]>
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> | |
0d536673 | 7 | #include <stdio_ext.h> |
11c3a366 | 8 | #include <stdlib.h> |
b6b609db | 9 | #include <string.h> |
11c3a366 | 10 | |
b5efdb8a | 11 | #include "alloc-util.h" |
8409f688 | 12 | #include "escape.h" |
07630cea | 13 | #include "gunicode.h" |
c30a49b2 | 14 | #include "locale-util.h" |
11c3a366 | 15 | #include "macro.h" |
b11d6a7b | 16 | #include "string-util.h" |
b4766d5f | 17 | #include "terminal-util.h" |
07630cea LP |
18 | #include "utf8.h" |
19 | #include "util.h" | |
c7e03d2e | 20 | #include "fileio.h" |
07630cea LP |
21 | |
22 | int strcmp_ptr(const char *a, const char *b) { | |
23 | ||
24 | /* Like strcmp(), but tries to make sense of NULL pointers */ | |
25 | if (a && b) | |
26 | return strcmp(a, b); | |
27 | ||
28 | if (!a && b) | |
29 | return -1; | |
30 | ||
31 | if (a && !b) | |
32 | return 1; | |
33 | ||
34 | return 0; | |
35 | } | |
36 | ||
37 | char* endswith(const char *s, const char *postfix) { | |
38 | size_t sl, pl; | |
39 | ||
40 | assert(s); | |
41 | assert(postfix); | |
42 | ||
43 | sl = strlen(s); | |
44 | pl = strlen(postfix); | |
45 | ||
46 | if (pl == 0) | |
47 | return (char*) s + sl; | |
48 | ||
49 | if (sl < pl) | |
50 | return NULL; | |
51 | ||
52 | if (memcmp(s + sl - pl, postfix, pl) != 0) | |
53 | return NULL; | |
54 | ||
55 | return (char*) s + sl - pl; | |
56 | } | |
57 | ||
58 | char* endswith_no_case(const char *s, const char *postfix) { | |
59 | size_t sl, pl; | |
60 | ||
61 | assert(s); | |
62 | assert(postfix); | |
63 | ||
64 | sl = strlen(s); | |
65 | pl = strlen(postfix); | |
66 | ||
67 | if (pl == 0) | |
68 | return (char*) s + sl; | |
69 | ||
70 | if (sl < pl) | |
71 | return NULL; | |
72 | ||
73 | if (strcasecmp(s + sl - pl, postfix) != 0) | |
74 | return NULL; | |
75 | ||
76 | return (char*) s + sl - pl; | |
77 | } | |
78 | ||
79 | char* first_word(const char *s, const char *word) { | |
80 | size_t sl, wl; | |
81 | const char *p; | |
82 | ||
83 | assert(s); | |
84 | assert(word); | |
85 | ||
86 | /* Checks if the string starts with the specified word, either | |
87 | * followed by NUL or by whitespace. Returns a pointer to the | |
88 | * NUL or the first character after the whitespace. */ | |
89 | ||
90 | sl = strlen(s); | |
91 | wl = strlen(word); | |
92 | ||
93 | if (sl < wl) | |
94 | return NULL; | |
95 | ||
96 | if (wl == 0) | |
97 | return (char*) s; | |
98 | ||
99 | if (memcmp(s, word, wl) != 0) | |
100 | return NULL; | |
101 | ||
102 | p = s + wl; | |
103 | if (*p == 0) | |
104 | return (char*) p; | |
105 | ||
106 | if (!strchr(WHITESPACE, *p)) | |
107 | return NULL; | |
108 | ||
109 | p += strspn(p, WHITESPACE); | |
110 | return (char*) p; | |
111 | } | |
112 | ||
113 | static size_t strcspn_escaped(const char *s, const char *reject) { | |
114 | bool escaped = false; | |
115 | int n; | |
116 | ||
117 | for (n=0; s[n]; n++) { | |
118 | if (escaped) | |
119 | escaped = false; | |
120 | else if (s[n] == '\\') | |
121 | escaped = true; | |
122 | else if (strchr(reject, s[n])) | |
123 | break; | |
124 | } | |
125 | ||
126 | /* if s ends in \, return index of previous char */ | |
127 | return n - escaped; | |
128 | } | |
129 | ||
130 | /* Split a string into words. */ | |
8059aa9c | 131 | const char* split(const char **state, size_t *l, const char *separator, SplitFlags flags) { |
07630cea LP |
132 | const char *current; |
133 | ||
134 | current = *state; | |
135 | ||
136 | if (!*current) { | |
137 | assert(**state == '\0'); | |
138 | return NULL; | |
139 | } | |
140 | ||
141 | current += strspn(current, separator); | |
142 | if (!*current) { | |
143 | *state = current; | |
144 | return NULL; | |
145 | } | |
146 | ||
8059aa9c | 147 | if (flags & SPLIT_QUOTES && strchr("\'\"", *current)) { |
07630cea LP |
148 | char quotechars[2] = {*current, '\0'}; |
149 | ||
150 | *l = strcspn_escaped(current + 1, quotechars); | |
151 | if (current[*l + 1] == '\0' || current[*l + 1] != quotechars[0] || | |
152 | (current[*l + 2] && !strchr(separator, current[*l + 2]))) { | |
153 | /* right quote missing or garbage at the end */ | |
8059aa9c YW |
154 | if (flags & SPLIT_RELAX) { |
155 | *state = current + *l + 1 + (current[*l + 1] != '\0'); | |
156 | return current + 1; | |
157 | } | |
07630cea LP |
158 | *state = current; |
159 | return NULL; | |
160 | } | |
161 | *state = current++ + *l + 2; | |
8059aa9c | 162 | } else if (flags & SPLIT_QUOTES) { |
07630cea | 163 | *l = strcspn_escaped(current, separator); |
8059aa9c | 164 | if (current[*l] && !strchr(separator, current[*l]) && !(flags & SPLIT_RELAX)) { |
07630cea LP |
165 | /* unfinished escape */ |
166 | *state = current; | |
167 | return NULL; | |
168 | } | |
169 | *state = current + *l; | |
170 | } else { | |
171 | *l = strcspn(current, separator); | |
172 | *state = current + *l; | |
173 | } | |
174 | ||
175 | return current; | |
176 | } | |
177 | ||
178 | char *strnappend(const char *s, const char *suffix, size_t b) { | |
179 | size_t a; | |
180 | char *r; | |
181 | ||
182 | if (!s && !suffix) | |
183 | return strdup(""); | |
184 | ||
185 | if (!s) | |
186 | return strndup(suffix, b); | |
187 | ||
188 | if (!suffix) | |
189 | return strdup(s); | |
190 | ||
191 | assert(s); | |
192 | assert(suffix); | |
193 | ||
194 | a = strlen(s); | |
195 | if (b > ((size_t) -1) - a) | |
196 | return NULL; | |
197 | ||
198 | r = new(char, a+b+1); | |
199 | if (!r) | |
200 | return NULL; | |
201 | ||
202 | memcpy(r, s, a); | |
203 | memcpy(r+a, suffix, b); | |
204 | r[a+b] = 0; | |
205 | ||
206 | return r; | |
207 | } | |
208 | ||
209 | char *strappend(const char *s, const char *suffix) { | |
7bf7ce28 | 210 | return strnappend(s, suffix, strlen_ptr(suffix)); |
07630cea LP |
211 | } |
212 | ||
605405c6 | 213 | char *strjoin_real(const char *x, ...) { |
07630cea LP |
214 | va_list ap; |
215 | size_t l; | |
216 | char *r, *p; | |
217 | ||
218 | va_start(ap, x); | |
219 | ||
220 | if (x) { | |
221 | l = strlen(x); | |
222 | ||
223 | for (;;) { | |
224 | const char *t; | |
225 | size_t n; | |
226 | ||
227 | t = va_arg(ap, const char *); | |
228 | if (!t) | |
229 | break; | |
230 | ||
231 | n = strlen(t); | |
232 | if (n > ((size_t) -1) - l) { | |
233 | va_end(ap); | |
234 | return NULL; | |
235 | } | |
236 | ||
237 | l += n; | |
238 | } | |
239 | } else | |
240 | l = 0; | |
241 | ||
242 | va_end(ap); | |
243 | ||
244 | r = new(char, l+1); | |
245 | if (!r) | |
246 | return NULL; | |
247 | ||
248 | if (x) { | |
249 | p = stpcpy(r, x); | |
250 | ||
251 | va_start(ap, x); | |
252 | ||
253 | for (;;) { | |
254 | const char *t; | |
255 | ||
256 | t = va_arg(ap, const char *); | |
257 | if (!t) | |
258 | break; | |
259 | ||
260 | p = stpcpy(p, t); | |
261 | } | |
262 | ||
263 | va_end(ap); | |
264 | } else | |
265 | r[0] = 0; | |
266 | ||
267 | return r; | |
268 | } | |
269 | ||
270 | char *strstrip(char *s) { | |
7546145e LP |
271 | if (!s) |
272 | return NULL; | |
273 | ||
0a6ffc5c | 274 | /* Drops trailing whitespace. Modifies the string in place. Returns pointer to first non-space character */ |
07630cea | 275 | |
0a6ffc5c | 276 | return delete_trailing_chars(skip_leading_chars(s, WHITESPACE), WHITESPACE); |
07630cea LP |
277 | } |
278 | ||
279 | char *delete_chars(char *s, const char *bad) { | |
280 | char *f, *t; | |
281 | ||
7546145e LP |
282 | /* Drops all specified bad characters, regardless where in the string */ |
283 | ||
284 | if (!s) | |
285 | return NULL; | |
286 | ||
287 | if (!bad) | |
288 | bad = WHITESPACE; | |
07630cea LP |
289 | |
290 | for (f = s, t = s; *f; f++) { | |
291 | if (strchr(bad, *f)) | |
292 | continue; | |
293 | ||
294 | *(t++) = *f; | |
295 | } | |
296 | ||
297 | *t = 0; | |
298 | ||
299 | return s; | |
300 | } | |
301 | ||
7546145e LP |
302 | char *delete_trailing_chars(char *s, const char *bad) { |
303 | char *p, *c = s; | |
304 | ||
305 | /* Drops all specified bad characters, at the end of the string */ | |
306 | ||
307 | if (!s) | |
308 | return NULL; | |
309 | ||
310 | if (!bad) | |
311 | bad = WHITESPACE; | |
312 | ||
313 | for (p = s; *p; p++) | |
314 | if (!strchr(bad, *p)) | |
315 | c = p + 1; | |
316 | ||
317 | *c = 0; | |
318 | ||
319 | return s; | |
320 | } | |
321 | ||
07630cea LP |
322 | char *truncate_nl(char *s) { |
323 | assert(s); | |
324 | ||
325 | s[strcspn(s, NEWLINE)] = 0; | |
326 | return s; | |
327 | } | |
328 | ||
b577e3d5 LP |
329 | char ascii_tolower(char x) { |
330 | ||
331 | if (x >= 'A' && x <= 'Z') | |
332 | return x - 'A' + 'a'; | |
333 | ||
334 | return x; | |
335 | } | |
336 | ||
846b8fc3 LP |
337 | char ascii_toupper(char x) { |
338 | ||
339 | if (x >= 'a' && x <= 'z') | |
340 | return x - 'a' + 'A'; | |
341 | ||
342 | return x; | |
343 | } | |
344 | ||
07630cea LP |
345 | char *ascii_strlower(char *t) { |
346 | char *p; | |
347 | ||
348 | assert(t); | |
349 | ||
350 | for (p = t; *p; p++) | |
b577e3d5 LP |
351 | *p = ascii_tolower(*p); |
352 | ||
353 | return t; | |
354 | } | |
355 | ||
846b8fc3 LP |
356 | char *ascii_strupper(char *t) { |
357 | char *p; | |
358 | ||
359 | assert(t); | |
360 | ||
361 | for (p = t; *p; p++) | |
362 | *p = ascii_toupper(*p); | |
363 | ||
364 | return t; | |
365 | } | |
366 | ||
b577e3d5 LP |
367 | char *ascii_strlower_n(char *t, size_t n) { |
368 | size_t i; | |
369 | ||
370 | if (n <= 0) | |
371 | return t; | |
372 | ||
373 | for (i = 0; i < n; i++) | |
374 | t[i] = ascii_tolower(t[i]); | |
07630cea LP |
375 | |
376 | return t; | |
377 | } | |
522d85ae LP |
378 | |
379 | int ascii_strcasecmp_n(const char *a, const char *b, size_t n) { | |
380 | ||
381 | for (; n > 0; a++, b++, n--) { | |
382 | int x, y; | |
383 | ||
384 | x = (int) (uint8_t) ascii_tolower(*a); | |
385 | y = (int) (uint8_t) ascii_tolower(*b); | |
386 | ||
387 | if (x != y) | |
388 | return x - y; | |
389 | } | |
390 | ||
391 | return 0; | |
392 | } | |
c1749834 LP |
393 | |
394 | int ascii_strcasecmp_nn(const char *a, size_t n, const char *b, size_t m) { | |
395 | int r; | |
396 | ||
397 | r = ascii_strcasecmp_n(a, b, MIN(n, m)); | |
398 | if (r != 0) | |
399 | return r; | |
400 | ||
6dd91b36 | 401 | return CMP(n, m); |
c1749834 | 402 | } |
07630cea LP |
403 | |
404 | bool chars_intersect(const char *a, const char *b) { | |
405 | const char *p; | |
406 | ||
407 | /* Returns true if any of the chars in a are in b. */ | |
408 | for (p = a; *p; p++) | |
409 | if (strchr(b, *p)) | |
410 | return true; | |
411 | ||
412 | return false; | |
413 | } | |
414 | ||
415 | bool string_has_cc(const char *p, const char *ok) { | |
416 | const char *t; | |
417 | ||
418 | assert(p); | |
419 | ||
420 | /* | |
421 | * Check if a string contains control characters. If 'ok' is | |
422 | * non-NULL it may be a string containing additional CCs to be | |
423 | * considered OK. | |
424 | */ | |
425 | ||
426 | for (t = p; *t; t++) { | |
427 | if (ok && strchr(ok, *t)) | |
428 | continue; | |
429 | ||
430 | if (*t > 0 && *t < ' ') | |
431 | return true; | |
432 | ||
433 | if (*t == 127) | |
434 | return true; | |
435 | } | |
436 | ||
437 | return false; | |
438 | } | |
439 | ||
8409f688 ZJS |
440 | static int write_ellipsis(char *buf, bool unicode) { |
441 | if (unicode || is_locale_utf8()) { | |
442 | buf[0] = 0xe2; /* tri-dot ellipsis: … */ | |
443 | buf[1] = 0x80; | |
444 | buf[2] = 0xa6; | |
445 | } else { | |
446 | buf[0] = '.'; | |
447 | buf[1] = '.'; | |
448 | buf[2] = '.'; | |
449 | } | |
450 | ||
451 | return 3; | |
452 | } | |
453 | ||
07630cea | 454 | static char *ascii_ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { |
9924aef6 ZJS |
455 | size_t x, need_space, suffix_len; |
456 | char *t; | |
07630cea LP |
457 | |
458 | assert(s); | |
459 | assert(percent <= 100); | |
c30a49b2 | 460 | assert(new_length != (size_t) -1); |
07630cea | 461 | |
c30a49b2 | 462 | if (old_length <= new_length) |
07630cea LP |
463 | return strndup(s, old_length); |
464 | ||
c30a49b2 LP |
465 | /* Special case short ellipsations */ |
466 | switch (new_length) { | |
467 | ||
468 | case 0: | |
469 | return strdup(""); | |
470 | ||
471 | case 1: | |
472 | if (is_locale_utf8()) | |
473 | return strdup("…"); | |
474 | else | |
475 | return strdup("."); | |
476 | ||
477 | case 2: | |
478 | if (!is_locale_utf8()) | |
479 | return strdup(".."); | |
480 | ||
481 | break; | |
482 | ||
483 | default: | |
484 | break; | |
485 | } | |
486 | ||
487 | /* Calculate how much space the ellipsis will take up. If we are in UTF-8 mode we only need space for one | |
488 | * character ("…"), otherwise for three characters ("..."). Note that in both cases we need 3 bytes of storage, | |
489 | * either for the UTF-8 encoded character or for three ASCII characters. */ | |
490 | need_space = is_locale_utf8() ? 1 : 3; | |
491 | ||
9924aef6 ZJS |
492 | t = new(char, new_length+3); |
493 | if (!t) | |
07630cea LP |
494 | return NULL; |
495 | ||
c30a49b2 | 496 | assert(new_length >= need_space); |
07630cea | 497 | |
c30a49b2 LP |
498 | x = ((new_length - need_space) * percent + 50) / 100; |
499 | assert(x <= new_length - need_space); | |
07630cea | 500 | |
9924aef6 ZJS |
501 | memcpy(t, s, x); |
502 | write_ellipsis(t + x, false); | |
503 | suffix_len = new_length - x - need_space; | |
504 | memcpy(t + x + 3, s + old_length - suffix_len, suffix_len); | |
505 | *(t + x + 3 + suffix_len) = '\0'; | |
07630cea | 506 | |
9924aef6 | 507 | return t; |
07630cea LP |
508 | } |
509 | ||
510 | char *ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 511 | size_t x, k, len, len2; |
07630cea | 512 | const char *i, *j; |
c30a49b2 | 513 | char *e; |
c932fb71 | 514 | int r; |
07630cea | 515 | |
c30a49b2 LP |
516 | /* Note that 'old_length' refers to bytes in the string, while 'new_length' refers to character cells taken up |
517 | * on screen. This distinction doesn't matter for ASCII strings, but it does matter for non-ASCII UTF-8 | |
518 | * strings. | |
519 | * | |
520 | * Ellipsation is done in a locale-dependent way: | |
521 | * 1. If the string passed in is fully ASCII and the current locale is not UTF-8, three dots are used ("...") | |
522 | * 2. Otherwise, a unicode ellipsis is used ("…") | |
523 | * | |
524 | * In other words: you'll get a unicode ellipsis as soon as either the string contains non-ASCII characters or | |
525 | * the current locale is UTF-8. | |
526 | */ | |
527 | ||
07630cea LP |
528 | assert(s); |
529 | assert(percent <= 100); | |
ddbc9319 LP |
530 | |
531 | if (new_length == (size_t) -1) | |
532 | return strndup(s, old_length); | |
533 | ||
c30a49b2 LP |
534 | if (new_length == 0) |
535 | return strdup(""); | |
07630cea | 536 | |
c30a49b2 | 537 | /* If no multibyte characters use ascii_ellipsize_mem for speed */ |
21e4e3e0 | 538 | if (ascii_is_valid_n(s, old_length)) |
07630cea LP |
539 | return ascii_ellipsize_mem(s, old_length, new_length, percent); |
540 | ||
c30a49b2 LP |
541 | x = ((new_length - 1) * percent) / 100; |
542 | assert(x <= new_length - 1); | |
07630cea LP |
543 | |
544 | k = 0; | |
9924aef6 | 545 | for (i = s; i < s + old_length; i = utf8_next_char(i)) { |
c932fb71 | 546 | char32_t c; |
9924aef6 | 547 | int w; |
07630cea | 548 | |
c932fb71 SL |
549 | r = utf8_encoded_to_unichar(i, &c); |
550 | if (r < 0) | |
07630cea | 551 | return NULL; |
07630cea | 552 | |
9924aef6 ZJS |
553 | w = unichar_iswide(c) ? 2 : 1; |
554 | if (k + w <= x) | |
555 | k += w; | |
556 | else | |
557 | break; | |
558 | } | |
07630cea | 559 | |
9924aef6 | 560 | for (j = s + old_length; j > i; ) { |
c932fb71 | 561 | char32_t c; |
9924aef6 ZJS |
562 | int w; |
563 | const char *jj; | |
07630cea | 564 | |
9924aef6 ZJS |
565 | jj = utf8_prev_char(j); |
566 | r = utf8_encoded_to_unichar(jj, &c); | |
c932fb71 | 567 | if (r < 0) |
07630cea | 568 | return NULL; |
9924aef6 ZJS |
569 | |
570 | w = unichar_iswide(c) ? 2 : 1; | |
571 | if (k + w <= new_length) { | |
572 | k += w; | |
573 | j = jj; | |
574 | } else | |
575 | break; | |
07630cea LP |
576 | } |
577 | assert(i <= j); | |
578 | ||
579 | /* we don't actually need to ellipsize */ | |
580 | if (i == j) | |
9924aef6 | 581 | return memdup_suffix0(s, old_length); |
07630cea | 582 | |
9924aef6 ZJS |
583 | /* make space for ellipsis, if possible */ |
584 | if (j < s + old_length) | |
585 | j = utf8_next_char(j); | |
586 | else if (i > s) | |
587 | i = utf8_prev_char(i); | |
07630cea LP |
588 | |
589 | len = i - s; | |
590 | len2 = s + old_length - j; | |
591 | e = new(char, len + 3 + len2 + 1); | |
592 | if (!e) | |
593 | return NULL; | |
594 | ||
595 | /* | |
596 | printf("old_length=%zu new_length=%zu x=%zu len=%u len2=%u k=%u\n", | |
597 | old_length, new_length, x, len, len2, k); | |
598 | */ | |
599 | ||
600 | memcpy(e, s, len); | |
8409f688 | 601 | write_ellipsis(e + len, true); |
9924aef6 ZJS |
602 | memcpy(e + len + 3, j, len2); |
603 | *(e + len + 3 + len2) = '\0'; | |
07630cea LP |
604 | |
605 | return e; | |
606 | } | |
607 | ||
8409f688 ZJS |
608 | char *cellescape(char *buf, size_t len, const char *s) { |
609 | /* Escape and ellipsize s into buffer buf of size len. Only non-control ASCII | |
610 | * characters are copied as they are, everything else is escaped. The result | |
611 | * is different then if escaping and ellipsization was performed in two | |
612 | * separate steps, because each sequence is either stored in full or skipped. | |
613 | * | |
614 | * This function should be used for logging about strings which expected to | |
615 | * be plain ASCII in a safe way. | |
616 | * | |
617 | * An ellipsis will be used if s is too long. It was always placed at the | |
618 | * very end. | |
619 | */ | |
620 | ||
61f6e276 LP |
621 | size_t i = 0, last_char_width[4] = {}, k = 0, j; |
622 | ||
623 | assert(len > 0); /* at least a terminating NUL */ | |
8409f688 | 624 | |
61f6e276 LP |
625 | for (;;) { |
626 | char four[4]; | |
627 | int w; | |
8409f688 | 628 | |
61f6e276 | 629 | if (*s == 0) /* terminating NUL detected? then we are done! */ |
8409f688 | 630 | goto done; |
61f6e276 LP |
631 | |
632 | w = cescape_char(*s, four); | |
633 | if (i + w + 1 > len) /* This character doesn't fit into the buffer anymore? In that case let's | |
634 | * ellipsize at the previous location */ | |
635 | break; | |
636 | ||
637 | /* OK, there was space, let's add this escaped character to the buffer */ | |
638 | memcpy(buf + i, four, w); | |
639 | i += w; | |
640 | ||
641 | /* And remember its width in the ring buffer */ | |
642 | last_char_width[k] = w; | |
643 | k = (k + 1) % 4; | |
644 | ||
645 | s++; | |
8409f688 ZJS |
646 | } |
647 | ||
61f6e276 LP |
648 | /* Ellipsation is necessary. This means we might need to truncate the string again to make space for 4 |
649 | * characters ideally, but the buffer is shorter than that in the first place take what we can get */ | |
650 | for (j = 0; j < ELEMENTSOF(last_char_width); j++) { | |
651 | ||
652 | if (i + 4 <= len) /* nice, we reached our space goal */ | |
653 | break; | |
654 | ||
655 | k = k == 0 ? 3 : k - 1; | |
656 | if (last_char_width[k] == 0) /* bummer, we reached the beginning of the strings */ | |
657 | break; | |
658 | ||
659 | assert(i >= last_char_width[k]); | |
660 | i -= last_char_width[k]; | |
8409f688 ZJS |
661 | } |
662 | ||
61f6e276 LP |
663 | if (i + 4 <= len) /* yay, enough space */ |
664 | i += write_ellipsis(buf + i, false); | |
665 | else if (i + 3 <= len) { /* only space for ".." */ | |
666 | buf[i++] = '.'; | |
667 | buf[i++] = '.'; | |
668 | } else if (i + 2 <= len) /* only space for a single "." */ | |
669 | buf[i++] = '.'; | |
670 | else | |
671 | assert(i + 1 <= len); | |
672 | ||
8409f688 ZJS |
673 | done: |
674 | buf[i] = '\0'; | |
675 | return buf; | |
676 | } | |
677 | ||
2d5dece8 | 678 | bool nulstr_contains(const char *nulstr, const char *needle) { |
07630cea LP |
679 | const char *i; |
680 | ||
681 | if (!nulstr) | |
682 | return false; | |
683 | ||
684 | NULSTR_FOREACH(i, nulstr) | |
685 | if (streq(i, needle)) | |
686 | return true; | |
687 | ||
688 | return false; | |
689 | } | |
690 | ||
691 | char* strshorten(char *s, size_t l) { | |
692 | assert(s); | |
693 | ||
47b33c7d | 694 | if (strnlen(s, l+1) > l) |
07630cea LP |
695 | s[l] = 0; |
696 | ||
697 | return s; | |
698 | } | |
699 | ||
700 | char *strreplace(const char *text, const char *old_string, const char *new_string) { | |
9d73565a LP |
701 | size_t l, old_len, new_len, allocated = 0; |
702 | char *t, *ret = NULL; | |
07630cea | 703 | const char *f; |
07630cea | 704 | |
07630cea LP |
705 | assert(old_string); |
706 | assert(new_string); | |
707 | ||
9d73565a LP |
708 | if (!text) |
709 | return NULL; | |
710 | ||
07630cea LP |
711 | old_len = strlen(old_string); |
712 | new_len = strlen(new_string); | |
713 | ||
714 | l = strlen(text); | |
9d73565a | 715 | if (!GREEDY_REALLOC(ret, allocated, l+1)) |
07630cea LP |
716 | return NULL; |
717 | ||
718 | f = text; | |
9d73565a | 719 | t = ret; |
07630cea | 720 | while (*f) { |
07630cea LP |
721 | size_t d, nl; |
722 | ||
723 | if (!startswith(f, old_string)) { | |
724 | *(t++) = *(f++); | |
725 | continue; | |
726 | } | |
727 | ||
9d73565a | 728 | d = t - ret; |
07630cea | 729 | nl = l - old_len + new_len; |
9d73565a LP |
730 | |
731 | if (!GREEDY_REALLOC(ret, allocated, nl + 1)) | |
732 | return mfree(ret); | |
07630cea LP |
733 | |
734 | l = nl; | |
9d73565a | 735 | t = ret + d; |
07630cea LP |
736 | |
737 | t = stpcpy(t, new_string); | |
738 | f += old_len; | |
739 | } | |
740 | ||
741 | *t = 0; | |
9d73565a | 742 | return ret; |
07630cea LP |
743 | } |
744 | ||
3042bbeb | 745 | static void advance_offsets(ssize_t diff, size_t offsets[static 2], size_t shift[static 2], size_t size) { |
b4766d5f ZJS |
746 | if (!offsets) |
747 | return; | |
748 | ||
749 | if ((size_t) diff < offsets[0]) | |
750 | shift[0] += size; | |
751 | if ((size_t) diff < offsets[1]) | |
752 | shift[1] += size; | |
753 | } | |
754 | ||
755 | char *strip_tab_ansi(char **ibuf, size_t *_isz, size_t highlight[2]) { | |
07630cea LP |
756 | const char *i, *begin = NULL; |
757 | enum { | |
758 | STATE_OTHER, | |
759 | STATE_ESCAPE, | |
695a944c LP |
760 | STATE_CSI, |
761 | STATE_CSO, | |
07630cea LP |
762 | } state = STATE_OTHER; |
763 | char *obuf = NULL; | |
b4766d5f | 764 | size_t osz = 0, isz, shift[2] = {}; |
07630cea LP |
765 | FILE *f; |
766 | ||
767 | assert(ibuf); | |
768 | assert(*ibuf); | |
769 | ||
695a944c LP |
770 | /* This does three things: |
771 | * | |
772 | * 1. Replaces TABs by 8 spaces | |
773 | * 2. Strips ANSI color sequences (a subset of CSI), i.e. ESC '[' … 'm' sequences | |
774 | * 3. Strips ANSI operating system sequences (CSO), i.e. ESC ']' … BEL sequences | |
775 | * | |
776 | * Everything else will be left as it is. In particular other ANSI sequences are left as they are, as are any | |
777 | * other special characters. Truncated ANSI sequences are left-as is too. This call is supposed to suppress the | |
778 | * most basic formatting noise, but nothing else. | |
779 | * | |
780 | * Why care for CSO sequences? Well, to undo what terminal_urlify() and friends generate. */ | |
07630cea LP |
781 | |
782 | isz = _isz ? *_isz : strlen(*ibuf); | |
783 | ||
784 | f = open_memstream(&obuf, &osz); | |
785 | if (!f) | |
786 | return NULL; | |
787 | ||
0d536673 LP |
788 | /* Note we turn off internal locking on f for performance reasons. It's safe to do so since we created f here |
789 | * and it doesn't leave our scope. */ | |
790 | ||
791 | (void) __fsetlocking(f, FSETLOCKING_BYCALLER); | |
db3f45e2 | 792 | |
07630cea LP |
793 | for (i = *ibuf; i < *ibuf + isz + 1; i++) { |
794 | ||
795 | switch (state) { | |
796 | ||
797 | case STATE_OTHER: | |
798 | if (i >= *ibuf + isz) /* EOT */ | |
799 | break; | |
800 | else if (*i == '\x1B') | |
801 | state = STATE_ESCAPE; | |
b4766d5f | 802 | else if (*i == '\t') { |
0d536673 | 803 | fputs(" ", f); |
b4766d5f ZJS |
804 | advance_offsets(i - *ibuf, highlight, shift, 7); |
805 | } else | |
0d536673 | 806 | fputc(*i, f); |
b4766d5f | 807 | |
07630cea LP |
808 | break; |
809 | ||
810 | case STATE_ESCAPE: | |
811 | if (i >= *ibuf + isz) { /* EOT */ | |
0d536673 | 812 | fputc('\x1B', f); |
b4766d5f | 813 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea | 814 | break; |
695a944c LP |
815 | } else if (*i == '[') { /* ANSI CSI */ |
816 | state = STATE_CSI; | |
817 | begin = i + 1; | |
818 | } else if (*i == ']') { /* ANSI CSO */ | |
819 | state = STATE_CSO; | |
07630cea LP |
820 | begin = i + 1; |
821 | } else { | |
0d536673 LP |
822 | fputc('\x1B', f); |
823 | fputc(*i, f); | |
b4766d5f | 824 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea LP |
825 | state = STATE_OTHER; |
826 | } | |
827 | ||
828 | break; | |
829 | ||
695a944c | 830 | case STATE_CSI: |
07630cea | 831 | |
695a944c LP |
832 | if (i >= *ibuf + isz || /* EOT … */ |
833 | !strchr("01234567890;m", *i)) { /* … or invalid chars in sequence */ | |
0d536673 LP |
834 | fputc('\x1B', f); |
835 | fputc('[', f); | |
b4766d5f | 836 | advance_offsets(i - *ibuf, highlight, shift, 2); |
07630cea LP |
837 | state = STATE_OTHER; |
838 | i = begin-1; | |
839 | } else if (*i == 'm') | |
840 | state = STATE_OTHER; | |
695a944c LP |
841 | |
842 | break; | |
843 | ||
844 | case STATE_CSO: | |
845 | ||
846 | if (i >= *ibuf + isz || /* EOT … */ | |
847 | (*i != '\a' && (uint8_t) *i < 32U) || (uint8_t) *i > 126U) { /* … or invalid chars in sequence */ | |
848 | fputc('\x1B', f); | |
849 | fputc(']', f); | |
850 | advance_offsets(i - *ibuf, highlight, shift, 2); | |
851 | state = STATE_OTHER; | |
852 | i = begin-1; | |
853 | } else if (*i == '\a') | |
854 | state = STATE_OTHER; | |
855 | ||
07630cea LP |
856 | break; |
857 | } | |
858 | } | |
859 | ||
c7e03d2e | 860 | if (fflush_and_check(f) < 0) { |
07630cea | 861 | fclose(f); |
6b430fdb | 862 | return mfree(obuf); |
07630cea LP |
863 | } |
864 | ||
865 | fclose(f); | |
866 | ||
867 | free(*ibuf); | |
868 | *ibuf = obuf; | |
869 | ||
870 | if (_isz) | |
871 | *_isz = osz; | |
872 | ||
b4766d5f ZJS |
873 | if (highlight) { |
874 | highlight[0] += shift[0]; | |
875 | highlight[1] += shift[1]; | |
876 | } | |
877 | ||
07630cea LP |
878 | return obuf; |
879 | } | |
880 | ||
bb8ad9ea LP |
881 | char *strextend_with_separator(char **x, const char *separator, ...) { |
882 | bool need_separator; | |
883 | size_t f, l, l_separator; | |
07630cea | 884 | char *r, *p; |
bb8ad9ea | 885 | va_list ap; |
07630cea LP |
886 | |
887 | assert(x); | |
888 | ||
7bf7ce28 | 889 | l = f = strlen_ptr(*x); |
07630cea | 890 | |
bb8ad9ea LP |
891 | need_separator = !isempty(*x); |
892 | l_separator = strlen_ptr(separator); | |
893 | ||
894 | va_start(ap, separator); | |
07630cea LP |
895 | for (;;) { |
896 | const char *t; | |
897 | size_t n; | |
898 | ||
899 | t = va_arg(ap, const char *); | |
900 | if (!t) | |
901 | break; | |
902 | ||
903 | n = strlen(t); | |
bb8ad9ea LP |
904 | |
905 | if (need_separator) | |
906 | n += l_separator; | |
907 | ||
07630cea LP |
908 | if (n > ((size_t) -1) - l) { |
909 | va_end(ap); | |
910 | return NULL; | |
911 | } | |
912 | ||
913 | l += n; | |
bb8ad9ea | 914 | need_separator = true; |
07630cea LP |
915 | } |
916 | va_end(ap); | |
917 | ||
bb8ad9ea LP |
918 | need_separator = !isempty(*x); |
919 | ||
07630cea LP |
920 | r = realloc(*x, l+1); |
921 | if (!r) | |
922 | return NULL; | |
923 | ||
924 | p = r + f; | |
925 | ||
bb8ad9ea | 926 | va_start(ap, separator); |
07630cea LP |
927 | for (;;) { |
928 | const char *t; | |
929 | ||
930 | t = va_arg(ap, const char *); | |
931 | if (!t) | |
932 | break; | |
933 | ||
bb8ad9ea LP |
934 | if (need_separator && separator) |
935 | p = stpcpy(p, separator); | |
936 | ||
07630cea | 937 | p = stpcpy(p, t); |
bb8ad9ea LP |
938 | |
939 | need_separator = true; | |
07630cea LP |
940 | } |
941 | va_end(ap); | |
942 | ||
bb8ad9ea LP |
943 | assert(p == r + l); |
944 | ||
07630cea LP |
945 | *p = 0; |
946 | *x = r; | |
947 | ||
948 | return r + l; | |
949 | } | |
950 | ||
951 | char *strrep(const char *s, unsigned n) { | |
952 | size_t l; | |
953 | char *r, *p; | |
954 | unsigned i; | |
955 | ||
956 | assert(s); | |
957 | ||
958 | l = strlen(s); | |
959 | p = r = malloc(l * n + 1); | |
960 | if (!r) | |
961 | return NULL; | |
962 | ||
963 | for (i = 0; i < n; i++) | |
964 | p = stpcpy(p, s); | |
965 | ||
966 | *p = 0; | |
967 | return r; | |
968 | } | |
969 | ||
970 | int split_pair(const char *s, const char *sep, char **l, char **r) { | |
971 | char *x, *a, *b; | |
972 | ||
973 | assert(s); | |
974 | assert(sep); | |
975 | assert(l); | |
976 | assert(r); | |
977 | ||
978 | if (isempty(sep)) | |
979 | return -EINVAL; | |
980 | ||
981 | x = strstr(s, sep); | |
982 | if (!x) | |
983 | return -EINVAL; | |
984 | ||
985 | a = strndup(s, x - s); | |
986 | if (!a) | |
987 | return -ENOMEM; | |
988 | ||
989 | b = strdup(x + strlen(sep)); | |
990 | if (!b) { | |
991 | free(a); | |
992 | return -ENOMEM; | |
993 | } | |
994 | ||
995 | *l = a; | |
996 | *r = b; | |
997 | ||
998 | return 0; | |
999 | } | |
1000 | ||
1001 | int free_and_strdup(char **p, const char *s) { | |
1002 | char *t; | |
1003 | ||
1004 | assert(p); | |
1005 | ||
7f546026 | 1006 | /* Replaces a string pointer with a strdup()ed new string, |
07630cea LP |
1007 | * possibly freeing the old one. */ |
1008 | ||
1009 | if (streq_ptr(*p, s)) | |
1010 | return 0; | |
1011 | ||
1012 | if (s) { | |
1013 | t = strdup(s); | |
1014 | if (!t) | |
1015 | return -ENOMEM; | |
1016 | } else | |
1017 | t = NULL; | |
1018 | ||
1019 | free(*p); | |
1020 | *p = t; | |
1021 | ||
1022 | return 1; | |
1023 | } | |
1024 | ||
7f546026 ZJS |
1025 | int free_and_strndup(char **p, const char *s, size_t l) { |
1026 | char *t; | |
1027 | ||
1028 | assert(p); | |
1029 | assert(s || l == 0); | |
1030 | ||
1031 | /* Replaces a string pointer with a strndup()ed new string, | |
1032 | * freeing the old one. */ | |
1033 | ||
1034 | if (!*p && !s) | |
1035 | return 0; | |
1036 | ||
1037 | if (*p && s && strneq(*p, s, l) && (l > strlen(*p) || (*p)[l] == '\0')) | |
1038 | return 0; | |
1039 | ||
1040 | if (s) { | |
1041 | t = strndup(s, l); | |
1042 | if (!t) | |
1043 | return -ENOMEM; | |
1044 | } else | |
1045 | t = NULL; | |
1046 | ||
1047 | free_and_replace(*p, t); | |
1048 | return 1; | |
1049 | } | |
1050 | ||
4b9545f1 | 1051 | #if !HAVE_EXPLICIT_BZERO |
b6b609db MB |
1052 | /* |
1053 | * Pointer to memset is volatile so that compiler must de-reference | |
1054 | * the pointer and can't assume that it points to any function in | |
1055 | * particular (such as memset, which it then might further "optimize") | |
1056 | * This approach is inspired by openssl's crypto/mem_clr.c. | |
1057 | */ | |
1058 | typedef void *(*memset_t)(void *,int,size_t); | |
9fe4ea21 | 1059 | |
b6b609db | 1060 | static volatile memset_t memset_func = memset; |
9fe4ea21 | 1061 | |
87f54463 LP |
1062 | void* explicit_bzero_safe(void *p, size_t l) { |
1063 | if (l > 0) | |
1064 | memset_func(p, '\0', l); | |
1065 | ||
1066 | return p; | |
9fe4ea21 | 1067 | } |
2d26d8e0 | 1068 | #endif |
9fe4ea21 | 1069 | |
9fe4ea21 | 1070 | char* string_erase(char *x) { |
07630cea | 1071 | if (!x) |
9fe4ea21 | 1072 | return NULL; |
07630cea LP |
1073 | |
1074 | /* A delicious drop of snake-oil! To be called on memory where | |
1075 | * we stored passphrases or so, after we used them. */ | |
87f54463 | 1076 | explicit_bzero_safe(x, strlen(x)); |
2d26d8e0 | 1077 | return x; |
07630cea LP |
1078 | } |
1079 | ||
1080 | char *string_free_erase(char *s) { | |
9fe4ea21 | 1081 | return mfree(string_erase(s)); |
07630cea | 1082 | } |
f3e2e81d LP |
1083 | |
1084 | bool string_is_safe(const char *p) { | |
1085 | const char *t; | |
1086 | ||
1087 | if (!p) | |
1088 | return false; | |
1089 | ||
1090 | for (t = p; *t; t++) { | |
1091 | if (*t > 0 && *t < ' ') /* no control characters */ | |
1092 | return false; | |
1093 | ||
1094 | if (strchr(QUOTES "\\\x7f", *t)) | |
1095 | return false; | |
1096 | } | |
1097 | ||
1098 | return true; | |
1099 | } |