+/*
+ * 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 "lexer_t.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 "adt/strset.h"
+#include "adt/util.h"
+#include "types.h"
+#include "type_t.h"
+#include "parser.h"
+#include "warning.h"
+#include "lang_features.h"
#include <assert.h>
#include <errno.h>
#include <string.h>
+#include <stdbool.h>
#include <ctype.h>
-//#define DEBUG_CHARS
-#define MAX_PUTBACK 3
-
-static int c;
-source_position_t source_position;
-static FILE *input;
-static char buf[1027];
-static const char *bufend;
-static const char *bufpos;
-static strset_t stringset;
-//static FILE **input_stack;
-//static char **buf_stack;
-
-static
-void error_prefix_at(const char *input_name, unsigned linenr)
-{
- fprintf(stderr, "%s:%d: Error: ", input_name, linenr);
-}
+#ifndef _WIN32
+#include <strings.h>
+#endif
-static
-void error_prefix()
+#define MAX_PUTBACK 16 // 3 would be enough, but 16 gives a nicer alignment
+#define BUF_SIZE 1024
+
+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;
+
+/**
+ * Prints a parse error message at the current token.
+ *
+ * @param msg the error message
+ */
+static void parse_error(const char *msg)
{
- error_prefix_at(source_position.input_name, 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_char()
+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);
-#ifdef DEBUG_CHARS
- printf("nchar '%c'\n", c);
-#endif
+ 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)
{
- char *p = (char*) bufpos - 1;
- bufpos--;
- assert(p >= buf);
- *p = pc;
-
-#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 NEWLINE \
+ '\r': \
+ next_char(); \
+ if (c == '\n') { \
+ case '\n': \
+ next_char(); \
+ } \
+ lexer_pos.lineno++; \
+ lexer_pos.colno = 1; \
+ goto newline; \
+ newline // Let it look like an ordinary case label.
-static
-int replace_trigraph(void)
+#define eat(c_type) (assert(c == c_type), next_char())
+
+static void maybe_concat_lines(void)
{
-#define MATCH_TRIGRAPH(ch,replacement) \
- case ch: \
- c = replacement; \
- return 1;
+ eat('\\');
+
+ switch (c) {
+ case NEWLINE:
+ return;
- switch(c) {
- MATCH_TRIGRAPH('=', '#')
- MATCH_TRIGRAPH('(', '[')
- MATCH_TRIGRAPH('/', '\\')
- MATCH_TRIGRAPH(')', ']')
- MATCH_TRIGRAPH('\'', '^')
- MATCH_TRIGRAPH('<', '{')
- MATCH_TRIGRAPH('!', '|')
- MATCH_TRIGRAPH('>', '}')
- MATCH_TRIGRAPH('-', '~')
default:
break;
}
- return 0;
-}
-
-#define SKIP_TRIGRAPHS(custom_putback, no_trigraph_code) \
- case '?': \
- next_char(); \
- if(c != '?') { \
- custom_putback; \
- put_back(c); \
- c = '?'; \
- no_trigraph_code; \
- } \
- next_char(); \
- if(replace_trigraph()) { \
- break; \
- } \
- custom_putback; \
- put_back('?'); \
- put_back(c); \
- c = '?'; \
- no_trigraph_code;
-
-#define EAT_NEWLINE(newline_code) \
- if(c == '\r') { \
- next_char(); \
- if(c == '\n') \
- next_char(); \
- source_position.linenr++; \
- newline_code; \
- } else if(c == '\n') { \
- next_char(); \
- source_position.linenr++; \
- newline_code; \
+ put_back(c);
+ 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 == '\\')) {
+ maybe_concat_lines();
+ return;
}
+ if (LIKELY(c != '?'))
+ return;
+
+ next_real_char();
+ if (LIKELY(c != '?')) {
+ put_back(c);
+ c = '?';
+ return;
+ }
+
+ 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;
+ }
+}
+
#define SYMBOL_CHARS \
+ case '$': if (!allow_dollar_in_symbol) goto dollar_sign; \
case 'a': \
case 'b': \
case 'c': \
case '8': \
case '9':
-static
-void parse_symbol(token_t *token)
+static bool is_universal_char_valid(utf32 const v)
{
- symbol_t *symbol;
- char *string;
+ /* C11 ยง6.4.3:2 */
+ if (v < 0xA0U && v != 0x24 && v != 0x40 && v != 0x60)
+ return false;
+ if (0xD800 <= v && v <= 0xDFFF)
+ return false;
+ return true;
+}
- obstack_1grow(&symbol_obstack, c);
- next_char();
+static int digit_value(utf32 digit);
- while(1) {
- switch(c) {
- case '\\':
+static utf32 parse_universal_char(unsigned const n_digits)
+{
+ utf32 v = 0;
+ for (unsigned k = n_digits; k != 0; --k) {
+ if (isxdigit(c)) {
+ v = 16 * v + digit_value(c);
next_char();
- EAT_NEWLINE(break;)
- goto end_symbol;
+ } else {
+ errorf(&lexer_pos, "short universal character name, expected %u more digits", k);
+ break;
+ }
+ }
+ if (!is_universal_char_valid(v)) {
+ errorf(&lexer_pos, "\\%c%0*X is not a valid universal character name", n_digits == 4 ? 'u' : 'U', (int)n_digits, v);
+ }
+ return v;
+}
+
+static bool is_universal_char_valid_identifier(utf32 const v)
+{
+ /* C11 Annex D.1 */
+ if ( v == 0x000A8) return true;
+ if ( v == 0x000AA) return true;
+ if ( v == 0x000AD) return true;
+ if ( v == 0x000AF) return true;
+ if (0x000B2 <= v && v <= 0x000B5) return true;
+ if (0x000B7 <= v && v <= 0x000BA) return true;
+ if (0x000BC <= v && v <= 0x000BE) return true;
+ if (0x000C0 <= v && v <= 0x000D6) return true;
+ if (0x000D8 <= v && v <= 0x000F6) return true;
+ if (0x000F8 <= v && v <= 0x000FF) return true;
+ if (0x00100 <= v && v <= 0x0167F) return true;
+ if (0x01681 <= v && v <= 0x0180D) return true;
+ if (0x0180F <= v && v <= 0x01FFF) return true;
+ if (0x0200B <= v && v <= 0x0200D) return true;
+ if (0x0202A <= v && v <= 0x0202E) return true;
+ if (0x0203F <= v && v <= 0x02040) return true;
+ if ( v == 0x02054) return true;
+ if (0x02060 <= v && v <= 0x0206F) return true;
+ if (0x02070 <= v && v <= 0x0218F) return true;
+ if (0x02460 <= v && v <= 0x024FF) return true;
+ if (0x02776 <= v && v <= 0x02793) return true;
+ if (0x02C00 <= v && v <= 0x02DFF) return true;
+ if (0x02E80 <= v && v <= 0x02FFF) return true;
+ if (0x03004 <= v && v <= 0x03007) return true;
+ if (0x03021 <= v && v <= 0x0302F) return true;
+ if (0x03031 <= v && v <= 0x0303F) return true;
+ if (0x03040 <= v && v <= 0x0D7FF) return true;
+ if (0x0F900 <= v && v <= 0x0FD3D) return true;
+ if (0x0FD40 <= v && v <= 0x0FDCF) return true;
+ if (0x0FDF0 <= v && v <= 0x0FE44) return true;
+ if (0x0FE47 <= v && v <= 0x0FFFD) return true;
+ if (0x10000 <= v && v <= 0x1FFFD) return true;
+ if (0x20000 <= v && v <= 0x2FFFD) return true;
+ if (0x30000 <= v && v <= 0x3FFFD) return true;
+ if (0x40000 <= v && v <= 0x4FFFD) return true;
+ if (0x50000 <= v && v <= 0x5FFFD) return true;
+ if (0x60000 <= v && v <= 0x6FFFD) return true;
+ if (0x70000 <= v && v <= 0x7FFFD) return true;
+ if (0x80000 <= v && v <= 0x8FFFD) return true;
+ if (0x90000 <= v && v <= 0x9FFFD) return true;
+ if (0xA0000 <= v && v <= 0xAFFFD) return true;
+ if (0xB0000 <= v && v <= 0xBFFFD) return true;
+ if (0xC0000 <= v && v <= 0xCFFFD) return true;
+ if (0xD0000 <= v && v <= 0xDFFFD) return true;
+ if (0xE0000 <= v && v <= 0xEFFFD) return true;
+ return false;
+}
+static bool is_universal_char_valid_identifier_start(utf32 const v)
+{
+ /* C11 Annex D.2 */
+ if (0x0300 <= v && v <= 0x036F) return false;
+ if (0x1DC0 <= v && v <= 0x1DFF) return false;
+ if (0x20D0 <= v && v <= 0x20FF) return false;
+ if (0xFE20 <= v && v <= 0xFE2F) return false;
+ return true;
+}
+
+/**
+ * Read a symbol from the input and build
+ * the lexer_token.
+ */
+static void parse_symbol(void)
+{
+ while (true) {
+ switch (c) {
DIGITS
SYMBOL_CHARS
- obstack_1grow(&symbol_obstack, c);
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
break;
- case '?':
+ case '\\':
next_char();
- if(c != '?') {
+ switch (c) {
+ {
+ unsigned n;
+ case 'U': n = 8; goto universal;
+ case 'u': n = 4; goto universal;
+universal:
+ next_char();
+ utf32 const v = parse_universal_char(n);
+ if (!is_universal_char_valid_identifier(v)) {
+ if (is_universal_char_valid(v)) {
+ errorf(&lexer_pos, "universal character \\%c%0*X is not valid in an identifier", n == 4 ? 'u' : 'U', (int)n, v);
+ }
+ } else if (obstack_object_size(&symbol_obstack) == 0 && !is_universal_char_valid_identifier_start(v)) {
+ errorf(&lexer_pos, "universal character \\%c%0*X is not valid as start of an identifier", n == 4 ? 'u' : 'U', (int)n, v);
+ } else {
+ obstack_grow_symbol(&symbol_obstack, v);
+ }
+ break;
+ }
+
+ default:
put_back(c);
- c = '?';
+ c = '\\';
goto end_symbol;
}
- next_char();
- if(replace_trigraph())
- break;
- put_back('?');
- put_back(c);
- c = '?';
- goto end_symbol;
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);
- token->type = symbol->ID;
- token->v.symbol = symbol;
+ lexer_token.kind = symbol->ID;
+ lexer_token.base.symbol = symbol;
- if(symbol->string != string) {
+ if (symbol->string != string) {
obstack_free(&symbol_obstack, string);
}
}
-static
-void parse_number_hex(token_t *token)
+static string_t identify_string(char *string, size_t len)
{
- assert(c == 'x' || c == 'X');
- next_char();
-
- if (!isdigit(c) &&
- !('A' <= c && c <= 'F') &&
- !('a' <= c && c <= 'f')) {
- parse_error("premature end of hex number literal");
- token->type = T_ERROR;
- return;
+ /* 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};
+}
- int value = 0;
- for(;;) {
- if (isdigit(c)) {
- 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 {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
+/**
+ * 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();
+ break;
+ default:
+ dollar_sign:
+ goto finish_suffix;
}
- next_char();
}
+finish_suffix:
+ if (obstack_object_size(&symbol_obstack) == 0) {
+ lexer_token.number.suffix.begin = NULL;
+ lexer_token.number.suffix.size = 0;
+ return;
+ }
+
+ obstack_1grow(&symbol_obstack, '\0');
+ size_t size = obstack_object_size(&symbol_obstack) - 1;
+ char *string = obstack_finish(&symbol_obstack);
+
+ lexer_token.number.suffix = identify_string(string, size);
}
-static
-void parse_number_oct(token_t *token)
+static void parse_exponent(void)
{
- assert(c == 'o' || c == 'O');
- next_char();
-
- int value = 0;
- for(;;) {
- if ('0' <= c && c <= '7') {
- value = 8 * value + c - '0';
- } else {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
- }
+ if (c == '-' || c == '+') {
+ obstack_1grow(&symbol_obstack, (char)c);
next_char();
}
+
+ if (isdigit(c)) {
+ do {
+ obstack_1grow(&symbol_obstack, (char)c);
+ next_char();
+ } while (isdigit(c));
+ } else {
+ errorf(&lexer_token.base.source_position, "exponent has no digits");
+ }
}
-static
-void parse_number_dec(token_t *token, int first_char)
+/**
+ * Parses a hex number including hex floats and set the
+ * lexer_token.
+ */
+static void parse_number_hex(void)
{
- int value = 0;
- if(first_char > 0) {
- assert(first_char >= '0' && first_char <= '9');
- value = first_char - '0';
+ bool is_float = false;
+ bool has_digits = false;
+
+ while (isxdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
- for(;;) {
- if (isdigit(c)) {
- value = 10 * value + c - '0';
- } else {
- token->type = T_INTEGER;
- token->v.intvalue = value;
- return;
+ if (c == '.') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+
+ while (isxdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
}
+ }
+ if (c == 'p' || c == 'P') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
+ parse_exponent();
+ } else if (is_float) {
+ errorf(&lexer_token.base.source_position,
+ "hexadecimal floatingpoint constant requires an exponent");
}
-}
+ obstack_1grow(&symbol_obstack, '\0');
-static
-void parse_number(token_t *token)
-{
- // TODO check for overflow
- // TODO check for various invalid inputs sequences
+ size_t size = obstack_object_size(&symbol_obstack) - 1;
+ char *string = obstack_finish(&symbol_obstack);
+ lexer_token.number.number = identify_string(string, size);
- if (c == '0') {
- next_char();
- switch (c) {
- case 'X':
- case 'x': parse_number_hex(token); break;
- case 'o':
- case 'O': parse_number_oct(token); break;
- default: parse_number_dec(token, '0');
- }
- } else {
- parse_number_dec(token, 0);
+ lexer_token.kind = is_float ? T_FLOATINGPOINT : T_INTEGER;
+
+ if (!has_digits) {
+ errorf(&lexer_token.base.source_position, "invalid number literal '%S'", &lexer_token.number.number);
+ lexer_token.number.number.begin = "0";
+ lexer_token.number.number.size = 1;
}
+
+ parse_number_suffix();
}
-static
-int parse_escape_sequence()
+static void parse_number_bin(void)
{
- while(1) {
- int ec = c;
+ bool has_digits = false;
+
+ while (c == '0' || c == '1') {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char)c);
next_char();
+ }
+ obstack_1grow(&symbol_obstack, '\0');
- switch(ec) {
- 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("hex escape sequences not implemented yet");
- return EOF;
- case '0':
- case '1':
- case '2':
- case '3':
- case '4':
- case '5':
- case '6':
- case '7':
- /* TODO parse octal number ... */
- parse_error("octal escape sequences not implemented yet");
- return EOF;
- case '?':
- if(c != '?') {
- return '?';
- }
- /* might be a trigraph */
- next_char();
- if(replace_trigraph()) {
- break;
- }
- put_back(c);
- c = '?';
- return '?';
+ size_t const size = obstack_object_size(&symbol_obstack) - 1;
+ char *const string = obstack_finish(&symbol_obstack);
+ lexer_token.number.number = identify_string(string, size);
+ lexer_token.kind = T_INTEGER;
- case EOF:
- parse_error("reached end of file while parsing escape sequence");
- return EOF;
- default:
- parse_error("unknown escape sequence");
- return EOF;
- }
+ if (!has_digits) {
+ errorf(&lexer_token.base.source_position, "invalid number literal '%S'", &lexer_token.number.number);
+ lexer_token.number.number.begin = "0";
+ lexer_token.number.number.size = 1;
}
+
+ parse_number_suffix();
}
-static
-void parse_string_literal(token_t *token)
+/**
+ * Returns true if the given char is a octal digit.
+ *
+ * @param char the character to check
+ */
+static bool is_octal_digit(utf32 chr)
{
- unsigned start_linenr = source_position.linenr;
- char *string;
- const char *result;
+ return '0' <= chr && chr <= '7';
+}
- assert(c == '"');
- next_char();
+/**
+ * Parses a number and sets the lexer_token.
+ */
+static void parse_number(void)
+{
+ bool is_float = false;
+ bool has_digits = false;
- while(1) {
- switch(c) {
- SKIP_TRIGRAPHS(,
- obstack_1grow(&symbol_obstack, '?');
+ assert(obstack_object_size(&symbol_obstack) == 0);
+ if (c == '0') {
+ obstack_1grow(&symbol_obstack, (char)c);
+ next_char();
+ if (c == 'x' || c == 'X') {
+ obstack_1grow(&symbol_obstack, (char)c);
next_char();
- break;
- )
-
- case '\\':
+ parse_number_hex();
+ return;
+ } else if (c == 'b' || c == 'B') {
+ /* GCC extension: binary constant 0x[bB][01]+. */
+ obstack_1grow(&symbol_obstack, (char)c);
next_char();
- EAT_NEWLINE(break;)
- int ec = parse_escape_sequence();
- obstack_1grow(&symbol_obstack, ec);
- break;
-
- case EOF:
- error_prefix_at(source_position.input_name, start_linenr);
- fprintf(stderr, "string has no end\n");
- token->type = T_ERROR;
+ parse_number_bin();
return;
+ }
+ has_digits = true;
+ }
- case '"':
- next_char();
- goto end_of_string;
+ while (isdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
+ next_char();
+ }
- default:
- obstack_1grow(&symbol_obstack, c);
+ if (c == '.') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, '.');
+ next_char();
+
+ while (isdigit(c)) {
+ has_digits = true;
+ obstack_1grow(&symbol_obstack, (char) c);
next_char();
- break;
}
}
+ if (c == 'e' || c == 'E') {
+ is_float = true;
+ obstack_1grow(&symbol_obstack, 'e');
+ next_char();
+ parse_exponent();
+ }
-end_of_string:
+ 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);
- /* TODO: concatenate multiple strings separated by whitespace... */
+ if (is_float) {
+ lexer_token.kind = T_FLOATINGPOINT;
+ } else {
+ lexer_token.kind = T_INTEGER;
+
+ if (string[0] == '0') {
+ /* 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);
+ }
+ }
+ }
- /* add finishing 0 to the string */
- obstack_1grow(&symbol_obstack, '\0');
- string = obstack_finish(&symbol_obstack);
+ if (!has_digits) {
+ errorf(&lexer_token.base.source_position, "invalid number literal '%S'",
+ &lexer_token.number.number);
+ }
- /* check if there is already a copy of the string */
- result = strset_insert(&stringset, string);
- if(result != string) {
- obstack_free(&symbol_obstack, string);
+ parse_number_suffix();
+}
+
+/**
+ * Returns the value of a digit.
+ * The only portable way to do it ...
+ */
+static int digit_value(utf32 const digit)
+{
+ 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");
}
+}
- token->type = T_STRING_LITERAL;
- token->v.string = result;
+/**
+ * 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));
+ utf32 value = digit_value(first_digit);
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + digit_value(c);
+ next_char();
+ if (!is_octal_digit(c)) return value;
+ value = 8 * value + digit_value(c);
+ next_char();
+ return value;
}
-#define MATCH_NEWLINE(code) \
- case '\r': \
- next_char(); \
- if(c == '\n') { \
- next_char(); \
- } \
- source_position.linenr++; \
- code; \
- case '\n': \
- next_char(); \
- source_position.linenr++; \
- code;
+/**
+ * Parses a hex character sequence.
+ */
+static utf32 parse_hex_sequence(void)
+{
+ utf32 value = 0;
+ while (isxdigit(c)) {
+ value = 16 * value + digit_value(c);
+ next_char();
+ }
+ return value;
+}
-static
-void parse_character_constant(token_t *token)
+/**
+ * Parse an escape sequence.
+ */
+static utf32 parse_escape_sequence(void)
{
- assert(c == '\'');
+ eat('\\');
+
+ utf32 const ec = c;
next_char();
- int found_char = 0;
- while(1) {
- switch(c) {
- SKIP_TRIGRAPHS(,
- found_char = '?';
- break;
- )
+ 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;
+ /* \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 '\\':
- next_char();
- EAT_NEWLINE(break;)
- found_char = '\\';
- break;
+ case 'U': return parse_universal_char(8);
+ case 'u': return parse_universal_char(4);
+
+ default:
+ break;
+ }
+ /* ยง6.4.4.4:8 footnote 64 */
+ parse_error("unknown escape sequence");
+ return EOF;
+}
- MATCH_NEWLINE(
- parse_error("newline while parsing character constant");
+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);
+}
+
+static void parse_string(utf32 const delim, token_kind_t const kind, string_encoding_t const enc, char const *const context)
+{
+ eat(delim);
+
+ while (true) {
+ switch (c) {
+ case '\\': {
+ utf32 const tc = parse_escape_sequence();
+ if (enc == STRING_ENCODING_CHAR) {
+ if (tc >= 0x100) {
+ warningf(WARN_OTHER, &lexer_pos, "escape sequence out of range");
+ }
+ obstack_1grow(&symbol_obstack, tc);
+ } else {
+ obstack_grow_symbol(&symbol_obstack, tc);
+ }
break;
- )
+ }
- case '\'':
- next_char();
- goto end_of_char_constant;
+ case NEWLINE:
+ errorf(&lexer_pos, "newline while parsing %s", context);
+ break;
case EOF:
- parse_error("EOF while parsing character constant");
- token->type = T_ERROR;
- return;
+ errorf(&lexer_token.base.source_position, "EOF while parsing %s", context);
+ goto end_of_string;
default:
- if(found_char != 0) {
- parse_error("more than 1 characters in character "
- "constant");
- goto end_of_char_constant;
+ if (c == delim) {
+ next_char();
+ goto end_of_string;
} else {
- found_char = c;
+ obstack_grow_symbol(&symbol_obstack, c);
next_char();
+ break;
}
- break;
}
}
-end_of_char_constant:
- token->type = T_INTEGER;
- token->v.intvalue = found_char;
+end_of_string:
+ obstack_1grow(&symbol_obstack, '\0');
+ size_t const size = obstack_object_size(&symbol_obstack) - 1;
+ char *const string = obstack_finish(&symbol_obstack);
+
+ lexer_token.kind = kind;
+ lexer_token.string.encoding = enc;
+ lexer_token.string.string = identify_string(string, size);
}
-static
-void skip_multiline_comment(void)
+/**
+ * Parse a string literal and set lexer_token.
+ */
+static void parse_string_literal(string_encoding_t const enc)
{
- unsigned start_linenr = source_position.linenr;
- int had_star = 0;
+ parse_string('"', T_STRING_LITERAL, enc, "string literal");
+}
- while(1) {
- switch(c) {
- case '*':
- next_char();
- had_star = 1;
- break;
+/**
+ * Parse a character constant and set lexer_token.
+ */
+static void parse_character_constant(string_encoding_t const enc)
+{
+ parse_string('\'', T_CHARACTER_CONSTANT, enc, "character constant");
+ if (lexer_token.string.string.size == 0) {
+ errorf(&lexer_token.base.source_position, "empty character constant");
+ }
+}
+/**
+ * Skip a multiline comment.
+ */
+static void skip_multiline_comment(void)
+{
+ while (true) {
+ switch (c) {
case '/':
next_char();
- if(had_star) {
- return;
+ if (c == '*') {
+ /* nested comment, warn here */
+ warningf(WARN_COMMENT, &lexer_pos, "'/*' within comment");
}
- had_star = 0;
- break;
-
- case '\\':
- next_char();
- EAT_NEWLINE(break;)
- had_star = 0;
break;
-
- case '?':
+ case '*':
next_char();
- if(c != '?') {
- had_star = 0;
- break;
+ if (c == '/') {
+ next_char();
+ return;
}
- next_char();
- if(replace_trigraph())
- break;
- put_back(c);
- c = '?';
- had_star = 0;
- /* we don't put back the 2nd ? as the comment text is discarded
- * anyway */
break;
- MATCH_NEWLINE(had_star = 0; break;)
+ case NEWLINE:
+ break;
- case EOF:
- error_prefix_at(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:
- had_star = 0;
next_char();
break;
}
}
}
-static
-void skip_line_comment(void)
+/**
+ * Skip a single line comment.
+ */
+static void skip_line_comment(void)
{
- while(1) {
- switch(c) {
- case '?':
- next_char();
- if(c != '?')
- break;
- next_char();
- if(replace_trigraph())
- break;
- put_back('?');
- /* we don't put back the 2nd ? as the comment text is discarded
- * anyway */
- break;
+ while (true) {
+ switch (c) {
+ case EOF:
+ return;
+
+ case '\n':
+ case '\r':
+ return;
case '\\':
next_char();
- if(c == '\n') {
- next_char();
- source_position.linenr++;
+ if (c == '\n' || c == '\r') {
+ warningf(WARN_COMMENT, &lexer_pos, "multi-line comment");
+ return;
}
break;
- case EOF:
- case '\r':
- case '\n':
- return;
-
default:
next_char();
break;
}
}
+/** The current preprocessor token. */
static token_t pp_token;
-static inline
-void next_pp_token(void)
+/**
+ * Read the next preprocessor token.
+ */
+static inline void next_pp_token(void)
{
- lexer_next_preprocessing_token(&pp_token);
+ lexer_next_preprocessing_token();
+ pp_token = lexer_token;
}
-static
-void eat_until_newline(void)
+/**
+ * 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)
-{
- error_prefix();
- fprintf(stderr, "#error directive: \n");
-
- /* parse pp-tokens until new-line */
-}
-
-static
-void define_directive(void)
+/**
+ * Parse the line directive.
+ */
+static void parse_line_directive(void)
{
- token_t temptoken;
-
- lexer_next_preprocessing_token(&temptoken);
- if(temptoken.type != T_IDENTIFIER) {
- parse_error("expected identifier after #define\n");
- eat_until_newline();
+ 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 && pp_token.string.encoding == STRING_ENCODING_CHAR) {
+ lexer_pos.input_name = pp_token.string.string.begin;
+ lexer_pos.is_system_header = false;
+ next_pp_token();
-static
-void ifdef_directive(int is_ifndef)
-{
- (void) is_ifndef;
- token_t temptoken;
- lexer_next_preprocessing_token(&temptoken);
- //expect_identifier();
- //extect_newline();
-}
+ /* 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();
+ }
+ }
-static
-void endif_directive(void)
-{
- //expect_newline();
+ eat_until_newline();
}
-static
-void parse_line_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)
{
- if(pp_token.type != T_INTEGER) {
- parse_error("expected integer");
- } else {
- source_position.linenr = pp_token.v.intvalue - 1;
- next_pp_token();
+ 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;
}
- if(pp_token.type == T_STRING_LITERAL) {
- source_position.input_name = pp_token.v.string;
+
+ stdc_pragma_kind_t kind = STDC_UNKNOWN;
+ if (pp_token.base.symbol->pp_ID == TP_STDC && c_mode & _C99) {
+ /* a STDC pragma */
next_pp_token();
- }
+ switch (pp_token.base.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) {
+ next_pp_token();
+ stdc_pragma_value_kind_t value;
+ switch (pp_token.base.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: value = STDC_VALUE_UNKNOWN; break;
+ }
+ if (value == STDC_VALUE_UNKNOWN) {
+ kind = STDC_UNKNOWN;
+ errorf(&pp_token.base.source_position, "bad STDC pragma argument");
+ }
+ }
+ }
eat_until_newline();
+ if (kind == STDC_UNKNOWN) {
+ warningf(WARN_UNKNOWN_PRAGMAS, &pp_token.base.source_position,
+ "encountered unknown #pragma");
+ }
}
-static
-void parse_preprocessor_identifier(void)
+/**
+ * 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;
-
- 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;
+ assert(pp_token.kind == T_IDENTIFIER);
+ switch (pp_token.base.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:
+ parse_pragma();
+ break;
+ case TP_error:
+ /* TODO; output the rest of the line */
+ parse_error("#error directive");
break;
}
}
-static
-void parse_preprocessor_directive()
+/**
+ * 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(); \
- 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) \
- SKIP_TRIGRAPHS(, \
- code; \
- ) \
- \
- case '\\': \
- next_char(); \
- EAT_NEWLINE(break;) \
- /* fallthrough */ \
default: \
- code; \
+ code \
+ return; \
} \
- } /* end of while(1) */ \
- break;
+ } /* end of while (true) */ \
#define ELSE(set_type) \
ELSE_CODE( \
- token->type = set_type; \
- return; \
+ lexer_token.kind = set_type; \
)
-void lexer_next_preprocessing_token(token_t *token)
+void lexer_next_preprocessing_token(void)
{
- while(1) {
- switch(c) {
+ while (true) {
+ lexer_token.base.source_position = lexer_pos;
+ lexer_token.base.symbol = NULL;
+
+ switch (c) {
case ' ':
case '\t':
next_char();
break;
- MATCH_NEWLINE(
- token->type = '\n';
+ case NEWLINE:
+ lexer_token.kind = '\n';
return;
- )
- SYMBOL_CHARS
- parse_symbol(token);
+ SYMBOL_CHARS {
+ parse_symbol();
+ /* might be a wide string ( L"string" ) */
+ string_encoding_t const enc = STRING_ENCODING_WIDE;
+ if (lexer_token.base.symbol == symbol_L) {
+ switch (c) {
+ case '"': parse_string_literal(enc); break;
+ case '\'': parse_character_constant(enc); break;
+ }
+ }
return;
+ }
DIGITS
- parse_number(token);
+ parse_number();
return;
case '"':
- parse_string_literal(token);
+ parse_string_literal(STRING_ENCODING_CHAR);
return;
case '\'':
- parse_character_constant(token);
- return;
-
- case '\\':
- next_char();
- if(c == '\n') {
- next_char();
- source_position.linenr++;
- break;
- } else {
- parse_error("unexpected '\\' found");
- token->type = T_ERROR;
- }
+ parse_character_constant(STRING_ENCODING_CHAR);
return;
case '.':
MAYBE_PROLOG
+ DIGITS
+ put_back(c);
+ c = '.';
+ parse_number();
+ return;
+
case '.':
MAYBE_PROLOG
MAYBE('.', T_DOTDOTDOT)
ELSE_CODE(
put_back(c);
c = '.';
- token->type = '.';
- return;
+ lexer_token.kind = '.';
)
ELSE('.')
case '&':
ELSE('+')
case '-':
MAYBE_PROLOG
+ MAYBE('>', T_MINUSGREATER)
MAYBE('-', T_MINUSMINUS)
MAYBE('=', T_MINUSEQUAL)
ELSE('-')
case '*':
next_char();
skip_multiline_comment();
- lexer_next_preprocessing_token(token);
+ lexer_next_preprocessing_token();
return;
case '/':
next_char();
skip_line_comment();
- lexer_next_preprocessing_token(token);
+ lexer_next_preprocessing_token();
return;
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 = '%';
- 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
MAYBE('=', T_LESSLESSEQUAL)
ELSE('<')
case '>':
MAYBE_PROLOG
+ MAYBE('=', T_GREATEREQUAL)
case '>':
MAYBE_PROLOG
MAYBE('=', T_GREATERGREATEREQUAL)
ELSE('|')
case ':':
MAYBE_PROLOG
- MAYBE('>', T_COLONGREATER)
+ MAYBE('>', ']')
+ MAYBE_MODE(':', T_COLONCOLON, _CXX)
ELSE(':')
case '=':
MAYBE_PROLOG
MAYBE('#', T_HASHHASH)
ELSE('#')
- case '?':
- next_char();
- /* just a simple ? */
- if(c != '?') {
- token->type = '?';
- return;
- }
- /* might be a trigraph */
+ case '\\':
next_char();
- if(replace_trigraph()) {
- break;
+ if (c == 'U' || c == 'u') {
+ put_back(c);
+ c = '\\';
+ parse_symbol();
+ } else {
+ lexer_token.kind = '\\';
}
- put_back(c);
- c = '?';
- token->type = '?';
return;
+ case '?':
case '[':
case ']':
case '(':
case '~':
case ';':
case ',':
- token->type = c;
+ lexer_token.kind = c;
next_char();
return;
case EOF:
- 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);
- token->type = T_ERROR;
- return;
+ break;
}
}
}
-void lexer_next_token(token_t *token)
+void lexer_next_token(void)
{
- while(1) {
- lexer_next_preprocessing_token(token);
- if(token->type == '\n') {
- do {
- lexer_next_preprocessing_token(token);
- } while(token->type == '\n');
+ lexer_next_preprocessing_token();
- if(token->type == '#') {
- parse_preprocessor_directive();
- continue;
- }
- }
- return;
+ while (lexer_token.kind == '\n') {
+newline_found:
+ lexer_next_preprocessing_token();
+ }
+
+ 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;
- source_position.linenr = 0;
- source_position.input_name = input_name;
+ lexer_pos.lineno += delta_lines;
+ lexer_pos.colno += delta_cols;
+ errorf(&lexer_pos, "%s", message);
+}
+
+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;
- /* we place a virtual '\n' at the beginning so the lexer knows we're at the
- * beginning of a line */
+ /* place a virtual \n at the beginning so the lexer knows that we're
+ * at the beginning of a line */
c = '\n';
}
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:%u\n", source_position.input_name,
+ source_position.lineno, (unsigned)source_position.colno);
fflush(stdout);
}