]>
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 | ||
07630cea LP |
115 | char *strnappend(const char *s, const char *suffix, size_t b) { |
116 | size_t a; | |
117 | char *r; | |
118 | ||
119 | if (!s && !suffix) | |
120 | return strdup(""); | |
121 | ||
122 | if (!s) | |
123 | return strndup(suffix, b); | |
124 | ||
125 | if (!suffix) | |
126 | return strdup(s); | |
127 | ||
128 | assert(s); | |
129 | assert(suffix); | |
130 | ||
131 | a = strlen(s); | |
132 | if (b > ((size_t) -1) - a) | |
133 | return NULL; | |
134 | ||
135 | r = new(char, a+b+1); | |
136 | if (!r) | |
137 | return NULL; | |
138 | ||
139 | memcpy(r, s, a); | |
140 | memcpy(r+a, suffix, b); | |
141 | r[a+b] = 0; | |
142 | ||
143 | return r; | |
144 | } | |
145 | ||
605405c6 | 146 | char *strjoin_real(const char *x, ...) { |
07630cea | 147 | va_list ap; |
020003f2 | 148 | size_t l = 0; |
07630cea LP |
149 | char *r, *p; |
150 | ||
151 | va_start(ap, x); | |
020003f2 RV |
152 | for (const char *t = x; t; t = va_arg(ap, const char *)) { |
153 | size_t n; | |
07630cea | 154 | |
020003f2 RV |
155 | n = strlen(t); |
156 | if (n > SIZE_MAX - l) { | |
157 | va_end(ap); | |
158 | return NULL; | |
07630cea | 159 | } |
020003f2 RV |
160 | l += n; |
161 | } | |
07630cea LP |
162 | va_end(ap); |
163 | ||
020003f2 | 164 | p = r = new(char, l+1); |
07630cea LP |
165 | if (!r) |
166 | return NULL; | |
167 | ||
020003f2 RV |
168 | va_start(ap, x); |
169 | for (const char *t = x; t; t = va_arg(ap, const char *)) | |
170 | p = stpcpy(p, t); | |
171 | va_end(ap); | |
07630cea | 172 | |
020003f2 | 173 | *p = 0; |
07630cea LP |
174 | |
175 | return r; | |
176 | } | |
177 | ||
178 | char *strstrip(char *s) { | |
7546145e LP |
179 | if (!s) |
180 | return NULL; | |
181 | ||
0a6ffc5c | 182 | /* Drops trailing whitespace. Modifies the string in place. Returns pointer to first non-space character */ |
07630cea | 183 | |
0a6ffc5c | 184 | return delete_trailing_chars(skip_leading_chars(s, WHITESPACE), WHITESPACE); |
07630cea LP |
185 | } |
186 | ||
187 | char *delete_chars(char *s, const char *bad) { | |
188 | char *f, *t; | |
189 | ||
7546145e LP |
190 | /* Drops all specified bad characters, regardless where in the string */ |
191 | ||
192 | if (!s) | |
193 | return NULL; | |
194 | ||
195 | if (!bad) | |
196 | bad = WHITESPACE; | |
07630cea LP |
197 | |
198 | for (f = s, t = s; *f; f++) { | |
199 | if (strchr(bad, *f)) | |
200 | continue; | |
201 | ||
202 | *(t++) = *f; | |
203 | } | |
204 | ||
205 | *t = 0; | |
206 | ||
207 | return s; | |
208 | } | |
209 | ||
7546145e LP |
210 | char *delete_trailing_chars(char *s, const char *bad) { |
211 | char *p, *c = s; | |
212 | ||
213 | /* Drops all specified bad characters, at the end of the string */ | |
214 | ||
215 | if (!s) | |
216 | return NULL; | |
217 | ||
218 | if (!bad) | |
219 | bad = WHITESPACE; | |
220 | ||
221 | for (p = s; *p; p++) | |
222 | if (!strchr(bad, *p)) | |
223 | c = p + 1; | |
224 | ||
225 | *c = 0; | |
226 | ||
227 | return s; | |
228 | } | |
229 | ||
07630cea LP |
230 | char *truncate_nl(char *s) { |
231 | assert(s); | |
232 | ||
233 | s[strcspn(s, NEWLINE)] = 0; | |
234 | return s; | |
235 | } | |
236 | ||
b577e3d5 LP |
237 | char ascii_tolower(char x) { |
238 | ||
239 | if (x >= 'A' && x <= 'Z') | |
240 | return x - 'A' + 'a'; | |
241 | ||
242 | return x; | |
243 | } | |
244 | ||
846b8fc3 LP |
245 | char ascii_toupper(char x) { |
246 | ||
247 | if (x >= 'a' && x <= 'z') | |
248 | return x - 'a' + 'A'; | |
249 | ||
250 | return x; | |
251 | } | |
252 | ||
07630cea LP |
253 | char *ascii_strlower(char *t) { |
254 | char *p; | |
255 | ||
256 | assert(t); | |
257 | ||
258 | for (p = t; *p; p++) | |
b577e3d5 LP |
259 | *p = ascii_tolower(*p); |
260 | ||
261 | return t; | |
262 | } | |
263 | ||
846b8fc3 LP |
264 | char *ascii_strupper(char *t) { |
265 | char *p; | |
266 | ||
267 | assert(t); | |
268 | ||
269 | for (p = t; *p; p++) | |
270 | *p = ascii_toupper(*p); | |
271 | ||
272 | return t; | |
273 | } | |
274 | ||
b577e3d5 LP |
275 | char *ascii_strlower_n(char *t, size_t n) { |
276 | size_t i; | |
277 | ||
278 | if (n <= 0) | |
279 | return t; | |
280 | ||
281 | for (i = 0; i < n; i++) | |
282 | t[i] = ascii_tolower(t[i]); | |
07630cea LP |
283 | |
284 | return t; | |
285 | } | |
522d85ae LP |
286 | |
287 | int ascii_strcasecmp_n(const char *a, const char *b, size_t n) { | |
288 | ||
289 | for (; n > 0; a++, b++, n--) { | |
290 | int x, y; | |
291 | ||
292 | x = (int) (uint8_t) ascii_tolower(*a); | |
293 | y = (int) (uint8_t) ascii_tolower(*b); | |
294 | ||
295 | if (x != y) | |
296 | return x - y; | |
297 | } | |
298 | ||
299 | return 0; | |
300 | } | |
c1749834 LP |
301 | |
302 | int ascii_strcasecmp_nn(const char *a, size_t n, const char *b, size_t m) { | |
303 | int r; | |
304 | ||
305 | r = ascii_strcasecmp_n(a, b, MIN(n, m)); | |
306 | if (r != 0) | |
307 | return r; | |
308 | ||
6dd91b36 | 309 | return CMP(n, m); |
c1749834 | 310 | } |
07630cea LP |
311 | |
312 | bool chars_intersect(const char *a, const char *b) { | |
313 | const char *p; | |
314 | ||
315 | /* Returns true if any of the chars in a are in b. */ | |
316 | for (p = a; *p; p++) | |
317 | if (strchr(b, *p)) | |
318 | return true; | |
319 | ||
320 | return false; | |
321 | } | |
322 | ||
323 | bool string_has_cc(const char *p, const char *ok) { | |
324 | const char *t; | |
325 | ||
326 | assert(p); | |
327 | ||
328 | /* | |
329 | * Check if a string contains control characters. If 'ok' is | |
330 | * non-NULL it may be a string containing additional CCs to be | |
331 | * considered OK. | |
332 | */ | |
333 | ||
334 | for (t = p; *t; t++) { | |
335 | if (ok && strchr(ok, *t)) | |
336 | continue; | |
337 | ||
338 | if (*t > 0 && *t < ' ') | |
339 | return true; | |
340 | ||
341 | if (*t == 127) | |
342 | return true; | |
343 | } | |
344 | ||
345 | return false; | |
346 | } | |
347 | ||
8409f688 ZJS |
348 | static int write_ellipsis(char *buf, bool unicode) { |
349 | if (unicode || is_locale_utf8()) { | |
350 | buf[0] = 0xe2; /* tri-dot ellipsis: … */ | |
351 | buf[1] = 0x80; | |
352 | buf[2] = 0xa6; | |
353 | } else { | |
354 | buf[0] = '.'; | |
355 | buf[1] = '.'; | |
356 | buf[2] = '.'; | |
357 | } | |
358 | ||
359 | return 3; | |
360 | } | |
361 | ||
07630cea | 362 | static char *ascii_ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { |
9924aef6 ZJS |
363 | size_t x, need_space, suffix_len; |
364 | char *t; | |
07630cea LP |
365 | |
366 | assert(s); | |
367 | assert(percent <= 100); | |
c30a49b2 | 368 | assert(new_length != (size_t) -1); |
07630cea | 369 | |
c30a49b2 | 370 | if (old_length <= new_length) |
07630cea LP |
371 | return strndup(s, old_length); |
372 | ||
c30a49b2 LP |
373 | /* Special case short ellipsations */ |
374 | switch (new_length) { | |
375 | ||
376 | case 0: | |
377 | return strdup(""); | |
378 | ||
379 | case 1: | |
380 | if (is_locale_utf8()) | |
381 | return strdup("…"); | |
382 | else | |
383 | return strdup("."); | |
384 | ||
385 | case 2: | |
386 | if (!is_locale_utf8()) | |
387 | return strdup(".."); | |
388 | ||
389 | break; | |
390 | ||
391 | default: | |
392 | break; | |
393 | } | |
394 | ||
395 | /* Calculate how much space the ellipsis will take up. If we are in UTF-8 mode we only need space for one | |
396 | * character ("…"), otherwise for three characters ("..."). Note that in both cases we need 3 bytes of storage, | |
397 | * either for the UTF-8 encoded character or for three ASCII characters. */ | |
398 | need_space = is_locale_utf8() ? 1 : 3; | |
399 | ||
9924aef6 ZJS |
400 | t = new(char, new_length+3); |
401 | if (!t) | |
07630cea LP |
402 | return NULL; |
403 | ||
c30a49b2 | 404 | assert(new_length >= need_space); |
07630cea | 405 | |
c30a49b2 LP |
406 | x = ((new_length - need_space) * percent + 50) / 100; |
407 | assert(x <= new_length - need_space); | |
07630cea | 408 | |
9924aef6 ZJS |
409 | memcpy(t, s, x); |
410 | write_ellipsis(t + x, false); | |
411 | suffix_len = new_length - x - need_space; | |
412 | memcpy(t + x + 3, s + old_length - suffix_len, suffix_len); | |
413 | *(t + x + 3 + suffix_len) = '\0'; | |
07630cea | 414 | |
9924aef6 | 415 | return t; |
07630cea LP |
416 | } |
417 | ||
418 | char *ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 419 | size_t x, k, len, len2; |
07630cea | 420 | const char *i, *j; |
c30a49b2 | 421 | char *e; |
c932fb71 | 422 | int r; |
07630cea | 423 | |
c30a49b2 LP |
424 | /* Note that 'old_length' refers to bytes in the string, while 'new_length' refers to character cells taken up |
425 | * on screen. This distinction doesn't matter for ASCII strings, but it does matter for non-ASCII UTF-8 | |
426 | * strings. | |
427 | * | |
428 | * Ellipsation is done in a locale-dependent way: | |
429 | * 1. If the string passed in is fully ASCII and the current locale is not UTF-8, three dots are used ("...") | |
430 | * 2. Otherwise, a unicode ellipsis is used ("…") | |
431 | * | |
432 | * In other words: you'll get a unicode ellipsis as soon as either the string contains non-ASCII characters or | |
433 | * the current locale is UTF-8. | |
434 | */ | |
435 | ||
07630cea LP |
436 | assert(s); |
437 | assert(percent <= 100); | |
ddbc9319 LP |
438 | |
439 | if (new_length == (size_t) -1) | |
440 | return strndup(s, old_length); | |
441 | ||
c30a49b2 LP |
442 | if (new_length == 0) |
443 | return strdup(""); | |
07630cea | 444 | |
c30a49b2 | 445 | /* If no multibyte characters use ascii_ellipsize_mem for speed */ |
21e4e3e0 | 446 | if (ascii_is_valid_n(s, old_length)) |
07630cea LP |
447 | return ascii_ellipsize_mem(s, old_length, new_length, percent); |
448 | ||
c30a49b2 LP |
449 | x = ((new_length - 1) * percent) / 100; |
450 | assert(x <= new_length - 1); | |
07630cea LP |
451 | |
452 | k = 0; | |
9924aef6 | 453 | for (i = s; i < s + old_length; i = utf8_next_char(i)) { |
c932fb71 | 454 | char32_t c; |
9924aef6 | 455 | int w; |
07630cea | 456 | |
c932fb71 SL |
457 | r = utf8_encoded_to_unichar(i, &c); |
458 | if (r < 0) | |
07630cea | 459 | return NULL; |
07630cea | 460 | |
9924aef6 ZJS |
461 | w = unichar_iswide(c) ? 2 : 1; |
462 | if (k + w <= x) | |
463 | k += w; | |
464 | else | |
465 | break; | |
466 | } | |
07630cea | 467 | |
9924aef6 | 468 | for (j = s + old_length; j > i; ) { |
c932fb71 | 469 | char32_t c; |
9924aef6 ZJS |
470 | int w; |
471 | const char *jj; | |
07630cea | 472 | |
9924aef6 ZJS |
473 | jj = utf8_prev_char(j); |
474 | r = utf8_encoded_to_unichar(jj, &c); | |
c932fb71 | 475 | if (r < 0) |
07630cea | 476 | return NULL; |
9924aef6 ZJS |
477 | |
478 | w = unichar_iswide(c) ? 2 : 1; | |
479 | if (k + w <= new_length) { | |
480 | k += w; | |
481 | j = jj; | |
482 | } else | |
483 | break; | |
07630cea LP |
484 | } |
485 | assert(i <= j); | |
486 | ||
487 | /* we don't actually need to ellipsize */ | |
488 | if (i == j) | |
9924aef6 | 489 | return memdup_suffix0(s, old_length); |
07630cea | 490 | |
9924aef6 ZJS |
491 | /* make space for ellipsis, if possible */ |
492 | if (j < s + old_length) | |
493 | j = utf8_next_char(j); | |
494 | else if (i > s) | |
495 | i = utf8_prev_char(i); | |
07630cea LP |
496 | |
497 | len = i - s; | |
498 | len2 = s + old_length - j; | |
499 | e = new(char, len + 3 + len2 + 1); | |
500 | if (!e) | |
501 | return NULL; | |
502 | ||
503 | /* | |
504 | printf("old_length=%zu new_length=%zu x=%zu len=%u len2=%u k=%u\n", | |
505 | old_length, new_length, x, len, len2, k); | |
506 | */ | |
507 | ||
508 | memcpy(e, s, len); | |
8409f688 | 509 | write_ellipsis(e + len, true); |
9924aef6 ZJS |
510 | memcpy(e + len + 3, j, len2); |
511 | *(e + len + 3 + len2) = '\0'; | |
07630cea LP |
512 | |
513 | return e; | |
514 | } | |
515 | ||
8409f688 ZJS |
516 | char *cellescape(char *buf, size_t len, const char *s) { |
517 | /* Escape and ellipsize s into buffer buf of size len. Only non-control ASCII | |
518 | * characters are copied as they are, everything else is escaped. The result | |
519 | * is different then if escaping and ellipsization was performed in two | |
520 | * separate steps, because each sequence is either stored in full or skipped. | |
521 | * | |
522 | * This function should be used for logging about strings which expected to | |
523 | * be plain ASCII in a safe way. | |
524 | * | |
525 | * An ellipsis will be used if s is too long. It was always placed at the | |
526 | * very end. | |
527 | */ | |
528 | ||
61f6e276 LP |
529 | size_t i = 0, last_char_width[4] = {}, k = 0, j; |
530 | ||
531 | assert(len > 0); /* at least a terminating NUL */ | |
8409f688 | 532 | |
61f6e276 LP |
533 | for (;;) { |
534 | char four[4]; | |
535 | int w; | |
8409f688 | 536 | |
61f6e276 | 537 | if (*s == 0) /* terminating NUL detected? then we are done! */ |
8409f688 | 538 | goto done; |
61f6e276 LP |
539 | |
540 | w = cescape_char(*s, four); | |
541 | if (i + w + 1 > len) /* This character doesn't fit into the buffer anymore? In that case let's | |
542 | * ellipsize at the previous location */ | |
543 | break; | |
544 | ||
545 | /* OK, there was space, let's add this escaped character to the buffer */ | |
546 | memcpy(buf + i, four, w); | |
547 | i += w; | |
548 | ||
549 | /* And remember its width in the ring buffer */ | |
550 | last_char_width[k] = w; | |
551 | k = (k + 1) % 4; | |
552 | ||
553 | s++; | |
8409f688 ZJS |
554 | } |
555 | ||
61f6e276 LP |
556 | /* Ellipsation is necessary. This means we might need to truncate the string again to make space for 4 |
557 | * characters ideally, but the buffer is shorter than that in the first place take what we can get */ | |
558 | for (j = 0; j < ELEMENTSOF(last_char_width); j++) { | |
559 | ||
560 | if (i + 4 <= len) /* nice, we reached our space goal */ | |
561 | break; | |
562 | ||
563 | k = k == 0 ? 3 : k - 1; | |
564 | if (last_char_width[k] == 0) /* bummer, we reached the beginning of the strings */ | |
565 | break; | |
566 | ||
567 | assert(i >= last_char_width[k]); | |
568 | i -= last_char_width[k]; | |
8409f688 ZJS |
569 | } |
570 | ||
61f6e276 LP |
571 | if (i + 4 <= len) /* yay, enough space */ |
572 | i += write_ellipsis(buf + i, false); | |
573 | else if (i + 3 <= len) { /* only space for ".." */ | |
574 | buf[i++] = '.'; | |
575 | buf[i++] = '.'; | |
576 | } else if (i + 2 <= len) /* only space for a single "." */ | |
577 | buf[i++] = '.'; | |
578 | else | |
579 | assert(i + 1 <= len); | |
580 | ||
8409f688 ZJS |
581 | done: |
582 | buf[i] = '\0'; | |
583 | return buf; | |
584 | } | |
585 | ||
07630cea LP |
586 | char* strshorten(char *s, size_t l) { |
587 | assert(s); | |
588 | ||
47b33c7d | 589 | if (strnlen(s, l+1) > l) |
07630cea LP |
590 | s[l] = 0; |
591 | ||
592 | return s; | |
593 | } | |
594 | ||
595 | char *strreplace(const char *text, const char *old_string, const char *new_string) { | |
9d73565a LP |
596 | size_t l, old_len, new_len, allocated = 0; |
597 | char *t, *ret = NULL; | |
07630cea | 598 | const char *f; |
07630cea | 599 | |
07630cea LP |
600 | assert(old_string); |
601 | assert(new_string); | |
602 | ||
9d73565a LP |
603 | if (!text) |
604 | return NULL; | |
605 | ||
07630cea LP |
606 | old_len = strlen(old_string); |
607 | new_len = strlen(new_string); | |
608 | ||
609 | l = strlen(text); | |
9d73565a | 610 | if (!GREEDY_REALLOC(ret, allocated, l+1)) |
07630cea LP |
611 | return NULL; |
612 | ||
613 | f = text; | |
9d73565a | 614 | t = ret; |
07630cea | 615 | while (*f) { |
07630cea LP |
616 | size_t d, nl; |
617 | ||
618 | if (!startswith(f, old_string)) { | |
619 | *(t++) = *(f++); | |
620 | continue; | |
621 | } | |
622 | ||
9d73565a | 623 | d = t - ret; |
07630cea | 624 | nl = l - old_len + new_len; |
9d73565a LP |
625 | |
626 | if (!GREEDY_REALLOC(ret, allocated, nl + 1)) | |
627 | return mfree(ret); | |
07630cea LP |
628 | |
629 | l = nl; | |
9d73565a | 630 | t = ret + d; |
07630cea LP |
631 | |
632 | t = stpcpy(t, new_string); | |
633 | f += old_len; | |
634 | } | |
635 | ||
636 | *t = 0; | |
9d73565a | 637 | return ret; |
07630cea LP |
638 | } |
639 | ||
6fb05690 LP |
640 | static void advance_offsets( |
641 | ssize_t diff, | |
642 | size_t offsets[2], /* note: we can't use [static 2] here, since this may be NULL */ | |
643 | size_t shift[static 2], | |
644 | size_t size) { | |
645 | ||
b4766d5f ZJS |
646 | if (!offsets) |
647 | return; | |
648 | ||
6fb05690 LP |
649 | assert(shift); |
650 | ||
b4766d5f ZJS |
651 | if ((size_t) diff < offsets[0]) |
652 | shift[0] += size; | |
653 | if ((size_t) diff < offsets[1]) | |
654 | shift[1] += size; | |
655 | } | |
656 | ||
657 | char *strip_tab_ansi(char **ibuf, size_t *_isz, size_t highlight[2]) { | |
62a3fc6d | 658 | const char *begin = NULL; |
07630cea LP |
659 | enum { |
660 | STATE_OTHER, | |
661 | STATE_ESCAPE, | |
695a944c LP |
662 | STATE_CSI, |
663 | STATE_CSO, | |
07630cea LP |
664 | } state = STATE_OTHER; |
665 | char *obuf = NULL; | |
62a3fc6d | 666 | size_t osz = 0, isz, shift[2] = {}, n_carriage_returns = 0; |
07630cea LP |
667 | FILE *f; |
668 | ||
669 | assert(ibuf); | |
670 | assert(*ibuf); | |
671 | ||
695a944c LP |
672 | /* This does three things: |
673 | * | |
674 | * 1. Replaces TABs by 8 spaces | |
675 | * 2. Strips ANSI color sequences (a subset of CSI), i.e. ESC '[' … 'm' sequences | |
676 | * 3. Strips ANSI operating system sequences (CSO), i.e. ESC ']' … BEL sequences | |
62a3fc6d ZJS |
677 | * 4. Strip trailing \r characters (since they would "move the cursor", but have no |
678 | * other effect). | |
695a944c | 679 | * |
2fe21124 ZJS |
680 | * Everything else will be left as it is. In particular other ANSI sequences are left as they are, as |
681 | * are any other special characters. Truncated ANSI sequences are left-as is too. This call is | |
682 | * supposed to suppress the most basic formatting noise, but nothing else. | |
695a944c LP |
683 | * |
684 | * Why care for CSO sequences? Well, to undo what terminal_urlify() and friends generate. */ | |
07630cea LP |
685 | |
686 | isz = _isz ? *_isz : strlen(*ibuf); | |
687 | ||
2fe21124 ZJS |
688 | /* Note we turn off internal locking on f for performance reasons. It's safe to do so since we |
689 | * created f here and it doesn't leave our scope. */ | |
690 | f = open_memstream_unlocked(&obuf, &osz); | |
07630cea LP |
691 | if (!f) |
692 | return NULL; | |
693 | ||
62a3fc6d | 694 | for (const char *i = *ibuf; i < *ibuf + isz + 1; i++) { |
07630cea LP |
695 | |
696 | switch (state) { | |
697 | ||
698 | case STATE_OTHER: | |
699 | if (i >= *ibuf + isz) /* EOT */ | |
700 | break; | |
62a3fc6d ZJS |
701 | |
702 | if (*i == '\r') { | |
703 | n_carriage_returns++; | |
704 | break; | |
705 | } else if (*i == '\n') | |
706 | /* Ignore carriage returns before new line */ | |
707 | n_carriage_returns = 0; | |
708 | for (; n_carriage_returns > 0; n_carriage_returns--) | |
709 | fputc('\r', f); | |
710 | ||
711 | if (*i == '\x1B') | |
07630cea | 712 | state = STATE_ESCAPE; |
b4766d5f | 713 | else if (*i == '\t') { |
0d536673 | 714 | fputs(" ", f); |
b4766d5f ZJS |
715 | advance_offsets(i - *ibuf, highlight, shift, 7); |
716 | } else | |
0d536673 | 717 | fputc(*i, f); |
b4766d5f | 718 | |
07630cea LP |
719 | break; |
720 | ||
721 | case STATE_ESCAPE: | |
62a3fc6d ZJS |
722 | assert(n_carriage_returns == 0); |
723 | ||
07630cea | 724 | if (i >= *ibuf + isz) { /* EOT */ |
0d536673 | 725 | fputc('\x1B', f); |
b4766d5f | 726 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea | 727 | break; |
695a944c LP |
728 | } else if (*i == '[') { /* ANSI CSI */ |
729 | state = STATE_CSI; | |
730 | begin = i + 1; | |
731 | } else if (*i == ']') { /* ANSI CSO */ | |
732 | state = STATE_CSO; | |
07630cea LP |
733 | begin = i + 1; |
734 | } else { | |
0d536673 LP |
735 | fputc('\x1B', f); |
736 | fputc(*i, f); | |
b4766d5f | 737 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea LP |
738 | state = STATE_OTHER; |
739 | } | |
740 | ||
741 | break; | |
742 | ||
695a944c | 743 | case STATE_CSI: |
62a3fc6d | 744 | assert(n_carriage_returns == 0); |
07630cea | 745 | |
695a944c LP |
746 | if (i >= *ibuf + isz || /* EOT … */ |
747 | !strchr("01234567890;m", *i)) { /* … or invalid chars in sequence */ | |
0d536673 LP |
748 | fputc('\x1B', f); |
749 | fputc('[', f); | |
b4766d5f | 750 | advance_offsets(i - *ibuf, highlight, shift, 2); |
07630cea LP |
751 | state = STATE_OTHER; |
752 | i = begin-1; | |
753 | } else if (*i == 'm') | |
754 | state = STATE_OTHER; | |
695a944c LP |
755 | |
756 | break; | |
757 | ||
758 | case STATE_CSO: | |
62a3fc6d | 759 | assert(n_carriage_returns == 0); |
695a944c LP |
760 | |
761 | if (i >= *ibuf + isz || /* EOT … */ | |
762 | (*i != '\a' && (uint8_t) *i < 32U) || (uint8_t) *i > 126U) { /* … or invalid chars in sequence */ | |
763 | fputc('\x1B', f); | |
764 | fputc(']', f); | |
765 | advance_offsets(i - *ibuf, highlight, shift, 2); | |
766 | state = STATE_OTHER; | |
767 | i = begin-1; | |
768 | } else if (*i == '\a') | |
769 | state = STATE_OTHER; | |
770 | ||
07630cea LP |
771 | break; |
772 | } | |
773 | } | |
774 | ||
c7e03d2e | 775 | if (fflush_and_check(f) < 0) { |
07630cea | 776 | fclose(f); |
6b430fdb | 777 | return mfree(obuf); |
07630cea | 778 | } |
07630cea LP |
779 | fclose(f); |
780 | ||
6fb05690 | 781 | free_and_replace(*ibuf, obuf); |
07630cea LP |
782 | |
783 | if (_isz) | |
784 | *_isz = osz; | |
785 | ||
b4766d5f ZJS |
786 | if (highlight) { |
787 | highlight[0] += shift[0]; | |
788 | highlight[1] += shift[1]; | |
789 | } | |
790 | ||
6fb05690 | 791 | return *ibuf; |
07630cea LP |
792 | } |
793 | ||
bb8ad9ea LP |
794 | char *strextend_with_separator(char **x, const char *separator, ...) { |
795 | bool need_separator; | |
796 | size_t f, l, l_separator; | |
07630cea | 797 | char *r, *p; |
bb8ad9ea | 798 | va_list ap; |
07630cea LP |
799 | |
800 | assert(x); | |
801 | ||
7bf7ce28 | 802 | l = f = strlen_ptr(*x); |
07630cea | 803 | |
bb8ad9ea LP |
804 | need_separator = !isempty(*x); |
805 | l_separator = strlen_ptr(separator); | |
806 | ||
807 | va_start(ap, separator); | |
07630cea LP |
808 | for (;;) { |
809 | const char *t; | |
810 | size_t n; | |
811 | ||
812 | t = va_arg(ap, const char *); | |
813 | if (!t) | |
814 | break; | |
815 | ||
816 | n = strlen(t); | |
bb8ad9ea LP |
817 | |
818 | if (need_separator) | |
819 | n += l_separator; | |
820 | ||
07630cea LP |
821 | if (n > ((size_t) -1) - l) { |
822 | va_end(ap); | |
823 | return NULL; | |
824 | } | |
825 | ||
826 | l += n; | |
bb8ad9ea | 827 | need_separator = true; |
07630cea LP |
828 | } |
829 | va_end(ap); | |
830 | ||
bb8ad9ea LP |
831 | need_separator = !isempty(*x); |
832 | ||
07630cea LP |
833 | r = realloc(*x, l+1); |
834 | if (!r) | |
835 | return NULL; | |
836 | ||
837 | p = r + f; | |
838 | ||
bb8ad9ea | 839 | va_start(ap, separator); |
07630cea LP |
840 | for (;;) { |
841 | const char *t; | |
842 | ||
843 | t = va_arg(ap, const char *); | |
844 | if (!t) | |
845 | break; | |
846 | ||
bb8ad9ea LP |
847 | if (need_separator && separator) |
848 | p = stpcpy(p, separator); | |
849 | ||
07630cea | 850 | p = stpcpy(p, t); |
bb8ad9ea LP |
851 | |
852 | need_separator = true; | |
07630cea LP |
853 | } |
854 | va_end(ap); | |
855 | ||
bb8ad9ea LP |
856 | assert(p == r + l); |
857 | ||
07630cea LP |
858 | *p = 0; |
859 | *x = r; | |
860 | ||
861 | return r + l; | |
862 | } | |
863 | ||
864 | char *strrep(const char *s, unsigned n) { | |
865 | size_t l; | |
866 | char *r, *p; | |
867 | unsigned i; | |
868 | ||
869 | assert(s); | |
870 | ||
871 | l = strlen(s); | |
872 | p = r = malloc(l * n + 1); | |
873 | if (!r) | |
874 | return NULL; | |
875 | ||
876 | for (i = 0; i < n; i++) | |
877 | p = stpcpy(p, s); | |
878 | ||
879 | *p = 0; | |
880 | return r; | |
881 | } | |
882 | ||
883 | int split_pair(const char *s, const char *sep, char **l, char **r) { | |
884 | char *x, *a, *b; | |
885 | ||
886 | assert(s); | |
887 | assert(sep); | |
888 | assert(l); | |
889 | assert(r); | |
890 | ||
891 | if (isempty(sep)) | |
892 | return -EINVAL; | |
893 | ||
894 | x = strstr(s, sep); | |
895 | if (!x) | |
896 | return -EINVAL; | |
897 | ||
898 | a = strndup(s, x - s); | |
899 | if (!a) | |
900 | return -ENOMEM; | |
901 | ||
902 | b = strdup(x + strlen(sep)); | |
903 | if (!b) { | |
904 | free(a); | |
905 | return -ENOMEM; | |
906 | } | |
907 | ||
908 | *l = a; | |
909 | *r = b; | |
910 | ||
911 | return 0; | |
912 | } | |
913 | ||
914 | int free_and_strdup(char **p, const char *s) { | |
915 | char *t; | |
916 | ||
917 | assert(p); | |
918 | ||
7f546026 | 919 | /* Replaces a string pointer with a strdup()ed new string, |
07630cea LP |
920 | * possibly freeing the old one. */ |
921 | ||
922 | if (streq_ptr(*p, s)) | |
923 | return 0; | |
924 | ||
925 | if (s) { | |
926 | t = strdup(s); | |
927 | if (!t) | |
928 | return -ENOMEM; | |
929 | } else | |
930 | t = NULL; | |
931 | ||
932 | free(*p); | |
933 | *p = t; | |
934 | ||
935 | return 1; | |
936 | } | |
937 | ||
7f546026 ZJS |
938 | int free_and_strndup(char **p, const char *s, size_t l) { |
939 | char *t; | |
940 | ||
941 | assert(p); | |
942 | assert(s || l == 0); | |
943 | ||
944 | /* Replaces a string pointer with a strndup()ed new string, | |
945 | * freeing the old one. */ | |
946 | ||
947 | if (!*p && !s) | |
948 | return 0; | |
949 | ||
950 | if (*p && s && strneq(*p, s, l) && (l > strlen(*p) || (*p)[l] == '\0')) | |
951 | return 0; | |
952 | ||
953 | if (s) { | |
954 | t = strndup(s, l); | |
955 | if (!t) | |
956 | return -ENOMEM; | |
957 | } else | |
958 | t = NULL; | |
959 | ||
960 | free_and_replace(*p, t); | |
961 | return 1; | |
962 | } | |
963 | ||
f3e2e81d LP |
964 | bool string_is_safe(const char *p) { |
965 | const char *t; | |
966 | ||
967 | if (!p) | |
968 | return false; | |
969 | ||
839d1b20 LP |
970 | /* Checks if the specified string contains no quotes or control characters */ |
971 | ||
f3e2e81d LP |
972 | for (t = p; *t; t++) { |
973 | if (*t > 0 && *t < ' ') /* no control characters */ | |
974 | return false; | |
975 | ||
976 | if (strchr(QUOTES "\\\x7f", *t)) | |
977 | return false; | |
978 | } | |
979 | ||
980 | return true; | |
981 | } | |
53caaffd LP |
982 | |
983 | char* string_erase(char *x) { | |
984 | if (!x) | |
985 | return NULL; | |
986 | ||
987 | /* A delicious drop of snake-oil! To be called on memory where we stored passphrases or so, after we | |
988 | * used them. */ | |
989 | explicit_bzero_safe(x, strlen(x)); | |
990 | return x; | |
991 | } | |
8dd6491e LP |
992 | |
993 | int string_truncate_lines(const char *s, size_t n_lines, char **ret) { | |
994 | const char *p = s, *e = s; | |
995 | bool truncation_applied = false; | |
996 | char *copy; | |
997 | size_t n = 0; | |
998 | ||
999 | assert(s); | |
1000 | ||
1001 | /* Truncate after the specified number of lines. Returns > 0 if a truncation was applied or == 0 if | |
1002 | * there were fewer lines in the string anyway. Trailing newlines on input are ignored, and not | |
1003 | * generated either. */ | |
1004 | ||
1005 | for (;;) { | |
1006 | size_t k; | |
1007 | ||
1008 | k = strcspn(p, "\n"); | |
1009 | ||
1010 | if (p[k] == 0) { | |
1011 | if (k == 0) /* final empty line */ | |
1012 | break; | |
1013 | ||
1014 | if (n >= n_lines) /* above threshold */ | |
1015 | break; | |
1016 | ||
1017 | e = p + k; /* last line to include */ | |
1018 | break; | |
1019 | } | |
1020 | ||
1021 | assert(p[k] == '\n'); | |
1022 | ||
1023 | if (n >= n_lines) | |
1024 | break; | |
1025 | ||
1026 | if (k > 0) | |
1027 | e = p + k; | |
1028 | ||
1029 | p += k + 1; | |
1030 | n++; | |
1031 | } | |
1032 | ||
1033 | /* e points after the last character we want to keep */ | |
1034 | if (isempty(e)) | |
1035 | copy = strdup(s); | |
1036 | else { | |
1037 | if (!in_charset(e, "\n")) /* We only consider things truncated if we remove something that | |
1038 | * isn't a new-line or a series of them */ | |
1039 | truncation_applied = true; | |
1040 | ||
1041 | copy = strndup(s, e - s); | |
1042 | } | |
1043 | if (!copy) | |
1044 | return -ENOMEM; | |
1045 | ||
1046 | *ret = copy; | |
1047 | return truncation_applied; | |
1048 | } | |
f6857fa6 LP |
1049 | |
1050 | int string_extract_line(const char *s, size_t i, char **ret) { | |
1051 | const char *p = s; | |
1052 | size_t c = 0; | |
1053 | ||
1054 | /* Extract the i'nth line from the specified string. Returns > 0 if there are more lines after that, | |
1055 | * and == 0 if we are looking at the last line or already beyond the last line. As special | |
1056 | * optimization, if the first line is requested and the string only consists of one line we return | |
1057 | * NULL, indicating the input string should be used as is, and avoid a memory allocation for a very | |
1058 | * common case. */ | |
1059 | ||
1060 | for (;;) { | |
1061 | const char *q; | |
1062 | ||
1063 | q = strchr(p, '\n'); | |
1064 | if (i == c) { | |
1065 | /* The line we are looking for! */ | |
1066 | ||
1067 | if (q) { | |
1068 | char *m; | |
1069 | ||
1070 | m = strndup(p, q - p); | |
1071 | if (!m) | |
1072 | return -ENOMEM; | |
1073 | ||
1074 | *ret = m; | |
1075 | return !isempty(q + 1); /* more coming? */ | |
1076 | } else { | |
1077 | if (p == s) | |
1078 | *ret = NULL; /* Just use the input string */ | |
1079 | else { | |
1080 | char *m; | |
1081 | ||
1082 | m = strdup(p); | |
1083 | if (!m) | |
1084 | return -ENOMEM; | |
1085 | ||
1086 | *ret = m; | |
1087 | } | |
1088 | ||
1089 | return 0; /* The end */ | |
1090 | } | |
1091 | } | |
1092 | ||
1093 | if (!q) { | |
1094 | char *m; | |
1095 | ||
1096 | /* No more lines, return empty line */ | |
1097 | ||
1098 | m = strdup(""); | |
1099 | if (!m) | |
1100 | return -ENOMEM; | |
1101 | ||
1102 | *ret = m; | |
1103 | return 0; /* The end */ | |
1104 | } | |
1105 | ||
1106 | p = q + 1; | |
1107 | c++; | |
1108 | } | |
1109 | } | |
53cd7f33 | 1110 | |
46bf625a | 1111 | int string_contains_word_strv(const char *string, const char *separators, char **words, const char **ret_word) { |
53cd7f33 ZJS |
1112 | /* In the default mode with no separators specified, we split on whitespace and |
1113 | * don't coalesce separators. */ | |
1114 | const ExtractFlags flags = separators ? EXTRACT_DONT_COALESCE_SEPARATORS : 0; | |
1115 | ||
46bf625a ZJS |
1116 | const char *found = NULL; |
1117 | ||
53cd7f33 ZJS |
1118 | for (const char *p = string;;) { |
1119 | _cleanup_free_ char *w = NULL; | |
1120 | int r; | |
1121 | ||
1122 | r = extract_first_word(&p, &w, separators, flags); | |
1123 | if (r < 0) | |
1124 | return r; | |
1125 | if (r == 0) | |
46bf625a ZJS |
1126 | break; |
1127 | ||
1128 | found = strv_find(words, w); | |
1129 | if (found) | |
1130 | break; | |
53cd7f33 | 1131 | } |
46bf625a ZJS |
1132 | |
1133 | if (ret_word) | |
1134 | *ret_word = found; | |
1135 | return !!found; | |
53cd7f33 | 1136 | } |