]>
Commit | Line | Data |
---|---|---|
83ffe9cd | 1 | // Copyright (C) 2020-2023 Free Software Foundation, Inc. |
18f6990f JP |
2 | |
3 | // This file is part of GCC. | |
4 | ||
5 | // GCC is free software; you can redistribute it and/or modify it under | |
6 | // the terms of the GNU General Public License as published by the Free | |
7 | // Software Foundation; either version 3, or (at your option) any later | |
8 | // version. | |
9 | ||
10 | // GCC is distributed in the hope that it will be useful, but WITHOUT ANY | |
11 | // WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
12 | // FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
13 | // for more details. | |
14 | ||
15 | // You should have received a copy of the GNU General Public License | |
16 | // along with GCC; see the file COPYING3. If not see | |
17 | // <http://www.gnu.org/licenses/>. | |
18 | ||
19 | #include "rust-system.h" | |
20 | #include "rust-lex.h" | |
21 | #include "rust-diagnostics.h" | |
22 | #include "rust-linemap.h" | |
23 | #include "rust-session-manager.h" | |
24 | #include "safe-ctype.h" | |
25 | ||
26 | namespace Rust { | |
27 | // TODO: move to separate compilation unit? | |
28 | // overload += for uint32_t to allow 32-bit encoded utf-8 to be added | |
29 | std::string & | |
30 | operator+= (std::string &str, Codepoint char32) | |
31 | { | |
32 | if (char32.value < 0x80) | |
33 | { | |
34 | str += static_cast<char> (char32.value); | |
35 | } | |
36 | else if (char32.value < (0x1F + 1) << (1 * 6)) | |
37 | { | |
38 | str += static_cast<char> (0xC0 | ((char32.value >> 6) & 0x1F)); | |
39 | str += static_cast<char> (0x80 | ((char32.value >> 0) & 0x3F)); | |
40 | } | |
41 | else if (char32.value < (0x0F + 1) << (2 * 6)) | |
42 | { | |
43 | str += static_cast<char> (0xE0 | ((char32.value >> 12) & 0x0F)); | |
44 | str += static_cast<char> (0x80 | ((char32.value >> 6) & 0x3F)); | |
45 | str += static_cast<char> (0x80 | ((char32.value >> 0) & 0x3F)); | |
46 | } | |
47 | else if (char32.value < (0x07 + 1) << (3 * 6)) | |
48 | { | |
49 | str += static_cast<char> (0xF0 | ((char32.value >> 18) & 0x07)); | |
50 | str += static_cast<char> (0x80 | ((char32.value >> 12) & 0x3F)); | |
51 | str += static_cast<char> (0x80 | ((char32.value >> 6) & 0x3F)); | |
52 | str += static_cast<char> (0x80 | ((char32.value >> 0) & 0x3F)); | |
53 | } | |
54 | else | |
55 | { | |
56 | rust_debug ("Invalid unicode codepoint found: '%u' ", char32.value); | |
57 | } | |
58 | return str; | |
59 | } | |
60 | ||
61 | std::string | |
62 | Codepoint::as_string () | |
63 | { | |
64 | std::string str; | |
65 | ||
66 | // str += Codepoint (value); | |
67 | str += *this; | |
68 | ||
69 | return str; | |
70 | } | |
71 | ||
72 | /* Includes all allowable float digits EXCEPT _ and . as that needs lookahead | |
73 | * for handling. */ | |
74 | bool | |
75 | is_float_digit (char number) | |
76 | { | |
77 | return ISDIGIT (number) || number == 'E' || number == 'e'; | |
78 | } | |
79 | ||
80 | /* Basically ISXDIGIT from safe-ctype but may change if Rust's encoding or | |
81 | * whatever is different */ | |
82 | bool | |
83 | is_x_digit (char number) | |
84 | { | |
85 | return ISXDIGIT (number); | |
86 | } | |
87 | ||
88 | bool | |
89 | is_octal_digit (char number) | |
90 | { | |
91 | return number >= '0' && number <= '7'; | |
92 | } | |
93 | ||
94 | bool | |
95 | is_bin_digit (char number) | |
96 | { | |
97 | return number == '0' || number == '1'; | |
98 | } | |
99 | ||
100 | bool | |
101 | check_valid_float_dot_end (char character) | |
102 | { | |
103 | return character != '.' && character != '_' && !ISALPHA (character); | |
104 | } | |
105 | ||
106 | // ISSPACE from safe-ctype but may change in future | |
107 | bool | |
108 | is_whitespace (char character) | |
109 | { | |
110 | return ISSPACE (character); | |
111 | } | |
112 | ||
113 | bool | |
114 | is_non_decimal_int_literal_separator (char character) | |
115 | { | |
116 | return character == 'x' || character == 'o' || character == 'b'; | |
117 | } | |
118 | ||
119 | Lexer::Lexer (const std::string &input) | |
120 | : input (RAIIFile::create_error ()), current_line (1), current_column (1), | |
121 | line_map (nullptr), raw_input_source (new BufferInputSource (input, 0)), | |
122 | input_queue{*raw_input_source}, token_queue (TokenSource (this)) | |
123 | {} | |
124 | ||
125 | Lexer::Lexer (const char *filename, RAIIFile file_input, Linemap *linemap) | |
126 | : input (std::move (file_input)), current_line (1), current_column (1), | |
127 | line_map (linemap), | |
128 | raw_input_source (new FileInputSource (input.get_raw ())), | |
129 | input_queue{*raw_input_source}, token_queue (TokenSource (this)) | |
130 | { | |
131 | // inform line_table that file is being entered and is in line 1 | |
132 | if (linemap) | |
133 | line_map->start_file (filename, current_line); | |
134 | } | |
135 | ||
136 | Lexer::~Lexer () | |
137 | { | |
138 | /* ok apparently stop (which is equivalent of original code in destructor) is | |
139 | * meant to be called after all files have finished parsing, for cleanup. On | |
140 | * the other hand, actual code that it calls to leave a certain line map is | |
141 | * mentioned in GCC docs as being useful for "just leaving an included header" | |
142 | * and stuff like that, so this line mapping functionality may need fixing. | |
143 | * FIXME: find out whether this occurs. */ | |
144 | ||
145 | // line_map->stop(); | |
146 | } | |
147 | ||
148 | /* TODO: need to optimise somehow to avoid the virtual function call in the | |
149 | * tight loop. Best idea at the moment is CRTP, but that might make lexer | |
150 | * implementation annoying when storing the "base class" (i.e. would need | |
151 | * template parameter everywhere), although in practice it would mostly just | |
152 | * look ugly and make enclosing classes like Parser also require a type | |
153 | * parameter. At this point a macro might be better. OK I guess macros can be | |
154 | * replaced by constexpr if or something if possible. */ | |
155 | Location | |
156 | Lexer::get_current_location () | |
157 | { | |
158 | if (line_map) | |
159 | return line_map->get_location (current_column); | |
160 | else | |
161 | // If we have no linemap, we're lexing something without proper locations | |
162 | return Location (); | |
163 | } | |
164 | ||
165 | int | |
166 | Lexer::peek_input (int n) | |
167 | { | |
168 | return input_queue.peek (n); | |
169 | } | |
170 | ||
171 | int | |
172 | Lexer::peek_input () | |
173 | { | |
174 | return peek_input (0); | |
175 | } | |
176 | ||
177 | void | |
178 | Lexer::skip_input (int n) | |
179 | { | |
180 | input_queue.skip (n); | |
181 | } | |
182 | ||
183 | void | |
184 | Lexer::skip_input () | |
185 | { | |
186 | skip_input (0); | |
187 | } | |
188 | ||
189 | void | |
190 | Lexer::replace_current_token (TokenPtr replacement) | |
191 | { | |
192 | token_queue.replace_current_value (replacement); | |
193 | ||
194 | rust_debug ("called 'replace_current_token' - this is deprecated"); | |
195 | } | |
196 | ||
197 | /* shitty anonymous namespace that can only be accessed inside the compilation | |
198 | * unit - used for classify_keyword binary search in sorted array of keywords | |
199 | * created with x-macros. */ | |
200 | namespace { | |
201 | // TODO: make constexpr when update to c++20 | |
202 | const std::string keyword_index[] = { | |
203 | #define RS_TOKEN(x, y) | |
204 | #define RS_TOKEN_KEYWORD(name, keyword) keyword, | |
205 | RS_TOKEN_LIST | |
206 | #undef RS_TOKEN_KEYWORD | |
207 | #undef RS_TOKEN | |
208 | }; | |
209 | ||
210 | constexpr TokenId keyword_keys[] = { | |
211 | #define RS_TOKEN(x, y) | |
212 | #define RS_TOKEN_KEYWORD(name, keyword) name, | |
213 | RS_TOKEN_LIST | |
214 | #undef RS_TOKEN_KEYWORD | |
215 | #undef RS_TOKEN | |
216 | }; | |
217 | ||
218 | constexpr int num_keywords = sizeof (keyword_index) / sizeof (*keyword_index); | |
219 | } // namespace | |
220 | ||
221 | /* Determines whether the string passed in is a keyword or not. If it is, it | |
222 | * returns the keyword name. */ | |
223 | TokenId | |
224 | Lexer::classify_keyword (const std::string &str) | |
225 | { | |
226 | const std::string *last = keyword_index + num_keywords; | |
227 | const std::string *idx = std::lower_bound (keyword_index, last, str); | |
228 | ||
229 | if (idx == last || str != *idx) | |
230 | return IDENTIFIER; | |
231 | ||
232 | // TODO: possibly replace this x-macro system with something like hash map? | |
233 | ||
234 | // We now have the expected token ID of the reserved keyword. However, some | |
235 | // keywords are reserved starting in certain editions. For example, `try` is | |
236 | // only a reserved keyword in editions >=2018. The language might gain new | |
237 | // reserved keywords in the future. | |
238 | // | |
239 | // https://doc.rust-lang.org/reference/keywords.html#reserved-keywords | |
240 | auto id = keyword_keys[idx - keyword_index]; | |
241 | ||
242 | // `try` is not a reserved keyword before 2018 | |
243 | if (Session::get_instance ().options.get_edition () | |
244 | == CompileOptions::Edition::E2015 | |
245 | && id == TRY) | |
246 | return IDENTIFIER; | |
247 | ||
248 | return id; | |
249 | } | |
250 | ||
251 | TokenPtr | |
252 | Lexer::build_token () | |
253 | { | |
254 | // loop to go through multiple characters to build a single token | |
255 | while (true) | |
256 | { | |
257 | Location loc = get_current_location (); | |
258 | current_char = peek_input (); | |
259 | skip_input (); | |
260 | ||
261 | // detect UTF8 bom | |
262 | // | |
263 | // Must be the first thing on the first line. | |
264 | // There might be an optional BOM (Byte Order Mark), which for UTF-8 is | |
265 | // the three bytes 0xEF, 0xBB and 0xBF. These can simply be skipped. | |
266 | if (current_line == 1 && current_column == 1 && current_char == 0xef | |
267 | && peek_input () == 0xbb && peek_input (1) == 0xbf) | |
268 | { | |
269 | skip_input (1); | |
270 | current_char = peek_input (); | |
271 | skip_input (); | |
272 | } | |
273 | ||
274 | // detect shebang | |
275 | // Must be the first thing on the first line, starting with #! | |
276 | // But since an attribute can also start with an #! we don't count it as a | |
277 | // shebang line when after any whitespace or comments there is a [. If it | |
278 | // is a shebang line we simple drop the line. Otherwise we don't consume | |
279 | // any characters and fall through to the real tokenizer. | |
280 | if (current_line == 1 && current_column == 1 && current_char == '#' | |
281 | && peek_input () == '!') | |
282 | { | |
283 | int n = 1; | |
284 | while (true) | |
285 | { | |
286 | int next_char = peek_input (n); | |
287 | if (is_whitespace (next_char)) | |
288 | n++; | |
289 | else if ((next_char == '/' && peek_input (n + 1) == '/' | |
290 | && peek_input (n + 2) != '!' | |
291 | && peek_input (n + 2) != '/') | |
292 | || (next_char == '/' && peek_input (n + 1) == '/' | |
293 | && peek_input (n + 2) == '/' | |
294 | && peek_input (n + 3) == '/')) | |
295 | { | |
296 | // two // or four //// | |
297 | // A single line comment | |
298 | // (but not an inner or outer doc comment) | |
299 | n += 2; | |
300 | next_char = peek_input (n); | |
301 | while (next_char != '\n' && next_char != EOF) | |
302 | { | |
303 | n++; | |
304 | next_char = peek_input (n); | |
305 | } | |
306 | if (next_char == '\n') | |
307 | n++; | |
308 | } | |
309 | else if (next_char == '/' && peek_input (n + 1) == '*' | |
310 | && peek_input (n + 2) == '*' | |
311 | && peek_input (n + 3) == '/') | |
312 | { | |
313 | /**/ | |
314 | n += 4; | |
315 | } | |
316 | else if (next_char == '/' && peek_input (n + 1) == '*' | |
317 | && peek_input (n + 2) == '*' && peek_input (n + 3) == '*' | |
318 | && peek_input (n + 4) == '/') | |
319 | { | |
320 | /***/ | |
321 | n += 5; | |
322 | } | |
323 | else if ((next_char == '/' && peek_input (n + 1) == '*' | |
324 | && peek_input (n + 2) != '*' | |
325 | && peek_input (n + 2) != '!') | |
326 | || (next_char == '/' && peek_input (n + 1) == '*' | |
327 | && peek_input (n + 2) == '*' | |
328 | && peek_input (n + 3) == '*')) | |
329 | { | |
330 | // one /* or three /*** | |
331 | // Start of a block comment | |
332 | // (but not an inner or outer doc comment) | |
333 | n += 2; | |
334 | int level = 1; | |
335 | while (level > 0) | |
336 | { | |
337 | if (peek_input (n) == EOF) | |
338 | break; | |
339 | else if (peek_input (n) == '/' | |
340 | && peek_input (n + 1) == '*') | |
341 | { | |
342 | n += 2; | |
343 | level += 1; | |
344 | } | |
345 | else if (peek_input (n) == '*' | |
346 | && peek_input (n + 1) == '/') | |
347 | { | |
348 | n += 2; | |
349 | level -= 1; | |
350 | } | |
351 | else | |
352 | n++; | |
353 | } | |
354 | } | |
355 | else if (next_char != '[') | |
356 | { | |
357 | // definitely shebang, ignore the first line | |
358 | while (current_char != '\n' && current_char != EOF) | |
359 | { | |
360 | current_char = peek_input (); | |
361 | skip_input (); | |
362 | } | |
363 | ||
364 | // newline | |
365 | current_line++; | |
366 | current_column = 1; | |
367 | // tell line_table that new line starts | |
368 | start_line (current_line, max_column_hint); | |
369 | break; | |
370 | } | |
371 | else | |
372 | break; /* Definitely not a shebang line. */ | |
373 | } | |
374 | } | |
375 | ||
376 | // return end of file token if end of file | |
377 | if (current_char == EOF) | |
378 | return Token::make (END_OF_FILE, loc); | |
379 | ||
380 | // if not end of file, start tokenising | |
381 | switch (current_char) | |
382 | { | |
383 | /* ignore whitespace characters for tokens but continue updating | |
384 | * location */ | |
385 | case '\n': // newline | |
386 | current_line++; | |
387 | current_column = 1; | |
388 | // tell line_table that new line starts | |
389 | start_line (current_line, max_column_hint); | |
390 | continue; | |
391 | case '\r': // cr | |
392 | // Ignore, we expect a newline (lf) soon. | |
393 | continue; | |
394 | case ' ': // space | |
395 | current_column++; | |
396 | continue; | |
397 | case '\t': // tab | |
398 | // width of a tab is not well-defined, assume 8 spaces | |
399 | current_column += 8; | |
400 | continue; | |
401 | ||
402 | // punctuation - actual tokens | |
403 | case '=': | |
404 | if (peek_input () == '>') | |
405 | { | |
406 | // match arm arrow | |
407 | skip_input (); | |
408 | current_column += 2; | |
409 | ||
410 | return Token::make (MATCH_ARROW, loc); | |
411 | } | |
412 | else if (peek_input () == '=') | |
413 | { | |
414 | // equality operator | |
415 | skip_input (); | |
416 | current_column += 2; | |
417 | ||
418 | return Token::make (EQUAL_EQUAL, loc); | |
419 | } | |
420 | else | |
421 | { | |
422 | // assignment operator | |
423 | current_column++; | |
424 | return Token::make (EQUAL, loc); | |
425 | } | |
426 | case '(': | |
427 | current_column++; | |
428 | return Token::make (LEFT_PAREN, loc); | |
429 | case '-': | |
430 | if (peek_input () == '>') | |
431 | { | |
432 | // return type specifier | |
433 | skip_input (); | |
434 | current_column += 2; | |
435 | ||
436 | return Token::make (RETURN_TYPE, loc); | |
437 | } | |
438 | else if (peek_input () == '=') | |
439 | { | |
440 | // minus-assign | |
441 | skip_input (); | |
442 | current_column += 2; | |
443 | ||
444 | return Token::make (MINUS_EQ, loc); | |
445 | } | |
446 | else | |
447 | { | |
448 | // minus | |
449 | current_column++; | |
450 | return Token::make (MINUS, loc); | |
451 | } | |
452 | case '+': | |
453 | if (peek_input () == '=') | |
454 | { | |
455 | // add-assign | |
456 | skip_input (); | |
457 | current_column += 2; | |
458 | ||
459 | return Token::make (PLUS_EQ, loc); | |
460 | } | |
461 | else | |
462 | { | |
463 | // add | |
464 | current_column++; | |
465 | return Token::make (PLUS, loc); | |
466 | } | |
467 | case ')': | |
468 | current_column++; | |
469 | return Token::make (RIGHT_PAREN, loc); | |
470 | case ';': | |
471 | current_column++; | |
472 | return Token::make (SEMICOLON, loc); | |
473 | case '*': | |
474 | if (peek_input () == '=') | |
475 | { | |
476 | // multiplication-assign | |
477 | skip_input (); | |
478 | current_column += 2; | |
479 | ||
480 | return Token::make (ASTERISK_EQ, loc); | |
481 | } | |
482 | else | |
483 | { | |
484 | // multiplication | |
485 | current_column++; | |
486 | return Token::make (ASTERISK, loc); | |
487 | } | |
488 | case ',': | |
489 | current_column++; | |
490 | return Token::make (COMMA, loc); | |
491 | case '/': | |
492 | if (peek_input () == '=') | |
493 | { | |
494 | // division-assign | |
495 | skip_input (); | |
496 | current_column += 2; | |
497 | ||
498 | return Token::make (DIV_EQ, loc); | |
499 | } | |
500 | else if ((peek_input () == '/' && peek_input (1) != '!' | |
501 | && peek_input (1) != '/') | |
502 | || (peek_input () == '/' && peek_input (1) == '/' | |
503 | && peek_input (2) == '/')) | |
504 | { | |
505 | // two // or four //// | |
506 | // single line comment | |
507 | // (but not an inner or outer doc comment) | |
508 | skip_input (); | |
509 | current_column += 2; | |
510 | current_char = peek_input (); | |
511 | ||
512 | // basically ignore until line finishes | |
513 | while (current_char != '\n' && current_char != EOF) | |
514 | { | |
515 | skip_input (); | |
516 | current_column++; // not used | |
517 | current_char = peek_input (); | |
518 | } | |
519 | continue; | |
520 | } | |
521 | else if (peek_input () == '/' | |
522 | && (peek_input (1) == '!' || peek_input (1) == '/')) | |
523 | { | |
524 | /* single line doc comment, inner or outer. */ | |
525 | bool is_inner = peek_input (1) == '!'; | |
526 | skip_input (1); | |
527 | current_column += 3; | |
528 | ||
529 | std::string str; | |
530 | str.reserve (32); | |
531 | current_char = peek_input (); | |
532 | while (current_char != '\n') | |
533 | { | |
534 | skip_input (); | |
535 | if (current_char == '\r') | |
536 | { | |
537 | char next_char = peek_input (); | |
538 | if (next_char == '\n') | |
539 | { | |
540 | current_char = '\n'; | |
541 | break; | |
542 | } | |
543 | rust_error_at ( | |
544 | loc, "Isolated CR %<\\r%> not allowed in doc comment"); | |
545 | current_char = next_char; | |
546 | continue; | |
547 | } | |
548 | if (current_char == EOF) | |
549 | { | |
550 | rust_error_at ( | |
551 | loc, "unexpected EOF while looking for end of comment"); | |
552 | break; | |
553 | } | |
554 | str += current_char; | |
555 | current_char = peek_input (); | |
556 | } | |
557 | skip_input (); | |
558 | current_line++; | |
559 | current_column = 1; | |
560 | // tell line_table that new line starts | |
561 | start_line (current_line, max_column_hint); | |
562 | ||
563 | str.shrink_to_fit (); | |
564 | if (is_inner) | |
565 | return Token::make_inner_doc_comment (loc, std::move (str)); | |
566 | else | |
567 | return Token::make_outer_doc_comment (loc, std::move (str)); | |
568 | } | |
569 | else if (peek_input () == '*' && peek_input (1) == '*' | |
570 | && peek_input (2) == '/') | |
571 | { | |
572 | /**/ | |
573 | skip_input (2); | |
574 | current_column += 4; | |
575 | continue; | |
576 | } | |
577 | else if (peek_input () == '*' && peek_input (1) == '*' | |
578 | && peek_input (2) == '*' && peek_input (3) == '/') | |
579 | { | |
580 | /***/ | |
581 | skip_input (3); | |
582 | current_column += 5; | |
583 | continue; | |
584 | } | |
585 | else if ((peek_input () == '*' && peek_input (1) != '!' | |
586 | && peek_input (1) != '*') | |
587 | || (peek_input () == '*' && peek_input (1) == '*' | |
588 | && peek_input (2) == '*')) | |
589 | { | |
590 | // one /* or three /*** | |
591 | // block comment | |
592 | // (but not an inner or outer doc comment) | |
593 | skip_input (); | |
594 | current_column += 2; | |
595 | ||
596 | int level = 1; | |
597 | while (level > 0) | |
598 | { | |
599 | current_char = peek_input (); | |
600 | ||
601 | if (current_char == EOF) | |
602 | { | |
603 | rust_error_at ( | |
604 | loc, "unexpected EOF while looking for end of comment"); | |
605 | break; | |
606 | } | |
607 | ||
608 | // if /* found | |
609 | if (current_char == '/' && peek_input (1) == '*') | |
610 | { | |
611 | // skip /* characters | |
612 | skip_input (1); | |
613 | ||
614 | current_column += 2; | |
615 | ||
616 | level += 1; | |
617 | continue; | |
618 | } | |
619 | ||
620 | // ignore until */ is found | |
621 | if (current_char == '*' && peek_input (1) == '/') | |
622 | { | |
623 | // skip */ characters | |
624 | skip_input (1); | |
625 | ||
626 | current_column += 2; | |
627 | ||
628 | level -= 1; | |
629 | continue; | |
630 | } | |
631 | ||
632 | if (current_char == '\n') | |
633 | { | |
634 | skip_input (); | |
635 | current_line++; | |
636 | current_column = 1; | |
637 | // tell line_table that new line starts | |
638 | start_line (current_line, max_column_hint); | |
639 | continue; | |
640 | } | |
641 | ||
642 | skip_input (); | |
643 | current_column++; | |
644 | } | |
645 | ||
646 | // refresh new token | |
647 | continue; | |
648 | } | |
649 | else if (peek_input () == '*' | |
650 | && (peek_input (1) == '!' || peek_input (1) == '*')) | |
651 | { | |
652 | // block doc comment, inner /*! or outer /** | |
653 | bool is_inner = peek_input (1) == '!'; | |
654 | skip_input (1); | |
655 | current_column += 3; | |
656 | ||
657 | std::string str; | |
658 | str.reserve (96); | |
659 | ||
660 | int level = 1; | |
661 | while (level > 0) | |
662 | { | |
663 | current_char = peek_input (); | |
664 | ||
665 | if (current_char == EOF) | |
666 | { | |
667 | rust_error_at ( | |
668 | loc, "unexpected EOF while looking for end of comment"); | |
669 | break; | |
670 | } | |
671 | ||
672 | // if /* found | |
673 | if (current_char == '/' && peek_input (1) == '*') | |
674 | { | |
675 | // skip /* characters | |
676 | skip_input (1); | |
677 | current_column += 2; | |
678 | ||
679 | level += 1; | |
680 | str += "/*"; | |
681 | continue; | |
682 | } | |
683 | ||
684 | // ignore until */ is found | |
685 | if (current_char == '*' && peek_input (1) == '/') | |
686 | { | |
687 | // skip */ characters | |
688 | skip_input (1); | |
689 | current_column += 2; | |
690 | ||
691 | level -= 1; | |
692 | if (level > 0) | |
693 | str += "*/"; | |
694 | continue; | |
695 | } | |
696 | ||
697 | if (current_char == '\r' && peek_input (1) != '\n') | |
698 | rust_error_at ( | |
699 | loc, "Isolated CR %<\\r%> not allowed in doc comment"); | |
700 | ||
701 | if (current_char == '\n') | |
702 | { | |
703 | skip_input (); | |
704 | current_line++; | |
705 | current_column = 1; | |
706 | // tell line_table that new line starts | |
707 | start_line (current_line, max_column_hint); | |
708 | str += '\n'; | |
709 | continue; | |
710 | } | |
711 | ||
712 | str += current_char; | |
713 | skip_input (); | |
714 | current_column++; | |
715 | } | |
716 | ||
717 | str.shrink_to_fit (); | |
718 | if (is_inner) | |
719 | return Token::make_inner_doc_comment (loc, std::move (str)); | |
720 | else | |
721 | return Token::make_outer_doc_comment (loc, std::move (str)); | |
722 | } | |
723 | else | |
724 | { | |
725 | // division | |
726 | current_column++; | |
727 | return Token::make (DIV, loc); | |
728 | } | |
729 | case '%': | |
730 | if (peek_input () == '=') | |
731 | { | |
732 | // modulo-assign | |
733 | skip_input (); | |
734 | current_column += 2; | |
735 | ||
736 | return Token::make (PERCENT_EQ, loc); | |
737 | } | |
738 | else | |
739 | { | |
740 | // modulo | |
741 | current_column++; | |
742 | return Token::make (PERCENT, loc); | |
743 | } | |
744 | case '^': | |
745 | if (peek_input () == '=') | |
746 | { | |
747 | // xor-assign? | |
748 | skip_input (); | |
749 | current_column += 2; | |
750 | ||
751 | return Token::make (CARET_EQ, loc); | |
752 | } | |
753 | else | |
754 | { | |
755 | // xor? | |
756 | current_column++; | |
757 | return Token::make (CARET, loc); | |
758 | } | |
759 | case '<': | |
760 | if (peek_input () == '<') | |
761 | { | |
762 | if (peek_input (1) == '=') | |
763 | { | |
764 | // left-shift assign | |
765 | skip_input (1); | |
766 | current_column += 3; | |
767 | ||
768 | return Token::make (LEFT_SHIFT_EQ, loc); | |
769 | } | |
770 | else | |
771 | { | |
772 | // left-shift | |
773 | skip_input (); | |
774 | current_column += 2; | |
775 | ||
776 | return Token::make (LEFT_SHIFT, loc); | |
777 | } | |
778 | } | |
779 | else if (peek_input () == '=') | |
780 | { | |
781 | // smaller than or equal to | |
782 | skip_input (); | |
783 | current_column += 2; | |
784 | ||
785 | return Token::make (LESS_OR_EQUAL, loc); | |
786 | } | |
787 | else | |
788 | { | |
789 | // smaller than | |
790 | current_column++; | |
791 | return Token::make (LEFT_ANGLE, loc); | |
792 | } | |
793 | break; | |
794 | case '>': | |
795 | if (peek_input () == '>') | |
796 | { | |
797 | if (peek_input (1) == '=') | |
798 | { | |
799 | // right-shift-assign | |
800 | skip_input (1); | |
801 | current_column += 3; | |
802 | ||
803 | return Token::make (RIGHT_SHIFT_EQ, loc); | |
804 | } | |
805 | else | |
806 | { | |
807 | // right-shift | |
808 | skip_input (); | |
809 | current_column += 2; | |
810 | ||
811 | return Token::make (RIGHT_SHIFT, loc); | |
812 | } | |
813 | } | |
814 | else if (peek_input () == '=') | |
815 | { | |
816 | // larger than or equal to | |
817 | skip_input (); | |
818 | current_column += 2; | |
819 | ||
820 | return Token::make (GREATER_OR_EQUAL, loc); | |
821 | } | |
822 | else | |
823 | { | |
824 | // larger than | |
825 | current_column++; | |
826 | return Token::make (RIGHT_ANGLE, loc); | |
827 | } | |
828 | case ':': | |
829 | if (peek_input () == ':') | |
830 | { | |
831 | // scope resolution :: | |
832 | skip_input (); | |
833 | current_column += 2; | |
834 | ||
835 | return Token::make (SCOPE_RESOLUTION, loc); | |
836 | } | |
837 | else | |
838 | { | |
839 | // single colon : | |
840 | current_column++; | |
841 | return Token::make (COLON, loc); | |
842 | } | |
843 | case '!': | |
844 | // no special handling for macros in lexer? | |
845 | if (peek_input () == '=') | |
846 | { | |
847 | // not equal boolean operator | |
848 | skip_input (); | |
849 | current_column += 2; | |
850 | ||
851 | return Token::make (NOT_EQUAL, loc); | |
852 | } | |
853 | else | |
854 | { | |
855 | // not equal unary operator | |
856 | current_column++; | |
857 | ||
858 | return Token::make (EXCLAM, loc); | |
859 | } | |
860 | case '?': | |
861 | current_column++; | |
862 | return Token::make (QUESTION_MARK, loc); | |
863 | case '#': | |
864 | current_column++; | |
865 | return Token::make (HASH, loc); | |
866 | case '[': | |
867 | current_column++; | |
868 | return Token::make (LEFT_SQUARE, loc); | |
869 | case ']': | |
870 | current_column++; | |
871 | return Token::make (RIGHT_SQUARE, loc); | |
872 | case '{': | |
873 | current_column++; | |
874 | return Token::make (LEFT_CURLY, loc); | |
875 | case '}': | |
876 | current_column++; | |
877 | return Token::make (RIGHT_CURLY, loc); | |
878 | case '@': | |
879 | current_column++; | |
880 | return Token::make (PATTERN_BIND, loc); | |
881 | case '$': | |
882 | current_column++; | |
883 | return Token::make (DOLLAR_SIGN, loc); | |
884 | case '~': | |
885 | current_column++; | |
886 | return Token::make (TILDE, loc); | |
887 | case '\\': | |
888 | current_column++; | |
889 | return Token::make (BACKSLASH, loc); | |
890 | case '`': | |
891 | current_column++; | |
892 | return Token::make (BACKTICK, loc); | |
893 | case '|': | |
894 | if (peek_input () == '=') | |
895 | { | |
896 | // bitwise or-assign? | |
897 | skip_input (); | |
898 | current_column += 2; | |
899 | ||
900 | return Token::make (PIPE_EQ, loc); | |
901 | } | |
902 | else if (peek_input () == '|') | |
903 | { | |
904 | // logical or | |
905 | skip_input (); | |
906 | current_column += 2; | |
907 | ||
908 | return Token::make (OR, loc); | |
909 | } | |
910 | else | |
911 | { | |
912 | // bitwise or | |
913 | current_column++; | |
914 | ||
915 | return Token::make (PIPE, loc); | |
916 | } | |
917 | case '&': | |
918 | if (peek_input () == '=') | |
919 | { | |
920 | // bitwise and-assign? | |
921 | skip_input (); | |
922 | current_column += 2; | |
923 | ||
924 | return Token::make (AMP_EQ, loc); | |
925 | } | |
926 | else if (peek_input () == '&') | |
927 | { | |
928 | // logical and | |
929 | skip_input (); | |
930 | current_column += 2; | |
931 | ||
932 | return Token::make (LOGICAL_AND, loc); | |
933 | } | |
934 | else | |
935 | { | |
936 | // bitwise and/reference | |
937 | current_column++; | |
938 | ||
939 | return Token::make (AMP, loc); | |
940 | } | |
941 | case '.': | |
942 | if (peek_input () == '.') | |
943 | { | |
944 | if (peek_input (1) == '.') | |
945 | { | |
946 | // ellipsis | |
947 | skip_input (1); | |
948 | current_column += 3; | |
949 | ||
950 | return Token::make (ELLIPSIS, loc); | |
951 | } | |
952 | else if (peek_input (1) == '=') | |
953 | { | |
954 | // ..= | |
955 | skip_input (1); | |
956 | current_column += 3; | |
957 | ||
958 | return Token::make (DOT_DOT_EQ, loc); | |
959 | } | |
960 | else | |
961 | { | |
962 | // .. | |
963 | skip_input (); | |
964 | current_column += 2; | |
965 | ||
966 | return Token::make (DOT_DOT, loc); | |
967 | } | |
968 | } | |
969 | else /*if (!ISDIGIT (peek_input ()))*/ | |
970 | { | |
971 | // single dot . | |
972 | // Only if followed by a non-number - otherwise is float | |
973 | // nope, float cannot start with '.'. | |
974 | current_column++; | |
975 | return Token::make (DOT, loc); | |
976 | } | |
977 | } | |
978 | // TODO: special handling of _ in the lexer? instead of being identifier | |
979 | ||
980 | // byte character, byte string and raw byte string literals | |
981 | if (current_char == 'b') | |
982 | { | |
983 | if (peek_input () == '\'') | |
984 | return parse_byte_char (loc); | |
985 | else if (peek_input () == '"') | |
986 | return parse_byte_string (loc); | |
987 | else if (peek_input () == 'r' | |
988 | && (peek_input (1) == '#' || peek_input (1) == '"')) | |
989 | return parse_raw_byte_string (loc); | |
990 | } | |
991 | ||
992 | // raw identifiers and raw strings | |
993 | if (current_char == 'r') | |
994 | { | |
995 | int peek = peek_input (); | |
996 | int peek1 = peek_input (1); | |
997 | ||
998 | if (peek == '#' && (ISALPHA (peek1) || peek1 == '_')) | |
999 | { | |
1000 | TokenPtr raw_ident_ptr = parse_raw_identifier (loc); | |
1001 | if (raw_ident_ptr != nullptr) | |
1002 | return raw_ident_ptr; | |
1003 | else | |
1004 | continue; /* input got parsed, it just wasn't valid. An error | |
1005 | was produced. */ | |
1006 | } | |
1007 | else | |
1008 | { | |
1009 | TokenPtr maybe_raw_string_ptr = maybe_parse_raw_string (loc); | |
1010 | if (maybe_raw_string_ptr != nullptr) | |
1011 | return maybe_raw_string_ptr; | |
1012 | } | |
1013 | } | |
1014 | ||
1015 | // find identifiers and keywords | |
1016 | if (ISALPHA (current_char) || current_char == '_') | |
1017 | return parse_identifier_or_keyword (loc); | |
1018 | ||
1019 | // int and float literals | |
1020 | if (ISDIGIT (current_char)) | |
1021 | { // _ not allowed as first char | |
1022 | if (current_char == '0' | |
1023 | && is_non_decimal_int_literal_separator (peek_input ())) | |
1024 | { | |
1025 | // handle binary, octal, hex literals | |
1026 | TokenPtr non_dec_int_lit_ptr | |
1027 | = parse_non_decimal_int_literals (loc); | |
1028 | if (non_dec_int_lit_ptr != nullptr) | |
1029 | return non_dec_int_lit_ptr; | |
1030 | } | |
1031 | else | |
1032 | { | |
1033 | // handle decimals (integer or float) | |
1034 | TokenPtr decimal_or_float_ptr = parse_decimal_int_or_float (loc); | |
1035 | if (decimal_or_float_ptr != nullptr) | |
1036 | return decimal_or_float_ptr; | |
1037 | } | |
1038 | } | |
1039 | ||
1040 | // string literals | |
1041 | if (current_char == '"') | |
1042 | return parse_string (loc); | |
1043 | ||
1044 | // char literals and lifetime names | |
1045 | if (current_char == '\'') | |
1046 | { | |
1047 | TokenPtr char_or_lifetime_ptr = parse_char_or_lifetime (loc); | |
1048 | if (char_or_lifetime_ptr != nullptr) | |
1049 | return char_or_lifetime_ptr; | |
1050 | } | |
1051 | ||
1052 | // DEBUG: check for specific character problems: | |
1053 | if (current_char == '0') | |
1054 | rust_debug ("'0' uncaught before unexpected character"); | |
1055 | else if (current_char == ']') | |
1056 | rust_debug ("']' uncaught before unexpected character"); | |
1057 | else if (current_char == 0x5d) | |
1058 | rust_debug ("whatever 0x5d is (not '0' or ']') uncaught before " | |
1059 | "unexpected character"); | |
1060 | ||
1061 | // didn't match anything so error | |
1062 | rust_error_at (loc, "unexpected character %<%x%>", current_char); | |
1063 | current_column++; | |
1064 | } | |
1065 | } | |
1066 | ||
1067 | // Parses in a type suffix. | |
1068 | std::pair<PrimitiveCoreType, int> | |
1069 | Lexer::parse_in_type_suffix () | |
1070 | { | |
1071 | std::string suffix; | |
1072 | suffix.reserve (5); | |
1073 | ||
1074 | int additional_length_offset = 0; | |
1075 | ||
1076 | // get suffix | |
1077 | while (ISALPHA (current_char) || ISDIGIT (current_char) | |
1078 | || current_char == '_') | |
1079 | { | |
1080 | if (current_char == '_') | |
1081 | { | |
1082 | // don't add _ to suffix | |
1083 | skip_input (); | |
1084 | current_char = peek_input (); | |
1085 | ||
1086 | additional_length_offset++; | |
1087 | ||
1088 | continue; | |
1089 | } | |
1090 | ||
1091 | additional_length_offset++; | |
1092 | ||
1093 | suffix += current_char; | |
1094 | skip_input (); | |
1095 | current_char = peek_input (); | |
1096 | } | |
1097 | ||
1098 | if (suffix.empty ()) | |
1099 | { | |
1100 | // no type suffix: do nothing but also no error | |
1101 | return std::make_pair (CORETYPE_UNKNOWN, additional_length_offset); | |
1102 | } | |
1103 | else if (suffix == "f32") | |
1104 | { | |
1105 | return std::make_pair (CORETYPE_F32, additional_length_offset); | |
1106 | } | |
1107 | else if (suffix == "f64") | |
1108 | { | |
1109 | return std::make_pair (CORETYPE_F64, additional_length_offset); | |
1110 | } | |
1111 | else if (suffix == "i8") | |
1112 | { | |
1113 | return std::make_pair (CORETYPE_I8, additional_length_offset); | |
1114 | } | |
1115 | else if (suffix == "i16") | |
1116 | { | |
1117 | return std::make_pair (CORETYPE_I16, additional_length_offset); | |
1118 | } | |
1119 | else if (suffix == "i32") | |
1120 | { | |
1121 | return std::make_pair (CORETYPE_I32, additional_length_offset); | |
1122 | } | |
1123 | else if (suffix == "i64") | |
1124 | { | |
1125 | return std::make_pair (CORETYPE_I64, additional_length_offset); | |
1126 | } | |
1127 | else if (suffix == "i128") | |
1128 | { | |
1129 | return std::make_pair (CORETYPE_I128, additional_length_offset); | |
1130 | } | |
1131 | else if (suffix == "isize") | |
1132 | { | |
1133 | return std::make_pair (CORETYPE_ISIZE, additional_length_offset); | |
1134 | } | |
1135 | else if (suffix == "u8") | |
1136 | { | |
1137 | return std::make_pair (CORETYPE_U8, additional_length_offset); | |
1138 | } | |
1139 | else if (suffix == "u16") | |
1140 | { | |
1141 | return std::make_pair (CORETYPE_U16, additional_length_offset); | |
1142 | } | |
1143 | else if (suffix == "u32") | |
1144 | { | |
1145 | return std::make_pair (CORETYPE_U32, additional_length_offset); | |
1146 | } | |
1147 | else if (suffix == "u64") | |
1148 | { | |
1149 | return std::make_pair (CORETYPE_U64, additional_length_offset); | |
1150 | } | |
1151 | else if (suffix == "u128") | |
1152 | { | |
1153 | return std::make_pair (CORETYPE_U128, additional_length_offset); | |
1154 | } | |
1155 | else if (suffix == "usize") | |
1156 | { | |
1157 | return std::make_pair (CORETYPE_USIZE, additional_length_offset); | |
1158 | } | |
1159 | else | |
1160 | { | |
1161 | rust_error_at (get_current_location (), "unknown number suffix %qs", | |
1162 | suffix.c_str ()); | |
1163 | ||
1164 | return std::make_pair (CORETYPE_UNKNOWN, additional_length_offset); | |
1165 | } | |
1166 | } | |
1167 | ||
1168 | // Parses in the exponent part (if any) of a float literal. | |
1169 | std::pair<std::string, int> | |
1170 | Lexer::parse_in_exponent_part () | |
1171 | { | |
1172 | int additional_length_offset = 0; | |
1173 | std::string str; | |
1174 | if (current_char == 'E' || current_char == 'e') | |
1175 | { | |
1176 | // add exponent to string as strtod works with it | |
1177 | str += current_char; | |
1178 | skip_input (); | |
1179 | current_char = peek_input (); | |
1180 | ||
1181 | additional_length_offset++; | |
1182 | ||
1183 | // special - and + handling | |
1184 | if (current_char == '-') | |
1185 | { | |
1186 | str += '-'; | |
1187 | ||
1188 | skip_input (); | |
1189 | current_char = peek_input (); | |
1190 | ||
1191 | additional_length_offset++; | |
1192 | } | |
1193 | else if (current_char == '+') | |
1194 | { | |
1195 | // don't add + but still skip input | |
1196 | skip_input (); | |
1197 | current_char = peek_input (); | |
1198 | ||
1199 | additional_length_offset++; | |
1200 | } | |
1201 | ||
1202 | // parse another decimal number for exponent | |
1203 | auto str_length = parse_in_decimal (); | |
1204 | str += std::get<0> (str_length); | |
1205 | additional_length_offset += std::get<1> (str_length); | |
1206 | } | |
1207 | return std::make_pair (str, additional_length_offset); | |
1208 | } | |
1209 | ||
1210 | // Parses a decimal integer. | |
1211 | std::tuple<std::string, int, bool> | |
1212 | Lexer::parse_in_decimal () | |
1213 | { | |
1214 | /* A pure decimal contains only digits. */ | |
1215 | bool pure_decimal = true; | |
1216 | int additional_length_offset = 0; | |
1217 | std::string str; | |
1218 | while (ISDIGIT (current_char) || current_char == '_') | |
1219 | { | |
1220 | if (current_char == '_') | |
1221 | { | |
1222 | pure_decimal = false; | |
1223 | // don't add _ to number | |
1224 | skip_input (); | |
1225 | current_char = peek_input (); | |
1226 | ||
1227 | additional_length_offset++; | |
1228 | ||
1229 | continue; | |
1230 | } | |
1231 | ||
1232 | additional_length_offset++; | |
1233 | ||
1234 | str += current_char; | |
1235 | skip_input (); | |
1236 | current_char = peek_input (); | |
1237 | } | |
1238 | return std::make_tuple (str, additional_length_offset, pure_decimal); | |
1239 | } | |
1240 | ||
1241 | /* Parses escapes (and string continues) in "byte" strings and characters. Does | |
1242 | * not support unicode. */ | |
1243 | std::tuple<char, int, bool> | |
1244 | Lexer::parse_escape (char opening_char) | |
1245 | { | |
1246 | int additional_length_offset = 0; | |
1247 | char output_char = 0; | |
1248 | ||
1249 | // skip to actual letter | |
1250 | skip_input (); | |
1251 | current_char = peek_input (); | |
1252 | additional_length_offset++; | |
1253 | ||
1254 | switch (current_char) | |
1255 | { | |
1256 | case 'x': { | |
1257 | auto hex_escape_pair = parse_partial_hex_escape (); | |
1258 | long hexLong = hex_escape_pair.first; | |
1259 | additional_length_offset += hex_escape_pair.second; | |
1260 | ||
1261 | if (hexLong > 255 || hexLong < 0) | |
1262 | rust_error_at ( | |
1263 | get_current_location (), | |
1264 | "byte \\x escape %<\\x%x%> out of range - allows up to %<\\xFF%>", | |
1265 | static_cast<unsigned int> (hexLong)); | |
1266 | /* TODO: restore capital for escape output - gcc pretty-printer doesn't | |
1267 | * support %X directly */ | |
1268 | char hexChar = static_cast<char> (hexLong); | |
1269 | ||
1270 | output_char = hexChar; | |
1271 | } | |
1272 | break; | |
1273 | case 'n': | |
1274 | output_char = '\n'; | |
1275 | break; | |
1276 | case 'r': | |
1277 | output_char = '\r'; | |
1278 | break; | |
1279 | case 't': | |
1280 | output_char = '\t'; | |
1281 | break; | |
1282 | case '\\': | |
1283 | output_char = '\\'; | |
1284 | break; | |
1285 | case '0': | |
1286 | output_char = '\0'; | |
1287 | break; | |
1288 | case '\'': | |
1289 | output_char = '\''; | |
1290 | break; | |
1291 | case '"': | |
1292 | output_char = '"'; | |
1293 | break; | |
1294 | case 'u': | |
1295 | rust_error_at (get_current_location (), | |
1296 | "cannot have a unicode escape \\u in a byte %s", | |
1297 | opening_char == '\'' ? "character" : "string"); | |
1298 | // Try to parse it anyway, just to skip it | |
1299 | parse_partial_unicode_escape (); | |
1300 | return std::make_tuple (output_char, additional_length_offset, false); | |
1301 | case '\r': | |
1302 | case '\n': | |
1303 | // string continue | |
1304 | return std::make_tuple (0, parse_partial_string_continue (), true); | |
1305 | default: | |
1306 | rust_error_at (get_current_location (), | |
1307 | "unknown escape sequence %<\\%c%>", current_char); | |
1308 | // returns false if no parsing could be done | |
1309 | // return false; | |
1310 | return std::make_tuple (output_char, additional_length_offset, false); | |
1311 | break; | |
1312 | } | |
1313 | // all non-special cases (string continue) should skip their used char | |
1314 | skip_input (); | |
1315 | current_char = peek_input (); | |
1316 | additional_length_offset++; | |
1317 | ||
1318 | // returns true if parsing was successful | |
1319 | // return true; | |
1320 | return std::make_tuple (output_char, additional_length_offset, false); | |
1321 | } | |
1322 | ||
1323 | /* Parses an escape (or string continue) in a string or character. Supports | |
1324 | * unicode escapes. */ | |
1325 | std::tuple<Codepoint, int, bool> | |
1326 | Lexer::parse_utf8_escape (char opening_char) | |
1327 | { | |
1328 | Codepoint output_char; | |
1329 | int additional_length_offset = 0; | |
1330 | ||
1331 | // skip to actual letter | |
1332 | skip_input (); | |
1333 | current_char = peek_input (); | |
1334 | additional_length_offset++; | |
1335 | ||
1336 | switch (current_char) | |
1337 | { | |
1338 | case 'x': { | |
1339 | auto hex_escape_pair = parse_partial_hex_escape (); | |
1340 | long hexLong = hex_escape_pair.first; | |
1341 | additional_length_offset += hex_escape_pair.second; | |
1342 | ||
1343 | if (hexLong > 127 || hexLong < 0) | |
1344 | rust_error_at ( | |
1345 | get_current_location (), | |
1346 | "ascii \\x escape %<\\x%x%> out of range - allows up to %<\\x7F%>", | |
1347 | static_cast<unsigned int> (hexLong)); | |
1348 | /* TODO: restore capital for escape output - gcc pretty-printer doesn't | |
1349 | * support %X directly */ | |
1350 | char hexChar = static_cast<char> (hexLong); | |
1351 | ||
1352 | output_char = hexChar; | |
1353 | } | |
1354 | break; | |
1355 | case 'n': | |
1356 | output_char = '\n'; | |
1357 | break; | |
1358 | case 'r': | |
1359 | output_char = '\r'; | |
1360 | break; | |
1361 | case 't': | |
1362 | output_char = '\t'; | |
1363 | break; | |
1364 | case '\\': | |
1365 | output_char = '\\'; | |
1366 | break; | |
1367 | case '0': | |
1368 | output_char = '\0'; | |
1369 | break; | |
1370 | case '\'': | |
1371 | output_char = '\''; | |
1372 | break; | |
1373 | case '"': | |
1374 | output_char = '"'; | |
1375 | break; | |
1376 | case 'u': { | |
1377 | auto unicode_escape_pair = parse_partial_unicode_escape (); | |
1378 | output_char = unicode_escape_pair.first; | |
1379 | additional_length_offset += unicode_escape_pair.second; | |
1380 | ||
1381 | return std::make_tuple (output_char, additional_length_offset, false); | |
1382 | } | |
1383 | break; | |
1384 | case '\r': | |
1385 | case '\n': | |
1386 | // string continue | |
1387 | return std::make_tuple (0, parse_partial_string_continue (), true); | |
1388 | default: | |
1389 | rust_error_at (get_current_location (), | |
1390 | "unknown escape sequence %<\\%c%>", current_char); | |
1391 | // returns false if no parsing could be done | |
1392 | // return false; | |
1393 | return std::make_tuple (output_char, additional_length_offset, false); | |
1394 | break; | |
1395 | } | |
1396 | /* all non-special cases (unicode, string continue) should skip their used | |
1397 | * char */ | |
1398 | skip_input (); | |
1399 | current_char = peek_input (); | |
1400 | additional_length_offset++; | |
1401 | ||
1402 | // returns true if parsing was successful | |
1403 | // return true; | |
1404 | return std::make_tuple (output_char, additional_length_offset, false); | |
1405 | } | |
1406 | ||
1407 | // Parses the body of a string continue that has been found in an escape. | |
1408 | int | |
1409 | Lexer::parse_partial_string_continue () | |
1410 | { | |
1411 | int additional_length_offset = 1; | |
1412 | ||
1413 | // string continue | |
1414 | while (is_whitespace (current_char)) | |
1415 | { | |
1416 | if (current_char == '\n') | |
1417 | { | |
1418 | current_line++; | |
1419 | current_column = 1; | |
1420 | // tell line_table that new line starts | |
1421 | start_line (current_line, max_column_hint); | |
1422 | ||
1423 | // reset "length" | |
1424 | additional_length_offset = 1; | |
1425 | ||
1426 | // get next char | |
1427 | skip_input (); | |
1428 | current_char = peek_input (); | |
1429 | ||
1430 | continue; | |
1431 | } | |
1432 | ||
1433 | skip_input (); | |
1434 | current_char = peek_input (); | |
1435 | additional_length_offset++; | |
1436 | } | |
1437 | ||
1438 | return additional_length_offset; | |
1439 | } | |
1440 | ||
1441 | /* Parses the body of a '\x' escape. Note that it does not check that the number | |
1442 | * is valid and smaller than 255. */ | |
1443 | std::pair<long, int> | |
1444 | Lexer::parse_partial_hex_escape () | |
1445 | { | |
1446 | // hex char string (null-terminated) | |
1447 | char hexNum[3] = {0, 0, 0}; | |
1448 | ||
1449 | // first hex char | |
1450 | current_char = peek_input (1); | |
1451 | int additional_length_offset = 1; | |
1452 | ||
1453 | if (!is_x_digit (current_char)) | |
1454 | { | |
1455 | rust_error_at (get_current_location (), | |
1456 | "invalid character %<\\x%c%> in \\x sequence", | |
1457 | current_char); | |
1458 | return std::make_pair (0, 0); | |
1459 | } | |
1460 | hexNum[0] = current_char; | |
1461 | ||
1462 | // second hex char | |
1463 | skip_input (); | |
1464 | current_char = peek_input (1); | |
1465 | additional_length_offset++; | |
1466 | ||
1467 | if (!is_x_digit (current_char)) | |
1468 | { | |
1469 | rust_error_at (get_current_location (), | |
1470 | "invalid character %<\\x%c%c%> in \\x sequence", hexNum[0], | |
1471 | current_char); | |
1472 | return std::make_pair (0, 1); | |
1473 | } | |
1474 | skip_input (); | |
1475 | hexNum[1] = current_char; | |
1476 | ||
1477 | long hexLong = std::strtol (hexNum, nullptr, 16); | |
1478 | ||
1479 | return std::make_pair (hexLong, additional_length_offset); | |
1480 | } | |
1481 | ||
1482 | // Parses the body of a unicode escape. | |
1483 | std::pair<Codepoint, int> | |
1484 | Lexer::parse_partial_unicode_escape () | |
1485 | { | |
1486 | skip_input (); | |
1487 | current_char = peek_input (); | |
1488 | int additional_length_offset = 0; | |
1489 | ||
1490 | if (current_char != '{') | |
1491 | { | |
1492 | rust_error_at (get_current_location (), | |
1493 | "unicode escape should start with %<{%>"); | |
1494 | /* Skip what should probaby have been between brackets. */ | |
1495 | while (is_x_digit (current_char) || current_char == '_') | |
1496 | { | |
1497 | skip_input (); | |
1498 | current_char = peek_input (); | |
1499 | additional_length_offset++; | |
1500 | } | |
1501 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1502 | } | |
1503 | ||
1504 | skip_input (); | |
1505 | current_char = peek_input (); | |
1506 | additional_length_offset++; | |
1507 | ||
1508 | if (current_char == '_') | |
1509 | { | |
1510 | rust_error_at (get_current_location (), | |
1511 | "unicode escape cannot start with %<_%>"); | |
1512 | skip_input (); | |
1513 | current_char = peek_input (); | |
1514 | additional_length_offset++; | |
1515 | // fallthrough and try to parse the rest anyway | |
1516 | } | |
1517 | ||
1518 | // parse unicode escape - 1-6 hex digits | |
1519 | std::string num_str; | |
1520 | num_str.reserve (6); | |
1521 | ||
1522 | // loop through to add entire hex number to string | |
1523 | while (is_x_digit (current_char) || current_char == '_') | |
1524 | { | |
1525 | if (current_char == '_') | |
1526 | { | |
1527 | // don't add _ to number | |
1528 | skip_input (); | |
1529 | current_char = peek_input (); | |
1530 | ||
1531 | additional_length_offset++; | |
1532 | ||
1533 | continue; | |
1534 | } | |
1535 | ||
1536 | additional_length_offset++; | |
1537 | ||
1538 | // add raw hex numbers | |
1539 | num_str += current_char; | |
1540 | ||
1541 | skip_input (); | |
1542 | current_char = peek_input (); | |
1543 | } | |
1544 | ||
1545 | if (current_char == '}') | |
1546 | { | |
1547 | skip_input (); | |
1548 | current_char = peek_input (); | |
1549 | additional_length_offset++; | |
1550 | } | |
1551 | else | |
1552 | { | |
1553 | // actually an error, but allow propagation anyway Assume that | |
1554 | // wrong bracketm whitespace or single/double quotes are wrong | |
1555 | // termination, otherwise it is a wrong character, then skip to the actual | |
1556 | // terminator. | |
1557 | if (current_char == '{' || is_whitespace (current_char) | |
1558 | || current_char == '\'' || current_char == '"') | |
1559 | { | |
1560 | rust_error_at (get_current_location (), | |
1561 | "expected terminating %<}%> in unicode escape"); | |
1562 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1563 | } | |
1564 | else | |
1565 | { | |
1566 | rust_error_at (get_current_location (), | |
1567 | "invalid character %<%c%> in unicode escape", | |
1568 | current_char); | |
1569 | while (current_char != '}' && current_char != '{' | |
1570 | && !is_whitespace (current_char) && current_char != '\'' | |
1571 | && current_char != '"') | |
1572 | { | |
1573 | skip_input (); | |
1574 | current_char = peek_input (); | |
1575 | additional_length_offset++; | |
1576 | } | |
1577 | // Consume the actual closing bracket if found | |
1578 | if (current_char == '}') | |
1579 | { | |
1580 | skip_input (); | |
1581 | current_char = peek_input (); | |
1582 | additional_length_offset++; | |
1583 | } | |
1584 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1585 | } | |
1586 | } | |
1587 | ||
1588 | // ensure 1-6 hex characters | |
1589 | if (num_str.length () > 6 || num_str.length () < 1) | |
1590 | { | |
1591 | rust_error_at (get_current_location (), | |
1592 | "unicode escape should be between 1 and 6 hex " | |
1593 | "characters; it is %lu", | |
1594 | (unsigned long) num_str.length ()); | |
1595 | // return false; | |
1596 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1597 | } | |
1598 | ||
1599 | unsigned long hex_num = std::strtoul (num_str.c_str (), nullptr, 16); | |
1600 | ||
1601 | if (hex_num > 0xd7ff && hex_num < 0xe000) | |
1602 | { | |
1603 | rust_error_at ( | |
1604 | get_current_location (), | |
1605 | "unicode escape cannot be a surrogate value (D800 to DFFF)"); | |
1606 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1607 | } | |
1608 | ||
1609 | if (hex_num > 0x10ffff) | |
1610 | { | |
1611 | rust_error_at (get_current_location (), | |
1612 | "unicode escape cannot be larger than 10FFFF"); | |
1613 | return std::make_pair (Codepoint (0), additional_length_offset); | |
1614 | } | |
1615 | ||
1616 | // return true; | |
1617 | return std::make_pair (Codepoint (static_cast<uint32_t> (hex_num)), | |
1618 | additional_length_offset); | |
1619 | } | |
1620 | ||
1621 | // Parses a byte character. | |
1622 | TokenPtr | |
1623 | Lexer::parse_byte_char (Location loc) | |
1624 | { | |
1625 | skip_input (); | |
1626 | current_column++; | |
1627 | // make current char the next character | |
1628 | current_char = peek_input (); | |
1629 | ||
1630 | int length = 1; | |
1631 | ||
1632 | // char to save | |
1633 | char byte_char = 0; | |
1634 | ||
1635 | // detect escapes | |
1636 | if (current_char == '\\') | |
1637 | { | |
1638 | auto escape_length_pair = parse_escape ('\''); | |
1639 | byte_char = std::get<0> (escape_length_pair); | |
1640 | length += std::get<1> (escape_length_pair); | |
1641 | ||
1642 | current_char = peek_input (); | |
1643 | ||
1644 | if (current_char != '\'') | |
1645 | { | |
1646 | rust_error_at (get_current_location (), "unclosed %<byte char%>"); | |
1647 | } | |
1648 | ||
1649 | skip_input (); | |
1650 | current_char = peek_input (); | |
1651 | length++; // go to next char | |
1652 | } | |
1653 | else if (current_char != '\'') | |
1654 | { | |
1655 | // otherwise, get character from direct input character | |
1656 | byte_char = current_char; | |
1657 | ||
1658 | skip_input (); | |
1659 | current_char = peek_input (); | |
1660 | length++; | |
1661 | ||
1662 | if (current_char != '\'') | |
1663 | { | |
1664 | rust_error_at (get_current_location (), "unclosed %<byte char%>"); | |
1665 | } | |
1666 | ||
1667 | skip_input (); | |
1668 | current_char = peek_input (); | |
1669 | length++; // go to next char | |
1670 | } | |
1671 | else | |
1672 | { | |
1673 | rust_error_at (get_current_location (), | |
1674 | "no character inside %<%> for %<byte char%>"); | |
1675 | } | |
1676 | ||
1677 | current_column += length; | |
1678 | ||
1679 | return Token::make_byte_char (loc, byte_char); | |
1680 | } | |
1681 | ||
1682 | // Parses a byte string. | |
1683 | TokenPtr | |
1684 | Lexer::parse_byte_string (Location loc) | |
1685 | { | |
1686 | // byte string | |
1687 | ||
1688 | // skip quote character | |
1689 | skip_input (); | |
1690 | current_column++; | |
1691 | ||
1692 | std::string str; | |
1693 | str.reserve (16); // some sensible default | |
1694 | ||
1695 | int length = 1; | |
1696 | current_char = peek_input (); | |
1697 | ||
1698 | while (current_char != '"' && current_char != EOF) | |
1699 | { | |
1700 | if (current_char == '\\') | |
1701 | { | |
1702 | auto escape_length_pair = parse_escape ('"'); | |
1703 | char output_char = std::get<0> (escape_length_pair); | |
1704 | ||
1705 | if (output_char == 0 && std::get<2> (escape_length_pair)) | |
1706 | length = std::get<1> (escape_length_pair) - 1; | |
1707 | else | |
1708 | length += std::get<1> (escape_length_pair); | |
1709 | ||
1710 | if (output_char != 0 || !std::get<2> (escape_length_pair)) | |
1711 | str += output_char; | |
1712 | ||
1713 | continue; | |
1714 | } | |
1715 | ||
1716 | length++; | |
1717 | ||
1718 | str += current_char; | |
1719 | skip_input (); | |
1720 | current_char = peek_input (); | |
1721 | } | |
1722 | ||
1723 | current_column += length; | |
1724 | ||
1725 | if (current_char == '"') | |
1726 | { | |
1727 | current_column++; | |
1728 | ||
1729 | skip_input (); | |
1730 | current_char = peek_input (); | |
1731 | } | |
1732 | else if (current_char == EOF) | |
1733 | { | |
1734 | rust_error_at (get_current_location (), "unended byte string literal"); | |
1735 | return Token::make (END_OF_FILE, get_current_location ()); | |
1736 | } | |
1737 | else | |
1738 | { | |
1739 | gcc_unreachable (); | |
1740 | } | |
1741 | ||
1742 | str.shrink_to_fit (); | |
1743 | ||
1744 | return Token::make_byte_string (loc, std::move (str)); | |
1745 | } | |
1746 | ||
1747 | // Parses a raw byte string. | |
1748 | TokenPtr | |
1749 | Lexer::parse_raw_byte_string (Location loc) | |
1750 | { | |
1751 | // raw byte string literals | |
1752 | std::string str; | |
1753 | str.reserve (16); // some sensible default | |
1754 | ||
1755 | int length = 1; | |
1756 | int hash_count = 0; | |
1757 | ||
1758 | // get hash count at beginnning | |
1759 | skip_input (); | |
1760 | current_char = peek_input (); | |
1761 | length++; | |
1762 | while (current_char == '#') | |
1763 | { | |
1764 | hash_count++; | |
1765 | length++; | |
1766 | ||
1767 | skip_input (); | |
1768 | current_char = peek_input (); | |
1769 | } | |
1770 | ||
1771 | if (current_char != '"') | |
1772 | { | |
1773 | rust_error_at (get_current_location (), | |
1774 | "raw byte string has no opening %<\"%>"); | |
1775 | } | |
1776 | ||
1777 | skip_input (); | |
1778 | current_char = peek_input (); | |
1779 | length++; | |
1780 | ||
1781 | while (true) | |
1782 | { | |
1783 | if (current_char == '"') | |
1784 | { | |
1785 | bool enough_hashes = true; | |
1786 | ||
1787 | for (int i = 0; i < hash_count; i++) | |
1788 | { | |
1789 | if (peek_input (i + 1) != '#') | |
1790 | { | |
1791 | enough_hashes = false; | |
1792 | break; | |
1793 | } | |
1794 | } | |
1795 | ||
1796 | if (enough_hashes) | |
1797 | { | |
1798 | // skip enough input and peek enough input | |
1799 | skip_input (hash_count); | |
1800 | current_char = peek_input (); | |
1801 | length += hash_count + 1; | |
1802 | break; | |
1803 | } | |
1804 | } | |
1805 | ||
1806 | if ((unsigned char) current_char > 127) | |
1807 | { | |
1808 | rust_error_at (get_current_location (), | |
1809 | "character %<%c%> in raw byte string out of range", | |
1810 | current_char); | |
1811 | current_char = 0; | |
1812 | } | |
1813 | ||
1814 | length++; | |
1815 | ||
1816 | str += current_char; | |
1817 | skip_input (); | |
1818 | current_char = peek_input (); | |
1819 | } | |
1820 | ||
1821 | current_column += length; | |
1822 | ||
1823 | str.shrink_to_fit (); | |
1824 | ||
1825 | return Token::make_byte_string (loc, std::move (str)); | |
1826 | } | |
1827 | ||
1828 | // Parses a raw identifier. | |
1829 | TokenPtr | |
1830 | Lexer::parse_raw_identifier (Location loc) | |
1831 | { | |
1832 | // raw identifier | |
1833 | std::string str; | |
1834 | str.reserve (16); // default | |
1835 | ||
1836 | skip_input (); | |
1837 | current_char = peek_input (); | |
1838 | ||
1839 | current_column += 2; | |
1840 | ||
1841 | bool first_is_underscore = current_char == '_'; | |
1842 | ||
1843 | int length = 0; | |
1844 | current_char = peek_input (); | |
1845 | // loop through entire name | |
1846 | while (ISALPHA (current_char) || ISDIGIT (current_char) | |
1847 | || current_char == '_') | |
1848 | { | |
1849 | length++; | |
1850 | ||
1851 | str += current_char; | |
1852 | skip_input (); | |
1853 | current_char = peek_input (); | |
1854 | } | |
1855 | ||
1856 | current_column += length; | |
1857 | ||
1858 | // if just a single underscore, not an identifier | |
1859 | if (first_is_underscore && length == 1) | |
1860 | rust_error_at (get_current_location (), | |
1861 | "%<_%> is not a valid raw identifier"); | |
1862 | ||
1863 | if (str == "crate" || str == "extern" || str == "self" || str == "super" | |
1864 | || str == "Self") | |
1865 | { | |
1866 | rust_error_at (get_current_location (), | |
1867 | "%qs is a forbidden raw identifier", str.c_str ()); | |
1868 | ||
1869 | return nullptr; | |
1870 | } | |
1871 | else | |
1872 | { | |
1873 | str.shrink_to_fit (); | |
1874 | ||
1875 | return Token::make_identifier (loc, std::move (str)); | |
1876 | } | |
1877 | } | |
1878 | ||
1879 | // skip broken string input (unterminated strings) | |
1880 | void | |
1881 | Lexer::skip_broken_string_input (int current_char) | |
1882 | { | |
1883 | while (current_char != '"' && current_char != EOF) | |
1884 | { | |
1885 | if (current_char == '\n') | |
1886 | { | |
1887 | current_line++; | |
1888 | current_column = 1; | |
1889 | } | |
1890 | else | |
1891 | { | |
1892 | current_column++; | |
1893 | } | |
1894 | skip_input (); | |
1895 | current_char = peek_input (); | |
1896 | } | |
1897 | if (current_char == '"') | |
1898 | { | |
1899 | current_column++; | |
1900 | ||
1901 | skip_input (); | |
1902 | current_char = peek_input (); | |
1903 | } | |
1904 | rust_debug ("skipped to %d:%d due to bad quotes", current_line, | |
1905 | current_column); | |
1906 | } | |
1907 | ||
1908 | // Parses a unicode string. | |
1909 | TokenPtr | |
1910 | Lexer::parse_string (Location loc) | |
1911 | { | |
1912 | Codepoint current_char32; | |
1913 | ||
1914 | std::string str; | |
1915 | str.reserve (16); // some sensible default | |
1916 | ||
1917 | int length = 1; | |
1918 | current_char32 = peek_codepoint_input (); | |
1919 | ||
1920 | // FIXME: This fails if the input ends. How do we check for EOF? | |
1921 | while (current_char32.value != '"' && !current_char32.is_eof ()) | |
1922 | { | |
1923 | if (current_char32.value == '\\') | |
1924 | { | |
1925 | // parse escape | |
1926 | auto utf8_escape_pair = parse_utf8_escape ('\''); | |
1927 | current_char32 = std::get<0> (utf8_escape_pair); | |
1928 | ||
1929 | if (current_char32 == Codepoint (0) && std::get<2> (utf8_escape_pair)) | |
1930 | length = std::get<1> (utf8_escape_pair) - 1; | |
1931 | else | |
1932 | length += std::get<1> (utf8_escape_pair); | |
1933 | ||
1934 | if (current_char32 != Codepoint (0) | |
1935 | || !std::get<2> (utf8_escape_pair)) | |
1936 | str += current_char32; | |
1937 | ||
1938 | // required as parsing utf8 escape only changes current_char | |
1939 | current_char32 = peek_codepoint_input (); | |
1940 | ||
1941 | continue; | |
1942 | } | |
1943 | ||
1944 | length += get_input_codepoint_length (); | |
1945 | ||
1946 | str += current_char32; | |
1947 | skip_codepoint_input (); | |
1948 | current_char32 = peek_codepoint_input (); | |
1949 | } | |
1950 | ||
1951 | current_column += length; | |
1952 | ||
1953 | if (current_char32.value == '"') | |
1954 | { | |
1955 | current_column++; | |
1956 | ||
1957 | skip_input (); | |
1958 | current_char = peek_input (); | |
1959 | } | |
1960 | else if (current_char32.is_eof ()) | |
1961 | { | |
1962 | rust_error_at (get_current_location (), "unended string literal"); | |
1963 | return Token::make (END_OF_FILE, get_current_location ()); | |
1964 | } | |
1965 | else | |
1966 | { | |
1967 | gcc_unreachable (); | |
1968 | } | |
1969 | ||
1970 | str.shrink_to_fit (); | |
1971 | return Token::make_string (loc, std::move (str)); | |
1972 | } | |
1973 | ||
1974 | // Parses an identifier or keyword. | |
1975 | TokenPtr | |
1976 | Lexer::parse_identifier_or_keyword (Location loc) | |
1977 | { | |
1978 | std::string str; | |
1979 | str.reserve (16); // default | |
1980 | str += current_char; | |
1981 | ||
1982 | bool first_is_underscore = current_char == '_'; | |
1983 | ||
1984 | int length = 1; | |
1985 | current_char = peek_input (); | |
1986 | // loop through entire name | |
1987 | while (ISALPHA (current_char) || ISDIGIT (current_char) | |
1988 | || current_char == '_') | |
1989 | { | |
1990 | length++; | |
1991 | ||
1992 | str += current_char; | |
1993 | skip_input (); | |
1994 | current_char = peek_input (); | |
1995 | } | |
1996 | ||
1997 | current_column += length; | |
1998 | ||
1999 | // if just a single underscore, not an identifier | |
2000 | if (first_is_underscore && length == 1) | |
2001 | return Token::make (UNDERSCORE, loc); | |
2002 | ||
2003 | str.shrink_to_fit (); | |
2004 | ||
2005 | TokenId keyword = classify_keyword (str); | |
2006 | if (keyword == IDENTIFIER) | |
2007 | return Token::make_identifier (loc, std::move (str)); | |
2008 | else | |
2009 | return Token::make (keyword, loc); | |
2010 | } | |
2011 | ||
2012 | // Possibly returns a raw string token if it exists - otherwise returns null. | |
2013 | TokenPtr | |
2014 | Lexer::maybe_parse_raw_string (Location loc) | |
2015 | { | |
2016 | int peek_index = 0; | |
2017 | while (peek_input (peek_index) == '#') | |
2018 | peek_index++; | |
2019 | ||
2020 | if (peek_input (peek_index) == '"') | |
2021 | return parse_raw_string (loc, peek_index); | |
2022 | else | |
2023 | return nullptr; | |
2024 | } | |
2025 | ||
2026 | // Returns a raw string token. | |
2027 | TokenPtr | |
2028 | Lexer::parse_raw_string (Location loc, int initial_hash_count) | |
2029 | { | |
2030 | // raw string literals | |
2031 | std::string str; | |
2032 | str.reserve (16); // some sensible default | |
2033 | ||
2034 | int length = 1 + initial_hash_count; | |
2035 | ||
2036 | if (initial_hash_count > 0) | |
2037 | skip_input (initial_hash_count - 1); | |
2038 | ||
2039 | current_char = peek_input (); | |
2040 | ||
2041 | if (current_char != '"') | |
2042 | rust_error_at (get_current_location (), "raw string has no opening %<\"%>"); | |
2043 | ||
2044 | length++; | |
2045 | skip_input (); | |
2046 | Codepoint current_char32 = peek_codepoint_input (); | |
2047 | ||
2048 | while (!current_char32.is_eof ()) | |
2049 | { | |
2050 | if (current_char32.value == '"') | |
2051 | { | |
2052 | bool enough_hashes = true; | |
2053 | ||
2054 | for (int i = 0; i < initial_hash_count; i++) | |
2055 | { | |
2056 | if (peek_input (i + 1) != '#') | |
2057 | { | |
2058 | enough_hashes = false; | |
2059 | break; | |
2060 | } | |
2061 | } | |
2062 | ||
2063 | if (enough_hashes) | |
2064 | { | |
2065 | // skip enough input and peek enough input | |
2066 | skip_input (initial_hash_count); | |
2067 | current_char = peek_input (); | |
2068 | length += initial_hash_count + 1; | |
2069 | break; | |
2070 | } | |
2071 | } | |
2072 | ||
2073 | length++; | |
2074 | ||
2075 | str += current_char32; | |
2076 | skip_codepoint_input (); | |
2077 | current_char32 = peek_codepoint_input (); | |
2078 | } | |
2079 | ||
2080 | current_column += length; | |
2081 | ||
2082 | str.shrink_to_fit (); | |
2083 | ||
2084 | return Token::make_string (loc, std::move (str)); | |
2085 | } | |
2086 | ||
2087 | template <typename IsDigitFunc> | |
2088 | TokenPtr | |
2089 | Lexer::parse_non_decimal_int_literal (Location loc, IsDigitFunc is_digit_func, | |
2090 | std::string existent_str, int base) | |
2091 | { | |
2092 | int length = 1; | |
2093 | ||
2094 | skip_input (); | |
2095 | current_char = peek_input (); | |
2096 | ||
2097 | length++; | |
2098 | ||
2099 | // loop through to add entire number to string | |
2100 | while (is_digit_func (current_char) || current_char == '_') | |
2101 | { | |
2102 | if (current_char == '_') | |
2103 | { | |
2104 | // don't add _ to number | |
2105 | skip_input (); | |
2106 | current_char = peek_input (); | |
2107 | ||
2108 | length++; | |
2109 | ||
2110 | continue; | |
2111 | } | |
2112 | ||
2113 | length++; | |
2114 | ||
2115 | // add raw numbers | |
2116 | existent_str += current_char; | |
2117 | skip_input (); | |
2118 | current_char = peek_input (); | |
2119 | } | |
2120 | ||
2121 | // convert value to decimal representation | |
2122 | long dec_num = std::strtol (existent_str.c_str (), nullptr, base); | |
2123 | ||
2124 | existent_str = std::to_string (dec_num); | |
2125 | ||
2126 | // parse in type suffix if it exists | |
2127 | auto type_suffix_pair = parse_in_type_suffix (); | |
2128 | PrimitiveCoreType type_hint = type_suffix_pair.first; | |
2129 | length += type_suffix_pair.second; | |
2130 | ||
2131 | current_column += length; | |
2132 | ||
2133 | if (type_hint == CORETYPE_F32 || type_hint == CORETYPE_F64) | |
2134 | { | |
2135 | rust_error_at (get_current_location (), | |
2136 | "invalid type suffix %qs for integer (%s) literal", | |
2137 | get_type_hint_string (type_hint), | |
2138 | base == 16 | |
2139 | ? "hex" | |
2140 | : (base == 8 ? "octal" | |
2141 | : (base == 2 ? "binary" | |
2142 | : "<insert unknown base>"))); | |
2143 | return nullptr; | |
2144 | } | |
2145 | return Token::make_int (loc, std::move (existent_str), type_hint); | |
2146 | } | |
2147 | ||
2148 | // Parses a hex, binary or octal int literal. | |
2149 | TokenPtr | |
2150 | Lexer::parse_non_decimal_int_literals (Location loc) | |
2151 | { | |
2152 | std::string str; | |
2153 | str.reserve (16); // some sensible default | |
2154 | str += current_char; | |
2155 | ||
2156 | current_char = peek_input (); | |
2157 | ||
2158 | if (current_char == 'x') | |
2159 | { | |
2160 | // hex (integer only) | |
2161 | return parse_non_decimal_int_literal (loc, is_x_digit, str + "x", 16); | |
2162 | } | |
2163 | else if (current_char == 'o') | |
2164 | { | |
2165 | // octal (integer only) | |
2166 | return parse_non_decimal_int_literal (loc, is_octal_digit, | |
2167 | std::move (str), 8); | |
2168 | } | |
2169 | else if (current_char == 'b') | |
2170 | { | |
2171 | // binary (integer only) | |
2172 | return parse_non_decimal_int_literal (loc, is_bin_digit, std::move (str), | |
2173 | 2); | |
2174 | } | |
2175 | else | |
2176 | { | |
2177 | return nullptr; | |
2178 | } | |
2179 | } | |
2180 | ||
2181 | // Parses a decimal-based int literal or float literal. | |
2182 | TokenPtr | |
2183 | Lexer::parse_decimal_int_or_float (Location loc) | |
2184 | { | |
2185 | std::string str; | |
2186 | str.reserve (16); // some sensible default | |
2187 | str += current_char; | |
2188 | ||
2189 | int length = 1; | |
2190 | bool first_zero = current_char == '0'; | |
2191 | ||
2192 | current_char = peek_input (); | |
2193 | ||
2194 | // parse initial decimal integer (or first integer part of float) literal | |
2195 | auto initial_decimal = parse_in_decimal (); | |
2196 | str += std::get<0> (initial_decimal); | |
2197 | length += std::get<1> (initial_decimal); | |
2198 | ||
2199 | // detect float literal | |
2200 | if (current_char == '.' && is_float_digit (peek_input (1))) | |
2201 | { | |
2202 | // float with a '.', parse another decimal into it | |
2203 | ||
2204 | // add . to str | |
2205 | str += current_char; | |
2206 | skip_input (); | |
2207 | current_char = peek_input (); | |
2208 | length++; | |
2209 | ||
2210 | // parse another decimal number for float | |
2211 | auto second_decimal = parse_in_decimal (); | |
2212 | str += std::get<0> (second_decimal); | |
2213 | length += std::get<1> (second_decimal); | |
2214 | ||
2215 | // parse in exponent part if it exists | |
2216 | auto exponent_pair = parse_in_exponent_part (); | |
2217 | str += exponent_pair.first; | |
2218 | length += exponent_pair.second; | |
2219 | ||
2220 | // parse in type suffix if it exists | |
2221 | auto type_suffix_pair = parse_in_type_suffix (); | |
2222 | PrimitiveCoreType type_hint = type_suffix_pair.first; | |
2223 | length += type_suffix_pair.second; | |
2224 | ||
2225 | if (type_hint != CORETYPE_F32 && type_hint != CORETYPE_F64 | |
2226 | && type_hint != CORETYPE_UNKNOWN) | |
2227 | { | |
2228 | rust_error_at (get_current_location (), | |
2229 | "invalid type suffix %qs for floating-point literal", | |
2230 | get_type_hint_string (type_hint)); | |
2231 | // ignore invalid type suffix as everything else seems fine | |
2232 | type_hint = CORETYPE_UNKNOWN; | |
2233 | } | |
2234 | ||
2235 | current_column += length; | |
2236 | ||
2237 | str.shrink_to_fit (); | |
2238 | return Token::make_float (loc, std::move (str), type_hint); | |
2239 | } | |
2240 | else if (current_char == '.' && check_valid_float_dot_end (peek_input (1))) | |
2241 | { | |
2242 | // float that is just an integer with a terminating '.' character | |
2243 | ||
2244 | // add . to str | |
2245 | str += current_char; | |
2246 | skip_input (); | |
2247 | current_char = peek_input (); | |
2248 | length++; | |
2249 | ||
2250 | // add a '0' after the . to prevent ambiguity | |
2251 | str += '0'; | |
2252 | ||
2253 | // type hint not allowed | |
2254 | ||
2255 | current_column += length; | |
2256 | ||
2257 | str.shrink_to_fit (); | |
2258 | return Token::make_float (loc, std::move (str), CORETYPE_UNKNOWN); | |
2259 | } | |
2260 | else if (current_char == 'E' || current_char == 'e') | |
2261 | { | |
2262 | // exponent float with no '.' character | |
2263 | ||
2264 | // parse exponent part | |
2265 | auto exponent_pair = parse_in_exponent_part (); | |
2266 | str += exponent_pair.first; | |
2267 | length += exponent_pair.second; | |
2268 | ||
2269 | // parse in type suffix if it exists | |
2270 | auto type_suffix_pair = parse_in_type_suffix (); | |
2271 | PrimitiveCoreType type_hint = type_suffix_pair.first; | |
2272 | length += type_suffix_pair.second; | |
2273 | ||
2274 | if (type_hint != CORETYPE_F32 && type_hint != CORETYPE_F64 | |
2275 | && type_hint != CORETYPE_UNKNOWN) | |
2276 | { | |
2277 | rust_error_at (get_current_location (), | |
2278 | "invalid type suffix %qs for floating-point literal", | |
2279 | get_type_hint_string (type_hint)); | |
2280 | // ignore invalid type suffix as everything else seems fine | |
2281 | type_hint = CORETYPE_UNKNOWN; | |
2282 | } | |
2283 | ||
2284 | current_column += length; | |
2285 | ||
2286 | str.shrink_to_fit (); | |
2287 | return Token::make_float (loc, std::move (str), type_hint); | |
2288 | } | |
2289 | else | |
2290 | { | |
2291 | // is an integer | |
2292 | ||
2293 | // parse in type suffix if it exists | |
2294 | auto type_suffix_pair = parse_in_type_suffix (); | |
2295 | PrimitiveCoreType type_hint = type_suffix_pair.first; | |
2296 | /* A "real" pure decimal doesn't have a suffix and no zero prefix. */ | |
2297 | if (type_hint == CORETYPE_UNKNOWN) | |
2298 | { | |
2299 | bool pure_decimal = std::get<2> (initial_decimal); | |
2300 | if (pure_decimal && (!first_zero || str.size () == 1)) | |
2301 | type_hint = CORETYPE_PURE_DECIMAL; | |
2302 | } | |
2303 | length += type_suffix_pair.second; | |
2304 | ||
2305 | current_column += length; | |
2306 | ||
2307 | str.shrink_to_fit (); | |
2308 | return Token::make_int (loc, std::move (str), type_hint); | |
2309 | } | |
2310 | } | |
2311 | ||
2312 | TokenPtr | |
2313 | Lexer::parse_char_or_lifetime (Location loc) | |
2314 | { | |
2315 | Codepoint current_char32; | |
2316 | ||
2317 | int length = 1; | |
2318 | ||
2319 | current_char32 = peek_codepoint_input (); | |
2320 | if (current_char32.is_eof ()) | |
2321 | return nullptr; | |
2322 | ||
2323 | // parse escaped char literal | |
2324 | if (current_char32.value == '\\') | |
2325 | { | |
2326 | // parse escape | |
2327 | auto utf8_escape_pair = parse_utf8_escape ('\''); | |
2328 | current_char32 = std::get<0> (utf8_escape_pair); | |
2329 | length += std::get<1> (utf8_escape_pair); | |
2330 | ||
2331 | if (peek_codepoint_input ().value != '\'') | |
2332 | { | |
2333 | rust_error_at (get_current_location (), "unended character literal"); | |
2334 | } | |
2335 | else | |
2336 | { | |
2337 | skip_codepoint_input (); | |
2338 | current_char = peek_input (); | |
2339 | length++; | |
2340 | } | |
2341 | ||
2342 | current_column += length; | |
2343 | ||
2344 | return Token::make_char (loc, current_char32); | |
2345 | } | |
2346 | else | |
2347 | { | |
2348 | skip_codepoint_input (); | |
2349 | ||
2350 | if (peek_codepoint_input ().value == '\'') | |
2351 | { | |
2352 | // parse non-escaped char literal | |
2353 | ||
2354 | // skip the ' character | |
2355 | skip_input (); | |
2356 | current_char = peek_input (); | |
2357 | ||
2358 | // TODO fix due to different widths of utf-8 chars? | |
2359 | current_column += 3; | |
2360 | ||
2361 | return Token::make_char (loc, current_char32); | |
2362 | } | |
2363 | else if (ISDIGIT (current_char32.value) || ISALPHA (current_char32.value) | |
2364 | || current_char32.value == '_') | |
2365 | { | |
2366 | // parse lifetime name | |
2367 | std::string str; | |
2368 | str += current_char32; | |
2369 | length++; | |
2370 | ||
2371 | current_char = peek_input (); | |
2372 | while (ISDIGIT (current_char) || ISALPHA (current_char) | |
2373 | || current_char == '_') | |
2374 | { | |
2375 | str += current_char; | |
2376 | skip_input (); | |
2377 | current_char = peek_input (); | |
2378 | length++; | |
2379 | } | |
2380 | ||
2381 | current_column += length; | |
2382 | ||
2383 | str.shrink_to_fit (); | |
2384 | return Token::make_lifetime (loc, std::move (str)); | |
2385 | } | |
2386 | else | |
2387 | { | |
2388 | rust_error_at ( | |
2389 | get_current_location (), | |
2390 | "expected %' after character constant in character literal"); | |
2391 | return nullptr; | |
2392 | } | |
2393 | } | |
2394 | } | |
2395 | ||
2396 | // Returns the length of the codepoint at the current position. | |
2397 | int | |
2398 | Lexer::get_input_codepoint_length () | |
2399 | { | |
2400 | uint8_t input = peek_input (); | |
2401 | ||
2402 | if ((int8_t) input == EOF) | |
2403 | return 0; | |
2404 | ||
2405 | if (input < 128) | |
2406 | { | |
2407 | // ascii -- 1 byte | |
2408 | // return input; | |
2409 | ||
2410 | return 1; | |
2411 | } | |
2412 | else if ((input & 0xC0) == 0x80) | |
2413 | { | |
2414 | // invalid (continuation; can't be first char) | |
2415 | // return 0xFFFE; | |
2416 | ||
2417 | return 0; | |
2418 | } | |
2419 | else if ((input & 0xE0) == 0xC0) | |
2420 | { | |
2421 | // 2 bytes | |
2422 | uint8_t input2 = peek_input (1); | |
2423 | if ((input2 & 0xC0) != 0x80) | |
2424 | return 0; | |
2425 | // return 0xFFFE; | |
2426 | ||
2427 | // uint32_t output = ((input & 0x1F) << 6) | ((input2 & 0x3F) << 0); | |
2428 | // return output; | |
2429 | return 2; | |
2430 | } | |
2431 | else if ((input & 0xF0) == 0xE0) | |
2432 | { | |
2433 | // 3 bytes | |
2434 | uint8_t input2 = peek_input (1); | |
2435 | if ((input2 & 0xC0) != 0x80) | |
2436 | return 0; | |
2437 | // return 0xFFFE; | |
2438 | ||
2439 | uint8_t input3 = peek_input (2); | |
2440 | if ((input3 & 0xC0) != 0x80) | |
2441 | return 0; | |
2442 | // return 0xFFFE; | |
2443 | ||
2444 | /*uint32_t output | |
2445 | = ((input & 0x0F) << 12) | ((input2 & 0x3F) << 6) | ((input3 & 0x3F) << | |
2446 | 0); return output;*/ | |
2447 | return 3; | |
2448 | } | |
2449 | else if ((input & 0xF8) == 0xF0) | |
2450 | { | |
2451 | // 4 bytes | |
2452 | uint8_t input2 = peek_input (1); | |
2453 | if ((input2 & 0xC0) != 0x80) | |
2454 | return 0; | |
2455 | // return 0xFFFE; | |
2456 | ||
2457 | uint8_t input3 = peek_input (2); | |
2458 | if ((input3 & 0xC0) != 0x80) | |
2459 | return 0; | |
2460 | // return 0xFFFE; | |
2461 | ||
2462 | uint8_t input4 = peek_input (3); | |
2463 | if ((input4 & 0xC0) != 0x80) | |
2464 | return 0; | |
2465 | // return 0xFFFE; | |
2466 | ||
2467 | /*uint32_t output = ((input & 0x07) << 18) | ((input2 & 0x3F) << 12) | |
2468 | | ((input3 & 0x3F) << 6) | ((input4 & 0x3F) << 0); | |
2469 | return output;*/ | |
2470 | return 4; | |
2471 | } | |
2472 | else | |
2473 | { | |
2474 | rust_error_at (get_current_location (), | |
2475 | "invalid UTF-8 [FIRST] (too long)"); | |
2476 | return 0; | |
2477 | } | |
2478 | } | |
2479 | ||
2480 | // Returns the codepoint at the current position. | |
2481 | Codepoint | |
2482 | Lexer::peek_codepoint_input () | |
2483 | { | |
2484 | uint8_t input = peek_input (); | |
2485 | ||
2486 | if ((int8_t) input == EOF) | |
2487 | return Codepoint::eof (); | |
2488 | ||
2489 | if (input < 128) | |
2490 | { | |
2491 | // ascii -- 1 byte | |
2492 | return {input}; | |
2493 | } | |
2494 | else if ((input & 0xC0) == 0x80) | |
2495 | { | |
2496 | // invalid (continuation; can't be first char) | |
2497 | return {0xFFFE}; | |
2498 | } | |
2499 | else if ((input & 0xE0) == 0xC0) | |
2500 | { | |
2501 | // 2 bytes | |
2502 | uint8_t input2 = peek_input (1); | |
2503 | if ((input2 & 0xC0) != 0x80) | |
2504 | return {0xFFFE}; | |
2505 | ||
2506 | uint32_t output = ((input & 0x1F) << 6) | ((input2 & 0x3F) << 0); | |
2507 | return {output}; | |
2508 | } | |
2509 | else if ((input & 0xF0) == 0xE0) | |
2510 | { | |
2511 | // 3 bytes | |
2512 | uint8_t input2 = peek_input (1); | |
2513 | if ((input2 & 0xC0) != 0x80) | |
2514 | return {0xFFFE}; | |
2515 | ||
2516 | uint8_t input3 = peek_input (2); | |
2517 | if ((input3 & 0xC0) != 0x80) | |
2518 | return {0xFFFE}; | |
2519 | ||
2520 | uint32_t output = ((input & 0x0F) << 12) | ((input2 & 0x3F) << 6) | |
2521 | | ((input3 & 0x3F) << 0); | |
2522 | return {output}; | |
2523 | } | |
2524 | else if ((input & 0xF8) == 0xF0) | |
2525 | { | |
2526 | // 4 bytes | |
2527 | uint8_t input2 = peek_input (1); | |
2528 | if ((input2 & 0xC0) != 0x80) | |
2529 | return {0xFFFE}; | |
2530 | ||
2531 | uint8_t input3 = peek_input (2); | |
2532 | if ((input3 & 0xC0) != 0x80) | |
2533 | return {0xFFFE}; | |
2534 | ||
2535 | uint8_t input4 = peek_input (3); | |
2536 | if ((input4 & 0xC0) != 0x80) | |
2537 | return {0xFFFE}; | |
2538 | ||
2539 | uint32_t output = ((input & 0x07) << 18) | ((input2 & 0x3F) << 12) | |
2540 | | ((input3 & 0x3F) << 6) | ((input4 & 0x3F) << 0); | |
2541 | return {output}; | |
2542 | } | |
2543 | else | |
2544 | { | |
2545 | rust_error_at (get_current_location (), | |
2546 | "invalid UTF-8 [SECND] (too long)"); | |
2547 | return {0xFFFE}; | |
2548 | } | |
2549 | } | |
2550 | ||
2551 | void | |
2552 | Lexer::skip_codepoint_input () | |
2553 | { | |
2554 | int toSkip = get_input_codepoint_length (); | |
2555 | gcc_assert (toSkip >= 1); | |
2556 | ||
2557 | skip_input (toSkip - 1); | |
2558 | } | |
2559 | ||
2560 | int | |
2561 | Lexer::test_get_input_codepoint_n_length (int n_start_offset) | |
2562 | { | |
2563 | uint8_t input = peek_input (n_start_offset); | |
2564 | ||
2565 | if (input < 128) | |
2566 | { | |
2567 | // ascii -- 1 byte | |
2568 | // return input; | |
2569 | return 1; | |
2570 | } | |
2571 | else if ((input & 0xC0) == 0x80) | |
2572 | { | |
2573 | // invalid (continuation; can't be first char) | |
2574 | // return 0xFFFE; | |
2575 | return 0; | |
2576 | } | |
2577 | else if ((input & 0xE0) == 0xC0) | |
2578 | { | |
2579 | // 2 bytes | |
2580 | uint8_t input2 = peek_input (n_start_offset + 1); | |
2581 | if ((input2 & 0xC0) != 0x80) | |
2582 | // return 0xFFFE; | |
2583 | return 0; | |
2584 | ||
2585 | // uint32_t output = ((input & 0x1F) << 6) | ((input2 & 0x3F) << 0); | |
2586 | // return output; | |
2587 | return 2; | |
2588 | } | |
2589 | else if ((input & 0xF0) == 0xE0) | |
2590 | { | |
2591 | // 3 bytes | |
2592 | uint8_t input2 = peek_input (n_start_offset + 1); | |
2593 | if ((input2 & 0xC0) != 0x80) | |
2594 | // return 0xFFFE; | |
2595 | return 0; | |
2596 | ||
2597 | uint8_t input3 = peek_input (n_start_offset + 2); | |
2598 | if ((input3 & 0xC0) != 0x80) | |
2599 | // return 0xFFFE; | |
2600 | return 0; | |
2601 | ||
2602 | /*uint32_t output | |
2603 | = ((input & 0x0F) << 12) | ((input2 & 0x3F) << 6) | ((input3 & 0x3F) << | |
2604 | 0); return output;*/ | |
2605 | return 3; | |
2606 | } | |
2607 | else if ((input & 0xF8) == 0xF0) | |
2608 | { | |
2609 | // 4 bytes | |
2610 | uint8_t input2 = peek_input (n_start_offset + 1); | |
2611 | if ((input2 & 0xC0) != 0x80) | |
2612 | // return 0xFFFE; | |
2613 | return 0; | |
2614 | ||
2615 | uint8_t input3 = peek_input (n_start_offset + 2); | |
2616 | if ((input3 & 0xC0) != 0x80) | |
2617 | // return 0xFFFE; | |
2618 | return 0; | |
2619 | ||
2620 | uint8_t input4 = peek_input (n_start_offset + 3); | |
2621 | if ((input4 & 0xC0) != 0x80) | |
2622 | // return 0xFFFE; | |
2623 | return 0; | |
2624 | ||
2625 | /*uint32_t output = ((input & 0x07) << 18) | ((input2 & 0x3F) << 12) | |
2626 | | ((input3 & 0x3F) << 6) | ((input4 & 0x3F) << 0); | |
2627 | return output;*/ | |
2628 | return 4; | |
2629 | } | |
2630 | else | |
2631 | { | |
2632 | rust_error_at (get_current_location (), | |
2633 | "invalid UTF-8 [THIRD] (too long)"); | |
2634 | return 0; | |
2635 | } | |
2636 | } | |
2637 | ||
2638 | // peeks the codepoint input at n codepoints ahead of current codepoint - try | |
2639 | // not to use | |
2640 | Codepoint | |
2641 | Lexer::test_peek_codepoint_input (int n) | |
2642 | { | |
2643 | int totalOffset = 0; | |
2644 | ||
2645 | // add up all offsets into total offset? does this do what I want? | |
2646 | for (int i = 0; i < n; i++) | |
2647 | { | |
2648 | totalOffset += test_get_input_codepoint_n_length (totalOffset); | |
2649 | } | |
2650 | // issues: this would have (at least) O(n) lookup time, not O(1) like the | |
2651 | // rest? | |
2652 | ||
2653 | // TODO: implement if still needed | |
2654 | ||
2655 | // error out of function as it is not implemented | |
2656 | gcc_assert (1 == 0); | |
2657 | return {0}; | |
2658 | /* | |
2659 | uint8_t input = peek_input(); | |
2660 | ||
2661 | if (input < 128) { | |
2662 | // ascii -- 1 byte | |
2663 | return input; | |
2664 | } else if ((input & 0xC0) == 0x80) { | |
2665 | // invalid (continuation; can't be first char) | |
2666 | return 0xFFFE; | |
2667 | } else if ((input & 0xE0) == 0xC0) { | |
2668 | // 2 bytes | |
2669 | uint8_t input2 = peek_input(1); | |
2670 | if ((input2 & 0xC0) != 0x80) | |
2671 | return 0xFFFE; | |
2672 | ||
2673 | uint32_t output = ((input & 0x1F) << 6) | ((input2 & 0x3F) << 0); | |
2674 | return output; | |
2675 | } else if ((input & 0xF0) == 0xE0) { | |
2676 | // 3 bytes | |
2677 | uint8_t input2 = peek_input(1); | |
2678 | if ((input2 & 0xC0) != 0x80) | |
2679 | return 0xFFFE; | |
2680 | ||
2681 | uint8_t input3 = peek_input(2); | |
2682 | if ((input3 & 0xC0) != 0x80) | |
2683 | return 0xFFFE; | |
2684 | ||
2685 | uint32_t output | |
2686 | = ((input & 0x0F) << 12) | ((input2 & 0x3F) << 6) | ((input3 & | |
2687 | 0x3F) << 0); return output; } else if ((input & 0xF8) == 0xF0) { | |
2688 | // 4 bytes | |
2689 | uint8_t input2 = peek_input(1); | |
2690 | if ((input2 & 0xC0) != 0x80) | |
2691 | return 0xFFFE; | |
2692 | ||
2693 | uint8_t input3 = peek_input(2); | |
2694 | if ((input3 & 0xC0) != 0x80) | |
2695 | return 0xFFFE; | |
2696 | ||
2697 | uint8_t input4 = peek_input(3); | |
2698 | if ((input4 & 0xC0) != 0x80) | |
2699 | return 0xFFFE; | |
2700 | ||
2701 | uint32_t output = ((input & 0x07) << 18) | ((input2 & 0x3F) << 12) | |
2702 | | ((input3 & 0x3F) << 6) | ((input4 & 0x3F) << | |
2703 | 0); return output; } else { rust_error_at(get_current_location(), "invalid | |
2704 | UTF-8 (too long)"); return 0xFFFE; | |
2705 | }*/ | |
2706 | } | |
2707 | ||
2708 | void | |
2709 | Lexer::split_current_token (TokenId new_left, TokenId new_right) | |
2710 | { | |
2711 | /* TODO: assert that this TokenId is a "simple token" like punctuation and not | |
2712 | * like "IDENTIFIER"? */ | |
2713 | Location current_loc = peek_token ()->get_locus (); | |
2714 | TokenPtr new_left_tok = Token::make (new_left, current_loc); | |
2715 | TokenPtr new_right_tok = Token::make (new_right, current_loc + 1); | |
2716 | ||
2717 | token_queue.replace_current_value (std::move (new_left_tok)); | |
2718 | token_queue.insert (1, std::move (new_right_tok)); | |
2719 | } | |
2720 | ||
2721 | void | |
2722 | Lexer::start_line (int current_line, int current_column) | |
2723 | { | |
2724 | if (line_map) | |
2725 | line_map->start_line (current_line, current_column); | |
2726 | } | |
2727 | ||
2728 | } // namespace Rust |