X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;f=lexer.c;h=247083c399645429fd1efbcd897c27c2865cff10;hb=979fce13621246b706781035786b5d11a8e2f608;hp=8968131cb1240c679acd468486a09be0955899dc;hpb=a5c4f4ba3d3a3a73b7c09c14ea8918f77180a7c4;p=cparser diff --git a/lexer.c b/lexer.c index 8968131..247083c 100644 --- a/lexer.c +++ b/lexer.c @@ -1,304 +1,1080 @@ +/* + * 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 "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 "target_architecture.h" +#include "parser.h" +#include "warning.h" +#include "lang_features.h" #include #include #include +#include #include +#ifndef _WIN32 +#include +#endif + //#define DEBUG_CHARS -#define MAX_PUTBACK 2 +#define MAX_PUTBACK 3 +#define BUF_SIZE 1024 + +#if defined(_WIN32) || defined(__CYGWIN__) +/* No strtold on windows and no replacement yet */ +#define strtold(s, e) strtod(s, e) +#endif -static -void error_prefix_at(lexer_t *this, const char *input_name, unsigned linenr) +typedef unsigned int utf32; + +static utf32 c; +token_t lexer_token; +symbol_t *symbol_L; +static FILE *input; +static utf32 buf[BUF_SIZE + MAX_PUTBACK]; +static const utf32 *bufend; +static const utf32 *bufpos; +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) { - (void) this; - fprintf(stderr, "%s:%d: Error: ", input_name, linenr); + errorf(&lexer_token.source_position, "%s", msg); } -static -void error_prefix(lexer_t *this) +/** + * Prints an internal error message at the current token. + * + * @param msg the error message + */ +static NORETURN internal_error(const char *msg) { - error_prefix_at(this, this->source_position.input_name, - this->source_position.linenr); + internal_errorf(&lexer_token.source_position, "%s", msg); +} + +static size_t read_block(unsigned char *const read_buf, size_t const n) +{ + size_t const s = fread(read_buf, 1, n, input); + if (s == 0) { + if (ferror(input)) + parse_error("read from input failed"); + buf[MAX_PUTBACK] = EOF; + bufpos = buf + MAX_PUTBACK; + bufend = buf + MAX_PUTBACK + 1; + } + return s; } -static -void parse_error(lexer_t *this, const char *msg) +static void decode_iso_8859_1(void) { - error_prefix(this); - fprintf(stderr, "%s\n", msg); + unsigned char read_buf[BUF_SIZE]; + size_t const s = read_block(read_buf, sizeof(read_buf)); + if (s == 0) + return; + + unsigned char const *src = read_buf; + unsigned char const *end = read_buf + s; + utf32 *dst = buf + MAX_PUTBACK; + while (src != end) + *dst++ = *src++; + + bufpos = buf + MAX_PUTBACK; + bufend = dst; } -static inline -void next_char(lexer_t *this) +static void decode_iso_8859_15(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; + unsigned char read_buf[BUF_SIZE]; + size_t const s = read_block(read_buf, sizeof(read_buf)); + if (s == 0) + return; + + unsigned char const *src = read_buf; + unsigned char const *end = read_buf + s; + utf32 *dst = buf + MAX_PUTBACK; + while (src != end) { + utf32 tc = *src++; + switch (tc) { + case 0xA4: tc = 0x20AC; break; // € + case 0xA6: tc = 0x0160; break; // Š + case 0xA8: tc = 0x0161; break; // š + case 0xB4: tc = 0x017D; break; // Ž + case 0xB8: tc = 0x017E; break; // ž + case 0xBC: tc = 0x0152; break; // Œ + case 0xBD: tc = 0x0153; break; // œ + case 0xBE: tc = 0x0178; break; // Ÿ + } + *dst++ = tc; + } + + bufpos = buf + MAX_PUTBACK; + bufend = dst; +} + +static void decode_utf8(void) +{ + static utf32 part_decoded_min_code; + static utf32 part_decoded_char; + static size_t part_decoded_rest_len; + + do { + unsigned char read_buf[BUF_SIZE]; + size_t const s = read_block(read_buf, sizeof(read_buf)); + if (s == 0) { + if (part_decoded_rest_len > 0) + parse_error("incomplete input char at end of input"); return; } - this->bufpos = this->buf + MAX_PUTBACK; - this->bufend = this->buf + MAX_PUTBACK + s; + + unsigned char const *src = read_buf; + unsigned char const *end = read_buf + s; + utf32 *dst = buf + MAX_PUTBACK; + utf32 decoded; + utf32 min_code; + + if (part_decoded_rest_len != 0) { + min_code = part_decoded_min_code; + decoded = part_decoded_char; + size_t const rest_len = part_decoded_rest_len; + part_decoded_rest_len = 0; + switch (rest_len) { + case 4: goto realign; + case 3: goto three_more; + case 2: goto two_more; + default: goto one_more; + } + } + + while (src != end) { + if ((*src & 0x80) == 0) { + decoded = *src++; + } else if ((*src & 0xE0) == 0xC0) { + min_code = 0x80; + decoded = *src++ & 0x1F; +one_more: + if (src == end) { + part_decoded_min_code = min_code; + part_decoded_char = decoded; + part_decoded_rest_len = 1; + break; + } + if ((*src & 0xC0) == 0x80) { + decoded = (decoded << 6) | (*src++ & 0x3F); + } else { + goto invalid_char; + } + if (decoded < min_code || + decoded > 0x10FFFF || + (0xD800 <= decoded && decoded < 0xE000) || // high/low surrogates + (0xFDD0 <= decoded && decoded < 0xFDF0) || // noncharacters + (decoded & 0xFFFE) == 0xFFFE) { // noncharacters + parse_error("invalid byte sequence in input"); + } + } else if ((*src & 0xF0) == 0xE0) { + min_code = 0x800; + decoded = *src++ & 0x0F; +two_more: + if (src == end) { + part_decoded_min_code = min_code; + part_decoded_char = decoded; + part_decoded_rest_len = 2; + break; + } + if ((*src & 0xC0) == 0x80) { + decoded = (decoded << 6) | (*src++ & 0x3F); + } else { + goto invalid_char; + } + goto one_more; + } else if ((*src & 0xF8) == 0xF0) { + min_code = 0x10000; + decoded = *src++ & 0x07; +three_more: + if (src == end) { + part_decoded_min_code = min_code; + part_decoded_char = decoded; + part_decoded_rest_len = 3; + break; + } + if ((*src & 0xC0) == 0x80) { + decoded = (decoded << 6) | (*src++ & 0x3F); + } else { + goto invalid_char; + } + goto two_more; + } else { +invalid_char: + parse_error("invalid byte sequence in input"); +realign: + do { + ++src; + if (src == end) { + part_decoded_rest_len = 4; + break; + } + } while ((*src & 0xC0) == 0x80 || (*src & 0xF8) == 0xF8); + continue; + } + *dst++ = decoded; + } + + bufpos = buf + MAX_PUTBACK; + bufend = dst; + } while (bufpos == bufend); +} + +static void decode_windows_1252(void) +{ + unsigned char read_buf[BUF_SIZE]; + size_t const s = read_block(read_buf, sizeof(read_buf)); + if (s == 0) + return; + + unsigned char const *src = read_buf; + unsigned char const *end = read_buf + s; + utf32 *dst = buf + MAX_PUTBACK; + while (src != end) { + utf32 tc = *src++; + switch (tc) { + case 0x80: tc = 0x20AC; break; // € + case 0x82: tc = 0x201A; break; // ‚ + case 0x83: tc = 0x0192; break; // ƒ + case 0x84: tc = 0x201E; break; // „ + case 0x85: tc = 0x2026; break; // … + case 0x86: tc = 0x2020; break; // † + case 0x87: tc = 0x2021; break; // ‡ + case 0x88: tc = 0x02C6; break; // ˆ + case 0x89: tc = 0x2030; break; // ‰ + case 0x8A: tc = 0x0160; break; // Š + case 0x8B: tc = 0x2039; break; // ‹ + case 0x8C: tc = 0x0152; break; // Œ + case 0x8E: tc = 0x017D; break; // Ž + case 0x91: tc = 0x2018; break; // ‘ + case 0x92: tc = 0x2019; break; // ’ + case 0x93: tc = 0x201C; break; // “ + case 0x94: tc = 0x201D; break; // ” + case 0x95: tc = 0x2022; break; // • + case 0x96: tc = 0x2013; break; // – + case 0x97: tc = 0x2014; break; // — + case 0x98: tc = 0x02DC; break; // ˜ + case 0x99: tc = 0x2122; break; // ™ + case 0x9A: tc = 0x0161; break; // š + case 0x9B: tc = 0x203A; break; // › + case 0x9C: tc = 0x0153; break; // œ + case 0x9E: tc = 0x017E; break; // ž + case 0x9F: tc = 0x0178; break; // Ÿ + } + *dst++ = tc; } - this->c = *(this->bufpos); -#ifdef DEBUG_CHARS - printf("nchar '%c'\n", this->c); -#endif + + bufpos = buf + MAX_PUTBACK; + bufend = dst; } -static inline -void put_back(lexer_t *this, int c) +typedef void (*decoder_t)(void); + +static decoder_t decoder = decode_utf8; + +typedef struct named_decoder_t { + char const *name; + decoder_t decoder; +} named_decoder_t; + +static named_decoder_t const decoders[] = { + { "CP819", decode_iso_8859_1 }, // offical alias + { "IBM819", decode_iso_8859_1 }, // offical alias + { "ISO-8859-1", decode_iso_8859_1 }, // offical alias + { "ISO-8859-15", decode_iso_8859_15 }, // offical name + { "ISO8859-1", decode_iso_8859_1 }, + { "ISO8859-15", decode_iso_8859_15 }, + { "ISO_8859-1", decode_iso_8859_1 }, // offical alias + { "ISO_8859-15", decode_iso_8859_15 }, // offical alias + { "ISO_8859-1:1987", decode_iso_8859_1 }, // offical name + { "Latin-9", decode_iso_8859_15 }, // offical alias + { "UTF-8", decode_utf8 }, // offical name + { "csISOLatin1", decode_iso_8859_1 }, // offical alias + { "cp1252", decode_windows_1252 }, + { "iso-ir-100", decode_iso_8859_1 }, // offical alias + { "l1", decode_iso_8859_1 }, // offical alias + { "latin1", decode_iso_8859_1 }, // offical alias + { "windows-1252", decode_windows_1252 }, // official name + + { NULL, NULL } +}; + +void select_input_encoding(char const* const encoding) { - char *p = (char*) this->bufpos - 1; - this->bufpos--; - assert(p >= this->buf); - *p = c; + for (named_decoder_t const *i = decoders; i->name != NULL; ++i) { + if (strcasecmp(encoding, i->name) != 0) + continue; + decoder = i->decoder; + return; + } + fprintf(stderr, "error: input encoding \"%s\" not supported\n", encoding); +} + +static inline void next_real_char(void) +{ + assert(bufpos <= bufend); + if (bufpos >= bufend) { + if (input == NULL) { + c = EOF; + return; + } + decoder(); + } + c = *bufpos++; +} + +/** + * 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); + *(--bufpos - buf + buf) = pc; #ifdef DEBUG_CHARS - printf("putback '%c'\n", c); + printf("putback '%lc'\n", pc); #endif } -static -int replace_trigraph(lexer_t *this) -{ -#define MATCH_TRIGRAPH(ch,replacement) \ - case ch: \ - this->c = replacement; \ - return 1; - - switch(this->c) { - MATCH_TRIGRAPH('=', '#') - MATCH_TRIGRAPH('(', '[') - MATCH_TRIGRAPH('/', '\\') - MATCH_TRIGRAPH(')', ']') - MATCH_TRIGRAPH('\'', '^') - MATCH_TRIGRAPH('<', '{') - MATCH_TRIGRAPH('!', '|') - MATCH_TRIGRAPH('>', '}') - MATCH_TRIGRAPH('-', '~') +static inline void next_char(void); + +#define MATCH_NEWLINE(code) \ + case '\r': \ + next_char(); \ + if(c == '\n') { \ + next_char(); \ + } \ + lexer_token.source_position.linenr++; \ + code \ + case '\n': \ + next_char(); \ + lexer_token.source_position.linenr++; \ + code + +#define eat(c_type) do { assert(c == c_type); next_char(); } while(0) + +static void maybe_concat_lines(void) +{ + eat('\\'); + + switch(c) { + MATCH_NEWLINE(return;) + + default: + break; + } + + 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(); + goto end_of_next_char; + } + + if(LIKELY(c != '?')) + goto end_of_next_char; + + next_real_char(); + if(LIKELY(c != '?')) { + put_back(c); + c = '?'; + goto end_of_next_char; + } + + next_real_char(); + switch(c) { + case '=': c = '#'; break; + case '(': c = '['; break; + case '/': c = '\\'; maybe_concat_lines(); break; + case ')': c = ']'; break; + case '\'': c = '^'; break; + case '<': c = '{'; break; + case '!': c = '|'; break; + case '>': c = '}'; break; + case '-': c = '~'; break; default: + put_back(c); + put_back('?'); + c = '?'; break; } - return 0; +end_of_next_char:; +#ifdef DEBUG_CHARS + printf("nchar '%c'\n", c); +#endif } -#define SKIP_TRIGRAPHS(no_trigraph_code) \ - case '?': \ - next_char(this); \ - if(this->c != '?') { \ - put_back(this, this->c); \ - this->c = '?'; \ - no_trigraph_code; \ - } \ - next_char(this); \ - if(replace_trigraph(this)) \ - break; \ - put_back(this, '?'); \ - put_back(this, this->c); \ - this->c = '?'; \ - no_trigraph_code; \ +#define 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 '_': -static -void parse_symbol(lexer_t *this, token_t *token) +#define DIGITS \ + case '0': \ + case '1': \ + case '2': \ + case '3': \ + case '4': \ + case '5': \ + case '6': \ + case '7': \ + 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, this->c); - next_char(this); + obstack_1grow(&symbol_obstack, (char) c); + next_char(); while(1) { - switch(this->c) { - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; - } - - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - obstack_1grow(&symbol_obstack, this->c); - next_char(this); + switch(c) { + DIGITS + SYMBOL_CHARS + obstack_1grow(&symbol_obstack, (char) c); + next_char(); break; - case '?': - next_char(this); - if(this->c != '?') { - put_back(this, this->c); - this->c = '?'; - goto end_symbol; - } - next_char(this); - if(replace_trigraph(this)) - break; - put_back(this, '?'); - put_back(this, this->c); - this->c = '?'; - goto end_symbol; - default: +dollar_sign: goto end_symbol; } } + end_symbol: obstack_1grow(&symbol_obstack, '\0'); string = obstack_finish(&symbol_obstack); symbol = symbol_table_insert(string); - if(symbol->ID > 0) { - token->type = symbol->ID; - } else { - token->type = T_IDENTIFIER; - } - token->v.symbol = symbol; + lexer_token.type = symbol->ID; + lexer_token.v.symbol = symbol; if(symbol->string != string) { obstack_free(&symbol_obstack, string); } } -#if 0 -static -preprocessor_token_type_t parse_pp_symbol(lexer_t *this) +static void parse_integer_suffix(bool is_oct_hex) { - do { - obstack_1grow(&symbol_obstack, this->c); - next_char(this); - } while(is_ident_char(this->c)); - obstack_1grow(&symbol_obstack, '\0'); + bool is_unsigned = false; + bool min_long = false; + bool min_longlong = false; + bool not_traditional = false; + int pos = 0; + char suffix[4]; + + if (c == 'U' || c == 'u') { + not_traditional = true; + suffix[pos++] = toupper(c); + is_unsigned = true; + next_char(); + if (c == 'L' || c == 'l') { + suffix[pos++] = toupper(c); + min_long = true; + next_char(); + if (c == 'L' || c == 'l') { + suffix[pos++] = toupper(c); + min_longlong = true; + next_char(); + } + } + } else if (c == 'l' || c == 'L') { + suffix[pos++] = toupper(c); + min_long = true; + next_char(); + if (c == 'l' || c == 'L') { + not_traditional = true; + suffix[pos++] = toupper(c); + min_longlong = true; + next_char(); + if (c == 'u' || c == 'U') { + suffix[pos++] = toupper(c); + is_unsigned = true; + next_char(); + } + } else if (c == 'u' || c == 'U') { + not_traditional = true; + suffix[pos++] = toupper(c); + is_unsigned = true; + next_char(); + lexer_token.datatype = type_unsigned_long; + } + } - char *string = obstack_finish(&symbol_obstack); - symbol_t *symbol = preprocessor_symbol_table_find(string); - obstack_free(&symbol_obstack, string); + if (warning.traditional && not_traditional) { + suffix[pos] = '\0'; + warningf(&lexer_token.source_position, + "traditional C rejects the '%s' suffix", 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 && (unsigned long long)v <= (unsigned long long)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; + } +} + +static void parse_floating_suffix(void) +{ + switch(c) { + /* TODO: do something useful with the suffixes... */ + case 'f': + case 'F': + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects the 'F' suffix"); + } + next_char(); + lexer_token.datatype = type_float; + break; + case 'l': + case 'L': + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects the 'F' suffix"); + } + next_char(); + lexer_token.datatype = type_long_double; + break; + default: + lexer_token.datatype = type_double; + break; + } +} - if(symbol == 0) - return TP_ERROR; +/** + * A replacement for strtoull. Only those parts needed for + * our parser are implemented. + */ +static unsigned long long parse_int_string(const char *s, const char **endptr, int base) +{ + unsigned long long v = 0; - return symbol->ID; + 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; } -#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; + assert(c == 'x' || c == 'X'); + next_char(); + + obstack_1grow(&symbol_obstack, '0'); + obstack_1grow(&symbol_obstack, 'x'); - 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; + while(isxdigit(c)) { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + } + + if (c == '.') { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + + while (isxdigit(c)) { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + } + is_float = true; + } + if (c == 'p' || c == 'P') { + 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(); + } + is_float = true; + } + + obstack_1grow(&symbol_obstack, '\0'); + char *string = obstack_finish(&symbol_obstack); + if(*string == '\0') { + parse_error("invalid hex number"); + lexer_token.type = T_ERROR; + obstack_free(&symbol_obstack, string); return; } - 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 (is_float) { + char *endptr; + lexer_token.type = T_FLOATINGPOINT; + lexer_token.v.floatvalue = strtold(string, &endptr); + + if(*endptr != '\0') { + parse_error("invalid hex float literal"); + } + + parse_floating_suffix(); + } else { + const char *endptr; + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = parse_int_string(string + 2, &endptr, 16); + if(*endptr != '\0') { + parse_error("hex number literal too long"); } - next_char(this); + parse_integer_suffix(true); } + + obstack_free(&symbol_obstack, string); } -static -void parse_number_oct(lexer_t *this, token_t *token) +/** + * Returns true if the given char is a octal digit. + * + * @param char the character to check + */ +static inline bool is_octal_digit(utf32 chr) { - assert(this->c == 'o' || this->c == 'O'); - next_char(this); + switch(chr) { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + return true; + default: + return 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); +/** + * Parses a octal number and set the lexer_token. + */ +static void parse_number_oct(void) +{ + while(is_octal_digit(c)) { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); } + obstack_1grow(&symbol_obstack, '\0'); + char *string = obstack_finish(&symbol_obstack); + + const char *endptr; + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = parse_int_string(string, &endptr, 8); + if(*endptr != '\0') { + parse_error("octal number literal too long"); + } + + obstack_free(&symbol_obstack, string); + parse_integer_suffix(true); } -static -void parse_number_dec(lexer_t *this, token_t *token, int first_char) +/** + * Parses a decimal including float number and set the + * lexer_token. + */ +static void parse_number_dec(void) { - int value = 0; - if(first_char > 0) { - assert(first_char >= '0' && first_char <= '9'); - value = first_char - '0'; + bool is_float = false; + while (isdigit(c)) { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); } - for(;;) { - if (isdigit(this->c)) { - value = 10 * value + this->c - '0'; - } else { - token->type = T_INTEGER; - token->v.intvalue = value; - return; + if (c == '.') { + obstack_1grow(&symbol_obstack, '.'); + next_char(); + + while (isdigit(c)) { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + } + is_float = true; + } + if(c == 'e' || c == 'E') { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + + if(c == '-' || c == '+') { + obstack_1grow(&symbol_obstack, (char) c); + next_char(); + } + + 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"); + } + + 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"); } - next_char(this); + + parse_integer_suffix(false); } + obstack_free(&symbol_obstack, string); } -static -void parse_number(lexer_t *this, token_t *token) +/** + * Parses a number and sets the lexer_token. + */ +static void parse_number(void) { - // TODO check for overflow - // TODO check for various invalid inputs sequences - - if (this->c == '0') { - next_char(this); - switch (this->c) { + if (c == '0') { + next_char(); + switch (c) { case 'X': - case 'x': parse_number_hex(this, token); break; - case 'o': - case 'O': parse_number_oct(this, token); break; - default: parse_number_dec(this, token, '0'); + case 'x': + parse_number_hex(); + break; + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + parse_number_oct(); + break; + case '8': + case '9': + next_char(); + parse_error("invalid octal number"); + lexer_token.type = T_ERROR; + return; + case '.': + case 'e': + case 'E': + default: + obstack_1grow(&symbol_obstack, '0'); + parse_number_dec(); + return; } } else { - parse_number_dec(this, token, 0); + parse_number_dec(); } } -static -int parse_escape_sequence(lexer_t *this) +/** + * 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'; @@ -306,77 +1082,183 @@ 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 */ + /* FALLTHROUGH */ default: - parse_error(this, "unknown escape sequence\n"); + /* §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) +/** + * Concatenate two strings. + */ +string_t concat_strings(const string_t *const s1, const string_t *const s2) { - unsigned start_linenr = this->source_position.linenr; - char *string; - const char *result; + const size_t len1 = s1->size - 1; + const size_t len2 = s2->size - 1; - assert(this->c == '"'); - next_char(this); + char *const concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1); + memcpy(concat, s1->begin, len1); + memcpy(concat + len1, s2->begin, len2 + 1); - while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - obstack_1grow(&symbol_obstack, '?'); - next_char(this); - break; - ) + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects string constant concatenation"); + } +#if 0 /* TODO hash */ + const char *result = strset_insert(&stringset, concat); + if(result != concat) { + obstack_free(&symbol_obstack, concat); + } - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; + return result; +#else + return (string_t){ concat, len1 + len2 + 1 }; +#endif +} + +/** + * Concatenate a string and a wide string. + */ +wide_string_t concat_string_wide_string(const string_t *const s1, const wide_string_t *const s2) +{ + const size_t len1 = s1->size - 1; + const size_t len2 = s2->size - 1; + + wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat)); + const char *const src = s1->begin; + for (size_t i = 0; i != len1; ++i) { + concat[i] = src[i]; + } + memcpy(concat + len1, s2->begin, (len2 + 1) * sizeof(*concat)); + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects string constant concatenation"); + } + + return (wide_string_t){ concat, len1 + len2 + 1 }; +} + +/** + * Concatenate two wide strings. + */ +wide_string_t concat_wide_strings(const wide_string_t *const s1, const wide_string_t *const s2) +{ + const size_t len1 = s1->size - 1; + const size_t len2 = s2->size - 1; + + wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat)); + memcpy(concat, s1->begin, len1 * sizeof(*concat)); + memcpy(concat + len1, s2->begin, (len2 + 1) * sizeof(*concat)); + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects string constant concatenation"); + } + + return (wide_string_t){ concat, len1 + len2 + 1 }; +} + +/** + * Concatenate a wide string and a string. + */ +wide_string_t concat_wide_string_string(const wide_string_t *const s1, const string_t *const s2) +{ + const size_t len1 = s1->size - 1; + const size_t len2 = s2->size - 1; + + wchar_rep_t *const concat = obstack_alloc(&symbol_obstack, (len1 + len2 + 1) * sizeof(*concat)); + memcpy(concat, s1->begin, len1 * sizeof(*concat)); + const char *const src = s2->begin; + wchar_rep_t *const dst = concat + len1; + for (size_t i = 0; i != len2 + 1; ++i) { + dst[i] = src[i]; + } + if (warning.traditional) { + warningf(&lexer_token.source_position, + "traditional C rejects string constant concatenation"); + } + + return (wide_string_t){ concat, len1 + len2 + 1 }; +} + +static void grow_symbol(utf32 const tc) +{ + struct obstack *const o = &symbol_obstack; + if (tc < 0x80U) { + obstack_1grow(o, tc); + } else if (tc < 0x800) { + obstack_1grow(o, 0xC0 | (tc >> 6)); + obstack_1grow(o, 0x80 | (tc & 0x3F)); + } else if (tc < 0x10000) { + obstack_1grow(o, 0xE0 | ( tc >> 12)); + obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F)); + obstack_1grow(o, 0x80 | ( tc & 0x3F)); + } else { + obstack_1grow(o, 0xF0 | ( tc >> 18)); + obstack_1grow(o, 0x80 | ((tc >> 12) & 0x3F)); + obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F)); + obstack_1grow(o, 0x80 | ( tc & 0x3F)); + } +} + +/** + * Parse a string literal and set lexer_token. + */ +static void parse_string_literal(void) +{ + const unsigned start_linenr = lexer_token.source_position.linenr; + + eat('"'); + + while(1) { + switch(c) { + case '\\': { + utf32 const tc = parse_escape_sequence(); + if (tc >= 0x100) { + warningf(&lexer_token.source_position, + "escape sequence out of range"); } - int c = parse_escape_sequence(this); - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, tc); 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; + case EOF: { + source_position_t source_position; + source_position.input_name = lexer_token.source_position.input_name; + source_position.linenr = start_linenr; + errorf(&source_position, "string has no end"); + lexer_token.type = T_ERROR; return; + } case '"': - next_char(this); + next_char(); goto end_of_string; default: - obstack_1grow(&symbol_obstack, this->c); - next_char(this); + grow_symbol(c); + next_char(); break; } } @@ -387,279 +1269,578 @@ end_of_string: /* add finishing 0 to the string */ obstack_1grow(&symbol_obstack, '\0'); - string = obstack_finish(&symbol_obstack); + const size_t size = (size_t)obstack_object_size(&symbol_obstack); + const char *const string = obstack_finish(&symbol_obstack); +#if 0 /* TODO hash */ /* check if there is already a copy of the string */ - result = strset_insert(&this->stringset, string); + result = strset_insert(&stringset, string); if(result != string) { obstack_free(&symbol_obstack, string); } +#else + const char *const result = string; +#endif - token->type = T_STRING_LITERAL; - token->v.string = result; + lexer_token.type = T_STRING_LITERAL; + lexer_token.v.string.begin = result; + lexer_token.v.string.size = size; } -static -void parse_character_constant(lexer_t *this, token_t *token) +/** + * Parse a wide character constant and set lexer_token. + */ +static void parse_wide_character_constant(void) { - assert(this->c == '\''); - next_char(this); + const unsigned start_linenr = lexer_token.source_position.linenr; + + eat('\''); - int found_char = 0; while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - found_char = '?'; + switch(c) { + case '\\': { + wchar_rep_t tc = parse_escape_sequence(); + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + break; + } + + MATCH_NEWLINE( + parse_error("newline while parsing character constant"); break; ) - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; + case '\'': + next_char(); + goto end_of_wide_char_constant; + + case EOF: { + source_position_t source_position = lexer_token.source_position; + source_position.linenr = start_linenr; + errorf(&source_position, "EOF while parsing character constant"); + lexer_token.type = T_ERROR; + return; + } + + default: { + wchar_rep_t tc = (wchar_rep_t) c; + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + next_char(); + break; + } + } + } + +end_of_wide_char_constant:; + size_t size = (size_t) obstack_object_size(&symbol_obstack); + assert(size % sizeof(wchar_rep_t) == 0); + size /= sizeof(wchar_rep_t); + + const wchar_rep_t *string = obstack_finish(&symbol_obstack); + + lexer_token.type = T_WIDE_CHARACTER_CONSTANT; + lexer_token.v.wide_string.begin = string; + lexer_token.v.wide_string.size = size; + lexer_token.datatype = type_wchar_t; +} + +/** + * Parse a wide string literal and set lexer_token. + */ +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: { + source_position_t source_position; + source_position.input_name = lexer_token.source_position.input_name; + source_position.linenr = start_linenr; + errorf(&source_position, "string has no end"); + lexer_token.type = T_ERROR; + return; + } + + case '"': + next_char(); + goto end_of_string; + + default: { + wchar_rep_t tc = c; + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + next_char(); + break; + } + } + } + +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 character constant and set lexer_token. + */ +static void parse_character_constant(void) +{ + const unsigned start_linenr = lexer_token.source_position.linenr; + + eat('\''); + + while(1) { + switch(c) { + case '\\': { + utf32 const tc = parse_escape_sequence(); + if (tc >= 0x100) { + warningf(&lexer_token.source_position, + "escape sequence out of range"); } - found_char = '\\'; + obstack_1grow(&symbol_obstack, tc); break; + } - case '\n': - next_char(this); - parse_error(this, "newline while parsing character constant"); - this->source_position.linenr++; + MATCH_NEWLINE( + parse_error("newline while parsing character constant"); break; + ) case '\'': - next_char(this); + next_char(); goto end_of_char_constant; - case EOF: - parse_error(this, "EOF while parsing character constant"); - token->type = T_ERROR; + case EOF: { + source_position_t source_position; + source_position.input_name = lexer_token.source_position.input_name; + source_position.linenr = start_linenr; + errorf(&source_position, "EOF while parsing character constant"); + lexer_token.type = T_ERROR; return; + } default: - if(found_char != 0) { - parse_error(this, "more than 1 characters in character " - "constant"); - goto end_of_char_constant; - } else { - found_char = this->c; - next_char(this); - } + grow_symbol(c); + next_char(); break; + } } -end_of_char_constant: - token->type = T_INTEGER; - token->v.intvalue = found_char; +end_of_char_constant:; + const size_t size = (size_t)obstack_object_size(&symbol_obstack); + const char *const string = obstack_finish(&symbol_obstack); + + lexer_token.type = T_CHARACTER_CONSTANT; + lexer_token.v.string.begin = string; + lexer_token.v.string.size = size; + lexer_token.datatype = c_mode & _CXX && size == 1 ? type_char : type_int; } -static -void skip_multiline_comment(lexer_t *this) +/** + * Skip a multiline comment. + */ +static void skip_multiline_comment(void) { - unsigned start_linenr = this->source_position.linenr; + unsigned start_linenr = lexer_token.source_position.linenr; while(1) { - switch(this->c) { + switch(c) { + case '/': + next_char(); + if (c == '*') { + /* nested comment, warn here */ + if (warning.comment) { + warningf(&lexer_token.source_position, "'/*' within comment"); + } + } + break; case '*': - next_char(this); - if(this->c == '/') { - next_char(this); + next_char(); + if(c == '/') { + next_char(); 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; - 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"); + MATCH_NEWLINE(break;) + + case EOF: { + source_position_t source_position; + source_position.input_name = lexer_token.source_position.input_name; + source_position.linenr = start_linenr; + errorf(&source_position, "at end of file while looking for comment end"); return; + } + default: - next_char(this); + next_char(); break; } } } -static -void skip_line_comment(lexer_t *this) +/** + * Skip a single line comment. + */ +static void skip_line_comment(void) { 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; + switch(c) { + case EOF: + return; + + 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') { + if (warning.comment) + warningf(&lexer_token.source_position, "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.type != '\n' && pp_token.type != T_EOF) { + next_pp_token(); } } -#endif -void preprocessor_next_token(lexer_t *this, token_t *token) +/** + * Handle the define directive. + */ +static void define_directive(void) { - /* skip whitespaces */ - while(this->c == ' ' || this->c == '\t' || this->c == '\r') { - next_char(this); + lexer_next_preprocessing_token(); + if(lexer_token.type != T_IDENTIFIER) { + parse_error("expected identifier after #define\n"); + eat_until_newline(); } +} - switch(this->c) { - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, token); - } +/** + * Handle the ifdef directive. + */ +static void ifdef_directive(int is_ifndef) +{ + (void) is_ifndef; + lexer_next_preprocessing_token(); + //expect_identifier(); + //extect_newline(); } -void lexer_next_token(lexer_t *this, token_t *token) +/** + * Handle the endif directive. + */ +static void endif_directive(void) { - while(1) { - switch(this->c) { - case ' ': - case '\t': - case '\r': - next_char(this); - break; + //expect_newline(); +} - case '\n': - this->source_position.linenr++; - next_char(this); - break; +/** + * Parse the line directive. + */ +static void parse_line_directive(void) +{ + if(pp_token.type != T_INTEGER) { + parse_error("expected integer"); + } else { + lexer_token.source_position.linenr = (unsigned int)(pp_token.v.intvalue - 1); + next_pp_token(); + } + if(pp_token.type == T_STRING_LITERAL) { + lexer_token.source_position.input_name = pp_token.v.string.begin; + next_pp_token(); + } - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, token); - return; + eat_until_newline(); +} - case '0' ... '9': - parse_number(this, token); - return; +/** + * 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) +{ + bool unknown_pragma = true; + + next_pp_token(); + if (pp_token.v.symbol->pp_ID == TP_STDC) { + stdc_pragma_kind_t kind = STDC_UNKNOWN; + /* a STDC pragma */ + if (c_mode & _C99) { + next_pp_token(); + + switch (pp_token.v.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.v.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.source_position, "bad STDC pragma argument"); + } + } + } + } else { + unknown_pragma = true; + } + eat_until_newline(); + if (unknown_pragma && warning.unknown_pragmas) { + warningf(&pp_token.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.type == T_IDENTIFIER); + symbol_t *symbol = pp_token.v.symbol; - case '\'': - parse_character_constant(this, token); - return; + switch(symbol->pp_ID) { + case TP_include: + printf("include - enable header name parsing!\n"); + break; + case TP_define: + define_directive(); + break; + case TP_ifdef: + ifdef_directive(0); + break; + case TP_ifndef: + ifdef_directive(1); + break; + case TP_endif: + endif_directive(); + break; + case TP_line: + next_pp_token(); + parse_line_directive(); + break; + case TP_if: + case TP_else: + case TP_elif: + case TP_undef: + case TP_error: + /* TODO; output the rest of the line */ + parse_error("#error directive: "); + break; + case TP_pragma: + parse_pragma(); + break; + } +} - 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; +/** + * Parse a preprocessor directive. + */ +static void parse_preprocessor_directive(void) +{ + next_pp_token(); + + switch(pp_token.type) { + 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); \ + next_char(); \ while(1) { \ - switch(this->c) { + switch(c) { #define MAYBE(ch, set_type) \ case ch: \ - next_char(this); \ - token->type = set_type; \ + next_char(); \ + lexer_token.type = set_type; \ return; -#define ELSE_CODE(code) \ - SKIP_TRIGRAPHS( \ - code; \ - ) \ - \ - case '\\': \ - next_char(this); \ - if(this->c == '\n') { \ - next_char(this); \ - this->source_position.linenr++; \ - break; \ +/* must use this as last thing */ +#define MAYBE_MODE(ch, set_type, mode) \ + case ch: \ + if (c_mode & mode) { \ + next_char(); \ + lexer_token.type = set_type; \ + return; \ } \ - /* fallthrough */ \ + /* fallthrough */ + +#define ELSE_CODE(code) \ default: \ - code; \ + code \ } \ } /* end of while(1) */ \ break; #define ELSE(set_type) \ ELSE_CODE( \ - token->type = set_type; \ + lexer_token.type = set_type; \ return; \ ) +void lexer_next_preprocessing_token(void) +{ + while(1) { + switch(c) { + case ' ': + case '\t': + next_char(); + break; + + MATCH_NEWLINE( + lexer_token.type = '\n'; + return; + ) + + SYMBOL_CHARS + parse_symbol(); + /* might be a wide string ( L"string" ) */ + if (lexer_token.v.symbol == symbol_L) { + switch (c) { + case '"': parse_wide_string_literal(); break; + case '\'': parse_wide_character_constant(); break; + } + } + return; + + DIGITS + parse_number(); + return; + + case '"': + parse_string_literal(); + return; + + case '\'': + parse_character_constant(); + return; + case '.': MAYBE_PROLOG + DIGITS + put_back(c); + c = '.'; + parse_number_dec(); + return; + case '.': MAYBE_PROLOG MAYBE('.', T_DOTDOTDOT) ELSE_CODE( - put_back(this, this->c); - this->c = '.'; - token->type = '.'; + put_back(c); + c = '.'; + lexer_token.type = '.'; return; ) ELSE('.') @@ -679,6 +1860,7 @@ 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 +1872,38 @@ 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; + put_back(c); + c = '%'; + lexer_token.type = '#'; return; ) - 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 +1911,7 @@ 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 +1928,8 @@ 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 +1938,9 @@ 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; - } - put_back(this, this->c); - this->c = '?'; - token->type = '?'; - return; - case '[': case ']': case '(': @@ -791,47 +1950,89 @@ void lexer_next_token(lexer_t *this, token_t *token) case '~': case ';': case ',': - token->type = this->c; - next_char(this); + case '\\': + lexer_token.type = c; + next_char(); return; case EOF: - token->type = T_EOF; + lexer_token.type = T_EOF; return; default: - next_char(this); - error_prefix(this); - fprintf(stderr, "unknown character '%c' found\n", this->c); - token->type = T_ERROR; +dollar_sign: + errorf(&lexer_token.source_position, "unknown character '%c' found", c); + next_char(); + lexer_token.type = T_ERROR; return; } } } -void lexer_init(lexer_t *this, FILE *stream, const char *input_name) +void lexer_next_token(void) +{ + lexer_next_preprocessing_token(); + + while (lexer_token.type == '\n') { +newline_found: + lexer_next_preprocessing_token(); + } + + if (lexer_token.type == '#') { + parse_preprocessor_directive(); + goto newline_found; + } +} + +void init_lexer(void) { - memset(this, 0, sizeof(this[0])); + strset_init(&stringset); + symbol_L = symbol_table_insert("L"); +} - this->input = stream; +void lexer_open_stream(FILE *stream, const char *input_name) +{ + input = stream; + lexer_token.source_position.linenr = 0; + lexer_token.source_position.input_name = input_name; - this->source_position.linenr = 0; - this->source_position.input_name = input_name; - strset_init(&this->stringset); + 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_open_buffer(const char *buffer, size_t len, const char *input_name) +{ + input = NULL; + lexer_token.source_position.linenr = 0; + lexer_token.source_position.input_name = input_name; + +#if 0 // TODO + bufpos = buffer; + bufend = buffer + len; +#else + (void)buffer; + (void)len; + panic("builtin lexing not done yet"); +#endif - /* we place a virtual '\n' at the beginning so the lexer knows we're at the - * beginning of a line */ - this->c = '\n'; + /* place a virtual \n at the beginning so the lexer knows that we're + * at the beginning of a line */ + c = '\n'; } -void lexer_destroy(lexer_t *this) +void exit_lexer(void) { - (void) this; + strset_destroy(&stringset); } static __attribute__((unused)) void dbg_pos(const source_position_t source_position) { - fprintf(stdout, "%s:%d\n", source_position.input_name, source_position.linenr); + fprintf(stdout, "%s:%u\n", source_position.input_name, + source_position.linenr); fflush(stdout); }