X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;f=lexer.c;h=14526b95694a642b9c3557c2656fd770d72beab1;hb=ef26133e70bcbc508d2991bafef373f5ca667ff2;hp=8968131cb1240c679acd468486a09be0955899dc;hpb=a5c4f4ba3d3a3a73b7c09c14ea8918f77180a7c4;p=cparser diff --git a/lexer.c b/lexer.c index 8968131..14526b9 100644 --- a/lexer.c +++ b/lexer.c @@ -1,304 +1,715 @@ +/* + * This file is part of cparser. + * Copyright (C) 2007-2009 Matthias Braun + * + * 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 -#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 #include #include +#include #include -//#define DEBUG_CHARS -#define MAX_PUTBACK 2 - -static -void error_prefix_at(lexer_t *this, const char *input_name, unsigned linenr) -{ - (void) this; - fprintf(stderr, "%s:%d: Error: ", input_name, linenr); -} +#ifndef _WIN32 +#include +#endif -static -void error_prefix(lexer_t *this) +#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(this, this->source_position.input_name, - this->source_position.linenr); + errorf(&lexer_pos, "%s", msg); } -static -void parse_error(lexer_t *this, 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(this); - fprintf(stderr, "%s\n", msg); + internal_errorf(&lexer_pos, "%s", 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); - if(s == 0) { - this->c = EOF; + assert(bufpos <= bufend); + if (bufpos >= bufend) { + size_t n = decode(input, input_buf+MAX_PUTBACK, BUF_SIZE); + if (n == 0) { + c = EOF; return; } - this->bufpos = this->buf + MAX_PUTBACK; - this->bufend = this->buf + MAX_PUTBACK + s; + bufpos = input_buf + MAX_PUTBACK; + bufend = bufpos + n; } - this->c = *(this->bufpos); -#ifdef DEBUG_CHARS - printf("nchar '%c'\n", this->c); -#endif + c = *bufpos++; + ++lexer_pos.colno; } -static inline -void put_back(lexer_t *this, int c) +/** + * 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*) this->bufpos - 1; - this->bufpos--; - assert(p >= this->buf); - *p = c; + *(--bufpos - input_buf + input_buf) = pc; + --lexer_pos.colno; +} -#ifdef DEBUG_CHARS - printf("putback '%c'\n", c); -#endif +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. + +#define eat(c_type) (assert(c == c_type), next_char()) + +static void maybe_concat_lines(void) +{ + eat('\\'); + + switch (c) { + case NEWLINE: + return; + + default: + break; + } + + put_back(c); + c = '\\'; } -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('-', '~') +/** + * 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 '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 '_': - return 0; +#define DIGITS \ + case '0': \ + case '1': \ + case '2': \ + case '3': \ + case '4': \ + case '5': \ + case '6': \ + case '7': \ + case '8': \ + case '9': + +static bool is_universal_char_valid(utf32 const v) +{ + /* C11 §6.4.3:2 */ + if (v < 0xA0U && v != 0x24 && v != 0x40 && v != 0x60) + return false; + if (0xD800 <= v && v <= 0xDFFF) + return false; + return true; } -#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; \ +static int digit_value(utf32 digit); -static -void parse_symbol(lexer_t *this, token_t *token) +static utf32 parse_universal_char(unsigned const n_digits) { - symbol_t *symbol; - char *string; + utf32 v = 0; + for (unsigned k = n_digits; k != 0; --k) { + if (isxdigit(c)) { + v = 16 * v + digit_value(c); + next_char(); + } 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; +} - obstack_1grow(&symbol_obstack, this->c); - next_char(this); +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, (char) c); + next_char(); + break; - while(1) { - switch(this->c) { case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; + next_char(); + 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; } - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - obstack_1grow(&symbol_obstack, this->c); - next_char(this); - break; - - case '?': - next_char(this); - if(this->c != '?') { - put_back(this, this->c); - this->c = '?'; + default: + put_back(c); + 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: +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); - if(symbol->ID > 0) { - token->type = symbol->ID; - } else { - token->type = T_IDENTIFIER; - } - 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 string_t identify_string(char *string, size_t len) +{ + /* TODO hash */ #if 0 -static -preprocessor_token_type_t parse_pp_symbol(lexer_t *this) + 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}; +} + +/** + * parse suffixes like 'LU' or 'f' after numbers + */ +static void parse_number_suffix(void) { - do { - obstack_1grow(&symbol_obstack, this->c); - next_char(this); - } while(is_ident_char(this->c)); + 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; + } + } +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); - char *string = obstack_finish(&symbol_obstack); - symbol_t *symbol = preprocessor_symbol_table_find(string); - obstack_free(&symbol_obstack, string); + lexer_token.number.suffix = identify_string(string, size); +} - if(symbol == 0) - return TP_ERROR; +static void parse_exponent(void) +{ + if (c == '-' || c == '+') { + obstack_1grow(&symbol_obstack, (char)c); + next_char(); + } - return symbol->ID; + 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"); + } } -#endif -static -void parse_number_hex(lexer_t *this, token_t *token) +/** + * Parses a hex number including hex floats and set the + * lexer_token. + */ +static void parse_number_hex(void) { - assert(this->c == 'x' || this->c == 'X'); - next_char(this); + bool is_float = false; + bool has_digits = false; - 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)) { + has_digits = true; + obstack_1grow(&symbol_obstack, (char) c); + next_char(); } - 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; + 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(); } - next_char(this); } + 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'); + + 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 : 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 -void parse_number_oct(lexer_t *this, token_t *token) +static void parse_number_bin(void) { - assert(this->c == 'o' || this->c == 'O'); - next_char(this); + bool has_digits = false; - 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); + while (c == '0' || c == '1') { + has_digits = true; + obstack_1grow(&symbol_obstack, (char)c); + next_char(); + } + obstack_1grow(&symbol_obstack, '\0'); + + 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; + + 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_number_dec(lexer_t *this, token_t *token, int first_char) +/** + * Returns true if the given char is a octal digit. + * + * @param char the character to check + */ +static bool is_octal_digit(utf32 chr) { - int value = 0; - if(first_char > 0) { - assert(first_char >= '0' && first_char <= '9'); - value = first_char - '0'; - } + return '0' <= chr && chr <= '7'; +} - for(;;) { - if (isdigit(this->c)) { - value = 10 * value + this->c - '0'; - } else { - token->type = T_INTEGER; - token->v.intvalue = value; +/** + * Parses a number and sets the lexer_token. + */ +static void parse_number(void) +{ + bool is_float = false; + bool has_digits = false; + + 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(); + 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(); + parse_number_bin(); return; } - next_char(this); + has_digits = true; } -} -static -void parse_number(lexer_t *this, token_t *token) -{ - // TODO check for overflow - // TODO check for various invalid inputs sequences + while (isdigit(c)) { + has_digits = true; + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + } - if (this->c == '0') { - next_char(this); - switch (this->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'); + 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(); } + } + if (c == 'e' || c == 'E') { + is_float = true; + obstack_1grow(&symbol_obstack, 'e'); + next_char(); + parse_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); + + if (is_float) { + lexer_token.kind = T_FLOATINGPOINT; } else { - parse_number_dec(this, token, 0); + 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); + } + } + } + + if (!has_digits) { + errorf(&lexer_token.base.source_position, "invalid number literal '%S'", + &lexer_token.number.number); } + + parse_number_suffix(); } -static -int parse_escape_sequence(lexer_t *this) +/** + * Returns the value of a digit. + * The only portable way to do it ... + */ +static int digit_value(utf32 const digit) { - int c = this->c; - next_char(this); + 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"); + } +} - switch(c) { - case '"': return '"'; - case '\'': return'\''; - case '\\': - if(this->c == '\n') { - this->source_position.linenr++; - next_char(this); - return parse_escape_sequence(this); - } - return '\\'; +/** + * 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; +} + +/** + * 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; +} + +/** + * Parse an escape sequence. + */ +static utf32 parse_escape_sequence(void) +{ + eat('\\'); + + utf32 const 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'; @@ -306,361 +717,442 @@ int parse_escape_sequence(lexer_t *this) 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)) { - return parse_escape_sequence(this); - } - put_back(this, this->c); - this->c = '?'; - return '?'; - + 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(this, "reached end of file while parsing escape sequence"); + 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 §6.11.4, 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': return parse_universal_char(8); + case 'u': return parse_universal_char(4); + default: - parse_error(this, "unknown escape sequence\n"); - return EOF; + break; } + /* §6.4.4.4:8 footnote 64 */ + parse_error("unknown escape sequence"); + return EOF; } -static -void parse_string_literal(lexer_t *this, token_t *token) +string_t make_string(const char *string) { - unsigned start_linenr = this->source_position.linenr; - char *string; - const char *result; + size_t len = strlen(string) + 1; + char *const space = obstack_alloc(&symbol_obstack, len); + memcpy(space, string, len); - assert(this->c == '"'); - next_char(this); + return identify_string(space, len); +} - while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - obstack_1grow(&symbol_obstack, '?'); - next_char(this); +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(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; - } - int c = parse_escape_sequence(this); - obstack_1grow(&symbol_obstack, c); + case NEWLINE: + errorf(&lexer_pos, "newline while parsing %s", context); break; case EOF: - error_prefix_at(this, this->source_position.input_name, - start_linenr); - fprintf(stderr, "string has no end\n"); - token->type = T_ERROR; - return; - - case '"': - next_char(this); + errorf(&lexer_token.base.source_position, "EOF while parsing %s", context); goto end_of_string; default: - obstack_1grow(&symbol_obstack, this->c); - next_char(this); - break; + if (c == delim) { + next_char(); + goto end_of_string; + } else { + obstack_grow_symbol(&symbol_obstack, c); + next_char(); + break; + } } } end_of_string: - - /* TODO: concatenate multiple strings separated by whitespace... */ - - /* add finishing 0 to the string */ obstack_1grow(&symbol_obstack, '\0'); - string = obstack_finish(&symbol_obstack); + size_t const size = obstack_object_size(&symbol_obstack) - 1; + char *const string = obstack_finish(&symbol_obstack); - /* check if there is already a copy of the string */ - result = strset_insert(&this->stringset, string); - if(result != string) { - obstack_free(&symbol_obstack, string); - } + lexer_token.kind = kind; + lexer_token.string.encoding = enc; + lexer_token.string.string = identify_string(string, size); +} - token->type = T_STRING_LITERAL; - token->v.string = result; +/** + * Parse a string literal and set lexer_token. + */ +static void parse_string_literal(string_encoding_t const enc) +{ + parse_string('"', T_STRING_LITERAL, enc, "string literal"); } -static -void parse_character_constant(lexer_t *this, token_t *token) +/** + * Parse a character constant and set lexer_token. + */ +static void parse_character_constant(string_encoding_t const enc) { - assert(this->c == '\''); - next_char(this); + parse_string('\'', T_CHARACTER_CONSTANT, enc, "character constant"); + if (lexer_token.string.string.size == 0) { + errorf(&lexer_token.base.source_position, "empty character constant"); + } +} - int found_char = 0; - while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - found_char = '?'; +/** + * Skip a multiline comment. + */ +static void skip_multiline_comment(void) +{ + while (true) { + switch (c) { + case '/': + next_char(); + if (c == '*') { + /* nested comment, warn here */ + warningf(WARN_COMMENT, &lexer_pos, "'/*' within comment"); + } break; - ) - - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; + case '*': + next_char(); + if (c == '/') { + next_char(); + return; } - found_char = '\\'; break; - case '\n': - next_char(this); - parse_error(this, "newline while parsing character constant"); - this->source_position.linenr++; + case NEWLINE: break; - case '\'': - next_char(this); - goto end_of_char_constant; - - case EOF: - parse_error(this, "EOF while parsing character constant"); - token->type = T_ERROR; + case EOF: { + errorf(&lexer_token.base.source_position, + "at end of file while looking for comment end"); return; + } default: - if(found_char != 0) { - parse_error(this, "more than 1 characters in character " - "constant"); - goto end_of_char_constant; - } else { - found_char = this->c; - next_char(this); - } + next_char(); break; } } - -end_of_char_constant: - token->type = T_INTEGER; - token->v.intvalue = found_char; } -static -void skip_multiline_comment(lexer_t *this) +/** + * Skip a single line comment. + */ +static void skip_line_comment(void) { - unsigned start_linenr = this->source_position.linenr; - - while(1) { - switch(this->c) { - case '*': - next_char(this); - if(this->c == '/') { - next_char(this); - return; - } - break; - 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 */ - break; - - case '\n': - this->source_position.linenr++; - next_char(this); - break; + while (true) { + switch (c) { 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"); return; - default: - next_char(this); - break; - } - } -} -static -void skip_line_comment(lexer_t *this) -{ - 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 */ - break; + case '\n': + case '\r': + return; case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; + next_char(); + if (c == '\n' || c == '\r') { + warningf(WARN_COMMENT, &lexer_pos, "multi-line comment"); + return; } break; - case EOF: - case '\n': - return; default: - next_char(this); + next_char(); break; } } } -#if 0 -static -void eat_until_newline(lexer_t *this) +/** The current preprocessor token. */ +static token_t pp_token; + +/** + * Read the next preprocessor token. + */ +static inline void next_pp_token(void) { - while(this->c != '\n') { - next_char(this); - } + lexer_next_preprocessing_token(); + pp_token = lexer_token; } -#endif -#if 0 -static -void parse_preprocessor_directive(lexer_t *this, token_t *result_token) +/** + * Eat all preprocessor tokens until newline. + */ +static void eat_until_newline(void) { - (void) result_token; - /* skip whitespaces */ - while(this->c == ' ' || this->c == '\t' || this->c == '\r') { - next_char(this); + while (pp_token.kind != '\n' && pp_token.kind != T_EOF) { + next_pp_token(); } } -#endif -void preprocessor_next_token(lexer_t *this, token_t *token) +/** + * Parse the line directive. + */ +static void parse_line_directive(void) { - /* skip whitespaces */ - while(this->c == ' ' || this->c == '\t' || this->c == '\r') { - next_char(this); + 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(); } - - switch(this->c) { - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, 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(); + + /* 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(); + } } + + eat_until_newline(); } -void lexer_next_token(lexer_t *this, token_t *token) +/** + * 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) { - while(1) { - switch(this->c) { - case ' ': - case '\t': - case '\r': - next_char(this); - break; - - case '\n': - this->source_position.linenr++; - next_char(this); - break; + 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; + } - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, token); - return; + 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(); - case '0' ... '9': - parse_number(this, token); - return; + 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"); + } +} - case '"': - parse_string_literal(this, token); - return; +/** + * Parse a preprocessor non-null directive. + */ +static void parse_preprocessor_identifier(void) +{ + 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_pragma: + parse_pragma(); + break; + case TP_error: + /* TODO; output the rest of the line */ + parse_error("#error directive"); + break; + } +} - case '\'': - parse_character_constant(this, token); - return; +/** + * Parse a preprocessor directive. + */ +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.kind) { + case T_IDENTIFIER: + parse_preprocessor_identifier(); + break; + case T_INTEGER: + parse_line_directive(); + break; + case '\n': + /* NULL directive, see §6.10.7 */ + break; + default: + parse_error("invalid preprocessor directive"); + eat_until_newline(); + break; + } +} #define MAYBE_PROLOG \ - next_char(this); \ - while(1) { \ - switch(this->c) { + next_char(); \ + while (true) { \ + switch (c) { #define MAYBE(ch, set_type) \ case ch: \ - next_char(this); \ - token->type = set_type; \ + next_char(); \ + lexer_token.kind = 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; \ +/* 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 */ \ + /* fallthrough */ + +#define ELSE_CODE(code) \ 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(void) +{ + while (true) { + lexer_token.base.source_position = lexer_pos; + lexer_token.base.symbol = NULL; + + switch (c) { + case ' ': + case '\t': + next_char(); + break; + + case NEWLINE: + lexer_token.kind = '\n'; + return; + + 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(); + return; + + case '"': + parse_string_literal(STRING_ENCODING_CHAR); + return; + + case '\'': + 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(this, this->c); - this->c = '.'; - token->type = '.'; - return; + put_back(c); + c = '.'; + lexer_token.kind = '.'; ) ELSE('.') case '&': @@ -679,6 +1171,7 @@ void lexer_next_token(lexer_t *this, token_t *token) ELSE('+') case '-': MAYBE_PROLOG + MAYBE('>', T_MINUSGREATER) MAYBE('-', T_MINUSMINUS) MAYBE('=', T_MINUSEQUAL) ELSE('-') @@ -690,37 +1183,37 @@ void lexer_next_token(lexer_t *this, token_t *token) 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_PERCENTGREATER) + MAYBE('>', '}') MAYBE('=', T_PERCENTEQUAL) case ':': MAYBE_PROLOG case '%': MAYBE_PROLOG - MAYBE(':', T_PERCENTCOLONPERCENTCOLON) + MAYBE(':', T_HASHHASH) ELSE_CODE( - put_back(this, this->c); - this->c = '%'; - token->type = T_PERCENTCOLON; - return; + put_back(c); + c = '%'; + 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) @@ -728,6 +1221,7 @@ void lexer_next_token(lexer_t *this, token_t *token) ELSE('<') case '>': MAYBE_PROLOG + MAYBE('=', T_GREATEREQUAL) case '>': MAYBE_PROLOG MAYBE('=', T_GREATERGREATEREQUAL) @@ -744,7 +1238,8 @@ void lexer_next_token(lexer_t *this, token_t *token) ELSE('|') case ':': MAYBE_PROLOG - MAYBE('>', T_COLONGREATER) + MAYBE('>', ']') + MAYBE_MODE(':', T_COLONCOLON, _CXX) ELSE(':') case '=': MAYBE_PROLOG @@ -753,35 +1248,20 @@ void lexer_next_token(lexer_t *this, token_t *token) case '#': MAYBE_PROLOG MAYBE('#', T_HASHHASH) -#if 0 - else { - if(line_begin) { - parse_preprocessor_directive(this, token); - return; - } else { - token->type = '#'; - } -#else ELSE('#') -#endif - 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; + case '\\': + next_char(); + if (c == 'U' || c == 'u') { + put_back(c); + c = '\\'; + parse_symbol(); + } else { + lexer_token.kind = '\\'; } - put_back(this, this->c); - this->c = '?'; - token->type = '?'; return; + case '?': case '[': case ']': case '(': @@ -791,47 +1271,77 @@ void lexer_next_token(lexer_t *this, token_t *token) case '~': case ';': case ',': - token->type = this->c; - next_char(this); + lexer_token.kind = c; + next_char(); return; case EOF: - token->type = T_EOF; + lexer_token.kind = T_EOF; return; default: - next_char(this); - error_prefix(this); - fprintf(stderr, "unknown character '%c' found\n", this->c); - token->type = T_ERROR; - return; +dollar_sign: + errorf(&lexer_pos, "unknown character '%c' found", c); + next_char(); + break; } } } -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(); - this->input = stream; + while (lexer_token.kind == '\n') { +newline_found: + lexer_next_preprocessing_token(); + } - this->source_position.linenr = 0; - this->source_position.input_name = input_name; - strset_init(&this->stringset); + if (lexer_token.kind == '#') { + parse_preprocessor_directive(); + goto newline_found; + } +} + +void init_lexer(void) +{ + strset_init(&stringset); + symbol_L = symbol_table_insert("L"); +} - /* we place a virtual '\n' at the beginning so the lexer knows we're at the - * beginning of a line */ - this->c = '\n'; +static void input_error(unsigned delta_lines, unsigned delta_cols, + const char *message) +{ + 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; + + /* 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:%u\n", source_position.input_name, + source_position.lineno, (unsigned)source_position.colno); fflush(stdout); }