYBE('>', '}')
MAYBE('=', T_PERCENTEQUAL)
case ':':
MAYBE_PROLOG
case '%':
MAYBE_PROLOG
- MAYBE(':', T_PERCENTCOLONPERCENTCOLON)
+ MAYBE(':', T_HASHHASH)
ELSE_CODE(
put_back(c);
c = '%';
- lexer_token.type = T_PERCENTCOLON;
- return;
+ lexer_token.kind = '#';
)
- ELSE(T_PERCENTCOLON)
+ ELSE('#')
ELSE('%')
case '<':
MAYBE_PROLOG
- MAYBE(':', T_LESSCOLON)
- MAYBE('%', T_LESSPERCENT)
+ MAYBE(':', '[')
+ MAYBE('%', '{')
MAYBE('=', T_LESSEQUAL)
case '<':
MAYBE_PROLOG
ELSE('|')
case ':':
MAYBE_PROLOG
- MAYBE('>', T_COLONGREATER)
+ MAYBE('>', ']')
+ MAYBE_MODE(':', T_COLONCOLON, _CXX)
ELSE(':')
case '=':
MAYBE_PROLOG
case ';':
case ',':
case '\\':
- lexer_token.type = c;
+ lexer_token.kind = c;
next_char();
return;
case EOF:
- lexer_token.type = T_EOF;
+ lexer_token.kind = T_EOF;
return;
default:
+dollar_sign:
+ errorf(&lexer_pos, "unknown character '%c' found", c);
next_char();
- error_prefix();
- fprintf(stderr, "unknown character '%c' found\n", c);
- lexer_token.type = T_ERROR;
+ lexer_token.kind = T_ERROR;
return;
}
}
void lexer_next_token(void)
{
lexer_next_preprocessing_token();
- if(lexer_token.type != '\n')
- return;
+ while (lexer_token.kind == '\n') {
newline_found:
- do {
lexer_next_preprocessing_token();
- } while(lexer_token.type == '\n');
+ }
- if(lexer_token.type == '#') {
+ if (lexer_token.kind == '#') {
parse_preprocessor_directive();
goto newline_found;
}
void init_lexer(void)
{
strset_init(&stringset);
+ symbol_L = symbol_table_insert("L");
}
-void lexer_open_stream(FILE *stream, const char *input_name)
+static void input_error(unsigned delta_lines, unsigned delta_cols,
+ const char *message)
{
- input = stream;
- lexer_token.source_position.linenr = 0;
- lexer_token.source_position.input_name = input_name;
+ lexer_pos.lineno += delta_lines;
+ lexer_pos.colno += delta_cols;
+ errorf(&lexer_pos, "%s", message);
+}
- symbol_L = symbol_table_insert("L");
+void lexer_switch_input(input_t *new_input, const char *input_name)
+{
+ lexer_pos.lineno = 0;
+ lexer_pos.colno = 0;
+ lexer_pos.input_name = input_name;
+
+ set_input_error_callback(input_error);
+ input = new_input;
+ bufpos = NULL;
+ bufend = NULL;
/* place a virtual \n at the beginning so the lexer knows that we're
* at the beginning of a line */
static __attribute__((unused))
void dbg_pos(const source_position_t source_position)
{
- fprintf(stdout, "%s:%u\n", source_position.input_name,
- source_position.linenr);
+ fprintf(stdout, "%s:%u:%u\n", source_position.input_name,
+ source_position.lineno, (unsigned)source_position.colno);
fflush(stdout);
}
projects / cparser / blobdiff