]>
Commit | Line | Data |
---|---|---|
53e1b683 | 1 | /* SPDX-License-Identifier: LGPL-2.1+ */ |
07630cea | 2 | |
11c3a366 TA |
3 | #include <errno.h> |
4 | #include <stdarg.h> | |
5 | #include <stdint.h> | |
6 | #include <stdio.h> | |
7 | #include <stdlib.h> | |
8 | ||
b5efdb8a | 9 | #include "alloc-util.h" |
8409f688 | 10 | #include "escape.h" |
53cd7f33 | 11 | #include "extract-word.h" |
090a9c1e | 12 | #include "fileio.h" |
07630cea | 13 | #include "gunicode.h" |
c30a49b2 | 14 | #include "locale-util.h" |
11c3a366 | 15 | #include "macro.h" |
090a9c1e | 16 | #include "memory-util.h" |
b11d6a7b | 17 | #include "string-util.h" |
46bf625a | 18 | #include "strv.h" |
b4766d5f | 19 | #include "terminal-util.h" |
07630cea LP |
20 | #include "utf8.h" |
21 | #include "util.h" | |
07630cea LP |
22 | |
23 | int strcmp_ptr(const char *a, const char *b) { | |
07630cea | 24 | /* Like strcmp(), but tries to make sense of NULL pointers */ |
6cf3011c | 25 | |
07630cea LP |
26 | if (a && b) |
27 | return strcmp(a, b); | |
6cf3011c ZJS |
28 | return CMP(a, b); /* Direct comparison of pointers, one of which is NULL */ |
29 | } | |
07630cea | 30 | |
6cf3011c ZJS |
31 | int strcasecmp_ptr(const char *a, const char *b) { |
32 | /* Like strcasecmp(), but tries to make sense of NULL pointers */ | |
07630cea | 33 | |
6cf3011c ZJS |
34 | if (a && b) |
35 | return strcasecmp(a, b); | |
36 | return CMP(a, b); /* Direct comparison of pointers, one of which is NULL */ | |
07630cea LP |
37 | } |
38 | ||
39 | char* endswith(const char *s, const char *postfix) { | |
40 | size_t sl, pl; | |
41 | ||
42 | assert(s); | |
43 | assert(postfix); | |
44 | ||
45 | sl = strlen(s); | |
46 | pl = strlen(postfix); | |
47 | ||
48 | if (pl == 0) | |
49 | return (char*) s + sl; | |
50 | ||
51 | if (sl < pl) | |
52 | return NULL; | |
53 | ||
54 | if (memcmp(s + sl - pl, postfix, pl) != 0) | |
55 | return NULL; | |
56 | ||
57 | return (char*) s + sl - pl; | |
58 | } | |
59 | ||
60 | char* endswith_no_case(const char *s, const char *postfix) { | |
61 | size_t sl, pl; | |
62 | ||
63 | assert(s); | |
64 | assert(postfix); | |
65 | ||
66 | sl = strlen(s); | |
67 | pl = strlen(postfix); | |
68 | ||
69 | if (pl == 0) | |
70 | return (char*) s + sl; | |
71 | ||
72 | if (sl < pl) | |
73 | return NULL; | |
74 | ||
75 | if (strcasecmp(s + sl - pl, postfix) != 0) | |
76 | return NULL; | |
77 | ||
78 | return (char*) s + sl - pl; | |
79 | } | |
80 | ||
81 | char* first_word(const char *s, const char *word) { | |
82 | size_t sl, wl; | |
83 | const char *p; | |
84 | ||
85 | assert(s); | |
86 | assert(word); | |
87 | ||
88 | /* Checks if the string starts with the specified word, either | |
89 | * followed by NUL or by whitespace. Returns a pointer to the | |
90 | * NUL or the first character after the whitespace. */ | |
91 | ||
92 | sl = strlen(s); | |
93 | wl = strlen(word); | |
94 | ||
95 | if (sl < wl) | |
96 | return NULL; | |
97 | ||
98 | if (wl == 0) | |
99 | return (char*) s; | |
100 | ||
101 | if (memcmp(s, word, wl) != 0) | |
102 | return NULL; | |
103 | ||
104 | p = s + wl; | |
105 | if (*p == 0) | |
106 | return (char*) p; | |
107 | ||
108 | if (!strchr(WHITESPACE, *p)) | |
109 | return NULL; | |
110 | ||
111 | p += strspn(p, WHITESPACE); | |
112 | return (char*) p; | |
113 | } | |
114 | ||
115 | static size_t strcspn_escaped(const char *s, const char *reject) { | |
116 | bool escaped = false; | |
117 | int n; | |
118 | ||
29965a2a | 119 | for (n = 0; s[n] != '\0'; n++) { |
07630cea LP |
120 | if (escaped) |
121 | escaped = false; | |
122 | else if (s[n] == '\\') | |
123 | escaped = true; | |
124 | else if (strchr(reject, s[n])) | |
125 | break; | |
126 | } | |
127 | ||
29965a2a | 128 | return n; |
07630cea LP |
129 | } |
130 | ||
131 | /* Split a string into words. */ | |
5fed82c6 LP |
132 | const char* split( |
133 | const char **state, | |
134 | size_t *l, | |
135 | const char *separator, | |
136 | SplitFlags flags) { | |
137 | ||
07630cea LP |
138 | const char *current; |
139 | ||
5fed82c6 LP |
140 | assert(state); |
141 | assert(l); | |
142 | ||
143 | if (!separator) | |
144 | separator = WHITESPACE; | |
145 | ||
07630cea LP |
146 | current = *state; |
147 | ||
5fed82c6 | 148 | if (*current == '\0') /* already at the end? */ |
07630cea | 149 | return NULL; |
07630cea | 150 | |
5fed82c6 LP |
151 | current += strspn(current, separator); /* skip leading separators */ |
152 | if (*current == '\0') { /* at the end now? */ | |
07630cea LP |
153 | *state = current; |
154 | return NULL; | |
155 | } | |
156 | ||
5fed82c6 LP |
157 | if (FLAGS_SET(flags, SPLIT_QUOTES)) { |
158 | ||
159 | if (strchr(QUOTES, *current)) { | |
160 | /* We are looking at a quote */ | |
161 | *l = strcspn_escaped(current + 1, CHAR_TO_STR(*current)); | |
162 | if (current[*l + 1] != *current || | |
163 | (current[*l + 2] != 0 && !strchr(separator, current[*l + 2]))) { | |
164 | /* right quote missing or garbage at the end */ | |
165 | if (FLAGS_SET(flags, SPLIT_RELAX)) { | |
166 | *state = current + *l + 1 + (current[*l + 1] != '\0'); | |
167 | return current + 1; | |
168 | } | |
169 | *state = current; | |
170 | return NULL; | |
8059aa9c | 171 | } |
5fed82c6 LP |
172 | *state = current++ + *l + 2; |
173 | ||
174 | } else { | |
175 | /* We are looking at a something that is not a quote */ | |
176 | *l = strcspn_escaped(current, separator); | |
177 | if (current[*l] && !strchr(separator, current[*l]) && !FLAGS_SET(flags, SPLIT_RELAX)) { | |
178 | /* unfinished escape */ | |
179 | *state = current; | |
180 | return NULL; | |
181 | } | |
182 | *state = current + *l; | |
07630cea | 183 | } |
07630cea LP |
184 | } else { |
185 | *l = strcspn(current, separator); | |
186 | *state = current + *l; | |
187 | } | |
188 | ||
189 | return current; | |
190 | } | |
191 | ||
192 | char *strnappend(const char *s, const char *suffix, size_t b) { | |
193 | size_t a; | |
194 | char *r; | |
195 | ||
196 | if (!s && !suffix) | |
197 | return strdup(""); | |
198 | ||
199 | if (!s) | |
200 | return strndup(suffix, b); | |
201 | ||
202 | if (!suffix) | |
203 | return strdup(s); | |
204 | ||
205 | assert(s); | |
206 | assert(suffix); | |
207 | ||
208 | a = strlen(s); | |
209 | if (b > ((size_t) -1) - a) | |
210 | return NULL; | |
211 | ||
212 | r = new(char, a+b+1); | |
213 | if (!r) | |
214 | return NULL; | |
215 | ||
216 | memcpy(r, s, a); | |
217 | memcpy(r+a, suffix, b); | |
218 | r[a+b] = 0; | |
219 | ||
220 | return r; | |
221 | } | |
222 | ||
605405c6 | 223 | char *strjoin_real(const char *x, ...) { |
07630cea LP |
224 | va_list ap; |
225 | size_t l; | |
226 | char *r, *p; | |
227 | ||
228 | va_start(ap, x); | |
229 | ||
230 | if (x) { | |
231 | l = strlen(x); | |
232 | ||
233 | for (;;) { | |
234 | const char *t; | |
235 | size_t n; | |
236 | ||
237 | t = va_arg(ap, const char *); | |
238 | if (!t) | |
239 | break; | |
240 | ||
241 | n = strlen(t); | |
242 | if (n > ((size_t) -1) - l) { | |
243 | va_end(ap); | |
244 | return NULL; | |
245 | } | |
246 | ||
247 | l += n; | |
248 | } | |
249 | } else | |
250 | l = 0; | |
251 | ||
252 | va_end(ap); | |
253 | ||
254 | r = new(char, l+1); | |
255 | if (!r) | |
256 | return NULL; | |
257 | ||
258 | if (x) { | |
259 | p = stpcpy(r, x); | |
260 | ||
261 | va_start(ap, x); | |
262 | ||
263 | for (;;) { | |
264 | const char *t; | |
265 | ||
266 | t = va_arg(ap, const char *); | |
267 | if (!t) | |
268 | break; | |
269 | ||
270 | p = stpcpy(p, t); | |
271 | } | |
272 | ||
273 | va_end(ap); | |
274 | } else | |
275 | r[0] = 0; | |
276 | ||
277 | return r; | |
278 | } | |
279 | ||
280 | char *strstrip(char *s) { | |
7546145e LP |
281 | if (!s) |
282 | return NULL; | |
283 | ||
0a6ffc5c | 284 | /* Drops trailing whitespace. Modifies the string in place. Returns pointer to first non-space character */ |
07630cea | 285 | |
0a6ffc5c | 286 | return delete_trailing_chars(skip_leading_chars(s, WHITESPACE), WHITESPACE); |
07630cea LP |
287 | } |
288 | ||
289 | char *delete_chars(char *s, const char *bad) { | |
290 | char *f, *t; | |
291 | ||
7546145e LP |
292 | /* Drops all specified bad characters, regardless where in the string */ |
293 | ||
294 | if (!s) | |
295 | return NULL; | |
296 | ||
297 | if (!bad) | |
298 | bad = WHITESPACE; | |
07630cea LP |
299 | |
300 | for (f = s, t = s; *f; f++) { | |
301 | if (strchr(bad, *f)) | |
302 | continue; | |
303 | ||
304 | *(t++) = *f; | |
305 | } | |
306 | ||
307 | *t = 0; | |
308 | ||
309 | return s; | |
310 | } | |
311 | ||
7546145e LP |
312 | char *delete_trailing_chars(char *s, const char *bad) { |
313 | char *p, *c = s; | |
314 | ||
315 | /* Drops all specified bad characters, at the end of the string */ | |
316 | ||
317 | if (!s) | |
318 | return NULL; | |
319 | ||
320 | if (!bad) | |
321 | bad = WHITESPACE; | |
322 | ||
323 | for (p = s; *p; p++) | |
324 | if (!strchr(bad, *p)) | |
325 | c = p + 1; | |
326 | ||
327 | *c = 0; | |
328 | ||
329 | return s; | |
330 | } | |
331 | ||
07630cea LP |
332 | char *truncate_nl(char *s) { |
333 | assert(s); | |
334 | ||
335 | s[strcspn(s, NEWLINE)] = 0; | |
336 | return s; | |
337 | } | |
338 | ||
b577e3d5 LP |
339 | char ascii_tolower(char x) { |
340 | ||
341 | if (x >= 'A' && x <= 'Z') | |
342 | return x - 'A' + 'a'; | |
343 | ||
344 | return x; | |
345 | } | |
346 | ||
846b8fc3 LP |
347 | char ascii_toupper(char x) { |
348 | ||
349 | if (x >= 'a' && x <= 'z') | |
350 | return x - 'a' + 'A'; | |
351 | ||
352 | return x; | |
353 | } | |
354 | ||
07630cea LP |
355 | char *ascii_strlower(char *t) { |
356 | char *p; | |
357 | ||
358 | assert(t); | |
359 | ||
360 | for (p = t; *p; p++) | |
b577e3d5 LP |
361 | *p = ascii_tolower(*p); |
362 | ||
363 | return t; | |
364 | } | |
365 | ||
846b8fc3 LP |
366 | char *ascii_strupper(char *t) { |
367 | char *p; | |
368 | ||
369 | assert(t); | |
370 | ||
371 | for (p = t; *p; p++) | |
372 | *p = ascii_toupper(*p); | |
373 | ||
374 | return t; | |
375 | } | |
376 | ||
b577e3d5 LP |
377 | char *ascii_strlower_n(char *t, size_t n) { |
378 | size_t i; | |
379 | ||
380 | if (n <= 0) | |
381 | return t; | |
382 | ||
383 | for (i = 0; i < n; i++) | |
384 | t[i] = ascii_tolower(t[i]); | |
07630cea LP |
385 | |
386 | return t; | |
387 | } | |
522d85ae LP |
388 | |
389 | int ascii_strcasecmp_n(const char *a, const char *b, size_t n) { | |
390 | ||
391 | for (; n > 0; a++, b++, n--) { | |
392 | int x, y; | |
393 | ||
394 | x = (int) (uint8_t) ascii_tolower(*a); | |
395 | y = (int) (uint8_t) ascii_tolower(*b); | |
396 | ||
397 | if (x != y) | |
398 | return x - y; | |
399 | } | |
400 | ||
401 | return 0; | |
402 | } | |
c1749834 LP |
403 | |
404 | int ascii_strcasecmp_nn(const char *a, size_t n, const char *b, size_t m) { | |
405 | int r; | |
406 | ||
407 | r = ascii_strcasecmp_n(a, b, MIN(n, m)); | |
408 | if (r != 0) | |
409 | return r; | |
410 | ||
6dd91b36 | 411 | return CMP(n, m); |
c1749834 | 412 | } |
07630cea LP |
413 | |
414 | bool chars_intersect(const char *a, const char *b) { | |
415 | const char *p; | |
416 | ||
417 | /* Returns true if any of the chars in a are in b. */ | |
418 | for (p = a; *p; p++) | |
419 | if (strchr(b, *p)) | |
420 | return true; | |
421 | ||
422 | return false; | |
423 | } | |
424 | ||
425 | bool string_has_cc(const char *p, const char *ok) { | |
426 | const char *t; | |
427 | ||
428 | assert(p); | |
429 | ||
430 | /* | |
431 | * Check if a string contains control characters. If 'ok' is | |
432 | * non-NULL it may be a string containing additional CCs to be | |
433 | * considered OK. | |
434 | */ | |
435 | ||
436 | for (t = p; *t; t++) { | |
437 | if (ok && strchr(ok, *t)) | |
438 | continue; | |
439 | ||
440 | if (*t > 0 && *t < ' ') | |
441 | return true; | |
442 | ||
443 | if (*t == 127) | |
444 | return true; | |
445 | } | |
446 | ||
447 | return false; | |
448 | } | |
449 | ||
8409f688 ZJS |
450 | static int write_ellipsis(char *buf, bool unicode) { |
451 | if (unicode || is_locale_utf8()) { | |
452 | buf[0] = 0xe2; /* tri-dot ellipsis: … */ | |
453 | buf[1] = 0x80; | |
454 | buf[2] = 0xa6; | |
455 | } else { | |
456 | buf[0] = '.'; | |
457 | buf[1] = '.'; | |
458 | buf[2] = '.'; | |
459 | } | |
460 | ||
461 | return 3; | |
462 | } | |
463 | ||
07630cea | 464 | static char *ascii_ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { |
9924aef6 ZJS |
465 | size_t x, need_space, suffix_len; |
466 | char *t; | |
07630cea LP |
467 | |
468 | assert(s); | |
469 | assert(percent <= 100); | |
c30a49b2 | 470 | assert(new_length != (size_t) -1); |
07630cea | 471 | |
c30a49b2 | 472 | if (old_length <= new_length) |
07630cea LP |
473 | return strndup(s, old_length); |
474 | ||
c30a49b2 LP |
475 | /* Special case short ellipsations */ |
476 | switch (new_length) { | |
477 | ||
478 | case 0: | |
479 | return strdup(""); | |
480 | ||
481 | case 1: | |
482 | if (is_locale_utf8()) | |
483 | return strdup("…"); | |
484 | else | |
485 | return strdup("."); | |
486 | ||
487 | case 2: | |
488 | if (!is_locale_utf8()) | |
489 | return strdup(".."); | |
490 | ||
491 | break; | |
492 | ||
493 | default: | |
494 | break; | |
495 | } | |
496 | ||
497 | /* Calculate how much space the ellipsis will take up. If we are in UTF-8 mode we only need space for one | |
498 | * character ("…"), otherwise for three characters ("..."). Note that in both cases we need 3 bytes of storage, | |
499 | * either for the UTF-8 encoded character or for three ASCII characters. */ | |
500 | need_space = is_locale_utf8() ? 1 : 3; | |
501 | ||
9924aef6 ZJS |
502 | t = new(char, new_length+3); |
503 | if (!t) | |
07630cea LP |
504 | return NULL; |
505 | ||
c30a49b2 | 506 | assert(new_length >= need_space); |
07630cea | 507 | |
c30a49b2 LP |
508 | x = ((new_length - need_space) * percent + 50) / 100; |
509 | assert(x <= new_length - need_space); | |
07630cea | 510 | |
9924aef6 ZJS |
511 | memcpy(t, s, x); |
512 | write_ellipsis(t + x, false); | |
513 | suffix_len = new_length - x - need_space; | |
514 | memcpy(t + x + 3, s + old_length - suffix_len, suffix_len); | |
515 | *(t + x + 3 + suffix_len) = '\0'; | |
07630cea | 516 | |
9924aef6 | 517 | return t; |
07630cea LP |
518 | } |
519 | ||
520 | char *ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 521 | size_t x, k, len, len2; |
07630cea | 522 | const char *i, *j; |
c30a49b2 | 523 | char *e; |
c932fb71 | 524 | int r; |
07630cea | 525 | |
c30a49b2 LP |
526 | /* Note that 'old_length' refers to bytes in the string, while 'new_length' refers to character cells taken up |
527 | * on screen. This distinction doesn't matter for ASCII strings, but it does matter for non-ASCII UTF-8 | |
528 | * strings. | |
529 | * | |
530 | * Ellipsation is done in a locale-dependent way: | |
531 | * 1. If the string passed in is fully ASCII and the current locale is not UTF-8, three dots are used ("...") | |
532 | * 2. Otherwise, a unicode ellipsis is used ("…") | |
533 | * | |
534 | * In other words: you'll get a unicode ellipsis as soon as either the string contains non-ASCII characters or | |
535 | * the current locale is UTF-8. | |
536 | */ | |
537 | ||
07630cea LP |
538 | assert(s); |
539 | assert(percent <= 100); | |
ddbc9319 LP |
540 | |
541 | if (new_length == (size_t) -1) | |
542 | return strndup(s, old_length); | |
543 | ||
c30a49b2 LP |
544 | if (new_length == 0) |
545 | return strdup(""); | |
07630cea | 546 | |
c30a49b2 | 547 | /* If no multibyte characters use ascii_ellipsize_mem for speed */ |
21e4e3e0 | 548 | if (ascii_is_valid_n(s, old_length)) |
07630cea LP |
549 | return ascii_ellipsize_mem(s, old_length, new_length, percent); |
550 | ||
c30a49b2 LP |
551 | x = ((new_length - 1) * percent) / 100; |
552 | assert(x <= new_length - 1); | |
07630cea LP |
553 | |
554 | k = 0; | |
9924aef6 | 555 | for (i = s; i < s + old_length; i = utf8_next_char(i)) { |
c932fb71 | 556 | char32_t c; |
9924aef6 | 557 | int w; |
07630cea | 558 | |
c932fb71 SL |
559 | r = utf8_encoded_to_unichar(i, &c); |
560 | if (r < 0) | |
07630cea | 561 | return NULL; |
07630cea | 562 | |
9924aef6 ZJS |
563 | w = unichar_iswide(c) ? 2 : 1; |
564 | if (k + w <= x) | |
565 | k += w; | |
566 | else | |
567 | break; | |
568 | } | |
07630cea | 569 | |
9924aef6 | 570 | for (j = s + old_length; j > i; ) { |
c932fb71 | 571 | char32_t c; |
9924aef6 ZJS |
572 | int w; |
573 | const char *jj; | |
07630cea | 574 | |
9924aef6 ZJS |
575 | jj = utf8_prev_char(j); |
576 | r = utf8_encoded_to_unichar(jj, &c); | |
c932fb71 | 577 | if (r < 0) |
07630cea | 578 | return NULL; |
9924aef6 ZJS |
579 | |
580 | w = unichar_iswide(c) ? 2 : 1; | |
581 | if (k + w <= new_length) { | |
582 | k += w; | |
583 | j = jj; | |
584 | } else | |
585 | break; | |
07630cea LP |
586 | } |
587 | assert(i <= j); | |
588 | ||
589 | /* we don't actually need to ellipsize */ | |
590 | if (i == j) | |
9924aef6 | 591 | return memdup_suffix0(s, old_length); |
07630cea | 592 | |
9924aef6 ZJS |
593 | /* make space for ellipsis, if possible */ |
594 | if (j < s + old_length) | |
595 | j = utf8_next_char(j); | |
596 | else if (i > s) | |
597 | i = utf8_prev_char(i); | |
07630cea LP |
598 | |
599 | len = i - s; | |
600 | len2 = s + old_length - j; | |
601 | e = new(char, len + 3 + len2 + 1); | |
602 | if (!e) | |
603 | return NULL; | |
604 | ||
605 | /* | |
606 | printf("old_length=%zu new_length=%zu x=%zu len=%u len2=%u k=%u\n", | |
607 | old_length, new_length, x, len, len2, k); | |
608 | */ | |
609 | ||
610 | memcpy(e, s, len); | |
8409f688 | 611 | write_ellipsis(e + len, true); |
9924aef6 ZJS |
612 | memcpy(e + len + 3, j, len2); |
613 | *(e + len + 3 + len2) = '\0'; | |
07630cea LP |
614 | |
615 | return e; | |
616 | } | |
617 | ||
8409f688 ZJS |
618 | char *cellescape(char *buf, size_t len, const char *s) { |
619 | /* Escape and ellipsize s into buffer buf of size len. Only non-control ASCII | |
620 | * characters are copied as they are, everything else is escaped. The result | |
621 | * is different then if escaping and ellipsization was performed in two | |
622 | * separate steps, because each sequence is either stored in full or skipped. | |
623 | * | |
624 | * This function should be used for logging about strings which expected to | |
625 | * be plain ASCII in a safe way. | |
626 | * | |
627 | * An ellipsis will be used if s is too long. It was always placed at the | |
628 | * very end. | |
629 | */ | |
630 | ||
61f6e276 LP |
631 | size_t i = 0, last_char_width[4] = {}, k = 0, j; |
632 | ||
633 | assert(len > 0); /* at least a terminating NUL */ | |
8409f688 | 634 | |
61f6e276 LP |
635 | for (;;) { |
636 | char four[4]; | |
637 | int w; | |
8409f688 | 638 | |
61f6e276 | 639 | if (*s == 0) /* terminating NUL detected? then we are done! */ |
8409f688 | 640 | goto done; |
61f6e276 LP |
641 | |
642 | w = cescape_char(*s, four); | |
643 | if (i + w + 1 > len) /* This character doesn't fit into the buffer anymore? In that case let's | |
644 | * ellipsize at the previous location */ | |
645 | break; | |
646 | ||
647 | /* OK, there was space, let's add this escaped character to the buffer */ | |
648 | memcpy(buf + i, four, w); | |
649 | i += w; | |
650 | ||
651 | /* And remember its width in the ring buffer */ | |
652 | last_char_width[k] = w; | |
653 | k = (k + 1) % 4; | |
654 | ||
655 | s++; | |
8409f688 ZJS |
656 | } |
657 | ||
61f6e276 LP |
658 | /* Ellipsation is necessary. This means we might need to truncate the string again to make space for 4 |
659 | * characters ideally, but the buffer is shorter than that in the first place take what we can get */ | |
660 | for (j = 0; j < ELEMENTSOF(last_char_width); j++) { | |
661 | ||
662 | if (i + 4 <= len) /* nice, we reached our space goal */ | |
663 | break; | |
664 | ||
665 | k = k == 0 ? 3 : k - 1; | |
666 | if (last_char_width[k] == 0) /* bummer, we reached the beginning of the strings */ | |
667 | break; | |
668 | ||
669 | assert(i >= last_char_width[k]); | |
670 | i -= last_char_width[k]; | |
8409f688 ZJS |
671 | } |
672 | ||
61f6e276 LP |
673 | if (i + 4 <= len) /* yay, enough space */ |
674 | i += write_ellipsis(buf + i, false); | |
675 | else if (i + 3 <= len) { /* only space for ".." */ | |
676 | buf[i++] = '.'; | |
677 | buf[i++] = '.'; | |
678 | } else if (i + 2 <= len) /* only space for a single "." */ | |
679 | buf[i++] = '.'; | |
680 | else | |
681 | assert(i + 1 <= len); | |
682 | ||
8409f688 ZJS |
683 | done: |
684 | buf[i] = '\0'; | |
685 | return buf; | |
686 | } | |
687 | ||
07630cea LP |
688 | char* strshorten(char *s, size_t l) { |
689 | assert(s); | |
690 | ||
47b33c7d | 691 | if (strnlen(s, l+1) > l) |
07630cea LP |
692 | s[l] = 0; |
693 | ||
694 | return s; | |
695 | } | |
696 | ||
697 | char *strreplace(const char *text, const char *old_string, const char *new_string) { | |
9d73565a LP |
698 | size_t l, old_len, new_len, allocated = 0; |
699 | char *t, *ret = NULL; | |
07630cea | 700 | const char *f; |
07630cea | 701 | |
07630cea LP |
702 | assert(old_string); |
703 | assert(new_string); | |
704 | ||
9d73565a LP |
705 | if (!text) |
706 | return NULL; | |
707 | ||
07630cea LP |
708 | old_len = strlen(old_string); |
709 | new_len = strlen(new_string); | |
710 | ||
711 | l = strlen(text); | |
9d73565a | 712 | if (!GREEDY_REALLOC(ret, allocated, l+1)) |
07630cea LP |
713 | return NULL; |
714 | ||
715 | f = text; | |
9d73565a | 716 | t = ret; |
07630cea | 717 | while (*f) { |
07630cea LP |
718 | size_t d, nl; |
719 | ||
720 | if (!startswith(f, old_string)) { | |
721 | *(t++) = *(f++); | |
722 | continue; | |
723 | } | |
724 | ||
9d73565a | 725 | d = t - ret; |
07630cea | 726 | nl = l - old_len + new_len; |
9d73565a LP |
727 | |
728 | if (!GREEDY_REALLOC(ret, allocated, nl + 1)) | |
729 | return mfree(ret); | |
07630cea LP |
730 | |
731 | l = nl; | |
9d73565a | 732 | t = ret + d; |
07630cea LP |
733 | |
734 | t = stpcpy(t, new_string); | |
735 | f += old_len; | |
736 | } | |
737 | ||
738 | *t = 0; | |
9d73565a | 739 | return ret; |
07630cea LP |
740 | } |
741 | ||
6fb05690 LP |
742 | static void advance_offsets( |
743 | ssize_t diff, | |
744 | size_t offsets[2], /* note: we can't use [static 2] here, since this may be NULL */ | |
745 | size_t shift[static 2], | |
746 | size_t size) { | |
747 | ||
b4766d5f ZJS |
748 | if (!offsets) |
749 | return; | |
750 | ||
6fb05690 LP |
751 | assert(shift); |
752 | ||
b4766d5f ZJS |
753 | if ((size_t) diff < offsets[0]) |
754 | shift[0] += size; | |
755 | if ((size_t) diff < offsets[1]) | |
756 | shift[1] += size; | |
757 | } | |
758 | ||
759 | char *strip_tab_ansi(char **ibuf, size_t *_isz, size_t highlight[2]) { | |
62a3fc6d | 760 | const char *begin = NULL; |
07630cea LP |
761 | enum { |
762 | STATE_OTHER, | |
763 | STATE_ESCAPE, | |
695a944c LP |
764 | STATE_CSI, |
765 | STATE_CSO, | |
07630cea LP |
766 | } state = STATE_OTHER; |
767 | char *obuf = NULL; | |
62a3fc6d | 768 | size_t osz = 0, isz, shift[2] = {}, n_carriage_returns = 0; |
07630cea LP |
769 | FILE *f; |
770 | ||
771 | assert(ibuf); | |
772 | assert(*ibuf); | |
773 | ||
695a944c LP |
774 | /* This does three things: |
775 | * | |
776 | * 1. Replaces TABs by 8 spaces | |
777 | * 2. Strips ANSI color sequences (a subset of CSI), i.e. ESC '[' … 'm' sequences | |
778 | * 3. Strips ANSI operating system sequences (CSO), i.e. ESC ']' … BEL sequences | |
62a3fc6d ZJS |
779 | * 4. Strip trailing \r characters (since they would "move the cursor", but have no |
780 | * other effect). | |
695a944c | 781 | * |
2fe21124 ZJS |
782 | * Everything else will be left as it is. In particular other ANSI sequences are left as they are, as |
783 | * are any other special characters. Truncated ANSI sequences are left-as is too. This call is | |
784 | * supposed to suppress the most basic formatting noise, but nothing else. | |
695a944c LP |
785 | * |
786 | * Why care for CSO sequences? Well, to undo what terminal_urlify() and friends generate. */ | |
07630cea LP |
787 | |
788 | isz = _isz ? *_isz : strlen(*ibuf); | |
789 | ||
2fe21124 ZJS |
790 | /* Note we turn off internal locking on f for performance reasons. It's safe to do so since we |
791 | * created f here and it doesn't leave our scope. */ | |
792 | f = open_memstream_unlocked(&obuf, &osz); | |
07630cea LP |
793 | if (!f) |
794 | return NULL; | |
795 | ||
62a3fc6d | 796 | for (const char *i = *ibuf; i < *ibuf + isz + 1; i++) { |
07630cea LP |
797 | |
798 | switch (state) { | |
799 | ||
800 | case STATE_OTHER: | |
801 | if (i >= *ibuf + isz) /* EOT */ | |
802 | break; | |
62a3fc6d ZJS |
803 | |
804 | if (*i == '\r') { | |
805 | n_carriage_returns++; | |
806 | break; | |
807 | } else if (*i == '\n') | |
808 | /* Ignore carriage returns before new line */ | |
809 | n_carriage_returns = 0; | |
810 | for (; n_carriage_returns > 0; n_carriage_returns--) | |
811 | fputc('\r', f); | |
812 | ||
813 | if (*i == '\x1B') | |
07630cea | 814 | state = STATE_ESCAPE; |
b4766d5f | 815 | else if (*i == '\t') { |
0d536673 | 816 | fputs(" ", f); |
b4766d5f ZJS |
817 | advance_offsets(i - *ibuf, highlight, shift, 7); |
818 | } else | |
0d536673 | 819 | fputc(*i, f); |
b4766d5f | 820 | |
07630cea LP |
821 | break; |
822 | ||
823 | case STATE_ESCAPE: | |
62a3fc6d ZJS |
824 | assert(n_carriage_returns == 0); |
825 | ||
07630cea | 826 | if (i >= *ibuf + isz) { /* EOT */ |
0d536673 | 827 | fputc('\x1B', f); |
b4766d5f | 828 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea | 829 | break; |
695a944c LP |
830 | } else if (*i == '[') { /* ANSI CSI */ |
831 | state = STATE_CSI; | |
832 | begin = i + 1; | |
833 | } else if (*i == ']') { /* ANSI CSO */ | |
834 | state = STATE_CSO; | |
07630cea LP |
835 | begin = i + 1; |
836 | } else { | |
0d536673 LP |
837 | fputc('\x1B', f); |
838 | fputc(*i, f); | |
b4766d5f | 839 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea LP |
840 | state = STATE_OTHER; |
841 | } | |
842 | ||
843 | break; | |
844 | ||
695a944c | 845 | case STATE_CSI: |
62a3fc6d | 846 | assert(n_carriage_returns == 0); |
07630cea | 847 | |
695a944c LP |
848 | if (i >= *ibuf + isz || /* EOT … */ |
849 | !strchr("01234567890;m", *i)) { /* … or invalid chars in sequence */ | |
0d536673 LP |
850 | fputc('\x1B', f); |
851 | fputc('[', f); | |
b4766d5f | 852 | advance_offsets(i - *ibuf, highlight, shift, 2); |
07630cea LP |
853 | state = STATE_OTHER; |
854 | i = begin-1; | |
855 | } else if (*i == 'm') | |
856 | state = STATE_OTHER; | |
695a944c LP |
857 | |
858 | break; | |
859 | ||
860 | case STATE_CSO: | |
62a3fc6d | 861 | assert(n_carriage_returns == 0); |
695a944c LP |
862 | |
863 | if (i >= *ibuf + isz || /* EOT … */ | |
864 | (*i != '\a' && (uint8_t) *i < 32U) || (uint8_t) *i > 126U) { /* … or invalid chars in sequence */ | |
865 | fputc('\x1B', f); | |
866 | fputc(']', f); | |
867 | advance_offsets(i - *ibuf, highlight, shift, 2); | |
868 | state = STATE_OTHER; | |
869 | i = begin-1; | |
870 | } else if (*i == '\a') | |
871 | state = STATE_OTHER; | |
872 | ||
07630cea LP |
873 | break; |
874 | } | |
875 | } | |
876 | ||
c7e03d2e | 877 | if (fflush_and_check(f) < 0) { |
07630cea | 878 | fclose(f); |
6b430fdb | 879 | return mfree(obuf); |
07630cea | 880 | } |
07630cea LP |
881 | fclose(f); |
882 | ||
6fb05690 | 883 | free_and_replace(*ibuf, obuf); |
07630cea LP |
884 | |
885 | if (_isz) | |
886 | *_isz = osz; | |
887 | ||
b4766d5f ZJS |
888 | if (highlight) { |
889 | highlight[0] += shift[0]; | |
890 | highlight[1] += shift[1]; | |
891 | } | |
892 | ||
6fb05690 | 893 | return *ibuf; |
07630cea LP |
894 | } |
895 | ||
bb8ad9ea LP |
896 | char *strextend_with_separator(char **x, const char *separator, ...) { |
897 | bool need_separator; | |
898 | size_t f, l, l_separator; | |
07630cea | 899 | char *r, *p; |
bb8ad9ea | 900 | va_list ap; |
07630cea LP |
901 | |
902 | assert(x); | |
903 | ||
7bf7ce28 | 904 | l = f = strlen_ptr(*x); |
07630cea | 905 | |
bb8ad9ea LP |
906 | need_separator = !isempty(*x); |
907 | l_separator = strlen_ptr(separator); | |
908 | ||
909 | va_start(ap, separator); | |
07630cea LP |
910 | for (;;) { |
911 | const char *t; | |
912 | size_t n; | |
913 | ||
914 | t = va_arg(ap, const char *); | |
915 | if (!t) | |
916 | break; | |
917 | ||
918 | n = strlen(t); | |
bb8ad9ea LP |
919 | |
920 | if (need_separator) | |
921 | n += l_separator; | |
922 | ||
07630cea LP |
923 | if (n > ((size_t) -1) - l) { |
924 | va_end(ap); | |
925 | return NULL; | |
926 | } | |
927 | ||
928 | l += n; | |
bb8ad9ea | 929 | need_separator = true; |
07630cea LP |
930 | } |
931 | va_end(ap); | |
932 | ||
bb8ad9ea LP |
933 | need_separator = !isempty(*x); |
934 | ||
07630cea LP |
935 | r = realloc(*x, l+1); |
936 | if (!r) | |
937 | return NULL; | |
938 | ||
939 | p = r + f; | |
940 | ||
bb8ad9ea | 941 | va_start(ap, separator); |
07630cea LP |
942 | for (;;) { |
943 | const char *t; | |
944 | ||
945 | t = va_arg(ap, const char *); | |
946 | if (!t) | |
947 | break; | |
948 | ||
bb8ad9ea LP |
949 | if (need_separator && separator) |
950 | p = stpcpy(p, separator); | |
951 | ||
07630cea | 952 | p = stpcpy(p, t); |
bb8ad9ea LP |
953 | |
954 | need_separator = true; | |
07630cea LP |
955 | } |
956 | va_end(ap); | |
957 | ||
bb8ad9ea LP |
958 | assert(p == r + l); |
959 | ||
07630cea LP |
960 | *p = 0; |
961 | *x = r; | |
962 | ||
963 | return r + l; | |
964 | } | |
965 | ||
966 | char *strrep(const char *s, unsigned n) { | |
967 | size_t l; | |
968 | char *r, *p; | |
969 | unsigned i; | |
970 | ||
971 | assert(s); | |
972 | ||
973 | l = strlen(s); | |
974 | p = r = malloc(l * n + 1); | |
975 | if (!r) | |
976 | return NULL; | |
977 | ||
978 | for (i = 0; i < n; i++) | |
979 | p = stpcpy(p, s); | |
980 | ||
981 | *p = 0; | |
982 | return r; | |
983 | } | |
984 | ||
985 | int split_pair(const char *s, const char *sep, char **l, char **r) { | |
986 | char *x, *a, *b; | |
987 | ||
988 | assert(s); | |
989 | assert(sep); | |
990 | assert(l); | |
991 | assert(r); | |
992 | ||
993 | if (isempty(sep)) | |
994 | return -EINVAL; | |
995 | ||
996 | x = strstr(s, sep); | |
997 | if (!x) | |
998 | return -EINVAL; | |
999 | ||
1000 | a = strndup(s, x - s); | |
1001 | if (!a) | |
1002 | return -ENOMEM; | |
1003 | ||
1004 | b = strdup(x + strlen(sep)); | |
1005 | if (!b) { | |
1006 | free(a); | |
1007 | return -ENOMEM; | |
1008 | } | |
1009 | ||
1010 | *l = a; | |
1011 | *r = b; | |
1012 | ||
1013 | return 0; | |
1014 | } | |
1015 | ||
1016 | int free_and_strdup(char **p, const char *s) { | |
1017 | char *t; | |
1018 | ||
1019 | assert(p); | |
1020 | ||
7f546026 | 1021 | /* Replaces a string pointer with a strdup()ed new string, |
07630cea LP |
1022 | * possibly freeing the old one. */ |
1023 | ||
1024 | if (streq_ptr(*p, s)) | |
1025 | return 0; | |
1026 | ||
1027 | if (s) { | |
1028 | t = strdup(s); | |
1029 | if (!t) | |
1030 | return -ENOMEM; | |
1031 | } else | |
1032 | t = NULL; | |
1033 | ||
1034 | free(*p); | |
1035 | *p = t; | |
1036 | ||
1037 | return 1; | |
1038 | } | |
1039 | ||
7f546026 ZJS |
1040 | int free_and_strndup(char **p, const char *s, size_t l) { |
1041 | char *t; | |
1042 | ||
1043 | assert(p); | |
1044 | assert(s || l == 0); | |
1045 | ||
1046 | /* Replaces a string pointer with a strndup()ed new string, | |
1047 | * freeing the old one. */ | |
1048 | ||
1049 | if (!*p && !s) | |
1050 | return 0; | |
1051 | ||
1052 | if (*p && s && strneq(*p, s, l) && (l > strlen(*p) || (*p)[l] == '\0')) | |
1053 | return 0; | |
1054 | ||
1055 | if (s) { | |
1056 | t = strndup(s, l); | |
1057 | if (!t) | |
1058 | return -ENOMEM; | |
1059 | } else | |
1060 | t = NULL; | |
1061 | ||
1062 | free_and_replace(*p, t); | |
1063 | return 1; | |
1064 | } | |
1065 | ||
f3e2e81d LP |
1066 | bool string_is_safe(const char *p) { |
1067 | const char *t; | |
1068 | ||
1069 | if (!p) | |
1070 | return false; | |
1071 | ||
839d1b20 LP |
1072 | /* Checks if the specified string contains no quotes or control characters */ |
1073 | ||
f3e2e81d LP |
1074 | for (t = p; *t; t++) { |
1075 | if (*t > 0 && *t < ' ') /* no control characters */ | |
1076 | return false; | |
1077 | ||
1078 | if (strchr(QUOTES "\\\x7f", *t)) | |
1079 | return false; | |
1080 | } | |
1081 | ||
1082 | return true; | |
1083 | } | |
53caaffd LP |
1084 | |
1085 | char* string_erase(char *x) { | |
1086 | if (!x) | |
1087 | return NULL; | |
1088 | ||
1089 | /* A delicious drop of snake-oil! To be called on memory where we stored passphrases or so, after we | |
1090 | * used them. */ | |
1091 | explicit_bzero_safe(x, strlen(x)); | |
1092 | return x; | |
1093 | } | |
8dd6491e LP |
1094 | |
1095 | int string_truncate_lines(const char *s, size_t n_lines, char **ret) { | |
1096 | const char *p = s, *e = s; | |
1097 | bool truncation_applied = false; | |
1098 | char *copy; | |
1099 | size_t n = 0; | |
1100 | ||
1101 | assert(s); | |
1102 | ||
1103 | /* Truncate after the specified number of lines. Returns > 0 if a truncation was applied or == 0 if | |
1104 | * there were fewer lines in the string anyway. Trailing newlines on input are ignored, and not | |
1105 | * generated either. */ | |
1106 | ||
1107 | for (;;) { | |
1108 | size_t k; | |
1109 | ||
1110 | k = strcspn(p, "\n"); | |
1111 | ||
1112 | if (p[k] == 0) { | |
1113 | if (k == 0) /* final empty line */ | |
1114 | break; | |
1115 | ||
1116 | if (n >= n_lines) /* above threshold */ | |
1117 | break; | |
1118 | ||
1119 | e = p + k; /* last line to include */ | |
1120 | break; | |
1121 | } | |
1122 | ||
1123 | assert(p[k] == '\n'); | |
1124 | ||
1125 | if (n >= n_lines) | |
1126 | break; | |
1127 | ||
1128 | if (k > 0) | |
1129 | e = p + k; | |
1130 | ||
1131 | p += k + 1; | |
1132 | n++; | |
1133 | } | |
1134 | ||
1135 | /* e points after the last character we want to keep */ | |
1136 | if (isempty(e)) | |
1137 | copy = strdup(s); | |
1138 | else { | |
1139 | if (!in_charset(e, "\n")) /* We only consider things truncated if we remove something that | |
1140 | * isn't a new-line or a series of them */ | |
1141 | truncation_applied = true; | |
1142 | ||
1143 | copy = strndup(s, e - s); | |
1144 | } | |
1145 | if (!copy) | |
1146 | return -ENOMEM; | |
1147 | ||
1148 | *ret = copy; | |
1149 | return truncation_applied; | |
1150 | } | |
f6857fa6 LP |
1151 | |
1152 | int string_extract_line(const char *s, size_t i, char **ret) { | |
1153 | const char *p = s; | |
1154 | size_t c = 0; | |
1155 | ||
1156 | /* Extract the i'nth line from the specified string. Returns > 0 if there are more lines after that, | |
1157 | * and == 0 if we are looking at the last line or already beyond the last line. As special | |
1158 | * optimization, if the first line is requested and the string only consists of one line we return | |
1159 | * NULL, indicating the input string should be used as is, and avoid a memory allocation for a very | |
1160 | * common case. */ | |
1161 | ||
1162 | for (;;) { | |
1163 | const char *q; | |
1164 | ||
1165 | q = strchr(p, '\n'); | |
1166 | if (i == c) { | |
1167 | /* The line we are looking for! */ | |
1168 | ||
1169 | if (q) { | |
1170 | char *m; | |
1171 | ||
1172 | m = strndup(p, q - p); | |
1173 | if (!m) | |
1174 | return -ENOMEM; | |
1175 | ||
1176 | *ret = m; | |
1177 | return !isempty(q + 1); /* more coming? */ | |
1178 | } else { | |
1179 | if (p == s) | |
1180 | *ret = NULL; /* Just use the input string */ | |
1181 | else { | |
1182 | char *m; | |
1183 | ||
1184 | m = strdup(p); | |
1185 | if (!m) | |
1186 | return -ENOMEM; | |
1187 | ||
1188 | *ret = m; | |
1189 | } | |
1190 | ||
1191 | return 0; /* The end */ | |
1192 | } | |
1193 | } | |
1194 | ||
1195 | if (!q) { | |
1196 | char *m; | |
1197 | ||
1198 | /* No more lines, return empty line */ | |
1199 | ||
1200 | m = strdup(""); | |
1201 | if (!m) | |
1202 | return -ENOMEM; | |
1203 | ||
1204 | *ret = m; | |
1205 | return 0; /* The end */ | |
1206 | } | |
1207 | ||
1208 | p = q + 1; | |
1209 | c++; | |
1210 | } | |
1211 | } | |
53cd7f33 | 1212 | |
46bf625a | 1213 | int string_contains_word_strv(const char *string, const char *separators, char **words, const char **ret_word) { |
53cd7f33 ZJS |
1214 | /* In the default mode with no separators specified, we split on whitespace and |
1215 | * don't coalesce separators. */ | |
1216 | const ExtractFlags flags = separators ? EXTRACT_DONT_COALESCE_SEPARATORS : 0; | |
1217 | ||
46bf625a ZJS |
1218 | const char *found = NULL; |
1219 | ||
53cd7f33 ZJS |
1220 | for (const char *p = string;;) { |
1221 | _cleanup_free_ char *w = NULL; | |
1222 | int r; | |
1223 | ||
1224 | r = extract_first_word(&p, &w, separators, flags); | |
1225 | if (r < 0) | |
1226 | return r; | |
1227 | if (r == 0) | |
46bf625a ZJS |
1228 | break; |
1229 | ||
1230 | found = strv_find(words, w); | |
1231 | if (found) | |
1232 | break; | |
53cd7f33 | 1233 | } |
46bf625a ZJS |
1234 | |
1235 | if (ret_word) | |
1236 | *ret_word = found; | |
1237 | return !!found; | |
53cd7f33 | 1238 | } |