2 * This file is part of cparser.
3 * Copyright (C) 2007-2009 Matthias Braun <matze@braunis.de>
5 * This program is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU General Public License
7 * as published by the Free Software Foundation; either version 2
8 * of the License, or (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
22 #include "diagnostic.h"
26 #include "symbol_table_t.h"
27 #include "adt/error.h"
28 #include "adt/strset.h"
32 #include "target_architecture.h"
35 #include "lang_features.h"
51 #if defined(_WIN32) || defined(__CYGWIN__)
52 /* No strtold on windows and no replacement yet */
53 #define strtold(s, e) strtod(s, e)
60 static utf32 buf[BUF_SIZE + MAX_PUTBACK];
61 static const utf32 *bufend;
62 static const utf32 *bufpos;
63 static strset_t stringset;
64 bool allow_dollar_in_symbol = true;
67 * Prints a parse error message at the current token.
69 * @param msg the error message
71 static void parse_error(const char *msg)
73 errorf(&lexer_token.source_position, "%s", msg);
77 * Prints an internal error message at the current token.
79 * @param msg the error message
81 static NORETURN internal_error(const char *msg)
83 internal_errorf(&lexer_token.source_position, "%s", msg);
86 static size_t read_block(unsigned char *const read_buf, size_t const n)
88 size_t const s = fread(read_buf, 1, n, input);
91 parse_error("read from input failed");
92 buf[MAX_PUTBACK] = EOF;
93 bufpos = buf + MAX_PUTBACK;
94 bufend = buf + MAX_PUTBACK + 1;
99 static void decode_iso_8859_1(void)
101 unsigned char read_buf[BUF_SIZE];
102 size_t const s = read_block(read_buf, sizeof(read_buf));
106 unsigned char const *src = read_buf;
107 unsigned char const *end = read_buf + s;
108 utf32 *dst = buf + MAX_PUTBACK;
112 bufpos = buf + MAX_PUTBACK;
116 static void decode_iso_8859_15(void)
118 unsigned char read_buf[BUF_SIZE];
119 size_t const s = read_block(read_buf, sizeof(read_buf));
123 unsigned char const *src = read_buf;
124 unsigned char const *end = read_buf + s;
125 utf32 *dst = buf + MAX_PUTBACK;
129 case 0xA4: tc = 0x20AC; break; // €
130 case 0xA6: tc = 0x0160; break; // Š
131 case 0xA8: tc = 0x0161; break; // š
132 case 0xB4: tc = 0x017D; break; // Ž
133 case 0xB8: tc = 0x017E; break; // ž
134 case 0xBC: tc = 0x0152; break; // Œ
135 case 0xBD: tc = 0x0153; break; // œ
136 case 0xBE: tc = 0x0178; break; // Ÿ
141 bufpos = buf + MAX_PUTBACK;
145 static void decode_utf8(void)
147 static utf32 part_decoded_min_code;
148 static utf32 part_decoded_char;
149 static size_t part_decoded_rest_len;
152 unsigned char read_buf[BUF_SIZE];
153 size_t const s = read_block(read_buf, sizeof(read_buf));
155 if (part_decoded_rest_len > 0)
156 parse_error("incomplete input char at end of input");
160 unsigned char const *src = read_buf;
161 unsigned char const *end = read_buf + s;
162 utf32 *dst = buf + MAX_PUTBACK;
166 if (part_decoded_rest_len != 0) {
167 min_code = part_decoded_min_code;
168 decoded = part_decoded_char;
169 size_t const rest_len = part_decoded_rest_len;
170 part_decoded_rest_len = 0;
172 case 4: goto realign;
173 case 3: goto three_more;
174 case 2: goto two_more;
175 default: goto one_more;
180 if ((*src & 0x80) == 0) {
182 } else if ((*src & 0xE0) == 0xC0) {
184 decoded = *src++ & 0x1F;
187 part_decoded_min_code = min_code;
188 part_decoded_char = decoded;
189 part_decoded_rest_len = 1;
192 if ((*src & 0xC0) == 0x80) {
193 decoded = (decoded << 6) | (*src++ & 0x3F);
197 if (decoded < min_code ||
198 decoded > 0x10FFFF ||
199 (0xD800 <= decoded && decoded < 0xE000) || // high/low surrogates
200 (0xFDD0 <= decoded && decoded < 0xFDF0) || // noncharacters
201 (decoded & 0xFFFE) == 0xFFFE) { // noncharacters
202 parse_error("invalid byte sequence in input");
204 } else if ((*src & 0xF0) == 0xE0) {
206 decoded = *src++ & 0x0F;
209 part_decoded_min_code = min_code;
210 part_decoded_char = decoded;
211 part_decoded_rest_len = 2;
214 if ((*src & 0xC0) == 0x80) {
215 decoded = (decoded << 6) | (*src++ & 0x3F);
220 } else if ((*src & 0xF8) == 0xF0) {
222 decoded = *src++ & 0x07;
225 part_decoded_min_code = min_code;
226 part_decoded_char = decoded;
227 part_decoded_rest_len = 3;
230 if ((*src & 0xC0) == 0x80) {
231 decoded = (decoded << 6) | (*src++ & 0x3F);
238 parse_error("invalid byte sequence in input");
243 part_decoded_rest_len = 4;
246 } while ((*src & 0xC0) == 0x80 || (*src & 0xF8) == 0xF8);
252 bufpos = buf + MAX_PUTBACK;
254 } while (bufpos == bufend);
257 static void decode_windows_1252(void)
259 unsigned char read_buf[BUF_SIZE];
260 size_t const s = read_block(read_buf, sizeof(read_buf));
264 unsigned char const *src = read_buf;
265 unsigned char const *end = read_buf + s;
266 utf32 *dst = buf + MAX_PUTBACK;
270 case 0x80: tc = 0x20AC; break; // €
271 case 0x82: tc = 0x201A; break; // ‚
272 case 0x83: tc = 0x0192; break; // ƒ
273 case 0x84: tc = 0x201E; break; // „
274 case 0x85: tc = 0x2026; break; // …
275 case 0x86: tc = 0x2020; break; // †
276 case 0x87: tc = 0x2021; break; // ‡
277 case 0x88: tc = 0x02C6; break; // ˆ
278 case 0x89: tc = 0x2030; break; // ‰
279 case 0x8A: tc = 0x0160; break; // Š
280 case 0x8B: tc = 0x2039; break; // ‹
281 case 0x8C: tc = 0x0152; break; // Œ
282 case 0x8E: tc = 0x017D; break; // Ž
283 case 0x91: tc = 0x2018; break; // ‘
284 case 0x92: tc = 0x2019; break; // ’
285 case 0x93: tc = 0x201C; break; // “
286 case 0x94: tc = 0x201D; break; // ”
287 case 0x95: tc = 0x2022; break; // •
288 case 0x96: tc = 0x2013; break; // –
289 case 0x97: tc = 0x2014; break; // —
290 case 0x98: tc = 0x02DC; break; // ˜
291 case 0x99: tc = 0x2122; break; // ™
292 case 0x9A: tc = 0x0161; break; // š
293 case 0x9B: tc = 0x203A; break; // ›
294 case 0x9C: tc = 0x0153; break; // œ
295 case 0x9E: tc = 0x017E; break; // ž
296 case 0x9F: tc = 0x0178; break; // Ÿ
301 bufpos = buf + MAX_PUTBACK;
305 typedef void (*decoder_t)(void);
307 static decoder_t decoder = decode_utf8;
309 typedef struct named_decoder_t {
314 static named_decoder_t const decoders[] = {
315 { "CP819", decode_iso_8859_1 }, // offical alias
316 { "IBM819", decode_iso_8859_1 }, // offical alias
317 { "ISO-8859-1", decode_iso_8859_1 }, // offical alias
318 { "ISO-8859-15", decode_iso_8859_15 }, // offical name
319 { "ISO8859-1", decode_iso_8859_1 },
320 { "ISO8859-15", decode_iso_8859_15 },
321 { "ISO_8859-1", decode_iso_8859_1 }, // offical alias
322 { "ISO_8859-15", decode_iso_8859_15 }, // offical alias
323 { "ISO_8859-1:1987", decode_iso_8859_1 }, // offical name
324 { "Latin-9", decode_iso_8859_15 }, // offical alias
325 { "UTF-8", decode_utf8 }, // offical name
326 { "csISOLatin1", decode_iso_8859_1 }, // offical alias
327 { "cp1252", decode_windows_1252 },
328 { "iso-ir-100", decode_iso_8859_1 }, // offical alias
329 { "l1", decode_iso_8859_1 }, // offical alias
330 { "latin1", decode_iso_8859_1 }, // offical alias
331 { "windows-1252", decode_windows_1252 }, // official name
336 /** strcasecmp is not part of C99 so we need our own implementation here */
337 static int my_strcasecmp(const char *s1, const char *s2)
339 for ( ; *s1 != 0; ++s1, ++s2) {
340 if (tolower(*s1) != tolower(*s2))
343 return (unsigned char)*s1 - (unsigned char)*s2;
346 void select_input_encoding(char const* const encoding)
348 for (named_decoder_t const *i = decoders; i->name != NULL; ++i) {
349 if (my_strcasecmp(encoding, i->name) != 0)
351 decoder = i->decoder;
354 fprintf(stderr, "error: input encoding \"%s\" not supported\n", encoding);
357 static inline void next_real_char(void)
359 assert(bufpos <= bufend);
360 if (bufpos >= bufend) {
371 * Put a character back into the buffer.
373 * @param pc the character to put back
375 static inline void put_back(utf32 const pc)
377 assert(bufpos > buf);
378 *(--bufpos - buf + buf) = pc;
381 printf("putback '%lc'\n", pc);
385 static inline void next_char(void);
387 #define MATCH_NEWLINE(code) \
393 lexer_token.source_position.linenr++; \
397 lexer_token.source_position.linenr++; \
400 #define eat(c_type) do { assert(c == c_type); next_char(); } while (0)
402 static void maybe_concat_lines(void)
407 MATCH_NEWLINE(return;)
418 * Set c to the next input character, ie.
419 * after expanding trigraphs.
421 static inline void next_char(void)
425 /* filter trigraphs */
426 if (UNLIKELY(c == '\\')) {
427 maybe_concat_lines();
428 goto end_of_next_char;
431 if (LIKELY(c != '?'))
432 goto end_of_next_char;
435 if (LIKELY(c != '?')) {
438 goto end_of_next_char;
443 case '=': c = '#'; break;
444 case '(': c = '['; break;
445 case '/': c = '\\'; maybe_concat_lines(); break;
446 case ')': c = ']'; break;
447 case '\'': c = '^'; break;
448 case '<': c = '{'; break;
449 case '!': c = '|'; break;
450 case '>': c = '}'; break;
451 case '-': c = '~'; break;
461 printf("nchar '%c'\n", c);
465 #define SYMBOL_CHARS \
466 case '$': if (!allow_dollar_in_symbol) goto dollar_sign; \
534 * Read a symbol from the input and build
537 static void parse_symbol(void)
539 obstack_1grow(&symbol_obstack, (char) c);
546 obstack_1grow(&symbol_obstack, (char) c);
557 obstack_1grow(&symbol_obstack, '\0');
559 char *string = obstack_finish(&symbol_obstack);
560 symbol_t *symbol = symbol_table_insert(string);
562 lexer_token.type = symbol->ID;
563 lexer_token.symbol = symbol;
565 if (symbol->string != string) {
566 obstack_free(&symbol_obstack, string);
571 * parse suffixes like 'LU' or 'f' after numbers
573 static void parse_number_suffix(void)
575 assert(obstack_object_size(&symbol_obstack) == 0);
579 obstack_1grow(&symbol_obstack, (char) c);
588 if (obstack_object_size(&symbol_obstack) == 0) {
589 lexer_token.symbol = NULL;
593 obstack_1grow(&symbol_obstack, '\0');
594 char *string = obstack_finish(&symbol_obstack);
595 symbol_t *symbol = symbol_table_insert(string);
597 if (symbol->string != string) {
598 obstack_free(&symbol_obstack, string);
600 lexer_token.symbol = symbol;
603 static string_t identify_string(char *string, size_t len)
607 const char *result = strset_insert(&stringset, concat);
608 if (result != concat) {
609 obstack_free(&symbol_obstack, concat);
612 const char *result = string;
614 return (string_t) {result, len};
618 * Parses a hex number including hex floats and set the
621 static void parse_number_hex(void)
623 bool is_float = false;
624 bool has_digits = false;
626 assert(obstack_object_size(&symbol_obstack) == 0);
627 while (isxdigit(c)) {
629 obstack_1grow(&symbol_obstack, (char) c);
635 obstack_1grow(&symbol_obstack, (char) c);
638 while (isxdigit(c)) {
640 obstack_1grow(&symbol_obstack, (char) c);
644 if (c == 'p' || c == 'P') {
646 obstack_1grow(&symbol_obstack, (char) c);
649 if (c == '-' || c == '+') {
650 obstack_1grow(&symbol_obstack, (char) c);
654 while (isxdigit(c)) {
655 obstack_1grow(&symbol_obstack, (char) c);
658 } else if (is_float) {
659 errorf(&lexer_token.source_position,
660 "hexadecimal floatingpoint constant requires an exponent");
662 obstack_1grow(&symbol_obstack, '\0');
664 size_t size = obstack_object_size(&symbol_obstack) - 1;
665 char *string = obstack_finish(&symbol_obstack);
666 lexer_token.literal = identify_string(string, size);
669 is_float ? T_FLOATINGPOINT_HEXADECIMAL : T_INTEGER_HEXADECIMAL;
672 errorf(&lexer_token.source_position, "invalid number literal '0x%S'",
673 &lexer_token.literal);
674 lexer_token.literal.begin = "0";
675 lexer_token.literal.size = 1;
678 parse_number_suffix();
682 * Returns true if the given char is a octal digit.
684 * @param char the character to check
686 static bool is_octal_digit(utf32 chr)
704 * Parses a number and sets the lexer_token.
706 static void parse_number(void)
708 bool is_float = false;
709 bool has_digits = false;
711 assert(obstack_object_size(&symbol_obstack) == 0);
714 if (c == 'x' || c == 'X') {
721 obstack_1grow(&symbol_obstack, '0');
726 obstack_1grow(&symbol_obstack, (char) c);
732 obstack_1grow(&symbol_obstack, '.');
737 obstack_1grow(&symbol_obstack, (char) c);
741 if (c == 'e' || c == 'E') {
743 obstack_1grow(&symbol_obstack, 'e');
746 if (c == '-' || c == '+') {
747 obstack_1grow(&symbol_obstack, (char) c);
752 obstack_1grow(&symbol_obstack, (char) c);
757 obstack_1grow(&symbol_obstack, '\0');
758 size_t size = obstack_object_size(&symbol_obstack) - 1;
759 char *string = obstack_finish(&symbol_obstack);
760 lexer_token.literal = identify_string(string, size);
762 /* is it an octal number? */
764 lexer_token.type = T_FLOATINGPOINT;
765 } else if (string[0] == '0') {
766 lexer_token.type = T_INTEGER_OCTAL;
768 /* check for invalid octal digits */
769 for (size_t i= 0; i < size; ++i) {
771 if (t == '8' || t == '9')
772 errorf(&lexer_token.source_position,
773 "invalid digit '%c' in octal number", t);
776 lexer_token.type = T_INTEGER;
780 errorf(&lexer_token.source_position, "invalid number literal '%S'",
781 &lexer_token.literal);
784 parse_number_suffix();
788 * Returns the value of a digit.
789 * The only portable way to do it ...
791 static int digit_value(utf32 const digit)
817 internal_error("wrong character given");
822 * Parses an octal character sequence.
824 * @param first_digit the already read first digit
826 static utf32 parse_octal_sequence(utf32 const first_digit)
828 assert(is_octal_digit(first_digit));
829 utf32 value = digit_value(first_digit);
830 if (!is_octal_digit(c)) return value;
831 value = 8 * value + digit_value(c);
833 if (!is_octal_digit(c)) return value;
834 value = 8 * value + digit_value(c);
840 * Parses a hex character sequence.
842 static utf32 parse_hex_sequence(void)
845 while (isxdigit(c)) {
846 value = 16 * value + digit_value(c);
853 * Parse an escape sequence.
855 static utf32 parse_escape_sequence(void)
863 case '"': return '"';
864 case '\'': return '\'';
865 case '\\': return '\\';
866 case '?': return '\?';
867 case 'a': return '\a';
868 case 'b': return '\b';
869 case 'f': return '\f';
870 case 'n': return '\n';
871 case 'r': return '\r';
872 case 't': return '\t';
873 case 'v': return '\v';
875 return parse_hex_sequence();
884 return parse_octal_sequence(ec);
886 parse_error("reached end of file while parsing escape sequence");
888 /* \E is not documented, but handled, by GCC. It is acceptable according
889 * to §6.11.4, whereas \e is not. */
893 return 27; /* hopefully 27 is ALWAYS the code for ESCAPE */
897 parse_error("universal character parsing not implemented yet");
902 /* §6.4.4.4:8 footnote 64 */
903 parse_error("unknown escape sequence");
908 * Concatenate two strings.
910 string_t concat_strings(const string_t *const s1, const string_t *const s2)
912 const size_t len1 = s1->size - 1;
913 const size_t len2 = s2->size - 1;
915 char *const concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1);
916 memcpy(concat, s1->begin, len1);
917 memcpy(concat + len1, s2->begin, len2 + 1);
919 return identify_string(concat, len1 + len2 + 1);
922 string_t make_string(const char *string)
924 size_t len = strlen(string) + 1;
925 char *const space = obstack_alloc(&symbol_obstack, len);
926 memcpy(space, string, len);
928 return identify_string(space, len);
931 static void grow_symbol(utf32 const tc)
933 struct obstack *const o = &symbol_obstack;
935 obstack_1grow(o, tc);
936 } else if (tc < 0x800) {
937 obstack_1grow(o, 0xC0 | (tc >> 6));
938 obstack_1grow(o, 0x80 | (tc & 0x3F));
939 } else if (tc < 0x10000) {
940 obstack_1grow(o, 0xE0 | ( tc >> 12));
941 obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F));
942 obstack_1grow(o, 0x80 | ( tc & 0x3F));
944 obstack_1grow(o, 0xF0 | ( tc >> 18));
945 obstack_1grow(o, 0x80 | ((tc >> 12) & 0x3F));
946 obstack_1grow(o, 0x80 | ((tc >> 6) & 0x3F));
947 obstack_1grow(o, 0x80 | ( tc & 0x3F));
952 * Parse a string literal and set lexer_token.
954 static void parse_string_literal(void)
956 const unsigned start_linenr = lexer_token.source_position.linenr;
963 utf32 const tc = parse_escape_sequence();
965 warningf(&lexer_token.source_position,
966 "escape sequence out of range");
968 obstack_1grow(&symbol_obstack, tc);
973 source_position_t source_position;
974 source_position.input_name = lexer_token.source_position.input_name;
975 source_position.linenr = start_linenr;
976 errorf(&source_position, "string has no end");
977 lexer_token.type = T_ERROR;
994 /* TODO: concatenate multiple strings separated by whitespace... */
996 /* add finishing 0 to the string */
997 obstack_1grow(&symbol_obstack, '\0');
998 const size_t size = (size_t)obstack_object_size(&symbol_obstack);
999 char *string = obstack_finish(&symbol_obstack);
1001 lexer_token.type = T_STRING_LITERAL;
1002 lexer_token.literal = identify_string(string, size);
1006 * Parse a wide character constant and set lexer_token.
1008 static void parse_wide_character_constant(void)
1010 const unsigned start_linenr = lexer_token.source_position.linenr;
1017 const utf32 tc = parse_escape_sequence();
1023 parse_error("newline while parsing character constant");
1029 goto end_of_wide_char_constant;
1032 source_position_t source_position = lexer_token.source_position;
1033 source_position.linenr = start_linenr;
1034 errorf(&source_position, "EOF while parsing character constant");
1035 lexer_token.type = T_ERROR;
1046 end_of_wide_char_constant:;
1047 obstack_1grow(&symbol_obstack, '\0');
1048 size_t size = (size_t) obstack_object_size(&symbol_obstack) - 1;
1049 char *string = obstack_finish(&symbol_obstack);
1051 lexer_token.type = T_WIDE_CHARACTER_CONSTANT;
1052 lexer_token.literal = identify_string(string, size);
1056 * Parse a wide string literal and set lexer_token.
1058 static void parse_wide_string_literal(void)
1060 parse_string_literal();
1061 if (lexer_token.type == T_STRING_LITERAL)
1062 lexer_token.type = T_WIDE_STRING_LITERAL;
1066 * Parse a character constant and set lexer_token.
1068 static void parse_character_constant(void)
1070 const unsigned start_linenr = lexer_token.source_position.linenr;
1077 utf32 const tc = parse_escape_sequence();
1079 warningf(&lexer_token.source_position,
1080 "escape sequence out of range");
1082 obstack_1grow(&symbol_obstack, tc);
1087 parse_error("newline while parsing character constant");
1093 goto end_of_char_constant;
1096 source_position_t source_position;
1097 source_position.input_name = lexer_token.source_position.input_name;
1098 source_position.linenr = start_linenr;
1099 errorf(&source_position, "EOF while parsing character constant");
1100 lexer_token.type = T_ERROR;
1112 end_of_char_constant:;
1113 obstack_1grow(&symbol_obstack, '\0');
1114 const size_t size = (size_t)obstack_object_size(&symbol_obstack)-1;
1115 char *const string = obstack_finish(&symbol_obstack);
1117 lexer_token.type = T_CHARACTER_CONSTANT;
1118 lexer_token.literal = identify_string(string, size);
1122 * Skip a multiline comment.
1124 static void skip_multiline_comment(void)
1126 unsigned start_linenr = lexer_token.source_position.linenr;
1133 /* nested comment, warn here */
1134 if (warning.comment) {
1135 warningf(&lexer_token.source_position, "'/*' within comment");
1147 MATCH_NEWLINE(break;)
1150 source_position_t source_position;
1151 source_position.input_name = lexer_token.source_position.input_name;
1152 source_position.linenr = start_linenr;
1153 errorf(&source_position, "at end of file while looking for comment end");
1165 * Skip a single line comment.
1167 static void skip_line_comment(void)
1180 if (c == '\n' || c == '\r') {
1181 if (warning.comment)
1182 warningf(&lexer_token.source_position, "multi-line comment");
1194 /** The current preprocessor token. */
1195 static token_t pp_token;
1198 * Read the next preprocessor token.
1200 static inline void next_pp_token(void)
1202 lexer_next_preprocessing_token();
1203 pp_token = lexer_token;
1207 * Eat all preprocessor tokens until newline.
1209 static void eat_until_newline(void)
1211 while (pp_token.type != '\n' && pp_token.type != T_EOF) {
1217 * Handle the define directive.
1219 static void define_directive(void)
1221 lexer_next_preprocessing_token();
1222 if (lexer_token.type != T_IDENTIFIER) {
1223 parse_error("expected identifier after #define\n");
1224 eat_until_newline();
1229 * Handle the ifdef directive.
1231 static void ifdef_directive(int is_ifndef)
1234 lexer_next_preprocessing_token();
1235 //expect_identifier();
1240 * Handle the endif directive.
1242 static void endif_directive(void)
1248 * Parse the line directive.
1250 static void parse_line_directive(void)
1252 if (pp_token.type != T_INTEGER) {
1253 parse_error("expected integer");
1255 /* use offset -1 as this is about the next line */
1256 lexer_token.source_position.linenr = atoi(pp_token.literal.begin) - 1;
1259 if (pp_token.type == T_STRING_LITERAL) {
1260 lexer_token.source_position.input_name = pp_token.literal.begin;
1264 eat_until_newline();
1270 typedef enum stdc_pragma_kind_t {
1274 STDC_CX_LIMITED_RANGE
1275 } stdc_pragma_kind_t;
1278 * STDC pragma values.
1280 typedef enum stdc_pragma_value_kind_t {
1285 } stdc_pragma_value_kind_t;
1288 * Parse a pragma directive.
1290 static void parse_pragma(void)
1292 bool unknown_pragma = true;
1295 if (pp_token.symbol->pp_ID == TP_STDC) {
1296 stdc_pragma_kind_t kind = STDC_UNKNOWN;
1298 if (c_mode & _C99) {
1301 switch (pp_token.symbol->pp_ID) {
1302 case TP_FP_CONTRACT:
1303 kind = STDC_FP_CONTRACT;
1305 case TP_FENV_ACCESS:
1306 kind = STDC_FENV_ACCESS;
1308 case TP_CX_LIMITED_RANGE:
1309 kind = STDC_CX_LIMITED_RANGE;
1314 if (kind != STDC_UNKNOWN) {
1315 stdc_pragma_value_kind_t value = STDC_VALUE_UNKNOWN;
1317 switch (pp_token.symbol->pp_ID) {
1319 value = STDC_VALUE_ON;
1322 value = STDC_VALUE_OFF;
1325 value = STDC_VALUE_DEFAULT;
1330 if (value != STDC_VALUE_UNKNOWN) {
1331 unknown_pragma = false;
1333 errorf(&pp_token.source_position, "bad STDC pragma argument");
1338 unknown_pragma = true;
1340 eat_until_newline();
1341 if (unknown_pragma && warning.unknown_pragmas) {
1342 warningf(&pp_token.source_position, "encountered unknown #pragma");
1347 * Parse a preprocessor non-null directive.
1349 static void parse_preprocessor_identifier(void)
1351 assert(pp_token.type == T_IDENTIFIER);
1352 symbol_t *symbol = pp_token.symbol;
1354 switch (symbol->pp_ID) {
1356 printf("include - enable header name parsing!\n");
1372 parse_line_directive();
1379 /* TODO; output the rest of the line */
1380 parse_error("#error directive: ");
1389 * Parse a preprocessor directive.
1391 static void parse_preprocessor_directive(void)
1395 switch (pp_token.type) {
1397 parse_preprocessor_identifier();
1400 parse_line_directive();
1403 /* NULL directive, see §6.10.7 */
1406 parse_error("invalid preprocessor directive");
1407 eat_until_newline();
1412 #define MAYBE_PROLOG \
1417 #define MAYBE(ch, set_type) \
1420 lexer_token.type = set_type; \
1423 /* must use this as last thing */
1424 #define MAYBE_MODE(ch, set_type, mode) \
1426 if (c_mode & mode) { \
1428 lexer_token.type = set_type; \
1433 #define ELSE_CODE(code) \
1437 } /* end of while (true) */ \
1440 #define ELSE(set_type) \
1442 lexer_token.type = set_type; \
1446 void lexer_next_preprocessing_token(void)
1456 lexer_token.type = '\n';
1462 /* might be a wide string ( L"string" ) */
1463 if (lexer_token.symbol == symbol_L) {
1465 case '"': parse_wide_string_literal(); break;
1466 case '\'': parse_wide_character_constant(); break;
1476 parse_string_literal();
1480 parse_character_constant();
1493 MAYBE('.', T_DOTDOTDOT)
1497 lexer_token.type = '.';
1503 MAYBE('&', T_ANDAND)
1504 MAYBE('=', T_ANDEQUAL)
1508 MAYBE('=', T_ASTERISKEQUAL)
1512 MAYBE('+', T_PLUSPLUS)
1513 MAYBE('=', T_PLUSEQUAL)
1517 MAYBE('>', T_MINUSGREATER)
1518 MAYBE('-', T_MINUSMINUS)
1519 MAYBE('=', T_MINUSEQUAL)
1523 MAYBE('=', T_EXCLAMATIONMARKEQUAL)
1527 MAYBE('=', T_SLASHEQUAL)
1530 skip_multiline_comment();
1531 lexer_next_preprocessing_token();
1535 skip_line_comment();
1536 lexer_next_preprocessing_token();
1542 MAYBE('=', T_PERCENTEQUAL)
1547 MAYBE(':', T_HASHHASH)
1551 lexer_token.type = '#';
1560 MAYBE('=', T_LESSEQUAL)
1563 MAYBE('=', T_LESSLESSEQUAL)
1568 MAYBE('=', T_GREATEREQUAL)
1571 MAYBE('=', T_GREATERGREATEREQUAL)
1572 ELSE(T_GREATERGREATER)
1576 MAYBE('=', T_CARETEQUAL)
1580 MAYBE('=', T_PIPEEQUAL)
1581 MAYBE('|', T_PIPEPIPE)
1586 MAYBE_MODE(':', T_COLONCOLON, _CXX)
1590 MAYBE('=', T_EQUALEQUAL)
1594 MAYBE('#', T_HASHHASH)
1608 lexer_token.type = c;
1613 lexer_token.type = T_EOF;
1618 errorf(&lexer_token.source_position, "unknown character '%c' found", c);
1620 lexer_token.type = T_ERROR;
1626 void lexer_next_token(void)
1628 lexer_next_preprocessing_token();
1630 while (lexer_token.type == '\n') {
1632 lexer_next_preprocessing_token();
1635 if (lexer_token.type == '#') {
1636 parse_preprocessor_directive();
1641 void init_lexer(void)
1643 strset_init(&stringset);
1644 symbol_L = symbol_table_insert("L");
1647 void lexer_open_stream(FILE *stream, const char *input_name)
1650 lexer_token.source_position.linenr = 0;
1651 lexer_token.source_position.input_name = input_name;
1656 /* place a virtual \n at the beginning so the lexer knows that we're
1657 * at the beginning of a line */
1661 void lexer_open_buffer(const char *buffer, size_t len, const char *input_name)
1664 lexer_token.source_position.linenr = 0;
1665 lexer_token.source_position.input_name = input_name;
1669 bufend = buffer + len;
1671 /* place a virtual \n at the beginning so the lexer knows that we're
1672 * at the beginning of a line */
1677 panic("builtin lexing not done yet");
1681 void exit_lexer(void)
1683 strset_destroy(&stringset);
1686 static __attribute__((unused))
1687 void dbg_pos(const source_position_t source_position)
1689 fprintf(stdout, "%s:%u\n", source_position.input_name,
1690 source_position.linenr);