]>
Commit | Line | Data |
---|---|---|
447924ef | 1 | /* Data and functions related to line maps and input files. |
818ab71a | 2 | Copyright (C) 2004-2016 Free Software Foundation, Inc. |
447924ef JM |
3 | |
4 | This file is part of GCC. | |
5 | ||
6 | GCC is free software; you can redistribute it and/or modify it under | |
7 | the terms of the GNU General Public License as published by the Free | |
8 | Software Foundation; either version 3, or (at your option) any later | |
9 | version. | |
10 | ||
11 | GCC is distributed in the hope that it will be useful, but WITHOUT ANY | |
12 | WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
13 | FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
14 | for more details. | |
15 | ||
16 | You should have received a copy of the GNU General Public License | |
17 | along with GCC; see the file COPYING3. If not see | |
18 | <http://www.gnu.org/licenses/>. */ | |
19 | ||
20 | #include "config.h" | |
21 | #include "system.h" | |
22 | #include "coretypes.h" | |
23 | #include "intl.h" | |
ba4ad400 | 24 | #include "diagnostic-core.h" |
d9b950dd | 25 | #include "selftest.h" |
741d3be5 | 26 | #include "cpplib.h" |
7ecc3eb9 | 27 | |
a7d79e5c DM |
28 | #ifndef HAVE_ICONV |
29 | #define HAVE_ICONV 0 | |
30 | #endif | |
31 | ||
7ecc3eb9 DS |
32 | /* This is a cache used by get_next_line to store the content of a |
33 | file to be searched for file lines. */ | |
34 | struct fcache | |
35 | { | |
36 | /* These are information used to store a line boundary. */ | |
37 | struct line_info | |
38 | { | |
39 | /* The line number. It starts from 1. */ | |
40 | size_t line_num; | |
41 | ||
42 | /* The position (byte count) of the beginning of the line, | |
43 | relative to the file data pointer. This starts at zero. */ | |
44 | size_t start_pos; | |
45 | ||
46 | /* The position (byte count) of the last byte of the line. This | |
47 | normally points to the '\n' character, or to one byte after the | |
48 | last byte of the file, if the file doesn't contain a '\n' | |
49 | character. */ | |
50 | size_t end_pos; | |
51 | ||
52 | line_info (size_t l, size_t s, size_t e) | |
53 | : line_num (l), start_pos (s), end_pos (e) | |
54 | {} | |
55 | ||
56 | line_info () | |
57 | :line_num (0), start_pos (0), end_pos (0) | |
58 | {} | |
59 | }; | |
60 | ||
61 | /* The number of time this file has been accessed. This is used | |
62 | to designate which file cache to evict from the cache | |
63 | array. */ | |
64 | unsigned use_count; | |
65 | ||
66 | const char *file_path; | |
67 | ||
68 | FILE *fp; | |
69 | ||
70 | /* This points to the content of the file that we've read so | |
71 | far. */ | |
72 | char *data; | |
73 | ||
74 | /* The size of the DATA array above.*/ | |
75 | size_t size; | |
76 | ||
77 | /* The number of bytes read from the underlying file so far. This | |
78 | must be less (or equal) than SIZE above. */ | |
79 | size_t nb_read; | |
80 | ||
81 | /* The index of the beginning of the current line. */ | |
82 | size_t line_start_idx; | |
83 | ||
84 | /* The number of the previous line read. This starts at 1. Zero | |
85 | means we've read no line so far. */ | |
86 | size_t line_num; | |
87 | ||
88 | /* This is the total number of lines of the current file. At the | |
89 | moment, we try to get this information from the line map | |
90 | subsystem. Note that this is just a hint. When using the C++ | |
91 | front-end, this hint is correct because the input file is then | |
92 | completely tokenized before parsing starts; so the line map knows | |
93 | the number of lines before compilation really starts. For e.g, | |
94 | the C front-end, it can happen that we start emitting diagnostics | |
95 | before the line map has seen the end of the file. */ | |
96 | size_t total_lines; | |
97 | ||
98 | /* This is a record of the beginning and end of the lines we've seen | |
99 | while reading the file. This is useful to avoid walking the data | |
100 | from the beginning when we are asked to read a line that is | |
101 | before LINE_START_IDX above. Note that the maximum size of this | |
102 | record is fcache_line_record_size, so that the memory consumption | |
103 | doesn't explode. We thus scale total_lines down to | |
104 | fcache_line_record_size. */ | |
105 | vec<line_info, va_heap> line_record; | |
106 | ||
107 | fcache (); | |
108 | ~fcache (); | |
109 | }; | |
447924ef JM |
110 | |
111 | /* Current position in real source file. */ | |
112 | ||
3edf64aa | 113 | location_t input_location = UNKNOWN_LOCATION; |
447924ef JM |
114 | |
115 | struct line_maps *line_table; | |
116 | ||
7ecc3eb9 DS |
117 | static fcache *fcache_tab; |
118 | static const size_t fcache_tab_size = 16; | |
119 | static const size_t fcache_buffer_size = 4 * 1024; | |
120 | static const size_t fcache_line_record_size = 100; | |
121 | ||
84756fd4 DS |
122 | /* Expand the source location LOC into a human readable location. If |
123 | LOC resolves to a builtin location, the file name of the readable | |
7eb918cc DS |
124 | location is set to the string "<built-in>". If EXPANSION_POINT_P is |
125 | TRUE and LOC is virtual, then it is resolved to the expansion | |
126 | point of the involved macro. Otherwise, it is resolved to the | |
c4ca1a09 DS |
127 | spelling location of the token. |
128 | ||
129 | When resolving to the spelling location of the token, if the | |
130 | resulting location is for a built-in location (that is, it has no | |
131 | associated line/column) in the context of a macro expansion, the | |
132 | returned location is the first one (while unwinding the macro | |
133 | location towards its expansion point) that is in real source | |
134 | code. */ | |
7eb918cc DS |
135 | |
136 | static expanded_location | |
137 | expand_location_1 (source_location loc, | |
138 | bool expansion_point_p) | |
447924ef JM |
139 | { |
140 | expanded_location xloc; | |
0e50b624 | 141 | const line_map_ordinary *map; |
c4ca1a09 | 142 | enum location_resolution_kind lrk = LRK_MACRO_EXPANSION_POINT; |
5368224f DC |
143 | tree block = NULL; |
144 | ||
145 | if (IS_ADHOC_LOC (loc)) | |
146 | { | |
147 | block = LOCATION_BLOCK (loc); | |
148 | loc = LOCATION_LOCUS (loc); | |
149 | } | |
c4ca1a09 DS |
150 | |
151 | memset (&xloc, 0, sizeof (xloc)); | |
84756fd4 | 152 | |
c4ca1a09 DS |
153 | if (loc >= RESERVED_LOCATION_COUNT) |
154 | { | |
155 | if (!expansion_point_p) | |
156 | { | |
157 | /* We want to resolve LOC to its spelling location. | |
158 | ||
159 | But if that spelling location is a reserved location that | |
160 | appears in the context of a macro expansion (like for a | |
161 | location for a built-in token), let's consider the first | |
162 | location (toward the expansion point) that is not reserved; | |
163 | that is, the first location that is in real source code. */ | |
164 | loc = linemap_unwind_to_first_non_reserved_loc (line_table, | |
0e50b624 | 165 | loc, NULL); |
c4ca1a09 DS |
166 | lrk = LRK_SPELLING_LOCATION; |
167 | } | |
168 | loc = linemap_resolve_location (line_table, loc, | |
169 | lrk, &map); | |
170 | xloc = linemap_expand_location (line_table, map, loc); | |
171 | } | |
84756fd4 | 172 | |
5368224f | 173 | xloc.data = block; |
447924ef | 174 | if (loc <= BUILTINS_LOCATION) |
84756fd4 DS |
175 | xloc.file = loc == UNKNOWN_LOCATION ? NULL : _("<built-in>"); |
176 | ||
447924ef JM |
177 | return xloc; |
178 | } | |
64a1a422 | 179 | |
7ecc3eb9 DS |
180 | /* Initialize the set of cache used for files accessed by caret |
181 | diagnostic. */ | |
182 | ||
183 | static void | |
184 | diagnostic_file_cache_init (void) | |
185 | { | |
186 | if (fcache_tab == NULL) | |
187 | fcache_tab = new fcache[fcache_tab_size]; | |
188 | } | |
189 | ||
592f32fa | 190 | /* Free the resources used by the set of cache used for files accessed |
7ecc3eb9 DS |
191 | by caret diagnostic. */ |
192 | ||
193 | void | |
194 | diagnostic_file_cache_fini (void) | |
195 | { | |
196 | if (fcache_tab) | |
197 | { | |
198 | delete [] (fcache_tab); | |
199 | fcache_tab = NULL; | |
200 | } | |
201 | } | |
202 | ||
203 | /* Return the total lines number that have been read so far by the | |
204 | line map (in the preprocessor) so far. For languages like C++ that | |
205 | entirely preprocess the input file before starting to parse, this | |
206 | equals the actual number of lines of the file. */ | |
207 | ||
208 | static size_t | |
209 | total_lines_num (const char *file_path) | |
210 | { | |
211 | size_t r = 0; | |
212 | source_location l = 0; | |
213 | if (linemap_get_file_highest_location (line_table, file_path, &l)) | |
214 | { | |
215 | gcc_assert (l >= RESERVED_LOCATION_COUNT); | |
216 | expanded_location xloc = expand_location (l); | |
217 | r = xloc.line; | |
218 | } | |
219 | return r; | |
220 | } | |
221 | ||
222 | /* Lookup the cache used for the content of a given file accessed by | |
223 | caret diagnostic. Return the found cached file, or NULL if no | |
224 | cached file was found. */ | |
225 | ||
226 | static fcache* | |
227 | lookup_file_in_cache_tab (const char *file_path) | |
228 | { | |
229 | if (file_path == NULL) | |
230 | return NULL; | |
231 | ||
232 | diagnostic_file_cache_init (); | |
233 | ||
234 | /* This will contain the found cached file. */ | |
235 | fcache *r = NULL; | |
236 | for (unsigned i = 0; i < fcache_tab_size; ++i) | |
237 | { | |
238 | fcache *c = &fcache_tab[i]; | |
239 | if (c->file_path && !strcmp (c->file_path, file_path)) | |
240 | { | |
241 | ++c->use_count; | |
242 | r = c; | |
243 | } | |
244 | } | |
245 | ||
246 | if (r) | |
247 | ++r->use_count; | |
248 | ||
249 | return r; | |
250 | } | |
251 | ||
252 | /* Return the file cache that has been less used, recently, or the | |
253 | first empty one. If HIGHEST_USE_COUNT is non-null, | |
254 | *HIGHEST_USE_COUNT is set to the highest use count of the entries | |
255 | in the cache table. */ | |
256 | ||
257 | static fcache* | |
258 | evicted_cache_tab_entry (unsigned *highest_use_count) | |
259 | { | |
260 | diagnostic_file_cache_init (); | |
261 | ||
262 | fcache *to_evict = &fcache_tab[0]; | |
263 | unsigned huc = to_evict->use_count; | |
264 | for (unsigned i = 1; i < fcache_tab_size; ++i) | |
265 | { | |
266 | fcache *c = &fcache_tab[i]; | |
267 | bool c_is_empty = (c->file_path == NULL); | |
268 | ||
269 | if (c->use_count < to_evict->use_count | |
270 | || (to_evict->file_path && c_is_empty)) | |
271 | /* We evict C because it's either an entry with a lower use | |
272 | count or one that is empty. */ | |
273 | to_evict = c; | |
274 | ||
275 | if (huc < c->use_count) | |
276 | huc = c->use_count; | |
277 | ||
278 | if (c_is_empty) | |
279 | /* We've reached the end of the cache; subsequent elements are | |
280 | all empty. */ | |
281 | break; | |
282 | } | |
283 | ||
284 | if (highest_use_count) | |
285 | *highest_use_count = huc; | |
286 | ||
287 | return to_evict; | |
288 | } | |
289 | ||
290 | /* Create the cache used for the content of a given file to be | |
291 | accessed by caret diagnostic. This cache is added to an array of | |
292 | cache and can be retrieved by lookup_file_in_cache_tab. This | |
293 | function returns the created cache. Note that only the last | |
294 | fcache_tab_size files are cached. */ | |
295 | ||
296 | static fcache* | |
297 | add_file_to_cache_tab (const char *file_path) | |
298 | { | |
299 | ||
300 | FILE *fp = fopen (file_path, "r"); | |
317363b4 DS |
301 | if (fp == NULL) |
302 | return NULL; | |
7ecc3eb9 DS |
303 | |
304 | unsigned highest_use_count = 0; | |
305 | fcache *r = evicted_cache_tab_entry (&highest_use_count); | |
306 | r->file_path = file_path; | |
307 | if (r->fp) | |
308 | fclose (r->fp); | |
309 | r->fp = fp; | |
310 | r->nb_read = 0; | |
311 | r->line_start_idx = 0; | |
312 | r->line_num = 0; | |
313 | r->line_record.truncate (0); | |
314 | /* Ensure that this cache entry doesn't get evicted next time | |
315 | add_file_to_cache_tab is called. */ | |
316 | r->use_count = ++highest_use_count; | |
317 | r->total_lines = total_lines_num (file_path); | |
318 | ||
319 | return r; | |
320 | } | |
321 | ||
322 | /* Lookup the cache used for the content of a given file accessed by | |
323 | caret diagnostic. If no cached file was found, create a new cache | |
324 | for this file, add it to the array of cached file and return | |
325 | it. */ | |
326 | ||
327 | static fcache* | |
328 | lookup_or_add_file_to_cache_tab (const char *file_path) | |
329 | { | |
330 | fcache *r = lookup_file_in_cache_tab (file_path); | |
331 | if (r == NULL) | |
332 | r = add_file_to_cache_tab (file_path); | |
333 | return r; | |
334 | } | |
335 | ||
336 | /* Default constructor for a cache of file used by caret | |
337 | diagnostic. */ | |
338 | ||
339 | fcache::fcache () | |
340 | : use_count (0), file_path (NULL), fp (NULL), data (0), | |
341 | size (0), nb_read (0), line_start_idx (0), line_num (0), | |
342 | total_lines (0) | |
343 | { | |
344 | line_record.create (0); | |
345 | } | |
346 | ||
347 | /* Destructor for a cache of file used by caret diagnostic. */ | |
348 | ||
349 | fcache::~fcache () | |
350 | { | |
351 | if (fp) | |
352 | { | |
353 | fclose (fp); | |
354 | fp = NULL; | |
355 | } | |
356 | if (data) | |
357 | { | |
358 | XDELETEVEC (data); | |
359 | data = 0; | |
360 | } | |
361 | line_record.release (); | |
362 | } | |
363 | ||
364 | /* Returns TRUE iff the cache would need to be filled with data coming | |
365 | from the file. That is, either the cache is empty or full or the | |
366 | current line is empty. Note that if the cache is full, it would | |
367 | need to be extended and filled again. */ | |
368 | ||
369 | static bool | |
370 | needs_read (fcache *c) | |
371 | { | |
372 | return (c->nb_read == 0 | |
373 | || c->nb_read == c->size | |
374 | || (c->line_start_idx >= c->nb_read - 1)); | |
375 | } | |
376 | ||
377 | /* Return TRUE iff the cache is full and thus needs to be | |
378 | extended. */ | |
379 | ||
380 | static bool | |
381 | needs_grow (fcache *c) | |
382 | { | |
383 | return c->nb_read == c->size; | |
384 | } | |
385 | ||
386 | /* Grow the cache if it needs to be extended. */ | |
387 | ||
388 | static void | |
389 | maybe_grow (fcache *c) | |
9fec0042 | 390 | { |
7ecc3eb9 DS |
391 | if (!needs_grow (c)) |
392 | return; | |
393 | ||
394 | size_t size = c->size == 0 ? fcache_buffer_size : c->size * 2; | |
395 | c->data = XRESIZEVEC (char, c->data, size + 1); | |
396 | c->size = size; | |
397 | } | |
9fec0042 | 398 | |
7ecc3eb9 DS |
399 | /* Read more data into the cache. Extends the cache if need be. |
400 | Returns TRUE iff new data could be read. */ | |
401 | ||
402 | static bool | |
403 | read_data (fcache *c) | |
404 | { | |
405 | if (feof (c->fp) || ferror (c->fp)) | |
406 | return false; | |
407 | ||
408 | maybe_grow (c); | |
409 | ||
410 | char * from = c->data + c->nb_read; | |
411 | size_t to_read = c->size - c->nb_read; | |
412 | size_t nb_read = fread (from, 1, to_read, c->fp); | |
413 | ||
414 | if (ferror (c->fp)) | |
415 | return false; | |
416 | ||
417 | c->nb_read += nb_read; | |
418 | return !!nb_read; | |
419 | } | |
420 | ||
421 | /* Read new data iff the cache needs to be filled with more data | |
422 | coming from the file FP. Return TRUE iff the cache was filled with | |
423 | mode data. */ | |
424 | ||
425 | static bool | |
426 | maybe_read_data (fcache *c) | |
427 | { | |
428 | if (!needs_read (c)) | |
429 | return false; | |
430 | return read_data (c); | |
431 | } | |
432 | ||
433 | /* Read a new line from file FP, using C as a cache for the data | |
434 | coming from the file. Upon successful completion, *LINE is set to | |
435 | the beginning of the line found. Space for that line has been | |
436 | allocated in the cache thus *LINE has the same life time as C. | |
437 | *LINE_LEN is set to the length of the line. Note that the line | |
438 | does not contain any terminal delimiter. This function returns | |
439 | true if some data was read or process from the cache, false | |
440 | otherwise. Note that subsequent calls to get_next_line return the | |
441 | next lines of the file and might overwrite the content of | |
442 | *LINE. */ | |
443 | ||
444 | static bool | |
445 | get_next_line (fcache *c, char **line, ssize_t *line_len) | |
446 | { | |
447 | /* Fill the cache with data to process. */ | |
448 | maybe_read_data (c); | |
449 | ||
450 | size_t remaining_size = c->nb_read - c->line_start_idx; | |
451 | if (remaining_size == 0) | |
452 | /* There is no more data to process. */ | |
453 | return false; | |
454 | ||
455 | char *line_start = c->data + c->line_start_idx; | |
456 | ||
457 | char *next_line_start = NULL; | |
458 | size_t len = 0; | |
459 | char *line_end = (char *) memchr (line_start, '\n', remaining_size); | |
460 | if (line_end == NULL) | |
9fec0042 | 461 | { |
7ecc3eb9 DS |
462 | /* We haven't found the end-of-line delimiter in the cache. |
463 | Fill the cache with more data from the file and look for the | |
464 | '\n'. */ | |
465 | while (maybe_read_data (c)) | |
466 | { | |
467 | line_start = c->data + c->line_start_idx; | |
468 | remaining_size = c->nb_read - c->line_start_idx; | |
469 | line_end = (char *) memchr (line_start, '\n', remaining_size); | |
470 | if (line_end != NULL) | |
471 | { | |
472 | next_line_start = line_end + 1; | |
473 | break; | |
474 | } | |
475 | } | |
476 | if (line_end == NULL) | |
477 | /* We've loadded all the file into the cache and still no | |
478 | '\n'. Let's say the line ends up at one byte passed the | |
479 | end of the file. This is to stay consistent with the case | |
480 | of when the line ends up with a '\n' and line_end points to | |
481 | that terminal '\n'. That consistency is useful below in | |
482 | the len calculation. */ | |
483 | line_end = c->data + c->nb_read ; | |
9fec0042 | 484 | } |
7ecc3eb9 DS |
485 | else |
486 | next_line_start = line_end + 1; | |
487 | ||
488 | if (ferror (c->fp)) | |
489 | return -1; | |
490 | ||
491 | /* At this point, we've found the end of the of line. It either | |
492 | points to the '\n' or to one byte after the last byte of the | |
493 | file. */ | |
494 | gcc_assert (line_end != NULL); | |
9fec0042 | 495 | |
7ecc3eb9 DS |
496 | len = line_end - line_start; |
497 | ||
498 | if (c->line_start_idx < c->nb_read) | |
499 | *line = line_start; | |
500 | ||
501 | ++c->line_num; | |
502 | ||
503 | /* Before we update our line record, make sure the hint about the | |
504 | total number of lines of the file is correct. If it's not, then | |
505 | we give up recording line boundaries from now on. */ | |
506 | bool update_line_record = true; | |
507 | if (c->line_num > c->total_lines) | |
508 | update_line_record = false; | |
509 | ||
510 | /* Now update our line record so that re-reading lines from the | |
511 | before c->line_start_idx is faster. */ | |
512 | if (update_line_record | |
513 | && c->line_record.length () < fcache_line_record_size) | |
514 | { | |
515 | /* If the file lines fits in the line record, we just record all | |
516 | its lines ...*/ | |
517 | if (c->total_lines <= fcache_line_record_size | |
518 | && c->line_num > c->line_record.length ()) | |
519 | c->line_record.safe_push (fcache::line_info (c->line_num, | |
520 | c->line_start_idx, | |
521 | line_end - c->data)); | |
522 | else if (c->total_lines > fcache_line_record_size) | |
523 | { | |
524 | /* ... otherwise, we just scale total_lines down to | |
525 | (fcache_line_record_size lines. */ | |
526 | size_t n = (c->line_num * fcache_line_record_size) / c->total_lines; | |
527 | if (c->line_record.length () == 0 | |
528 | || n >= c->line_record.length ()) | |
529 | c->line_record.safe_push (fcache::line_info (c->line_num, | |
530 | c->line_start_idx, | |
531 | line_end - c->data)); | |
532 | } | |
533 | } | |
534 | ||
535 | /* Update c->line_start_idx so that it points to the next line to be | |
536 | read. */ | |
537 | if (next_line_start) | |
538 | c->line_start_idx = next_line_start - c->data; | |
539 | else | |
540 | /* We didn't find any terminal '\n'. Let's consider that the end | |
541 | of line is the end of the data in the cache. The next | |
542 | invocation of get_next_line will either read more data from the | |
543 | underlying file or return false early because we've reached the | |
544 | end of the file. */ | |
545 | c->line_start_idx = c->nb_read; | |
546 | ||
547 | *line_len = len; | |
548 | ||
549 | return true; | |
550 | } | |
551 | ||
552 | /* Reads the next line from FILE into *LINE. If *LINE is too small | |
553 | (or NULL) it is allocated (or extended) to have enough space to | |
554 | containe the line. *LINE_LENGTH must contain the size of the | |
555 | initial*LINE buffer. It's then updated by this function to the | |
556 | actual length of the returned line. Note that the returned line | |
557 | can contain several zero bytes. Also note that the returned string | |
558 | is allocated in static storage that is going to be re-used by | |
559 | subsequent invocations of read_line. */ | |
560 | ||
561 | static bool | |
562 | read_next_line (fcache *cache, char ** line, ssize_t *line_len) | |
563 | { | |
564 | char *l = NULL; | |
565 | ssize_t len = 0; | |
566 | ||
567 | if (!get_next_line (cache, &l, &len)) | |
568 | return false; | |
569 | ||
570 | if (*line == NULL) | |
571 | *line = XNEWVEC (char, len); | |
572 | else | |
573 | if (*line_len < len) | |
574 | *line = XRESIZEVEC (char, *line, len); | |
575 | ||
576 | memcpy (*line, l, len); | |
577 | *line_len = len; | |
578 | ||
579 | return true; | |
580 | } | |
581 | ||
582 | /* Consume the next bytes coming from the cache (or from its | |
583 | underlying file if there are remaining unread bytes in the file) | |
584 | until we reach the next end-of-line (or end-of-file). There is no | |
585 | copying from the cache involved. Return TRUE upon successful | |
586 | completion. */ | |
587 | ||
588 | static bool | |
589 | goto_next_line (fcache *cache) | |
590 | { | |
591 | char *l; | |
592 | ssize_t len; | |
593 | ||
594 | return get_next_line (cache, &l, &len); | |
595 | } | |
596 | ||
597 | /* Read an arbitrary line number LINE_NUM from the file cached in C. | |
598 | The line is copied into *LINE. *LINE_LEN must have been set to the | |
599 | length of *LINE. If *LINE is too small (or NULL) it's extended (or | |
600 | allocated) and *LINE_LEN is adjusted accordingly. *LINE ends up | |
601 | with a terminal zero byte and can contain additional zero bytes. | |
602 | This function returns bool if a line was read. */ | |
603 | ||
604 | static bool | |
605 | read_line_num (fcache *c, size_t line_num, | |
606 | char ** line, ssize_t *line_len) | |
607 | { | |
608 | gcc_assert (line_num > 0); | |
609 | ||
610 | if (line_num <= c->line_num) | |
9789a912 | 611 | { |
7ecc3eb9 DS |
612 | /* We've been asked to read lines that are before c->line_num. |
613 | So lets use our line record (if it's not empty) to try to | |
614 | avoid re-reading the file from the beginning again. */ | |
7f4d640c | 615 | |
7ecc3eb9 | 616 | if (c->line_record.is_empty ()) |
9fec0042 | 617 | { |
7ecc3eb9 DS |
618 | c->line_start_idx = 0; |
619 | c->line_num = 0; | |
620 | } | |
621 | else | |
622 | { | |
623 | fcache::line_info *i = NULL; | |
624 | if (c->total_lines <= fcache_line_record_size) | |
625 | { | |
626 | /* In languages where the input file is not totally | |
627 | preprocessed up front, the c->total_lines hint | |
628 | can be smaller than the number of lines of the | |
629 | file. In that case, only the first | |
630 | c->total_lines have been recorded. | |
631 | ||
632 | Otherwise, the first c->total_lines we've read have | |
633 | their start/end recorded here. */ | |
634 | i = (line_num <= c->total_lines) | |
635 | ? &c->line_record[line_num - 1] | |
636 | : &c->line_record[c->total_lines - 1]; | |
637 | gcc_assert (i->line_num <= line_num); | |
638 | } | |
639 | else | |
640 | { | |
641 | /* So the file had more lines than our line record | |
642 | size. Thus the number of lines we've recorded has | |
643 | been scaled down to fcache_line_reacord_size. Let's | |
644 | pick the start/end of the recorded line that is | |
645 | closest to line_num. */ | |
646 | size_t n = (line_num <= c->total_lines) | |
647 | ? line_num * fcache_line_record_size / c->total_lines | |
648 | : c ->line_record.length () - 1; | |
649 | if (n < c->line_record.length ()) | |
650 | { | |
651 | i = &c->line_record[n]; | |
652 | gcc_assert (i->line_num <= line_num); | |
653 | } | |
654 | } | |
655 | ||
656 | if (i && i->line_num == line_num) | |
657 | { | |
658 | /* We have the start/end of the line. Let's just copy | |
659 | it again and we are done. */ | |
660 | ssize_t len = i->end_pos - i->start_pos + 1; | |
661 | if (*line_len < len) | |
662 | *line = XRESIZEVEC (char, *line, len); | |
663 | memmove (*line, c->data + i->start_pos, len); | |
664 | (*line)[len - 1] = '\0'; | |
665 | *line_len = --len; | |
666 | return true; | |
667 | } | |
668 | ||
669 | if (i) | |
670 | { | |
671 | c->line_start_idx = i->start_pos; | |
672 | c->line_num = i->line_num - 1; | |
673 | } | |
674 | else | |
675 | { | |
676 | c->line_start_idx = 0; | |
677 | c->line_num = 0; | |
678 | } | |
9fec0042 | 679 | } |
9fec0042 | 680 | } |
7ecc3eb9 DS |
681 | |
682 | /* Let's walk from line c->line_num up to line_num - 1, without | |
683 | copying any line. */ | |
684 | while (c->line_num < line_num - 1) | |
685 | if (!goto_next_line (c)) | |
686 | return false; | |
687 | ||
688 | /* The line we want is the next one. Let's read and copy it back to | |
689 | the caller. */ | |
690 | return read_next_line (c, line, line_len); | |
9fec0042 MLI |
691 | } |
692 | ||
31bdd08a | 693 | /* Return the physical source line that corresponds to FILE_PATH/LINE in a |
9fec0042 | 694 | buffer that is statically allocated. The newline is replaced by |
7ecc3eb9 DS |
695 | the null character. Note that the line can contain several null |
696 | characters, so LINE_LEN, if non-null, points to the actual length | |
697 | of the line. */ | |
9fec0042 MLI |
698 | |
699 | const char * | |
31bdd08a | 700 | location_get_source_line (const char *file_path, int line, |
7ecc3eb9 | 701 | int *line_len) |
9fec0042 | 702 | { |
7ecc3eb9 DS |
703 | static char *buffer; |
704 | static ssize_t len; | |
705 | ||
31bdd08a | 706 | if (line == 0) |
367c8286 DS |
707 | return NULL; |
708 | ||
31bdd08a | 709 | fcache *c = lookup_or_add_file_to_cache_tab (file_path); |
367c8286 DS |
710 | if (c == NULL) |
711 | return NULL; | |
712 | ||
31bdd08a | 713 | bool read = read_line_num (c, line, &buffer, &len); |
9fec0042 | 714 | |
7ecc3eb9 DS |
715 | if (read && line_len) |
716 | *line_len = len; | |
9fec0042 | 717 | |
7ecc3eb9 | 718 | return read ? buffer : NULL; |
9fec0042 MLI |
719 | } |
720 | ||
c468587a DS |
721 | /* Test if the location originates from the spelling location of a |
722 | builtin-tokens. That is, return TRUE if LOC is a (possibly | |
723 | virtual) location of a built-in token that appears in the expansion | |
724 | list of a macro. Please note that this function also works on | |
725 | tokens that result from built-in tokens. For instance, the | |
726 | function would return true if passed a token "4" that is the result | |
727 | of the expansion of the built-in __LINE__ macro. */ | |
728 | bool | |
729 | is_location_from_builtin_token (source_location loc) | |
730 | { | |
0e50b624 | 731 | const line_map_ordinary *map = NULL; |
c468587a DS |
732 | loc = linemap_resolve_location (line_table, loc, |
733 | LRK_SPELLING_LOCATION, &map); | |
734 | return loc == BUILTINS_LOCATION; | |
735 | } | |
736 | ||
7eb918cc DS |
737 | /* Expand the source location LOC into a human readable location. If |
738 | LOC is virtual, it resolves to the expansion point of the involved | |
739 | macro. If LOC resolves to a builtin location, the file name of the | |
740 | readable location is set to the string "<built-in>". */ | |
741 | ||
742 | expanded_location | |
743 | expand_location (source_location loc) | |
744 | { | |
745 | return expand_location_1 (loc, /*expansion_point_p=*/true); | |
746 | } | |
747 | ||
748 | /* Expand the source location LOC into a human readable location. If | |
749 | LOC is virtual, it resolves to the expansion location of the | |
750 | relevant macro. If LOC resolves to a builtin location, the file | |
751 | name of the readable location is set to the string | |
752 | "<built-in>". */ | |
753 | ||
754 | expanded_location | |
755 | expand_location_to_spelling_point (source_location loc) | |
756 | { | |
e1f0c178 | 757 | return expand_location_1 (loc, /*expansion_point_p=*/false); |
7eb918cc DS |
758 | } |
759 | ||
8a645150 DM |
760 | /* The rich_location class within libcpp requires a way to expand |
761 | source_location instances, and relies on the client code | |
762 | providing a symbol named | |
763 | linemap_client_expand_location_to_spelling_point | |
764 | to do this. | |
765 | ||
766 | This is the implementation for libcommon.a (all host binaries), | |
767 | which simply calls into expand_location_to_spelling_point. */ | |
768 | ||
769 | expanded_location | |
770 | linemap_client_expand_location_to_spelling_point (source_location loc) | |
771 | { | |
772 | return expand_location_to_spelling_point (loc); | |
773 | } | |
774 | ||
775 | ||
e1f0c178 MLI |
776 | /* If LOCATION is in a system header and if it is a virtual location for |
777 | a token coming from the expansion of a macro, unwind it to the | |
778 | location of the expansion point of the macro. Otherwise, just return | |
70dc395a DS |
779 | LOCATION. |
780 | ||
781 | This is used for instance when we want to emit diagnostics about a | |
e1f0c178 MLI |
782 | token that may be located in a macro that is itself defined in a |
783 | system header, for example, for the NULL macro. In such a case, if | |
784 | LOCATION were passed directly to diagnostic functions such as | |
785 | warning_at, the diagnostic would be suppressed (unless | |
786 | -Wsystem-headers). */ | |
70dc395a DS |
787 | |
788 | source_location | |
789 | expansion_point_location_if_in_system_header (source_location location) | |
790 | { | |
791 | if (in_system_header_at (location)) | |
792 | location = linemap_resolve_location (line_table, location, | |
793 | LRK_MACRO_EXPANSION_POINT, | |
794 | NULL); | |
795 | return location; | |
796 | } | |
7eb918cc | 797 | |
79ce98bc MP |
798 | /* If LOCATION is a virtual location for a token coming from the expansion |
799 | of a macro, unwind to the location of the expansion point of the macro. */ | |
800 | ||
801 | source_location | |
802 | expansion_point_location (source_location location) | |
803 | { | |
804 | return linemap_resolve_location (line_table, location, | |
805 | LRK_MACRO_EXPANSION_POINT, NULL); | |
806 | } | |
807 | ||
a01fc549 DM |
808 | /* Given location LOC, strip away any packed range information |
809 | or ad-hoc information. */ | |
810 | ||
811 | location_t | |
812 | get_pure_location (location_t loc) | |
813 | { | |
814 | if (IS_ADHOC_LOC (loc)) | |
815 | loc | |
816 | = line_table->location_adhoc_data_map.data[loc & MAX_SOURCE_LOCATION].locus; | |
817 | ||
818 | if (loc >= LINEMAPS_MACRO_LOWEST_LOCATION (line_table)) | |
819 | return loc; | |
820 | ||
821 | if (loc < RESERVED_LOCATION_COUNT) | |
822 | return loc; | |
823 | ||
824 | const line_map *map = linemap_lookup (line_table, loc); | |
825 | const line_map_ordinary *ordmap = linemap_check_ordinary (map); | |
826 | ||
827 | return loc & ~((1 << ordmap->m_range_bits) - 1); | |
828 | } | |
829 | ||
830 | /* Construct a location with caret at CARET, ranging from START to | |
831 | finish e.g. | |
832 | ||
833 | 11111111112 | |
834 | 12345678901234567890 | |
835 | 522 | |
836 | 523 return foo + bar; | |
837 | ~~~~^~~~~ | |
838 | 524 | |
839 | ||
840 | The location's caret is at the "+", line 523 column 15, but starts | |
841 | earlier, at the "f" of "foo" at column 11. The finish is at the "r" | |
842 | of "bar" at column 19. */ | |
843 | ||
844 | location_t | |
845 | make_location (location_t caret, location_t start, location_t finish) | |
846 | { | |
847 | location_t pure_loc = get_pure_location (caret); | |
848 | source_range src_range; | |
849 | src_range.m_start = start; | |
850 | src_range.m_finish = finish; | |
851 | location_t combined_loc = COMBINE_LOCATION_DATA (line_table, | |
852 | pure_loc, | |
853 | src_range, | |
854 | NULL); | |
855 | return combined_loc; | |
856 | } | |
857 | ||
64a1a422 TT |
858 | #define ONE_K 1024 |
859 | #define ONE_M (ONE_K * ONE_K) | |
860 | ||
861 | /* Display a number as an integer multiple of either: | |
862 | - 1024, if said integer is >= to 10 K (in base 2) | |
863 | - 1024 * 1024, if said integer is >= 10 M in (base 2) | |
864 | */ | |
865 | #define SCALE(x) ((unsigned long) ((x) < 10 * ONE_K \ | |
866 | ? (x) \ | |
867 | : ((x) < 10 * ONE_M \ | |
868 | ? (x) / ONE_K \ | |
869 | : (x) / ONE_M))) | |
870 | ||
871 | /* For a given integer, display either: | |
872 | - the character 'k', if the number is higher than 10 K (in base 2) | |
873 | but strictly lower than 10 M (in base 2) | |
874 | - the character 'M' if the number is higher than 10 M (in base2) | |
875 | - the charcter ' ' if the number is strictly lower than 10 K */ | |
876 | #define STAT_LABEL(x) ((x) < 10 * ONE_K ? ' ' : ((x) < 10 * ONE_M ? 'k' : 'M')) | |
877 | ||
878 | /* Display an integer amount as multiple of 1K or 1M (in base 2). | |
879 | Display the correct unit (either k, M, or ' ') after the amout, as | |
880 | well. */ | |
881 | #define FORMAT_AMOUNT(size) SCALE (size), STAT_LABEL (size) | |
882 | ||
883 | /* Dump statistics to stderr about the memory usage of the line_table | |
884 | set of line maps. This also displays some statistics about macro | |
885 | expansion. */ | |
886 | ||
887 | void | |
888 | dump_line_table_statistics (void) | |
889 | { | |
890 | struct linemap_stats s; | |
d17687f6 | 891 | long total_used_map_size, |
64a1a422 TT |
892 | macro_maps_size, |
893 | total_allocated_map_size; | |
894 | ||
895 | memset (&s, 0, sizeof (s)); | |
896 | ||
897 | linemap_get_statistics (line_table, &s); | |
898 | ||
899 | macro_maps_size = s.macro_maps_used_size | |
900 | + s.macro_maps_locations_size; | |
901 | ||
902 | total_allocated_map_size = s.ordinary_maps_allocated_size | |
903 | + s.macro_maps_allocated_size | |
904 | + s.macro_maps_locations_size; | |
905 | ||
906 | total_used_map_size = s.ordinary_maps_used_size | |
907 | + s.macro_maps_used_size | |
908 | + s.macro_maps_locations_size; | |
909 | ||
d17687f6 | 910 | fprintf (stderr, "Number of expanded macros: %5ld\n", |
64a1a422 TT |
911 | s.num_expanded_macros); |
912 | if (s.num_expanded_macros != 0) | |
d17687f6 | 913 | fprintf (stderr, "Average number of tokens per macro expansion: %5ld\n", |
64a1a422 TT |
914 | s.num_macro_tokens / s.num_expanded_macros); |
915 | fprintf (stderr, | |
916 | "\nLine Table allocations during the " | |
917 | "compilation process\n"); | |
d17687f6 | 918 | fprintf (stderr, "Number of ordinary maps used: %5ld%c\n", |
64a1a422 TT |
919 | SCALE (s.num_ordinary_maps_used), |
920 | STAT_LABEL (s.num_ordinary_maps_used)); | |
d17687f6 | 921 | fprintf (stderr, "Ordinary map used size: %5ld%c\n", |
64a1a422 TT |
922 | SCALE (s.ordinary_maps_used_size), |
923 | STAT_LABEL (s.ordinary_maps_used_size)); | |
d17687f6 | 924 | fprintf (stderr, "Number of ordinary maps allocated: %5ld%c\n", |
64a1a422 TT |
925 | SCALE (s.num_ordinary_maps_allocated), |
926 | STAT_LABEL (s.num_ordinary_maps_allocated)); | |
d17687f6 | 927 | fprintf (stderr, "Ordinary maps allocated size: %5ld%c\n", |
64a1a422 TT |
928 | SCALE (s.ordinary_maps_allocated_size), |
929 | STAT_LABEL (s.ordinary_maps_allocated_size)); | |
d17687f6 | 930 | fprintf (stderr, "Number of macro maps used: %5ld%c\n", |
64a1a422 TT |
931 | SCALE (s.num_macro_maps_used), |
932 | STAT_LABEL (s.num_macro_maps_used)); | |
d17687f6 | 933 | fprintf (stderr, "Macro maps used size: %5ld%c\n", |
64a1a422 TT |
934 | SCALE (s.macro_maps_used_size), |
935 | STAT_LABEL (s.macro_maps_used_size)); | |
d17687f6 | 936 | fprintf (stderr, "Macro maps locations size: %5ld%c\n", |
64a1a422 TT |
937 | SCALE (s.macro_maps_locations_size), |
938 | STAT_LABEL (s.macro_maps_locations_size)); | |
d17687f6 | 939 | fprintf (stderr, "Macro maps size: %5ld%c\n", |
64a1a422 TT |
940 | SCALE (macro_maps_size), |
941 | STAT_LABEL (macro_maps_size)); | |
d17687f6 | 942 | fprintf (stderr, "Duplicated maps locations size: %5ld%c\n", |
64a1a422 TT |
943 | SCALE (s.duplicated_macro_maps_locations_size), |
944 | STAT_LABEL (s.duplicated_macro_maps_locations_size)); | |
d17687f6 | 945 | fprintf (stderr, "Total allocated maps size: %5ld%c\n", |
64a1a422 TT |
946 | SCALE (total_allocated_map_size), |
947 | STAT_LABEL (total_allocated_map_size)); | |
d17687f6 | 948 | fprintf (stderr, "Total used maps size: %5ld%c\n", |
64a1a422 TT |
949 | SCALE (total_used_map_size), |
950 | STAT_LABEL (total_used_map_size)); | |
ee015909 DM |
951 | fprintf (stderr, "Ad-hoc table size: %5ld%c\n", |
952 | SCALE (s.adhoc_table_size), | |
953 | STAT_LABEL (s.adhoc_table_size)); | |
954 | fprintf (stderr, "Ad-hoc table entries used: %5ld\n", | |
955 | s.adhoc_table_entries_used); | |
ebedc9a3 DM |
956 | fprintf (stderr, "optimized_ranges: %i\n", |
957 | line_table->num_optimized_ranges); | |
958 | fprintf (stderr, "unoptimized_ranges: %i\n", | |
959 | line_table->num_unoptimized_ranges); | |
ee015909 | 960 | |
64a1a422 TT |
961 | fprintf (stderr, "\n"); |
962 | } | |
ba4ad400 DM |
963 | |
964 | /* Get location one beyond the final location in ordinary map IDX. */ | |
965 | ||
966 | static source_location | |
967 | get_end_location (struct line_maps *set, unsigned int idx) | |
968 | { | |
969 | if (idx == LINEMAPS_ORDINARY_USED (set) - 1) | |
970 | return set->highest_location; | |
971 | ||
972 | struct line_map *next_map = LINEMAPS_ORDINARY_MAP_AT (set, idx + 1); | |
973 | return MAP_START_LOCATION (next_map); | |
974 | } | |
975 | ||
976 | /* Helper function for write_digit_row. */ | |
977 | ||
978 | static void | |
979 | write_digit (FILE *stream, int digit) | |
980 | { | |
981 | fputc ('0' + (digit % 10), stream); | |
982 | } | |
983 | ||
984 | /* Helper function for dump_location_info. | |
985 | Write a row of numbers to STREAM, numbering a source line, | |
986 | giving the units, tens, hundreds etc of the column number. */ | |
987 | ||
988 | static void | |
989 | write_digit_row (FILE *stream, int indent, | |
ebedc9a3 | 990 | const line_map_ordinary *map, |
ba4ad400 DM |
991 | source_location loc, int max_col, int divisor) |
992 | { | |
993 | fprintf (stream, "%*c", indent, ' '); | |
994 | fprintf (stream, "|"); | |
995 | for (int column = 1; column < max_col; column++) | |
996 | { | |
ebedc9a3 | 997 | source_location column_loc = loc + (column << map->m_range_bits); |
ba4ad400 DM |
998 | write_digit (stream, column_loc / divisor); |
999 | } | |
1000 | fprintf (stream, "\n"); | |
1001 | } | |
1002 | ||
1003 | /* Write a half-closed (START) / half-open (END) interval of | |
1004 | source_location to STREAM. */ | |
1005 | ||
1006 | static void | |
1007 | dump_location_range (FILE *stream, | |
1008 | source_location start, source_location end) | |
1009 | { | |
1010 | fprintf (stream, | |
1011 | " source_location interval: %u <= loc < %u\n", | |
1012 | start, end); | |
1013 | } | |
1014 | ||
1015 | /* Write a labelled description of a half-closed (START) / half-open (END) | |
1016 | interval of source_location to STREAM. */ | |
1017 | ||
1018 | static void | |
1019 | dump_labelled_location_range (FILE *stream, | |
1020 | const char *name, | |
1021 | source_location start, source_location end) | |
1022 | { | |
1023 | fprintf (stream, "%s\n", name); | |
1024 | dump_location_range (stream, start, end); | |
1025 | fprintf (stream, "\n"); | |
1026 | } | |
1027 | ||
1028 | /* Write a visualization of the locations in the line_table to STREAM. */ | |
1029 | ||
1030 | void | |
1031 | dump_location_info (FILE *stream) | |
1032 | { | |
1033 | /* Visualize the reserved locations. */ | |
1034 | dump_labelled_location_range (stream, "RESERVED LOCATIONS", | |
1035 | 0, RESERVED_LOCATION_COUNT); | |
1036 | ||
1037 | /* Visualize the ordinary line_map instances, rendering the sources. */ | |
1038 | for (unsigned int idx = 0; idx < LINEMAPS_ORDINARY_USED (line_table); idx++) | |
1039 | { | |
1040 | source_location end_location = get_end_location (line_table, idx); | |
1041 | /* half-closed: doesn't include this one. */ | |
1042 | ||
0e50b624 DM |
1043 | const line_map_ordinary *map |
1044 | = LINEMAPS_ORDINARY_MAP_AT (line_table, idx); | |
ba4ad400 DM |
1045 | fprintf (stream, "ORDINARY MAP: %i\n", idx); |
1046 | dump_location_range (stream, | |
1047 | MAP_START_LOCATION (map), end_location); | |
1048 | fprintf (stream, " file: %s\n", ORDINARY_MAP_FILE_NAME (map)); | |
1049 | fprintf (stream, " starting at line: %i\n", | |
1050 | ORDINARY_MAP_STARTING_LINE_NUMBER (map)); | |
ebedc9a3 DM |
1051 | fprintf (stream, " column and range bits: %i\n", |
1052 | map->m_column_and_range_bits); | |
ba4ad400 | 1053 | fprintf (stream, " column bits: %i\n", |
ebedc9a3 DM |
1054 | map->m_column_and_range_bits - map->m_range_bits); |
1055 | fprintf (stream, " range bits: %i\n", | |
1056 | map->m_range_bits); | |
ba4ad400 DM |
1057 | |
1058 | /* Render the span of source lines that this "map" covers. */ | |
1059 | for (source_location loc = MAP_START_LOCATION (map); | |
1060 | loc < end_location; | |
ebedc9a3 | 1061 | loc += (1 << map->m_range_bits) ) |
ba4ad400 | 1062 | { |
ebedc9a3 DM |
1063 | gcc_assert (pure_location_p (line_table, loc) ); |
1064 | ||
ba4ad400 DM |
1065 | expanded_location exploc |
1066 | = linemap_expand_location (line_table, map, loc); | |
1067 | ||
1068 | if (0 == exploc.column) | |
1069 | { | |
1070 | /* Beginning of a new source line: draw the line. */ | |
1071 | ||
1072 | int line_size; | |
31bdd08a DM |
1073 | const char *line_text = location_get_source_line (exploc.file, |
1074 | exploc.line, | |
1075 | &line_size); | |
ba4ad400 DM |
1076 | if (!line_text) |
1077 | break; | |
1078 | fprintf (stream, | |
1079 | "%s:%3i|loc:%5i|%.*s\n", | |
1080 | exploc.file, exploc.line, | |
1081 | loc, | |
1082 | line_size, line_text); | |
1083 | ||
1084 | /* "loc" is at column 0, which means "the whole line". | |
1085 | Render the locations *within* the line, by underlining | |
1086 | it, showing the source_location numeric values | |
1087 | at each column. */ | |
ebedc9a3 | 1088 | int max_col = (1 << map->m_column_and_range_bits) - 1; |
ba4ad400 DM |
1089 | if (max_col > line_size) |
1090 | max_col = line_size + 1; | |
1091 | ||
1092 | int indent = 14 + strlen (exploc.file); | |
1093 | ||
1094 | /* Thousands. */ | |
1095 | if (end_location > 999) | |
ebedc9a3 | 1096 | write_digit_row (stream, indent, map, loc, max_col, 1000); |
ba4ad400 DM |
1097 | |
1098 | /* Hundreds. */ | |
1099 | if (end_location > 99) | |
ebedc9a3 | 1100 | write_digit_row (stream, indent, map, loc, max_col, 100); |
ba4ad400 DM |
1101 | |
1102 | /* Tens. */ | |
ebedc9a3 | 1103 | write_digit_row (stream, indent, map, loc, max_col, 10); |
ba4ad400 DM |
1104 | |
1105 | /* Units. */ | |
ebedc9a3 | 1106 | write_digit_row (stream, indent, map, loc, max_col, 1); |
ba4ad400 DM |
1107 | } |
1108 | } | |
1109 | fprintf (stream, "\n"); | |
1110 | } | |
1111 | ||
1112 | /* Visualize unallocated values. */ | |
1113 | dump_labelled_location_range (stream, "UNALLOCATED LOCATIONS", | |
1114 | line_table->highest_location, | |
1115 | LINEMAPS_MACRO_LOWEST_LOCATION (line_table)); | |
1116 | ||
1117 | /* Visualize the macro line_map instances, rendering the sources. */ | |
1118 | for (unsigned int i = 0; i < LINEMAPS_MACRO_USED (line_table); i++) | |
1119 | { | |
1120 | /* Each macro map that is allocated owns source_location values | |
1121 | that are *lower* that the one before them. | |
1122 | Hence it's meaningful to view them either in order of ascending | |
1123 | source locations, or in order of ascending macro map index. */ | |
1124 | const bool ascending_source_locations = true; | |
1125 | unsigned int idx = (ascending_source_locations | |
1126 | ? (LINEMAPS_MACRO_USED (line_table) - (i + 1)) | |
1127 | : i); | |
0e50b624 | 1128 | const line_map_macro *map = LINEMAPS_MACRO_MAP_AT (line_table, idx); |
ba4ad400 DM |
1129 | fprintf (stream, "MACRO %i: %s (%u tokens)\n", |
1130 | idx, | |
1131 | linemap_map_get_macro_name (map), | |
1132 | MACRO_MAP_NUM_MACRO_TOKENS (map)); | |
1133 | dump_location_range (stream, | |
1134 | map->start_location, | |
1135 | (map->start_location | |
1136 | + MACRO_MAP_NUM_MACRO_TOKENS (map))); | |
1137 | inform (MACRO_MAP_EXPANSION_POINT_LOCATION (map), | |
1138 | "expansion point is location %i", | |
1139 | MACRO_MAP_EXPANSION_POINT_LOCATION (map)); | |
1140 | fprintf (stream, " map->start_location: %u\n", | |
1141 | map->start_location); | |
1142 | ||
1143 | fprintf (stream, " macro_locations:\n"); | |
1144 | for (unsigned int i = 0; i < MACRO_MAP_NUM_MACRO_TOKENS (map); i++) | |
1145 | { | |
1146 | source_location x = MACRO_MAP_LOCATIONS (map)[2 * i]; | |
1147 | source_location y = MACRO_MAP_LOCATIONS (map)[(2 * i) + 1]; | |
1148 | ||
1149 | /* linemap_add_macro_token encodes token numbers in an expansion | |
1150 | by putting them after MAP_START_LOCATION. */ | |
1151 | ||
1152 | /* I'm typically seeing 4 uninitialized entries at the end of | |
1153 | 0xafafafaf. | |
1154 | This appears to be due to macro.c:replace_args | |
1155 | adding 2 extra args for padding tokens; presumably there may | |
1156 | be a leading and/or trailing padding token injected, | |
1157 | each for 2 more location slots. | |
1158 | This would explain there being up to 4 source_locations slots | |
1159 | that may be uninitialized. */ | |
1160 | ||
1161 | fprintf (stream, " %u: %u, %u\n", | |
1162 | i, | |
1163 | x, | |
1164 | y); | |
1165 | if (x == y) | |
1166 | { | |
1167 | if (x < MAP_START_LOCATION (map)) | |
1168 | inform (x, "token %u has x-location == y-location == %u", i, x); | |
1169 | else | |
1170 | fprintf (stream, | |
1171 | "x-location == y-location == %u encodes token # %u\n", | |
1172 | x, x - MAP_START_LOCATION (map)); | |
1173 | } | |
1174 | else | |
1175 | { | |
1176 | inform (x, "token %u has x-location == %u", i, x); | |
1177 | inform (x, "token %u has y-location == %u", i, y); | |
1178 | } | |
1179 | } | |
1180 | fprintf (stream, "\n"); | |
1181 | } | |
1182 | ||
1183 | /* It appears that MAX_SOURCE_LOCATION itself is never assigned to a | |
1184 | macro map, presumably due to an off-by-one error somewhere | |
1185 | between the logic in linemap_enter_macro and | |
1186 | LINEMAPS_MACRO_LOWEST_LOCATION. */ | |
1187 | dump_labelled_location_range (stream, "MAX_SOURCE_LOCATION", | |
1188 | MAX_SOURCE_LOCATION, | |
1189 | MAX_SOURCE_LOCATION + 1); | |
1190 | ||
1191 | /* Visualize ad-hoc values. */ | |
1192 | dump_labelled_location_range (stream, "AD-HOC LOCATIONS", | |
1193 | MAX_SOURCE_LOCATION + 1, UINT_MAX); | |
1194 | } | |
d9b950dd | 1195 | |
88fa5555 DM |
1196 | /* string_concat's constructor. */ |
1197 | ||
1198 | string_concat::string_concat (int num, location_t *locs) | |
1199 | : m_num (num) | |
1200 | { | |
1201 | m_locs = ggc_vec_alloc <location_t> (num); | |
1202 | for (int i = 0; i < num; i++) | |
1203 | m_locs[i] = locs[i]; | |
1204 | } | |
1205 | ||
1206 | /* string_concat_db's constructor. */ | |
1207 | ||
1208 | string_concat_db::string_concat_db () | |
1209 | { | |
1210 | m_table = hash_map <location_hash, string_concat *>::create_ggc (64); | |
1211 | } | |
1212 | ||
1213 | /* Record that a string concatenation occurred, covering NUM | |
1214 | string literal tokens. LOCS is an array of size NUM, containing the | |
1215 | locations of the tokens. A copy of LOCS is taken. */ | |
1216 | ||
1217 | void | |
1218 | string_concat_db::record_string_concatenation (int num, location_t *locs) | |
1219 | { | |
1220 | gcc_assert (num > 1); | |
1221 | gcc_assert (locs); | |
1222 | ||
1223 | location_t key_loc = get_key_loc (locs[0]); | |
1224 | ||
1225 | string_concat *concat | |
1226 | = new (ggc_alloc <string_concat> ()) string_concat (num, locs); | |
1227 | m_table->put (key_loc, concat); | |
1228 | } | |
1229 | ||
1230 | /* Determine if LOC was the location of the the initial token of a | |
1231 | concatenation of string literal tokens. | |
1232 | If so, *OUT_NUM is written to with the number of tokens, and | |
1233 | *OUT_LOCS with the location of an array of locations of the | |
1234 | tokens, and return true. *OUT_LOCS is a borrowed pointer to | |
1235 | storage owned by the string_concat_db. | |
1236 | Otherwise, return false. */ | |
1237 | ||
1238 | bool | |
1239 | string_concat_db::get_string_concatenation (location_t loc, | |
1240 | int *out_num, | |
1241 | location_t **out_locs) | |
1242 | { | |
1243 | gcc_assert (out_num); | |
1244 | gcc_assert (out_locs); | |
1245 | ||
1246 | location_t key_loc = get_key_loc (loc); | |
1247 | ||
1248 | string_concat **concat = m_table->get (key_loc); | |
1249 | if (!concat) | |
1250 | return false; | |
1251 | ||
1252 | *out_num = (*concat)->m_num; | |
1253 | *out_locs =(*concat)->m_locs; | |
1254 | return true; | |
1255 | } | |
1256 | ||
1257 | /* Internal function. Canonicalize LOC into a form suitable for | |
1258 | use as a key within the database, stripping away macro expansion, | |
1259 | ad-hoc information, and range information, using the location of | |
1260 | the start of LOC within an ordinary linemap. */ | |
1261 | ||
1262 | location_t | |
1263 | string_concat_db::get_key_loc (location_t loc) | |
1264 | { | |
1265 | loc = linemap_resolve_location (line_table, loc, LRK_SPELLING_LOCATION, | |
1266 | NULL); | |
1267 | ||
1268 | loc = get_range_from_loc (line_table, loc).m_start; | |
1269 | ||
1270 | return loc; | |
1271 | } | |
1272 | ||
1273 | /* Helper class for use within get_substring_ranges_for_loc. | |
1274 | An vec of cpp_string with responsibility for releasing all of the | |
1275 | str->text for each str in the vector. */ | |
1276 | ||
1277 | class auto_cpp_string_vec : public auto_vec <cpp_string> | |
1278 | { | |
1279 | public: | |
1280 | auto_cpp_string_vec (int alloc) | |
1281 | : auto_vec <cpp_string> (alloc) {} | |
1282 | ||
1283 | ~auto_cpp_string_vec () | |
1284 | { | |
1285 | /* Clean up the copies within this vec. */ | |
1286 | int i; | |
1287 | cpp_string *str; | |
1288 | FOR_EACH_VEC_ELT (*this, i, str) | |
1289 | free (const_cast <unsigned char *> (str->text)); | |
1290 | } | |
1291 | }; | |
1292 | ||
1293 | /* Attempt to populate RANGES with source location information on the | |
1294 | individual characters within the string literal found at STRLOC. | |
1295 | If CONCATS is non-NULL, then any string literals that the token at | |
1296 | STRLOC was concatenated with are also added to RANGES. | |
1297 | ||
1298 | Return NULL if successful, or an error message if any errors occurred (in | |
1299 | which case RANGES may be only partially populated and should not | |
1300 | be used). | |
1301 | ||
1302 | This is implemented by re-parsing the relevant source line(s). */ | |
1303 | ||
1304 | static const char * | |
1305 | get_substring_ranges_for_loc (cpp_reader *pfile, | |
1306 | string_concat_db *concats, | |
1307 | location_t strloc, | |
1308 | enum cpp_ttype type, | |
1309 | cpp_substring_ranges &ranges) | |
1310 | { | |
1311 | gcc_assert (pfile); | |
1312 | ||
1313 | if (strloc == UNKNOWN_LOCATION) | |
1314 | return "unknown location"; | |
1315 | ||
1316 | /* If string concatenation has occurred at STRLOC, get the locations | |
1317 | of all of the literal tokens making up the compound string. | |
1318 | Otherwise, just use STRLOC. */ | |
1319 | int num_locs = 1; | |
1320 | location_t *strlocs = &strloc; | |
1321 | if (concats) | |
1322 | concats->get_string_concatenation (strloc, &num_locs, &strlocs); | |
1323 | ||
1324 | auto_cpp_string_vec strs (num_locs); | |
1325 | auto_vec <cpp_string_location_reader> loc_readers (num_locs); | |
1326 | for (int i = 0; i < num_locs; i++) | |
1327 | { | |
1328 | /* Get range of strloc. We will use it to locate the start and finish | |
1329 | of the literal token within the line. */ | |
1330 | source_range src_range = get_range_from_loc (line_table, strlocs[i]); | |
1331 | ||
1332 | if (src_range.m_start >= LINEMAPS_MACRO_LOWEST_LOCATION (line_table)) | |
1333 | /* If the string is within a macro expansion, we can't get at the | |
1334 | end location. */ | |
1335 | return "macro expansion"; | |
1336 | ||
1337 | if (src_range.m_start >= LINE_MAP_MAX_LOCATION_WITH_COLS) | |
1338 | /* If so, we can't reliably determine where the token started within | |
1339 | its line. */ | |
1340 | return "range starts after LINE_MAP_MAX_LOCATION_WITH_COLS"; | |
1341 | ||
1342 | if (src_range.m_finish >= LINE_MAP_MAX_LOCATION_WITH_COLS) | |
1343 | /* If so, we can't reliably determine where the token finished within | |
1344 | its line. */ | |
1345 | return "range ends after LINE_MAP_MAX_LOCATION_WITH_COLS"; | |
1346 | ||
1347 | expanded_location start | |
1348 | = expand_location_to_spelling_point (src_range.m_start); | |
1349 | expanded_location finish | |
1350 | = expand_location_to_spelling_point (src_range.m_finish); | |
1351 | if (start.file != finish.file) | |
1352 | return "range endpoints are in different files"; | |
1353 | if (start.line != finish.line) | |
1354 | return "range endpoints are on different lines"; | |
1355 | if (start.column > finish.column) | |
1356 | return "range endpoints are reversed"; | |
1357 | ||
1358 | int line_width; | |
1359 | const char *line = location_get_source_line (start.file, start.line, | |
1360 | &line_width); | |
1361 | if (line == NULL) | |
1362 | return "unable to read source line"; | |
1363 | ||
1364 | /* Determine the location of the literal (including quotes | |
1365 | and leading prefix chars, such as the 'u' in a u"" | |
1366 | token). */ | |
1367 | const char *literal = line + start.column - 1; | |
1368 | int literal_length = finish.column - start.column + 1; | |
1369 | ||
1370 | gcc_assert (line_width >= (start.column - 1 + literal_length)); | |
1371 | cpp_string from; | |
1372 | from.len = literal_length; | |
1373 | /* Make a copy of the literal, to avoid having to rely on | |
1374 | the lifetime of the copy of the line within the cache. | |
1375 | This will be released by the auto_cpp_string_vec dtor. */ | |
1376 | from.text = XDUPVEC (unsigned char, literal, literal_length); | |
1377 | strs.safe_push (from); | |
1378 | ||
1379 | /* For very long lines, a new linemap could have started | |
1380 | halfway through the token. | |
1381 | Ensure that the loc_reader uses the linemap of the | |
1382 | *end* of the token for its start location. */ | |
1383 | const line_map_ordinary *final_ord_map; | |
1384 | linemap_resolve_location (line_table, src_range.m_finish, | |
1385 | LRK_MACRO_EXPANSION_POINT, &final_ord_map); | |
1386 | location_t start_loc | |
1387 | = linemap_position_for_line_and_column (line_table, final_ord_map, | |
1388 | start.line, start.column); | |
1389 | ||
1390 | cpp_string_location_reader loc_reader (start_loc, line_table); | |
1391 | loc_readers.safe_push (loc_reader); | |
1392 | } | |
1393 | ||
1394 | /* Rerun cpp_interpret_string, or rather, a modified version of it. */ | |
1395 | const char *err = cpp_interpret_string_ranges (pfile, strs.address (), | |
1396 | loc_readers.address (), | |
1397 | num_locs, &ranges, type); | |
1398 | if (err) | |
1399 | return err; | |
1400 | ||
1401 | /* Success: "ranges" should now contain information on the string. */ | |
1402 | return NULL; | |
1403 | } | |
1404 | ||
65e736c0 DM |
1405 | /* Attempt to populate *OUT_LOC with source location information on the |
1406 | given characters within the string literal found at STRLOC. | |
1407 | CARET_IDX, START_IDX, and END_IDX refer to offsets within the execution | |
1408 | character set. | |
1409 | ||
1410 | For example, given CARET_IDX = 4, START_IDX = 3, END_IDX = 7 | |
1411 | and string literal "012345\n789" | |
1412 | *OUT_LOC is written to with: | |
1413 | "012345\n789" | |
1414 | ~^~~~~ | |
1415 | ||
88fa5555 DM |
1416 | If CONCATS is non-NULL, then any string literals that the token at |
1417 | STRLOC was concatenated with are also considered. | |
1418 | ||
1419 | This is implemented by re-parsing the relevant source line(s). | |
1420 | ||
1421 | Return NULL if successful, or an error message if any errors occurred. | |
1422 | Error messages are intended for GCC developers (to help debugging) rather | |
1423 | than for end-users. */ | |
1424 | ||
1425 | const char * | |
65e736c0 DM |
1426 | get_source_location_for_substring (cpp_reader *pfile, |
1427 | string_concat_db *concats, | |
1428 | location_t strloc, | |
1429 | enum cpp_ttype type, | |
1430 | int caret_idx, int start_idx, int end_idx, | |
1431 | source_location *out_loc) | |
1432 | { | |
1433 | gcc_checking_assert (caret_idx >= 0); | |
88fa5555 DM |
1434 | gcc_checking_assert (start_idx >= 0); |
1435 | gcc_checking_assert (end_idx >= 0); | |
65e736c0 | 1436 | gcc_assert (out_loc); |
88fa5555 DM |
1437 | |
1438 | cpp_substring_ranges ranges; | |
1439 | const char *err | |
1440 | = get_substring_ranges_for_loc (pfile, concats, strloc, type, ranges); | |
1441 | if (err) | |
1442 | return err; | |
1443 | ||
65e736c0 DM |
1444 | if (caret_idx >= ranges.get_num_ranges ()) |
1445 | return "caret_idx out of range"; | |
88fa5555 DM |
1446 | if (start_idx >= ranges.get_num_ranges ()) |
1447 | return "start_idx out of range"; | |
1448 | if (end_idx >= ranges.get_num_ranges ()) | |
1449 | return "end_idx out of range"; | |
1450 | ||
65e736c0 DM |
1451 | *out_loc = make_location (ranges.get_range (caret_idx).m_start, |
1452 | ranges.get_range (start_idx).m_start, | |
1453 | ranges.get_range (end_idx).m_finish); | |
1454 | return NULL; | |
1455 | } | |
1456 | ||
0e06d2b3 DM |
1457 | #if CHECKING_P |
1458 | ||
1459 | namespace selftest { | |
1460 | ||
1461 | /* Selftests of location handling. */ | |
1462 | ||
65e736c0 DM |
1463 | /* Attempt to populate *OUT_RANGE with source location information on the |
1464 | given character within the string literal found at STRLOC. | |
1465 | CHAR_IDX refers to an offset within the execution character set. | |
1466 | If CONCATS is non-NULL, then any string literals that the token at | |
1467 | STRLOC was concatenated with are also considered. | |
1468 | ||
1469 | This is implemented by re-parsing the relevant source line(s). | |
1470 | ||
1471 | Return NULL if successful, or an error message if any errors occurred. | |
1472 | Error messages are intended for GCC developers (to help debugging) rather | |
1473 | than for end-users. */ | |
1474 | ||
1475 | static const char * | |
1476 | get_source_range_for_char (cpp_reader *pfile, | |
1477 | string_concat_db *concats, | |
1478 | location_t strloc, | |
1479 | enum cpp_ttype type, | |
1480 | int char_idx, | |
1481 | source_range *out_range) | |
1482 | { | |
1483 | gcc_checking_assert (char_idx >= 0); | |
1484 | gcc_assert (out_range); | |
1485 | ||
1486 | cpp_substring_ranges ranges; | |
1487 | const char *err | |
1488 | = get_substring_ranges_for_loc (pfile, concats, strloc, type, ranges); | |
1489 | if (err) | |
1490 | return err; | |
1491 | ||
1492 | if (char_idx >= ranges.get_num_ranges ()) | |
1493 | return "char_idx out of range"; | |
1494 | ||
1495 | *out_range = ranges.get_range (char_idx); | |
88fa5555 DM |
1496 | return NULL; |
1497 | } | |
1498 | ||
65e736c0 | 1499 | /* As get_source_range_for_char, but write to *OUT the number |
88fa5555 DM |
1500 | of ranges that are available. */ |
1501 | ||
0e06d2b3 | 1502 | static const char * |
88fa5555 DM |
1503 | get_num_source_ranges_for_substring (cpp_reader *pfile, |
1504 | string_concat_db *concats, | |
1505 | location_t strloc, | |
1506 | enum cpp_ttype type, | |
1507 | int *out) | |
1508 | { | |
1509 | gcc_assert (out); | |
1510 | ||
1511 | cpp_substring_ranges ranges; | |
1512 | const char *err | |
1513 | = get_substring_ranges_for_loc (pfile, concats, strloc, type, ranges); | |
1514 | ||
1515 | if (err) | |
1516 | return err; | |
1517 | ||
1518 | *out = ranges.get_num_ranges (); | |
1519 | return NULL; | |
1520 | } | |
1521 | ||
d9b950dd DM |
1522 | /* Selftests of location handling. */ |
1523 | ||
741d3be5 DM |
1524 | /* Helper function for verifying location data: when location_t |
1525 | values are > LINE_MAP_MAX_LOCATION_WITH_COLS, they are treated | |
1526 | as having column 0. */ | |
1527 | ||
1528 | static bool | |
1529 | should_have_column_data_p (location_t loc) | |
1530 | { | |
1531 | if (IS_ADHOC_LOC (loc)) | |
1532 | loc = get_location_from_adhoc_loc (line_table, loc); | |
1533 | if (loc > LINE_MAP_MAX_LOCATION_WITH_COLS) | |
1534 | return false; | |
1535 | return true; | |
1536 | } | |
1537 | ||
1538 | /* Selftest for should_have_column_data_p. */ | |
1539 | ||
1540 | static void | |
1541 | test_should_have_column_data_p () | |
1542 | { | |
1543 | ASSERT_TRUE (should_have_column_data_p (RESERVED_LOCATION_COUNT)); | |
1544 | ASSERT_TRUE | |
1545 | (should_have_column_data_p (LINE_MAP_MAX_LOCATION_WITH_COLS)); | |
1546 | ASSERT_FALSE | |
1547 | (should_have_column_data_p (LINE_MAP_MAX_LOCATION_WITH_COLS + 1)); | |
1548 | } | |
1549 | ||
d9b950dd DM |
1550 | /* Verify the result of LOCATION_FILE/LOCATION_LINE/LOCATION_COLUMN |
1551 | on LOC. */ | |
1552 | ||
1553 | static void | |
1554 | assert_loceq (const char *exp_filename, int exp_linenum, int exp_colnum, | |
1555 | location_t loc) | |
1556 | { | |
1557 | ASSERT_STREQ (exp_filename, LOCATION_FILE (loc)); | |
1558 | ASSERT_EQ (exp_linenum, LOCATION_LINE (loc)); | |
741d3be5 DM |
1559 | /* If location_t values are sufficiently high, then column numbers |
1560 | will be unavailable and LOCATION_COLUMN (loc) will be 0. | |
1561 | When close to the threshold, column numbers *may* be present: if | |
1562 | the final linemap before the threshold contains a line that straddles | |
1563 | the threshold, locations in that line have column information. */ | |
1564 | if (should_have_column_data_p (loc)) | |
1565 | ASSERT_EQ (exp_colnum, LOCATION_COLUMN (loc)); | |
1566 | } | |
1567 | ||
1568 | /* Various selftests in this file involve constructing a line table | |
1569 | and one or more line maps within it. | |
1570 | ||
1571 | For maximum test coverage we want to run these tests with a variety | |
1572 | of situations: | |
1573 | - line_table->default_range_bits: some frontends use a non-zero value | |
1574 | and others use zero | |
1575 | - the fallback modes within line-map.c: there are various threshold | |
1576 | values for source_location/location_t beyond line-map.c changes | |
1577 | behavior (disabling of the range-packing optimization, disabling | |
1578 | of column-tracking). We can exercise these by starting the line_table | |
1579 | at interesting values at or near these thresholds. | |
1580 | ||
1581 | The following struct describes a particular case within our test | |
1582 | matrix. */ | |
1583 | ||
1584 | struct line_table_case | |
1585 | { | |
1586 | line_table_case (int default_range_bits, int base_location) | |
1587 | : m_default_range_bits (default_range_bits), | |
1588 | m_base_location (base_location) | |
1589 | {} | |
1590 | ||
1591 | int m_default_range_bits; | |
1592 | int m_base_location; | |
1593 | }; | |
1594 | ||
1595 | /* A class for overriding the global "line_table" within a selftest, | |
1596 | restoring its value afterwards. */ | |
1597 | ||
1598 | class temp_line_table | |
1599 | { | |
1600 | public: | |
1601 | temp_line_table (const line_table_case &); | |
1602 | ~temp_line_table (); | |
1603 | ||
1604 | private: | |
1605 | line_maps *m_old_line_table; | |
1606 | }; | |
1607 | ||
1608 | /* Constructor. Store the old value of line_table, and create a new | |
1609 | one, using the sitation described in CASE_. */ | |
1610 | ||
1611 | temp_line_table::temp_line_table (const line_table_case &case_) | |
1612 | : m_old_line_table (line_table) | |
1613 | { | |
1614 | line_table = ggc_alloc<line_maps> (); | |
1615 | linemap_init (line_table, BUILTINS_LOCATION); | |
1616 | line_table->reallocator = m_old_line_table->reallocator; | |
1617 | line_table->round_alloc_size = m_old_line_table->round_alloc_size; | |
1618 | line_table->default_range_bits = case_.m_default_range_bits; | |
1619 | if (case_.m_base_location) | |
1620 | { | |
1621 | line_table->highest_location = case_.m_base_location; | |
1622 | line_table->highest_line = case_.m_base_location; | |
1623 | } | |
1624 | } | |
1625 | ||
1626 | /* Destructor. Restore the old value of line_table. */ | |
1627 | ||
1628 | temp_line_table::~temp_line_table () | |
1629 | { | |
1630 | line_table = m_old_line_table; | |
d9b950dd DM |
1631 | } |
1632 | ||
1633 | /* Verify basic operation of ordinary linemaps. */ | |
1634 | ||
1635 | static void | |
741d3be5 | 1636 | test_accessing_ordinary_linemaps (const line_table_case &case_) |
d9b950dd | 1637 | { |
741d3be5 DM |
1638 | temp_line_table tmp_lt (case_); |
1639 | ||
d9b950dd DM |
1640 | /* Build a simple linemap describing some locations. */ |
1641 | linemap_add (line_table, LC_ENTER, false, "foo.c", 0); | |
1642 | ||
1643 | linemap_line_start (line_table, 1, 100); | |
1644 | location_t loc_a = linemap_position_for_column (line_table, 1); | |
1645 | location_t loc_b = linemap_position_for_column (line_table, 23); | |
1646 | ||
1647 | linemap_line_start (line_table, 2, 100); | |
1648 | location_t loc_c = linemap_position_for_column (line_table, 1); | |
1649 | location_t loc_d = linemap_position_for_column (line_table, 17); | |
1650 | ||
1651 | /* Example of a very long line. */ | |
1652 | linemap_line_start (line_table, 3, 2000); | |
1653 | location_t loc_e = linemap_position_for_column (line_table, 700); | |
1654 | ||
1655 | linemap_add (line_table, LC_LEAVE, false, NULL, 0); | |
1656 | ||
1657 | /* Multiple files. */ | |
1658 | linemap_add (line_table, LC_ENTER, false, "bar.c", 0); | |
1659 | linemap_line_start (line_table, 1, 200); | |
1660 | location_t loc_f = linemap_position_for_column (line_table, 150); | |
1661 | linemap_add (line_table, LC_LEAVE, false, NULL, 0); | |
1662 | ||
1663 | /* Verify that we can recover the location info. */ | |
1664 | assert_loceq ("foo.c", 1, 1, loc_a); | |
1665 | assert_loceq ("foo.c", 1, 23, loc_b); | |
1666 | assert_loceq ("foo.c", 2, 1, loc_c); | |
1667 | assert_loceq ("foo.c", 2, 17, loc_d); | |
1668 | assert_loceq ("foo.c", 3, 700, loc_e); | |
1669 | assert_loceq ("bar.c", 1, 150, loc_f); | |
1670 | ||
1671 | ASSERT_FALSE (is_location_from_builtin_token (loc_a)); | |
a01fc549 DM |
1672 | ASSERT_TRUE (pure_location_p (line_table, loc_a)); |
1673 | ||
1674 | /* Verify using make_location to build a range, and extracting data | |
1675 | back from it. */ | |
1676 | location_t range_c_b_d = make_location (loc_c, loc_b, loc_d); | |
1677 | ASSERT_FALSE (pure_location_p (line_table, range_c_b_d)); | |
1678 | ASSERT_EQ (loc_c, get_location_from_adhoc_loc (line_table, range_c_b_d)); | |
1679 | source_range src_range = get_range_from_loc (line_table, range_c_b_d); | |
1680 | ASSERT_EQ (loc_b, src_range.m_start); | |
1681 | ASSERT_EQ (loc_d, src_range.m_finish); | |
d9b950dd DM |
1682 | } |
1683 | ||
1684 | /* Verify various properties of UNKNOWN_LOCATION. */ | |
1685 | ||
1686 | static void | |
1687 | test_unknown_location () | |
1688 | { | |
1689 | ASSERT_EQ (NULL, LOCATION_FILE (UNKNOWN_LOCATION)); | |
1690 | ASSERT_EQ (0, LOCATION_LINE (UNKNOWN_LOCATION)); | |
1691 | ASSERT_EQ (0, LOCATION_COLUMN (UNKNOWN_LOCATION)); | |
1692 | } | |
1693 | ||
1694 | /* Verify various properties of BUILTINS_LOCATION. */ | |
1695 | ||
1696 | static void | |
1697 | test_builtins () | |
1698 | { | |
10d2fc23 | 1699 | assert_loceq (_("<built-in>"), 0, 0, BUILTINS_LOCATION); |
d9b950dd DM |
1700 | ASSERT_PRED1 (is_location_from_builtin_token, BUILTINS_LOCATION); |
1701 | } | |
1702 | ||
1703 | /* Verify reading of input files (e.g. for caret-based diagnostics). */ | |
1704 | ||
1705 | static void | |
1706 | test_reading_source_line () | |
1707 | { | |
85ecd05c | 1708 | /* Create a tempfile and write some text to it. */ |
741d3be5 DM |
1709 | temp_source_file tmp (SELFTEST_LOCATION, ".txt", |
1710 | "01234567890123456789\n" | |
1711 | "This is the test text\n" | |
1712 | "This is the 3rd line\n"); | |
85ecd05c DM |
1713 | |
1714 | /* Read back a specific line from the tempfile. */ | |
d9b950dd | 1715 | int line_size; |
741d3be5 DM |
1716 | const char *source_line = location_get_source_line (tmp.get_filename (), |
1717 | 2, &line_size); | |
d9b950dd | 1718 | ASSERT_TRUE (source_line != NULL); |
85ecd05c DM |
1719 | ASSERT_EQ (21, line_size); |
1720 | if (!strncmp ("This is the test text", | |
1721 | source_line, line_size)) | |
09765e3a | 1722 | ::selftest::pass (SELFTEST_LOCATION, |
d9b950dd DM |
1723 | "source_line matched expected value"); |
1724 | else | |
09765e3a | 1725 | ::selftest::fail (SELFTEST_LOCATION, |
d9b950dd | 1726 | "source_line did not match expected value"); |
85ecd05c | 1727 | |
d9b950dd DM |
1728 | } |
1729 | ||
741d3be5 DM |
1730 | /* Tests of lexing. */ |
1731 | ||
1732 | /* Verify that token TOK from PARSER has cpp_token_as_text | |
1733 | equal to EXPECTED_TEXT. */ | |
1734 | ||
1735 | #define ASSERT_TOKEN_AS_TEXT_EQ(PARSER, TOK, EXPECTED_TEXT) \ | |
1736 | SELFTEST_BEGIN_STMT \ | |
1737 | unsigned char *actual_txt = cpp_token_as_text ((PARSER), (TOK)); \ | |
1738 | ASSERT_STREQ ((EXPECTED_TEXT), (const char *)actual_txt); \ | |
1739 | SELFTEST_END_STMT | |
1740 | ||
1741 | /* Verify that TOK's src_loc is within EXP_FILENAME at EXP_LINENUM, | |
1742 | and ranges from EXP_START_COL to EXP_FINISH_COL. | |
1743 | Use LOC as the effective location of the selftest. */ | |
1744 | ||
1745 | static void | |
1746 | assert_token_loc_eq (const location &loc, | |
1747 | const cpp_token *tok, | |
1748 | const char *exp_filename, int exp_linenum, | |
1749 | int exp_start_col, int exp_finish_col) | |
1750 | { | |
1751 | location_t tok_loc = tok->src_loc; | |
1752 | ASSERT_STREQ_AT (loc, exp_filename, LOCATION_FILE (tok_loc)); | |
1753 | ASSERT_EQ_AT (loc, exp_linenum, LOCATION_LINE (tok_loc)); | |
1754 | ||
1755 | /* If location_t values are sufficiently high, then column numbers | |
1756 | will be unavailable. */ | |
1757 | if (!should_have_column_data_p (tok_loc)) | |
1758 | return; | |
1759 | ||
1760 | ASSERT_EQ_AT (loc, exp_start_col, LOCATION_COLUMN (tok_loc)); | |
1761 | source_range tok_range = get_range_from_loc (line_table, tok_loc); | |
1762 | ASSERT_EQ_AT (loc, exp_start_col, LOCATION_COLUMN (tok_range.m_start)); | |
1763 | ASSERT_EQ_AT (loc, exp_finish_col, LOCATION_COLUMN (tok_range.m_finish)); | |
1764 | } | |
1765 | ||
1766 | /* Use assert_token_loc_eq to verify the TOK->src_loc, using | |
1767 | SELFTEST_LOCATION as the effective location of the selftest. */ | |
1768 | ||
1769 | #define ASSERT_TOKEN_LOC_EQ(TOK, EXP_FILENAME, EXP_LINENUM, \ | |
1770 | EXP_START_COL, EXP_FINISH_COL) \ | |
1771 | assert_token_loc_eq (SELFTEST_LOCATION, (TOK), (EXP_FILENAME), \ | |
1772 | (EXP_LINENUM), (EXP_START_COL), (EXP_FINISH_COL)) | |
1773 | ||
1774 | /* Test of lexing a file using libcpp, verifying tokens and their | |
1775 | location information. */ | |
1776 | ||
1777 | static void | |
1778 | test_lexer (const line_table_case &case_) | |
1779 | { | |
1780 | /* Create a tempfile and write some text to it. */ | |
1781 | const char *content = | |
1782 | /*00000000011111111112222222222333333.3333444444444.455555555556 | |
1783 | 12345678901234567890123456789012345.6789012345678.901234567890. */ | |
1784 | ("test_name /* c-style comment */\n" | |
1785 | " \"test literal\"\n" | |
1786 | " // test c++-style comment\n" | |
1787 | " 42\n"); | |
1788 | temp_source_file tmp (SELFTEST_LOCATION, ".txt", content); | |
1789 | ||
1790 | temp_line_table tmp_lt (case_); | |
1791 | ||
1792 | cpp_reader *parser = cpp_create_reader (CLK_GNUC89, NULL, line_table); | |
1793 | ||
1794 | const char *fname = cpp_read_main_file (parser, tmp.get_filename ()); | |
1795 | ASSERT_NE (fname, NULL); | |
1796 | ||
1797 | /* Verify that we get the expected tokens back, with the correct | |
1798 | location information. */ | |
1799 | ||
1800 | location_t loc; | |
1801 | const cpp_token *tok; | |
1802 | tok = cpp_get_token_with_location (parser, &loc); | |
1803 | ASSERT_NE (tok, NULL); | |
1804 | ASSERT_EQ (tok->type, CPP_NAME); | |
1805 | ASSERT_TOKEN_AS_TEXT_EQ (parser, tok, "test_name"); | |
1806 | ASSERT_TOKEN_LOC_EQ (tok, tmp.get_filename (), 1, 1, 9); | |
1807 | ||
1808 | tok = cpp_get_token_with_location (parser, &loc); | |
1809 | ASSERT_NE (tok, NULL); | |
1810 | ASSERT_EQ (tok->type, CPP_STRING); | |
1811 | ASSERT_TOKEN_AS_TEXT_EQ (parser, tok, "\"test literal\""); | |
1812 | ASSERT_TOKEN_LOC_EQ (tok, tmp.get_filename (), 2, 35, 48); | |
1813 | ||
1814 | tok = cpp_get_token_with_location (parser, &loc); | |
1815 | ASSERT_NE (tok, NULL); | |
1816 | ASSERT_EQ (tok->type, CPP_NUMBER); | |
1817 | ASSERT_TOKEN_AS_TEXT_EQ (parser, tok, "42"); | |
1818 | ASSERT_TOKEN_LOC_EQ (tok, tmp.get_filename (), 4, 4, 5); | |
1819 | ||
1820 | tok = cpp_get_token_with_location (parser, &loc); | |
1821 | ASSERT_NE (tok, NULL); | |
1822 | ASSERT_EQ (tok->type, CPP_EOF); | |
1823 | ||
1824 | cpp_finish (parser, NULL); | |
1825 | cpp_destroy (parser); | |
1826 | } | |
1827 | ||
88fa5555 DM |
1828 | /* Forward decls. */ |
1829 | ||
1830 | struct lexer_test; | |
1831 | class lexer_test_options; | |
1832 | ||
1833 | /* A class for specifying options of a lexer_test. | |
1834 | The "apply" vfunc is called during the lexer_test constructor. */ | |
1835 | ||
1836 | class lexer_test_options | |
1837 | { | |
1838 | public: | |
1839 | virtual void apply (lexer_test &) = 0; | |
1840 | }; | |
1841 | ||
1842 | /* A struct for writing lexer tests. */ | |
1843 | ||
1844 | struct lexer_test | |
1845 | { | |
1846 | lexer_test (const line_table_case &case_, const char *content, | |
1847 | lexer_test_options *options); | |
1848 | ~lexer_test (); | |
1849 | ||
1850 | const cpp_token *get_token (); | |
1851 | ||
1852 | temp_source_file m_tempfile; | |
1853 | temp_line_table m_tmp_lt; | |
1854 | cpp_reader *m_parser; | |
1855 | string_concat_db m_concats; | |
1856 | }; | |
1857 | ||
1858 | /* Use an EBCDIC encoding for the execution charset, specifically | |
1859 | IBM1047-encoded (aka "EBCDIC 1047", or "Code page 1047"). | |
1860 | ||
1861 | This exercises iconv integration within libcpp. | |
1862 | Not every build of iconv supports the given charset, | |
1863 | so we need to flag this error and handle it gracefully. */ | |
1864 | ||
1865 | class ebcdic_execution_charset : public lexer_test_options | |
1866 | { | |
1867 | public: | |
1868 | ebcdic_execution_charset () : m_num_iconv_errors (0) | |
1869 | { | |
1870 | gcc_assert (s_singleton == NULL); | |
1871 | s_singleton = this; | |
1872 | } | |
1873 | ~ebcdic_execution_charset () | |
1874 | { | |
1875 | gcc_assert (s_singleton == this); | |
1876 | s_singleton = NULL; | |
1877 | } | |
1878 | ||
1879 | void apply (lexer_test &test) FINAL OVERRIDE | |
1880 | { | |
1881 | cpp_options *cpp_opts = cpp_get_options (test.m_parser); | |
1882 | cpp_opts->narrow_charset = "IBM1047"; | |
1883 | ||
1884 | cpp_callbacks *callbacks = cpp_get_callbacks (test.m_parser); | |
1885 | callbacks->error = on_error; | |
1886 | } | |
1887 | ||
1888 | static bool on_error (cpp_reader *pfile ATTRIBUTE_UNUSED, | |
1889 | int level ATTRIBUTE_UNUSED, | |
1890 | int reason ATTRIBUTE_UNUSED, | |
1891 | rich_location *richloc ATTRIBUTE_UNUSED, | |
1892 | const char *msgid, va_list *ap ATTRIBUTE_UNUSED) | |
1893 | ATTRIBUTE_FPTR_PRINTF(5,0) | |
1894 | { | |
1895 | gcc_assert (s_singleton); | |
1896 | /* Detect and record errors emitted by libcpp/charset.c:init_iconv_desc | |
1897 | when the local iconv build doesn't support the conversion. */ | |
1898 | if (strstr (msgid, "not supported by iconv")) | |
1899 | { | |
1900 | s_singleton->m_num_iconv_errors++; | |
1901 | return true; | |
1902 | } | |
1903 | ||
1904 | /* Otherwise, we have an unexpected error. */ | |
1905 | abort (); | |
1906 | } | |
1907 | ||
1908 | bool iconv_errors_occurred_p () const { return m_num_iconv_errors > 0; } | |
1909 | ||
1910 | private: | |
1911 | static ebcdic_execution_charset *s_singleton; | |
1912 | int m_num_iconv_errors; | |
1913 | }; | |
1914 | ||
1915 | ebcdic_execution_charset *ebcdic_execution_charset::s_singleton; | |
1916 | ||
1917 | /* Constructor. Override line_table with a new instance based on CASE_, | |
1918 | and write CONTENT to a tempfile. Create a cpp_reader, and use it to | |
1919 | start parsing the tempfile. */ | |
1920 | ||
1921 | lexer_test::lexer_test (const line_table_case &case_, const char *content, | |
1922 | lexer_test_options *options) : | |
1923 | /* Create a tempfile and write the text to it. */ | |
1924 | m_tempfile (SELFTEST_LOCATION, ".c", content), | |
1925 | m_tmp_lt (case_), | |
1926 | m_parser (cpp_create_reader (CLK_GNUC99, NULL, line_table)), | |
1927 | m_concats () | |
1928 | { | |
1929 | if (options) | |
1930 | options->apply (*this); | |
1931 | ||
1932 | cpp_init_iconv (m_parser); | |
1933 | ||
1934 | /* Parse the file. */ | |
1935 | const char *fname = cpp_read_main_file (m_parser, | |
1936 | m_tempfile.get_filename ()); | |
1937 | ASSERT_NE (fname, NULL); | |
1938 | } | |
1939 | ||
1940 | /* Destructor. Verify that the next token in m_parser is EOF. */ | |
1941 | ||
1942 | lexer_test::~lexer_test () | |
1943 | { | |
1944 | location_t loc; | |
1945 | const cpp_token *tok; | |
1946 | ||
1947 | tok = cpp_get_token_with_location (m_parser, &loc); | |
1948 | ASSERT_NE (tok, NULL); | |
1949 | ASSERT_EQ (tok->type, CPP_EOF); | |
1950 | ||
1951 | cpp_finish (m_parser, NULL); | |
1952 | cpp_destroy (m_parser); | |
1953 | } | |
1954 | ||
1955 | /* Get the next token from m_parser. */ | |
1956 | ||
1957 | const cpp_token * | |
1958 | lexer_test::get_token () | |
1959 | { | |
1960 | location_t loc; | |
1961 | const cpp_token *tok; | |
1962 | ||
1963 | tok = cpp_get_token_with_location (m_parser, &loc); | |
1964 | ASSERT_NE (tok, NULL); | |
1965 | return tok; | |
1966 | } | |
1967 | ||
1968 | /* Verify that locations within string literals are correctly handled. */ | |
1969 | ||
1970 | /* Verify get_source_range_for_substring for token(s) at STRLOC, | |
1971 | using the string concatenation database for TEST. | |
1972 | ||
1973 | Assert that the character at index IDX is on EXPECTED_LINE, | |
1974 | and that it begins at column EXPECTED_START_COL and ends at | |
1975 | EXPECTED_FINISH_COL (unless the locations are beyond | |
1976 | LINE_MAP_MAX_LOCATION_WITH_COLS, in which case don't check their | |
1977 | columns). */ | |
1978 | ||
1979 | static void | |
1980 | assert_char_at_range (const location &loc, | |
1981 | lexer_test& test, | |
1982 | location_t strloc, enum cpp_ttype type, int idx, | |
1983 | int expected_line, int expected_start_col, | |
1984 | int expected_finish_col) | |
1985 | { | |
1986 | cpp_reader *pfile = test.m_parser; | |
1987 | string_concat_db *concats = &test.m_concats; | |
1988 | ||
1989 | source_range actual_range; | |
1990 | const char *err | |
65e736c0 DM |
1991 | = get_source_range_for_char (pfile, concats, strloc, type, idx, |
1992 | &actual_range); | |
88fa5555 DM |
1993 | if (should_have_column_data_p (strloc)) |
1994 | ASSERT_EQ_AT (loc, NULL, err); | |
1995 | else | |
1996 | { | |
1997 | ASSERT_STREQ_AT (loc, | |
1998 | "range starts after LINE_MAP_MAX_LOCATION_WITH_COLS", | |
1999 | err); | |
2000 | return; | |
2001 | } | |
2002 | ||
2003 | int actual_start_line = LOCATION_LINE (actual_range.m_start); | |
2004 | ASSERT_EQ_AT (loc, expected_line, actual_start_line); | |
2005 | int actual_finish_line = LOCATION_LINE (actual_range.m_finish); | |
2006 | ASSERT_EQ_AT (loc, expected_line, actual_finish_line); | |
2007 | ||
2008 | if (should_have_column_data_p (actual_range.m_start)) | |
2009 | { | |
2010 | int actual_start_col = LOCATION_COLUMN (actual_range.m_start); | |
2011 | ASSERT_EQ_AT (loc, expected_start_col, actual_start_col); | |
2012 | } | |
2013 | if (should_have_column_data_p (actual_range.m_finish)) | |
2014 | { | |
2015 | int actual_finish_col = LOCATION_COLUMN (actual_range.m_finish); | |
2016 | ASSERT_EQ_AT (loc, expected_finish_col, actual_finish_col); | |
2017 | } | |
2018 | } | |
2019 | ||
2020 | /* Macro for calling assert_char_at_range, supplying SELFTEST_LOCATION for | |
2021 | the effective location of any errors. */ | |
2022 | ||
2023 | #define ASSERT_CHAR_AT_RANGE(LEXER_TEST, STRLOC, TYPE, IDX, EXPECTED_LINE, \ | |
2024 | EXPECTED_START_COL, EXPECTED_FINISH_COL) \ | |
2025 | assert_char_at_range (SELFTEST_LOCATION, (LEXER_TEST), (STRLOC), (TYPE), \ | |
2026 | (IDX), (EXPECTED_LINE), (EXPECTED_START_COL), \ | |
2027 | (EXPECTED_FINISH_COL)) | |
2028 | ||
2029 | /* Verify get_num_source_ranges_for_substring for token(s) at STRLOC, | |
2030 | using the string concatenation database for TEST. | |
2031 | ||
2032 | Assert that the token(s) at STRLOC contain EXPECTED_NUM_RANGES. */ | |
2033 | ||
2034 | static void | |
2035 | assert_num_substring_ranges (const location &loc, | |
2036 | lexer_test& test, | |
2037 | location_t strloc, | |
2038 | enum cpp_ttype type, | |
2039 | int expected_num_ranges) | |
2040 | { | |
2041 | cpp_reader *pfile = test.m_parser; | |
2042 | string_concat_db *concats = &test.m_concats; | |
2043 | ||
0e06d2b3 | 2044 | int actual_num_ranges = -1; |
88fa5555 DM |
2045 | const char *err |
2046 | = get_num_source_ranges_for_substring (pfile, concats, strloc, type, | |
2047 | &actual_num_ranges); | |
2048 | if (should_have_column_data_p (strloc)) | |
2049 | ASSERT_EQ_AT (loc, NULL, err); | |
2050 | else | |
2051 | { | |
2052 | ASSERT_STREQ_AT (loc, | |
2053 | "range starts after LINE_MAP_MAX_LOCATION_WITH_COLS", | |
2054 | err); | |
2055 | return; | |
2056 | } | |
2057 | ASSERT_EQ_AT (loc, expected_num_ranges, actual_num_ranges); | |
2058 | } | |
2059 | ||
2060 | /* Macro for calling assert_num_substring_ranges, supplying | |
2061 | SELFTEST_LOCATION for the effective location of any errors. */ | |
2062 | ||
2063 | #define ASSERT_NUM_SUBSTRING_RANGES(LEXER_TEST, STRLOC, TYPE, \ | |
2064 | EXPECTED_NUM_RANGES) \ | |
2065 | assert_num_substring_ranges (SELFTEST_LOCATION, (LEXER_TEST), (STRLOC), \ | |
2066 | (TYPE), (EXPECTED_NUM_RANGES)) | |
2067 | ||
2068 | ||
2069 | /* Verify that get_num_source_ranges_for_substring for token(s) at STRLOC | |
2070 | returns an error (using the string concatenation database for TEST). */ | |
2071 | ||
2072 | static void | |
2073 | assert_has_no_substring_ranges (const location &loc, | |
2074 | lexer_test& test, | |
2075 | location_t strloc, | |
2076 | enum cpp_ttype type, | |
2077 | const char *expected_err) | |
2078 | { | |
2079 | cpp_reader *pfile = test.m_parser; | |
2080 | string_concat_db *concats = &test.m_concats; | |
2081 | cpp_substring_ranges ranges; | |
2082 | const char *actual_err | |
2083 | = get_substring_ranges_for_loc (pfile, concats, strloc, | |
2084 | type, ranges); | |
2085 | if (should_have_column_data_p (strloc)) | |
2086 | ASSERT_STREQ_AT (loc, expected_err, actual_err); | |
2087 | else | |
2088 | ASSERT_STREQ_AT (loc, | |
2089 | "range starts after LINE_MAP_MAX_LOCATION_WITH_COLS", | |
2090 | actual_err); | |
2091 | } | |
2092 | ||
2093 | #define ASSERT_HAS_NO_SUBSTRING_RANGES(LEXER_TEST, STRLOC, TYPE, ERR) \ | |
2094 | assert_has_no_substring_ranges (SELFTEST_LOCATION, (LEXER_TEST), \ | |
2095 | (STRLOC), (TYPE), (ERR)) | |
2096 | ||
2097 | /* Lex a simple string literal. Verify the substring location data, before | |
2098 | and after running cpp_interpret_string on it. */ | |
2099 | ||
2100 | static void | |
2101 | test_lexer_string_locations_simple (const line_table_case &case_) | |
2102 | { | |
2103 | /* Digits 0-9 (with 0 at column 10), the simple way. | |
2104 | ....................000000000.11111111112.2222222223333333333 | |
2105 | ....................123456789.01234567890.1234567890123456789 | |
2106 | We add a trailing comment to ensure that we correctly locate | |
2107 | the end of the string literal token. */ | |
2108 | const char *content = " \"0123456789\" /* not a string */\n"; | |
2109 | lexer_test test (case_, content, NULL); | |
2110 | ||
2111 | /* Verify that we get the expected token back, with the correct | |
2112 | location information. */ | |
2113 | const cpp_token *tok = test.get_token (); | |
2114 | ASSERT_EQ (tok->type, CPP_STRING); | |
2115 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"0123456789\""); | |
2116 | ASSERT_TOKEN_LOC_EQ (tok, test.m_tempfile.get_filename (), 1, 9, 20); | |
2117 | ||
2118 | /* At this point in lexing, the quote characters are treated as part of | |
2119 | the string (they are stripped off by cpp_interpret_string). */ | |
2120 | ||
2121 | ASSERT_EQ (tok->val.str.len, 12); | |
2122 | ||
2123 | /* Verify that cpp_interpret_string works. */ | |
2124 | cpp_string dst_string; | |
2125 | const enum cpp_ttype type = CPP_STRING; | |
2126 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2127 | &dst_string, type); | |
2128 | ASSERT_TRUE (result); | |
2129 | ASSERT_STREQ ("0123456789", (const char *)dst_string.text); | |
2130 | free (const_cast <unsigned char *> (dst_string.text)); | |
2131 | ||
2132 | /* Verify ranges of individual characters. This no longer includes the | |
2133 | quotes. */ | |
2134 | for (int i = 0; i <= 9; i++) | |
2135 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, | |
2136 | 10 + i, 10 + i); | |
2137 | ||
2138 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 10); | |
2139 | } | |
2140 | ||
2141 | /* As test_lexer_string_locations_simple, but use an EBCDIC execution | |
2142 | encoding. */ | |
2143 | ||
2144 | static void | |
2145 | test_lexer_string_locations_ebcdic (const line_table_case &case_) | |
2146 | { | |
2147 | /* EBCDIC support requires iconv. */ | |
2148 | if (!HAVE_ICONV) | |
2149 | return; | |
2150 | ||
2151 | /* Digits 0-9 (with 0 at column 10), the simple way. | |
2152 | ....................000000000.11111111112.2222222223333333333 | |
2153 | ....................123456789.01234567890.1234567890123456789 | |
2154 | We add a trailing comment to ensure that we correctly locate | |
2155 | the end of the string literal token. */ | |
2156 | const char *content = " \"0123456789\" /* not a string */\n"; | |
2157 | ebcdic_execution_charset use_ebcdic; | |
2158 | lexer_test test (case_, content, &use_ebcdic); | |
2159 | ||
2160 | /* Verify that we get the expected token back, with the correct | |
2161 | location information. */ | |
2162 | const cpp_token *tok = test.get_token (); | |
2163 | ASSERT_EQ (tok->type, CPP_STRING); | |
2164 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"0123456789\""); | |
2165 | ASSERT_TOKEN_LOC_EQ (tok, test.m_tempfile.get_filename (), 1, 9, 20); | |
2166 | ||
2167 | /* At this point in lexing, the quote characters are treated as part of | |
2168 | the string (they are stripped off by cpp_interpret_string). */ | |
2169 | ||
2170 | ASSERT_EQ (tok->val.str.len, 12); | |
2171 | ||
2172 | /* The remainder of the test requires an iconv implementation that | |
2173 | can convert from UTF-8 to the EBCDIC encoding requested above. */ | |
2174 | if (use_ebcdic.iconv_errors_occurred_p ()) | |
2175 | return; | |
2176 | ||
2177 | /* Verify that cpp_interpret_string works. */ | |
2178 | cpp_string dst_string; | |
2179 | const enum cpp_ttype type = CPP_STRING; | |
2180 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2181 | &dst_string, type); | |
2182 | ASSERT_TRUE (result); | |
2183 | /* We should now have EBCDIC-encoded text, specifically | |
2184 | IBM1047-encoded (aka "EBCDIC 1047", or "Code page 1047"). | |
2185 | The digits 0-9 are encoded as 240-249 i.e. 0xf0-0xf9. */ | |
2186 | ASSERT_STREQ ("\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7\xf8\xf9", | |
2187 | (const char *)dst_string.text); | |
2188 | free (const_cast <unsigned char *> (dst_string.text)); | |
2189 | ||
2190 | /* Verify that we don't attempt to record substring location information | |
2191 | for such cases. */ | |
2192 | ASSERT_HAS_NO_SUBSTRING_RANGES | |
2193 | (test, tok->src_loc, type, | |
2194 | "execution character set != source character set"); | |
2195 | } | |
2196 | ||
2197 | /* Lex a string literal containing a hex-escaped character. | |
2198 | Verify the substring location data, before and after running | |
2199 | cpp_interpret_string on it. */ | |
2200 | ||
2201 | static void | |
2202 | test_lexer_string_locations_hex (const line_table_case &case_) | |
2203 | { | |
2204 | /* Digits 0-9, expressing digit 5 in ASCII as "\x35" | |
2205 | and with a space in place of digit 6, to terminate the escaped | |
2206 | hex code. | |
2207 | ....................000000000.111111.11112222. | |
2208 | ....................123456789.012345.67890123. */ | |
2209 | const char *content = " \"01234\\x35 789\"\n"; | |
2210 | lexer_test test (case_, content, NULL); | |
2211 | ||
2212 | /* Verify that we get the expected token back, with the correct | |
2213 | location information. */ | |
2214 | const cpp_token *tok = test.get_token (); | |
2215 | ASSERT_EQ (tok->type, CPP_STRING); | |
2216 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"01234\\x35 789\""); | |
2217 | ASSERT_TOKEN_LOC_EQ (tok, test.m_tempfile.get_filename (), 1, 9, 23); | |
2218 | ||
2219 | /* At this point in lexing, the quote characters are treated as part of | |
2220 | the string (they are stripped off by cpp_interpret_string). */ | |
2221 | ASSERT_EQ (tok->val.str.len, 15); | |
2222 | ||
2223 | /* Verify that cpp_interpret_string works. */ | |
2224 | cpp_string dst_string; | |
2225 | const enum cpp_ttype type = CPP_STRING; | |
2226 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2227 | &dst_string, type); | |
2228 | ASSERT_TRUE (result); | |
2229 | ASSERT_STREQ ("012345 789", (const char *)dst_string.text); | |
2230 | free (const_cast <unsigned char *> (dst_string.text)); | |
2231 | ||
2232 | /* Verify ranges of individual characters. This no longer includes the | |
2233 | quotes. */ | |
2234 | for (int i = 0; i <= 4; i++) | |
2235 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2236 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, 5, 1, 15, 18); | |
2237 | for (int i = 6; i <= 9; i++) | |
2238 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 13 + i, 13 + i); | |
2239 | ||
2240 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 10); | |
2241 | } | |
2242 | ||
2243 | /* Lex a string literal containing an octal-escaped character. | |
2244 | Verify the substring location data after running cpp_interpret_string | |
2245 | on it. */ | |
2246 | ||
2247 | static void | |
2248 | test_lexer_string_locations_oct (const line_table_case &case_) | |
2249 | { | |
2250 | /* Digits 0-9, expressing digit 5 in ASCII as "\065" | |
2251 | and with a space in place of digit 6, to terminate the escaped | |
2252 | octal code. | |
2253 | ....................000000000.111111.11112222.2222223333333333444 | |
2254 | ....................123456789.012345.67890123.4567890123456789012 */ | |
2255 | const char *content = " \"01234\\065 789\" /* not a string */\n"; | |
2256 | lexer_test test (case_, content, NULL); | |
2257 | ||
2258 | /* Verify that we get the expected token back, with the correct | |
2259 | location information. */ | |
2260 | const cpp_token *tok = test.get_token (); | |
2261 | ASSERT_EQ (tok->type, CPP_STRING); | |
2262 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"01234\\065 789\""); | |
2263 | ||
2264 | /* Verify that cpp_interpret_string works. */ | |
2265 | cpp_string dst_string; | |
2266 | const enum cpp_ttype type = CPP_STRING; | |
2267 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2268 | &dst_string, type); | |
2269 | ASSERT_TRUE (result); | |
2270 | ASSERT_STREQ ("012345 789", (const char *)dst_string.text); | |
2271 | free (const_cast <unsigned char *> (dst_string.text)); | |
2272 | ||
2273 | /* Verify ranges of individual characters. This no longer includes the | |
2274 | quotes. */ | |
2275 | for (int i = 0; i < 5; i++) | |
2276 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2277 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, 5, 1, 15, 18); | |
2278 | for (int i = 6; i <= 9; i++) | |
2279 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 13 + i, 13 + i); | |
2280 | ||
2281 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 10); | |
2282 | } | |
2283 | ||
2284 | /* Test of string literal containing letter escapes. */ | |
2285 | ||
2286 | static void | |
2287 | test_lexer_string_locations_letter_escape_1 (const line_table_case &case_) | |
2288 | { | |
2289 | /* The string "\tfoo\\\nbar" i.e. tab, "foo", backslash, newline, bar. | |
2290 | .....................000000000.1.11111.1.1.11222.22222223333333 | |
2291 | .....................123456789.0.12345.6.7.89012.34567890123456. */ | |
2292 | const char *content = (" \"\\tfoo\\\\\\nbar\" /* non-str */\n"); | |
2293 | lexer_test test (case_, content, NULL); | |
2294 | ||
2295 | /* Verify that we get the expected tokens back. */ | |
2296 | const cpp_token *tok = test.get_token (); | |
2297 | ASSERT_EQ (tok->type, CPP_STRING); | |
2298 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"\\tfoo\\\\\\nbar\""); | |
2299 | ||
2300 | /* Verify ranges of individual characters. */ | |
2301 | /* "\t". */ | |
2302 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2303 | 0, 1, 10, 11); | |
2304 | /* "foo". */ | |
2305 | for (int i = 1; i <= 3; i++) | |
2306 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2307 | i, 1, 11 + i, 11 + i); | |
2308 | /* "\\" and "\n". */ | |
2309 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2310 | 4, 1, 15, 16); | |
2311 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2312 | 5, 1, 17, 18); | |
2313 | ||
2314 | /* "bar". */ | |
2315 | for (int i = 6; i <= 8; i++) | |
2316 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2317 | i, 1, 13 + i, 13 + i); | |
2318 | ||
2319 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, 9); | |
2320 | } | |
2321 | ||
2322 | /* Another test of a string literal containing a letter escape. | |
2323 | Based on string seen in | |
2324 | printf ("%-%\n"); | |
2325 | in gcc.dg/format/c90-printf-1.c. */ | |
2326 | ||
2327 | static void | |
2328 | test_lexer_string_locations_letter_escape_2 (const line_table_case &case_) | |
2329 | { | |
2330 | /* .....................000000000.1111.11.1111.22222222223. | |
2331 | .....................123456789.0123.45.6789.01234567890. */ | |
2332 | const char *content = (" \"%-%\\n\" /* non-str */\n"); | |
2333 | lexer_test test (case_, content, NULL); | |
2334 | ||
2335 | /* Verify that we get the expected tokens back. */ | |
2336 | const cpp_token *tok = test.get_token (); | |
2337 | ASSERT_EQ (tok->type, CPP_STRING); | |
2338 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"%-%\\n\""); | |
2339 | ||
2340 | /* Verify ranges of individual characters. */ | |
2341 | /* "%-%". */ | |
2342 | for (int i = 0; i < 3; i++) | |
2343 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2344 | i, 1, 10 + i, 10 + i); | |
2345 | /* "\n". */ | |
2346 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2347 | 3, 1, 13, 14); | |
2348 | ||
2349 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, 4); | |
2350 | } | |
2351 | ||
2352 | /* Lex a string literal containing UCN 4 characters. | |
2353 | Verify the substring location data after running cpp_interpret_string | |
2354 | on it. */ | |
2355 | ||
2356 | static void | |
2357 | test_lexer_string_locations_ucn4 (const line_table_case &case_) | |
2358 | { | |
2359 | /* Digits 0-9, expressing digits 5 and 6 as Roman numerals expressed | |
2360 | as UCN 4. | |
2361 | ....................000000000.111111.111122.222222223.33333333344444 | |
2362 | ....................123456789.012345.678901.234567890.12345678901234 */ | |
2363 | const char *content = " \"01234\\u2174\\u2175789\" /* non-str */\n"; | |
2364 | lexer_test test (case_, content, NULL); | |
2365 | ||
2366 | /* Verify that we get the expected token back, with the correct | |
2367 | location information. */ | |
2368 | const cpp_token *tok = test.get_token (); | |
2369 | ASSERT_EQ (tok->type, CPP_STRING); | |
2370 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"01234\\u2174\\u2175789\""); | |
2371 | ||
2372 | /* Verify that cpp_interpret_string works. | |
2373 | The string should be encoded in the execution character | |
2374 | set. Assuming that that is UTF-8, we should have the following: | |
2375 | ----------- ---- ----- ------- ---------------- | |
2376 | Byte offset Byte Octal Unicode Source Column(s) | |
2377 | ----------- ---- ----- ------- ---------------- | |
2378 | 0 0x30 '0' 10 | |
2379 | 1 0x31 '1' 11 | |
2380 | 2 0x32 '2' 12 | |
2381 | 3 0x33 '3' 13 | |
2382 | 4 0x34 '4' 14 | |
2383 | 5 0xE2 \342 U+2174 15-20 | |
2384 | 6 0x85 \205 (cont) 15-20 | |
2385 | 7 0xB4 \264 (cont) 15-20 | |
2386 | 8 0xE2 \342 U+2175 21-26 | |
2387 | 9 0x85 \205 (cont) 21-26 | |
2388 | 10 0xB5 \265 (cont) 21-26 | |
2389 | 11 0x37 '7' 27 | |
2390 | 12 0x38 '8' 28 | |
2391 | 13 0x39 '9' 29 | |
2392 | ----------- ---- ----- ------- ---------------. */ | |
2393 | ||
2394 | cpp_string dst_string; | |
2395 | const enum cpp_ttype type = CPP_STRING; | |
2396 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2397 | &dst_string, type); | |
2398 | ASSERT_TRUE (result); | |
2399 | ASSERT_STREQ ("01234\342\205\264\342\205\265789", | |
2400 | (const char *)dst_string.text); | |
2401 | free (const_cast <unsigned char *> (dst_string.text)); | |
2402 | ||
2403 | /* Verify ranges of individual characters. This no longer includes the | |
2404 | quotes. | |
2405 | '01234'. */ | |
2406 | for (int i = 0; i <= 4; i++) | |
2407 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2408 | /* U+2174. */ | |
2409 | for (int i = 5; i <= 7; i++) | |
2410 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 15, 20); | |
2411 | /* U+2175. */ | |
2412 | for (int i = 8; i <= 10; i++) | |
2413 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 21, 26); | |
2414 | /* '789'. */ | |
2415 | for (int i = 11; i <= 13; i++) | |
2416 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 16 + i, 16 + i); | |
2417 | ||
2418 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 14); | |
2419 | } | |
2420 | ||
2421 | /* Lex a string literal containing UCN 8 characters. | |
2422 | Verify the substring location data after running cpp_interpret_string | |
2423 | on it. */ | |
2424 | ||
2425 | static void | |
2426 | test_lexer_string_locations_ucn8 (const line_table_case &case_) | |
2427 | { | |
2428 | /* Digits 0-9, expressing digits 5 and 6 as Roman numerals as UCN 8. | |
2429 | ....................000000000.111111.1111222222.2222333333333.344444 | |
2430 | ....................123456789.012345.6789012345.6789012345678.901234 */ | |
2431 | const char *content = " \"01234\\U00002174\\U00002175789\" /* */\n"; | |
2432 | lexer_test test (case_, content, NULL); | |
2433 | ||
2434 | /* Verify that we get the expected token back, with the correct | |
2435 | location information. */ | |
2436 | const cpp_token *tok = test.get_token (); | |
2437 | ASSERT_EQ (tok->type, CPP_STRING); | |
2438 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, | |
2439 | "\"01234\\U00002174\\U00002175789\""); | |
2440 | ||
2441 | /* Verify that cpp_interpret_string works. | |
2442 | The UTF-8 encoding of the string is identical to that from | |
2443 | the ucn4 testcase above; the only difference is the column | |
2444 | locations. */ | |
2445 | cpp_string dst_string; | |
2446 | const enum cpp_ttype type = CPP_STRING; | |
2447 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2448 | &dst_string, type); | |
2449 | ASSERT_TRUE (result); | |
2450 | ASSERT_STREQ ("01234\342\205\264\342\205\265789", | |
2451 | (const char *)dst_string.text); | |
2452 | free (const_cast <unsigned char *> (dst_string.text)); | |
2453 | ||
2454 | /* Verify ranges of individual characters. This no longer includes the | |
2455 | quotes. | |
2456 | '01234'. */ | |
2457 | for (int i = 0; i <= 4; i++) | |
2458 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2459 | /* U+2174. */ | |
2460 | for (int i = 5; i <= 7; i++) | |
2461 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 15, 24); | |
2462 | /* U+2175. */ | |
2463 | for (int i = 8; i <= 10; i++) | |
2464 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 25, 34); | |
2465 | /* '789' at columns 35-37 */ | |
2466 | for (int i = 11; i <= 13; i++) | |
2467 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 24 + i, 24 + i); | |
2468 | ||
2469 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 14); | |
2470 | } | |
2471 | ||
2472 | /* Fetch a big-endian 32-bit value and convert to host endianness. */ | |
2473 | ||
2474 | static uint32_t | |
2475 | uint32_from_big_endian (const uint32_t *ptr_be_value) | |
2476 | { | |
2477 | const unsigned char *buf = (const unsigned char *)ptr_be_value; | |
2478 | return (((uint32_t) buf[0] << 24) | |
2479 | | ((uint32_t) buf[1] << 16) | |
2480 | | ((uint32_t) buf[2] << 8) | |
2481 | | (uint32_t) buf[3]); | |
2482 | } | |
2483 | ||
2484 | /* Lex a wide string literal and verify that attempts to read substring | |
2485 | location data from it fail gracefully. */ | |
2486 | ||
2487 | static void | |
2488 | test_lexer_string_locations_wide_string (const line_table_case &case_) | |
2489 | { | |
2490 | /* Digits 0-9. | |
2491 | ....................000000000.11111111112.22222222233333 | |
2492 | ....................123456789.01234567890.12345678901234 */ | |
2493 | const char *content = " L\"0123456789\" /* non-str */\n"; | |
2494 | lexer_test test (case_, content, NULL); | |
2495 | ||
2496 | /* Verify that we get the expected token back, with the correct | |
2497 | location information. */ | |
2498 | const cpp_token *tok = test.get_token (); | |
2499 | ASSERT_EQ (tok->type, CPP_WSTRING); | |
2500 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "L\"0123456789\""); | |
2501 | ||
2502 | /* Verify that cpp_interpret_string works, using CPP_WSTRING. */ | |
2503 | cpp_string dst_string; | |
2504 | const enum cpp_ttype type = CPP_WSTRING; | |
2505 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2506 | &dst_string, type); | |
2507 | ASSERT_TRUE (result); | |
2508 | /* The cpp_reader defaults to big-endian with | |
2509 | CHAR_BIT * sizeof (int) for the wchar_precision, so dst_string should | |
2510 | now be encoded as UTF-32BE. */ | |
2511 | const uint32_t *be32_chars = (const uint32_t *)dst_string.text; | |
2512 | ASSERT_EQ ('0', uint32_from_big_endian (&be32_chars[0])); | |
2513 | ASSERT_EQ ('5', uint32_from_big_endian (&be32_chars[5])); | |
2514 | ASSERT_EQ ('9', uint32_from_big_endian (&be32_chars[9])); | |
2515 | ASSERT_EQ (0, uint32_from_big_endian (&be32_chars[10])); | |
2516 | free (const_cast <unsigned char *> (dst_string.text)); | |
2517 | ||
2518 | /* We don't yet support generating substring location information | |
2519 | for L"" strings. */ | |
2520 | ASSERT_HAS_NO_SUBSTRING_RANGES | |
2521 | (test, tok->src_loc, type, | |
2522 | "execution character set != source character set"); | |
2523 | } | |
2524 | ||
2525 | /* Fetch a big-endian 16-bit value and convert to host endianness. */ | |
2526 | ||
2527 | static uint16_t | |
2528 | uint16_from_big_endian (const uint16_t *ptr_be_value) | |
2529 | { | |
2530 | const unsigned char *buf = (const unsigned char *)ptr_be_value; | |
2531 | return ((uint16_t) buf[0] << 8) | (uint16_t) buf[1]; | |
2532 | } | |
2533 | ||
2534 | /* Lex a u"" string literal and verify that attempts to read substring | |
2535 | location data from it fail gracefully. */ | |
2536 | ||
2537 | static void | |
2538 | test_lexer_string_locations_string16 (const line_table_case &case_) | |
2539 | { | |
2540 | /* Digits 0-9. | |
2541 | ....................000000000.11111111112.22222222233333 | |
2542 | ....................123456789.01234567890.12345678901234 */ | |
2543 | const char *content = " u\"0123456789\" /* non-str */\n"; | |
2544 | lexer_test test (case_, content, NULL); | |
2545 | ||
2546 | /* Verify that we get the expected token back, with the correct | |
2547 | location information. */ | |
2548 | const cpp_token *tok = test.get_token (); | |
2549 | ASSERT_EQ (tok->type, CPP_STRING16); | |
2550 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "u\"0123456789\""); | |
2551 | ||
2552 | /* Verify that cpp_interpret_string works, using CPP_STRING16. */ | |
2553 | cpp_string dst_string; | |
2554 | const enum cpp_ttype type = CPP_STRING16; | |
2555 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2556 | &dst_string, type); | |
2557 | ASSERT_TRUE (result); | |
2558 | ||
2559 | /* The cpp_reader defaults to big-endian, so dst_string should | |
2560 | now be encoded as UTF-16BE. */ | |
2561 | const uint16_t *be16_chars = (const uint16_t *)dst_string.text; | |
2562 | ASSERT_EQ ('0', uint16_from_big_endian (&be16_chars[0])); | |
2563 | ASSERT_EQ ('5', uint16_from_big_endian (&be16_chars[5])); | |
2564 | ASSERT_EQ ('9', uint16_from_big_endian (&be16_chars[9])); | |
2565 | ASSERT_EQ (0, uint16_from_big_endian (&be16_chars[10])); | |
2566 | free (const_cast <unsigned char *> (dst_string.text)); | |
2567 | ||
2568 | /* We don't yet support generating substring location information | |
2569 | for L"" strings. */ | |
2570 | ASSERT_HAS_NO_SUBSTRING_RANGES | |
2571 | (test, tok->src_loc, type, | |
2572 | "execution character set != source character set"); | |
2573 | } | |
2574 | ||
2575 | /* Lex a U"" string literal and verify that attempts to read substring | |
2576 | location data from it fail gracefully. */ | |
2577 | ||
2578 | static void | |
2579 | test_lexer_string_locations_string32 (const line_table_case &case_) | |
2580 | { | |
2581 | /* Digits 0-9. | |
2582 | ....................000000000.11111111112.22222222233333 | |
2583 | ....................123456789.01234567890.12345678901234 */ | |
2584 | const char *content = " U\"0123456789\" /* non-str */\n"; | |
2585 | lexer_test test (case_, content, NULL); | |
2586 | ||
2587 | /* Verify that we get the expected token back, with the correct | |
2588 | location information. */ | |
2589 | const cpp_token *tok = test.get_token (); | |
2590 | ASSERT_EQ (tok->type, CPP_STRING32); | |
2591 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "U\"0123456789\""); | |
2592 | ||
2593 | /* Verify that cpp_interpret_string works, using CPP_STRING32. */ | |
2594 | cpp_string dst_string; | |
2595 | const enum cpp_ttype type = CPP_STRING32; | |
2596 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2597 | &dst_string, type); | |
2598 | ASSERT_TRUE (result); | |
2599 | ||
2600 | /* The cpp_reader defaults to big-endian, so dst_string should | |
2601 | now be encoded as UTF-32BE. */ | |
2602 | const uint32_t *be32_chars = (const uint32_t *)dst_string.text; | |
2603 | ASSERT_EQ ('0', uint32_from_big_endian (&be32_chars[0])); | |
2604 | ASSERT_EQ ('5', uint32_from_big_endian (&be32_chars[5])); | |
2605 | ASSERT_EQ ('9', uint32_from_big_endian (&be32_chars[9])); | |
2606 | ASSERT_EQ (0, uint32_from_big_endian (&be32_chars[10])); | |
2607 | free (const_cast <unsigned char *> (dst_string.text)); | |
2608 | ||
2609 | /* We don't yet support generating substring location information | |
2610 | for L"" strings. */ | |
2611 | ASSERT_HAS_NO_SUBSTRING_RANGES | |
2612 | (test, tok->src_loc, type, | |
2613 | "execution character set != source character set"); | |
2614 | } | |
2615 | ||
2616 | /* Lex a u8-string literal. | |
2617 | Verify the substring location data after running cpp_interpret_string | |
2618 | on it. */ | |
2619 | ||
2620 | static void | |
2621 | test_lexer_string_locations_u8 (const line_table_case &case_) | |
2622 | { | |
2623 | /* Digits 0-9. | |
2624 | ....................000000000.11111111112.22222222233333 | |
2625 | ....................123456789.01234567890.12345678901234 */ | |
2626 | const char *content = " u8\"0123456789\" /* non-str */\n"; | |
2627 | lexer_test test (case_, content, NULL); | |
2628 | ||
2629 | /* Verify that we get the expected token back, with the correct | |
2630 | location information. */ | |
2631 | const cpp_token *tok = test.get_token (); | |
2632 | ASSERT_EQ (tok->type, CPP_UTF8STRING); | |
2633 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "u8\"0123456789\""); | |
2634 | ||
2635 | /* Verify that cpp_interpret_string works. */ | |
2636 | cpp_string dst_string; | |
2637 | const enum cpp_ttype type = CPP_STRING; | |
2638 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2639 | &dst_string, type); | |
2640 | ASSERT_TRUE (result); | |
2641 | ASSERT_STREQ ("0123456789", (const char *)dst_string.text); | |
2642 | free (const_cast <unsigned char *> (dst_string.text)); | |
2643 | ||
2644 | /* Verify ranges of individual characters. This no longer includes the | |
2645 | quotes. */ | |
2646 | for (int i = 0; i <= 9; i++) | |
2647 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2648 | } | |
2649 | ||
2650 | /* Lex a string literal containing UTF-8 source characters. | |
2651 | Verify the substring location data after running cpp_interpret_string | |
2652 | on it. */ | |
2653 | ||
2654 | static void | |
2655 | test_lexer_string_locations_utf8_source (const line_table_case &case_) | |
2656 | { | |
2657 | /* This string literal is written out to the source file as UTF-8, | |
2658 | and is of the form "before mojibake after", where "mojibake" | |
2659 | is written as the following four unicode code points: | |
2660 | U+6587 CJK UNIFIED IDEOGRAPH-6587 | |
2661 | U+5B57 CJK UNIFIED IDEOGRAPH-5B57 | |
2662 | U+5316 CJK UNIFIED IDEOGRAPH-5316 | |
2663 | U+3051 HIRAGANA LETTER KE. | |
2664 | Each of these is 3 bytes wide when encoded in UTF-8, whereas the | |
2665 | "before" and "after" are 1 byte per unicode character. | |
2666 | ||
2667 | The numbering shown are "columns", which are *byte* numbers within | |
2668 | the line, rather than unicode character numbers. | |
2669 | ||
2670 | .................... 000000000.1111111. | |
2671 | .................... 123456789.0123456. */ | |
2672 | const char *content = (" \"before " | |
2673 | /* U+6587 CJK UNIFIED IDEOGRAPH-6587 | |
2674 | UTF-8: 0xE6 0x96 0x87 | |
2675 | C octal escaped UTF-8: \346\226\207 | |
2676 | "column" numbers: 17-19. */ | |
2677 | "\346\226\207" | |
2678 | ||
2679 | /* U+5B57 CJK UNIFIED IDEOGRAPH-5B57 | |
2680 | UTF-8: 0xE5 0xAD 0x97 | |
2681 | C octal escaped UTF-8: \345\255\227 | |
2682 | "column" numbers: 20-22. */ | |
2683 | "\345\255\227" | |
2684 | ||
2685 | /* U+5316 CJK UNIFIED IDEOGRAPH-5316 | |
2686 | UTF-8: 0xE5 0x8C 0x96 | |
2687 | C octal escaped UTF-8: \345\214\226 | |
2688 | "column" numbers: 23-25. */ | |
2689 | "\345\214\226" | |
2690 | ||
2691 | /* U+3051 HIRAGANA LETTER KE | |
2692 | UTF-8: 0xE3 0x81 0x91 | |
2693 | C octal escaped UTF-8: \343\201\221 | |
2694 | "column" numbers: 26-28. */ | |
2695 | "\343\201\221" | |
2696 | ||
2697 | /* column numbers 29 onwards | |
2698 | 2333333.33334444444444 | |
2699 | 9012345.67890123456789. */ | |
2700 | " after\" /* non-str */\n"); | |
2701 | lexer_test test (case_, content, NULL); | |
2702 | ||
2703 | /* Verify that we get the expected token back, with the correct | |
2704 | location information. */ | |
2705 | const cpp_token *tok = test.get_token (); | |
2706 | ASSERT_EQ (tok->type, CPP_STRING); | |
2707 | ASSERT_TOKEN_AS_TEXT_EQ | |
2708 | (test.m_parser, tok, | |
2709 | "\"before \346\226\207\345\255\227\345\214\226\343\201\221 after\""); | |
2710 | ||
2711 | /* Verify that cpp_interpret_string works. */ | |
2712 | cpp_string dst_string; | |
2713 | const enum cpp_ttype type = CPP_STRING; | |
2714 | bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1, | |
2715 | &dst_string, type); | |
2716 | ASSERT_TRUE (result); | |
2717 | ASSERT_STREQ | |
2718 | ("before \346\226\207\345\255\227\345\214\226\343\201\221 after", | |
2719 | (const char *)dst_string.text); | |
2720 | free (const_cast <unsigned char *> (dst_string.text)); | |
2721 | ||
2722 | /* Verify ranges of individual characters. This no longer includes the | |
2723 | quotes. | |
2724 | Assuming that both source and execution encodings are UTF-8, we have | |
2725 | a run of 25 octets in each. */ | |
2726 | for (int i = 0; i < 25; i++) | |
2727 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, type, i, 1, 10 + i, 10 + i); | |
2728 | ||
2729 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, type, 25); | |
2730 | } | |
2731 | ||
2732 | /* Test of string literal concatenation. */ | |
2733 | ||
2734 | static void | |
2735 | test_lexer_string_locations_concatenation_1 (const line_table_case &case_) | |
2736 | { | |
2737 | /* Digits 0-9. | |
2738 | .....................000000000.111111.11112222222222 | |
2739 | .....................123456789.012345.67890123456789. */ | |
2740 | const char *content = (" \"01234\" /* non-str */\n" | |
2741 | " \"56789\" /* non-str */\n"); | |
2742 | lexer_test test (case_, content, NULL); | |
2743 | ||
2744 | location_t input_locs[2]; | |
2745 | ||
2746 | /* Verify that we get the expected tokens back. */ | |
2747 | auto_vec <cpp_string> input_strings; | |
2748 | const cpp_token *tok_a = test.get_token (); | |
2749 | ASSERT_EQ (tok_a->type, CPP_STRING); | |
2750 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok_a, "\"01234\""); | |
2751 | input_strings.safe_push (tok_a->val.str); | |
2752 | input_locs[0] = tok_a->src_loc; | |
2753 | ||
2754 | const cpp_token *tok_b = test.get_token (); | |
2755 | ASSERT_EQ (tok_b->type, CPP_STRING); | |
2756 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok_b, "\"56789\""); | |
2757 | input_strings.safe_push (tok_b->val.str); | |
2758 | input_locs[1] = tok_b->src_loc; | |
2759 | ||
2760 | /* Verify that cpp_interpret_string works. */ | |
2761 | cpp_string dst_string; | |
2762 | const enum cpp_ttype type = CPP_STRING; | |
2763 | bool result = cpp_interpret_string (test.m_parser, | |
2764 | input_strings.address (), 2, | |
2765 | &dst_string, type); | |
2766 | ASSERT_TRUE (result); | |
2767 | ASSERT_STREQ ("0123456789", (const char *)dst_string.text); | |
2768 | free (const_cast <unsigned char *> (dst_string.text)); | |
2769 | ||
2770 | /* Simulate c-lex.c's lex_string in order to record concatenation. */ | |
2771 | test.m_concats.record_string_concatenation (2, input_locs); | |
2772 | ||
2773 | location_t initial_loc = input_locs[0]; | |
2774 | ||
2775 | for (int i = 0; i <= 4; i++) | |
2776 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, i, 1, 10 + i, 10 + i); | |
2777 | for (int i = 5; i <= 9; i++) | |
2778 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, i, 2, 5 + i, 5 + i); | |
2779 | ||
2780 | ASSERT_NUM_SUBSTRING_RANGES (test, initial_loc, type, 10); | |
2781 | } | |
2782 | ||
2783 | /* Another test of string literal concatenation. */ | |
2784 | ||
2785 | static void | |
2786 | test_lexer_string_locations_concatenation_2 (const line_table_case &case_) | |
2787 | { | |
2788 | /* Digits 0-9. | |
2789 | .....................000000000.111.11111112222222 | |
2790 | .....................123456789.012.34567890123456. */ | |
2791 | const char *content = (" \"01\" /* non-str */\n" | |
2792 | " \"23\" /* non-str */\n" | |
2793 | " \"45\" /* non-str */\n" | |
2794 | " \"67\" /* non-str */\n" | |
2795 | " \"89\" /* non-str */\n"); | |
2796 | lexer_test test (case_, content, NULL); | |
2797 | ||
2798 | auto_vec <cpp_string> input_strings; | |
2799 | location_t input_locs[5]; | |
2800 | ||
2801 | /* Verify that we get the expected tokens back. */ | |
2802 | for (int i = 0; i < 5; i++) | |
2803 | { | |
2804 | const cpp_token *tok = test.get_token (); | |
2805 | ASSERT_EQ (tok->type, CPP_STRING); | |
2806 | input_strings.safe_push (tok->val.str); | |
2807 | input_locs[i] = tok->src_loc; | |
2808 | } | |
2809 | ||
2810 | /* Verify that cpp_interpret_string works. */ | |
2811 | cpp_string dst_string; | |
2812 | const enum cpp_ttype type = CPP_STRING; | |
2813 | bool result = cpp_interpret_string (test.m_parser, | |
2814 | input_strings.address (), 5, | |
2815 | &dst_string, type); | |
2816 | ASSERT_TRUE (result); | |
2817 | ASSERT_STREQ ("0123456789", (const char *)dst_string.text); | |
2818 | free (const_cast <unsigned char *> (dst_string.text)); | |
2819 | ||
2820 | /* Simulate c-lex.c's lex_string in order to record concatenation. */ | |
2821 | test.m_concats.record_string_concatenation (5, input_locs); | |
2822 | ||
2823 | location_t initial_loc = input_locs[0]; | |
2824 | ||
2825 | /* Within ASSERT_CHAR_AT_RANGE (actually assert_char_at_range), we can | |
2826 | detect if the initial loc is after LINE_MAP_MAX_LOCATION_WITH_COLS | |
2827 | and expect get_source_range_for_substring to fail. | |
2828 | However, for a string concatenation test, we can have a case | |
2829 | where the initial string is fully before LINE_MAP_MAX_LOCATION_WITH_COLS, | |
2830 | but subsequent strings can be after it. | |
2831 | Attempting to detect this within assert_char_at_range | |
2832 | would overcomplicate the logic for the common test cases, so | |
2833 | we detect it here. */ | |
2834 | if (should_have_column_data_p (input_locs[0]) | |
2835 | && !should_have_column_data_p (input_locs[4])) | |
2836 | { | |
2837 | /* Verify that get_source_range_for_substring gracefully rejects | |
2838 | this case. */ | |
2839 | source_range actual_range; | |
2840 | const char *err | |
65e736c0 DM |
2841 | = get_source_range_for_char (test.m_parser, &test.m_concats, |
2842 | initial_loc, type, 0, &actual_range); | |
88fa5555 DM |
2843 | ASSERT_STREQ ("range starts after LINE_MAP_MAX_LOCATION_WITH_COLS", err); |
2844 | return; | |
2845 | } | |
2846 | ||
2847 | for (int i = 0; i < 5; i++) | |
2848 | for (int j = 0; j < 2; j++) | |
2849 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, (i * 2) + j, | |
2850 | i + 1, 10 + j, 10 + j); | |
2851 | ||
2852 | ASSERT_NUM_SUBSTRING_RANGES (test, initial_loc, type, 10); | |
2853 | } | |
2854 | ||
2855 | /* Another test of string literal concatenation, this time combined with | |
2856 | various kinds of escaped characters. */ | |
2857 | ||
2858 | static void | |
2859 | test_lexer_string_locations_concatenation_3 (const line_table_case &case_) | |
2860 | { | |
2861 | /* Digits 0-9, expressing digit 5 in ASCII as hex "\x35" | |
2862 | digit 6 in ASCII as octal "\066", concatenating multiple strings. */ | |
2863 | const char *content | |
2864 | /* .000000000.111111.111.1.2222.222.2.2233.333.3333.34444444444555 | |
2865 | .123456789.012345.678.9.0123.456.7.8901.234.5678.90123456789012. */ | |
2866 | = (" \"01234\" \"\\x35\" \"\\066\" \"789\" /* non-str */\n"); | |
2867 | lexer_test test (case_, content, NULL); | |
2868 | ||
2869 | auto_vec <cpp_string> input_strings; | |
2870 | location_t input_locs[4]; | |
2871 | ||
2872 | /* Verify that we get the expected tokens back. */ | |
2873 | for (int i = 0; i < 4; i++) | |
2874 | { | |
2875 | const cpp_token *tok = test.get_token (); | |
2876 | ASSERT_EQ (tok->type, CPP_STRING); | |
2877 | input_strings.safe_push (tok->val.str); | |
2878 | input_locs[i] = tok->src_loc; | |
2879 | } | |
2880 | ||
2881 | /* Verify that cpp_interpret_string works. */ | |
2882 | cpp_string dst_string; | |
2883 | const enum cpp_ttype type = CPP_STRING; | |
2884 | bool result = cpp_interpret_string (test.m_parser, | |
2885 | input_strings.address (), 4, | |
2886 | &dst_string, type); | |
2887 | ASSERT_TRUE (result); | |
2888 | ASSERT_STREQ ("0123456789", (const char *)dst_string.text); | |
2889 | free (const_cast <unsigned char *> (dst_string.text)); | |
2890 | ||
2891 | /* Simulate c-lex.c's lex_string in order to record concatenation. */ | |
2892 | test.m_concats.record_string_concatenation (4, input_locs); | |
2893 | ||
2894 | location_t initial_loc = input_locs[0]; | |
2895 | ||
2896 | for (int i = 0; i <= 4; i++) | |
2897 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, i, 1, 10 + i, 10 + i); | |
2898 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, 5, 1, 19, 22); | |
2899 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, 6, 1, 27, 30); | |
2900 | for (int i = 7; i <= 9; i++) | |
2901 | ASSERT_CHAR_AT_RANGE (test, initial_loc, type, i, 1, 28 + i, 28 + i); | |
2902 | ||
2903 | ASSERT_NUM_SUBSTRING_RANGES (test, initial_loc, type, 10); | |
2904 | } | |
2905 | ||
2906 | /* Test of string literal in a macro. */ | |
2907 | ||
2908 | static void | |
2909 | test_lexer_string_locations_macro (const line_table_case &case_) | |
2910 | { | |
2911 | /* Digits 0-9. | |
2912 | .....................0000000001111111111.22222222223. | |
2913 | .....................1234567890123456789.01234567890. */ | |
2914 | const char *content = ("#define MACRO \"0123456789\" /* non-str */\n" | |
2915 | " MACRO"); | |
2916 | lexer_test test (case_, content, NULL); | |
2917 | ||
2918 | /* Verify that we get the expected tokens back. */ | |
2919 | const cpp_token *tok = test.get_token (); | |
2920 | ASSERT_EQ (tok->type, CPP_PADDING); | |
2921 | ||
2922 | tok = test.get_token (); | |
2923 | ASSERT_EQ (tok->type, CPP_STRING); | |
2924 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"0123456789\""); | |
2925 | ||
2926 | /* Verify ranges of individual characters. We ought to | |
2927 | see columns within the macro definition. */ | |
2928 | for (int i = 0; i <= 9; i++) | |
2929 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
2930 | i, 1, 20 + i, 20 + i); | |
2931 | ||
2932 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, 10); | |
2933 | ||
2934 | tok = test.get_token (); | |
2935 | ASSERT_EQ (tok->type, CPP_PADDING); | |
2936 | } | |
2937 | ||
2938 | /* Test of stringification of a macro argument. */ | |
2939 | ||
2940 | static void | |
2941 | test_lexer_string_locations_stringified_macro_argument | |
2942 | (const line_table_case &case_) | |
2943 | { | |
2944 | /* .....................000000000111111111122222222223. | |
2945 | .....................123456789012345678901234567890. */ | |
2946 | const char *content = ("#define MACRO(X) #X /* non-str */\n" | |
2947 | "MACRO(foo)\n"); | |
2948 | lexer_test test (case_, content, NULL); | |
2949 | ||
2950 | /* Verify that we get the expected token back. */ | |
2951 | const cpp_token *tok = test.get_token (); | |
2952 | ASSERT_EQ (tok->type, CPP_PADDING); | |
2953 | ||
2954 | tok = test.get_token (); | |
2955 | ASSERT_EQ (tok->type, CPP_STRING); | |
2956 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "\"foo\""); | |
2957 | ||
2958 | /* We don't support getting the location of a stringified macro | |
2959 | argument. Verify that it fails gracefully. */ | |
2960 | ASSERT_HAS_NO_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, | |
2961 | "cpp_interpret_string_1 failed"); | |
2962 | ||
2963 | tok = test.get_token (); | |
2964 | ASSERT_EQ (tok->type, CPP_PADDING); | |
2965 | ||
2966 | tok = test.get_token (); | |
2967 | ASSERT_EQ (tok->type, CPP_PADDING); | |
2968 | } | |
2969 | ||
2970 | /* Ensure that we are fail gracefully if something attempts to pass | |
2971 | in a location that isn't a string literal token. Seen on this code: | |
2972 | ||
2973 | const char a[] = " %d "; | |
2974 | __builtin_printf (a, 0.5); | |
2975 | ^ | |
2976 | ||
2977 | when c-format.c erroneously used the indicated one-character | |
2978 | location as the format string location, leading to a read past the | |
2979 | end of a string buffer in cpp_interpret_string_1. */ | |
2980 | ||
2981 | static void | |
2982 | test_lexer_string_locations_non_string (const line_table_case &case_) | |
2983 | { | |
2984 | /* .....................000000000111111111122222222223. | |
2985 | .....................123456789012345678901234567890. */ | |
2986 | const char *content = (" a\n"); | |
2987 | lexer_test test (case_, content, NULL); | |
2988 | ||
2989 | /* Verify that we get the expected token back. */ | |
2990 | const cpp_token *tok = test.get_token (); | |
2991 | ASSERT_EQ (tok->type, CPP_NAME); | |
2992 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "a"); | |
2993 | ||
2994 | /* At this point, libcpp is attempting to interpret the name as a | |
2995 | string literal, despite it not starting with a quote. We don't detect | |
2996 | that, but we should at least fail gracefully. */ | |
2997 | ASSERT_HAS_NO_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, | |
2998 | "cpp_interpret_string_1 failed"); | |
2999 | } | |
3000 | ||
3001 | /* Ensure that we can read substring information for a token which | |
3002 | starts in one linemap and ends in another . Adapted from | |
3003 | gcc.dg/cpp/pr69985.c. */ | |
3004 | ||
3005 | static void | |
3006 | test_lexer_string_locations_long_line (const line_table_case &case_) | |
3007 | { | |
3008 | /* .....................000000.000111111111 | |
3009 | .....................123456.789012346789. */ | |
3010 | const char *content = ("/* A very long line, so that we start a new line map. */\n" | |
3011 | " \"0123456789012345678901234567890123456789" | |
3012 | "0123456789012345678901234567890123456789" | |
3013 | "0123456789012345678901234567890123456789" | |
3014 | "0123456789\"\n"); | |
3015 | ||
3016 | lexer_test test (case_, content, NULL); | |
3017 | ||
3018 | /* Verify that we get the expected token back. */ | |
3019 | const cpp_token *tok = test.get_token (); | |
3020 | ASSERT_EQ (tok->type, CPP_STRING); | |
3021 | ||
3022 | if (!should_have_column_data_p (line_table->highest_location)) | |
3023 | return; | |
3024 | ||
3025 | /* Verify ranges of individual characters. */ | |
3026 | ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, 130); | |
3027 | for (int i = 0; i < 130; i++) | |
3028 | ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING, | |
3029 | i, 2, 7 + i, 7 + i); | |
3030 | } | |
3031 | ||
3032 | /* Test of lexing char constants. */ | |
3033 | ||
3034 | static void | |
3035 | test_lexer_char_constants (const line_table_case &case_) | |
3036 | { | |
3037 | /* Various char constants. | |
3038 | .....................0000000001111111111.22222222223. | |
3039 | .....................1234567890123456789.01234567890. */ | |
3040 | const char *content = (" 'a'\n" | |
3041 | " u'a'\n" | |
3042 | " U'a'\n" | |
3043 | " L'a'\n" | |
3044 | " 'abc'\n"); | |
3045 | lexer_test test (case_, content, NULL); | |
3046 | ||
3047 | /* Verify that we get the expected tokens back. */ | |
3048 | /* 'a'. */ | |
3049 | const cpp_token *tok = test.get_token (); | |
3050 | ASSERT_EQ (tok->type, CPP_CHAR); | |
3051 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "'a'"); | |
3052 | ||
3053 | unsigned int chars_seen; | |
3054 | int unsignedp; | |
3055 | cppchar_t cc = cpp_interpret_charconst (test.m_parser, tok, | |
3056 | &chars_seen, &unsignedp); | |
3057 | ASSERT_EQ (cc, 'a'); | |
3058 | ASSERT_EQ (chars_seen, 1); | |
3059 | ||
3060 | /* u'a'. */ | |
3061 | tok = test.get_token (); | |
3062 | ASSERT_EQ (tok->type, CPP_CHAR16); | |
3063 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "u'a'"); | |
3064 | ||
3065 | /* U'a'. */ | |
3066 | tok = test.get_token (); | |
3067 | ASSERT_EQ (tok->type, CPP_CHAR32); | |
3068 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "U'a'"); | |
3069 | ||
3070 | /* L'a'. */ | |
3071 | tok = test.get_token (); | |
3072 | ASSERT_EQ (tok->type, CPP_WCHAR); | |
3073 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "L'a'"); | |
3074 | ||
3075 | /* 'abc' (c-char-sequence). */ | |
3076 | tok = test.get_token (); | |
3077 | ASSERT_EQ (tok->type, CPP_CHAR); | |
3078 | ASSERT_TOKEN_AS_TEXT_EQ (test.m_parser, tok, "'abc'"); | |
3079 | } | |
741d3be5 DM |
3080 | /* A table of interesting location_t values, giving one axis of our test |
3081 | matrix. */ | |
3082 | ||
3083 | static const location_t boundary_locations[] = { | |
3084 | /* Zero means "don't override the default values for a new line_table". */ | |
3085 | 0, | |
3086 | ||
3087 | /* An arbitrary non-zero value that isn't close to one of | |
3088 | the boundary values below. */ | |
3089 | 0x10000, | |
3090 | ||
3091 | /* Values near LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES. */ | |
3092 | LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES - 0x100, | |
3093 | LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES - 1, | |
3094 | LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES, | |
3095 | LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES + 1, | |
3096 | LINE_MAP_MAX_LOCATION_WITH_PACKED_RANGES + 0x100, | |
3097 | ||
3098 | /* Values near LINE_MAP_MAX_LOCATION_WITH_COLS. */ | |
3099 | LINE_MAP_MAX_LOCATION_WITH_COLS - 0x100, | |
3100 | LINE_MAP_MAX_LOCATION_WITH_COLS - 1, | |
3101 | LINE_MAP_MAX_LOCATION_WITH_COLS, | |
3102 | LINE_MAP_MAX_LOCATION_WITH_COLS + 1, | |
3103 | LINE_MAP_MAX_LOCATION_WITH_COLS + 0x100, | |
3104 | }; | |
3105 | ||
d9b950dd DM |
3106 | /* Run all of the selftests within this file. */ |
3107 | ||
3108 | void | |
3109 | input_c_tests () | |
3110 | { | |
741d3be5 | 3111 | test_should_have_column_data_p (); |
d9b950dd DM |
3112 | test_unknown_location (); |
3113 | test_builtins (); | |
741d3be5 DM |
3114 | |
3115 | /* As noted above in the description of struct line_table_case, | |
3116 | we want to explore a test matrix of interesting line_table | |
3117 | situations, running various selftests for each case within the | |
3118 | matrix. */ | |
3119 | ||
3120 | /* Run all tests with: | |
3121 | (a) line_table->default_range_bits == 0, and | |
3122 | (b) line_table->default_range_bits == 5. */ | |
3123 | int num_cases_tested = 0; | |
3124 | for (int default_range_bits = 0; default_range_bits <= 5; | |
3125 | default_range_bits += 5) | |
3126 | { | |
3127 | /* ...and use each of the "interesting" location values as | |
3128 | the starting location within line_table. */ | |
3129 | const int num_boundary_locations | |
3130 | = sizeof (boundary_locations) / sizeof (boundary_locations[0]); | |
3131 | for (int loc_idx = 0; loc_idx < num_boundary_locations; loc_idx++) | |
3132 | { | |
3133 | line_table_case c (default_range_bits, boundary_locations[loc_idx]); | |
3134 | ||
3135 | /* Run all tests for the given case within the test matrix. */ | |
3136 | test_accessing_ordinary_linemaps (c); | |
3137 | test_lexer (c); | |
88fa5555 DM |
3138 | test_lexer_string_locations_simple (c); |
3139 | test_lexer_string_locations_ebcdic (c); | |
3140 | test_lexer_string_locations_hex (c); | |
3141 | test_lexer_string_locations_oct (c); | |
3142 | test_lexer_string_locations_letter_escape_1 (c); | |
3143 | test_lexer_string_locations_letter_escape_2 (c); | |
3144 | test_lexer_string_locations_ucn4 (c); | |
3145 | test_lexer_string_locations_ucn8 (c); | |
3146 | test_lexer_string_locations_wide_string (c); | |
3147 | test_lexer_string_locations_string16 (c); | |
3148 | test_lexer_string_locations_string32 (c); | |
3149 | test_lexer_string_locations_u8 (c); | |
3150 | test_lexer_string_locations_utf8_source (c); | |
3151 | test_lexer_string_locations_concatenation_1 (c); | |
3152 | test_lexer_string_locations_concatenation_2 (c); | |
3153 | test_lexer_string_locations_concatenation_3 (c); | |
3154 | test_lexer_string_locations_macro (c); | |
3155 | test_lexer_string_locations_stringified_macro_argument (c); | |
3156 | test_lexer_string_locations_non_string (c); | |
3157 | test_lexer_string_locations_long_line (c); | |
3158 | test_lexer_char_constants (c); | |
741d3be5 DM |
3159 | |
3160 | num_cases_tested++; | |
3161 | } | |
3162 | } | |
3163 | ||
3164 | /* Verify that we fully covered the test matrix. */ | |
3165 | ASSERT_EQ (num_cases_tested, 2 * 12); | |
3166 | ||
d9b950dd DM |
3167 | test_reading_source_line (); |
3168 | } | |
3169 | ||
3170 | } // namespace selftest | |
3171 | ||
3172 | #endif /* CHECKING_P */ |