]>
Commit | Line | Data |
---|---|---|
53e1b683 | 1 | /* SPDX-License-Identifier: LGPL-2.1+ */ |
07630cea LP |
2 | /*** |
3 | This file is part of systemd. | |
4 | ||
5 | Copyright 2010 Lennart Poettering | |
07630cea LP |
6 | ***/ |
7 | ||
11c3a366 TA |
8 | #include <errno.h> |
9 | #include <stdarg.h> | |
10 | #include <stdint.h> | |
11 | #include <stdio.h> | |
0d536673 | 12 | #include <stdio_ext.h> |
11c3a366 | 13 | #include <stdlib.h> |
b6b609db | 14 | #include <string.h> |
11c3a366 | 15 | |
b5efdb8a | 16 | #include "alloc-util.h" |
07630cea | 17 | #include "gunicode.h" |
c30a49b2 | 18 | #include "locale-util.h" |
11c3a366 | 19 | #include "macro.h" |
b11d6a7b | 20 | #include "string-util.h" |
b4766d5f | 21 | #include "terminal-util.h" |
07630cea LP |
22 | #include "utf8.h" |
23 | #include "util.h" | |
c7e03d2e | 24 | #include "fileio.h" |
07630cea LP |
25 | |
26 | int strcmp_ptr(const char *a, const char *b) { | |
27 | ||
28 | /* Like strcmp(), but tries to make sense of NULL pointers */ | |
29 | if (a && b) | |
30 | return strcmp(a, b); | |
31 | ||
32 | if (!a && b) | |
33 | return -1; | |
34 | ||
35 | if (a && !b) | |
36 | return 1; | |
37 | ||
38 | return 0; | |
39 | } | |
40 | ||
41 | char* endswith(const char *s, const char *postfix) { | |
42 | size_t sl, pl; | |
43 | ||
44 | assert(s); | |
45 | assert(postfix); | |
46 | ||
47 | sl = strlen(s); | |
48 | pl = strlen(postfix); | |
49 | ||
50 | if (pl == 0) | |
51 | return (char*) s + sl; | |
52 | ||
53 | if (sl < pl) | |
54 | return NULL; | |
55 | ||
56 | if (memcmp(s + sl - pl, postfix, pl) != 0) | |
57 | return NULL; | |
58 | ||
59 | return (char*) s + sl - pl; | |
60 | } | |
61 | ||
62 | char* endswith_no_case(const char *s, const char *postfix) { | |
63 | size_t sl, pl; | |
64 | ||
65 | assert(s); | |
66 | assert(postfix); | |
67 | ||
68 | sl = strlen(s); | |
69 | pl = strlen(postfix); | |
70 | ||
71 | if (pl == 0) | |
72 | return (char*) s + sl; | |
73 | ||
74 | if (sl < pl) | |
75 | return NULL; | |
76 | ||
77 | if (strcasecmp(s + sl - pl, postfix) != 0) | |
78 | return NULL; | |
79 | ||
80 | return (char*) s + sl - pl; | |
81 | } | |
82 | ||
83 | char* first_word(const char *s, const char *word) { | |
84 | size_t sl, wl; | |
85 | const char *p; | |
86 | ||
87 | assert(s); | |
88 | assert(word); | |
89 | ||
90 | /* Checks if the string starts with the specified word, either | |
91 | * followed by NUL or by whitespace. Returns a pointer to the | |
92 | * NUL or the first character after the whitespace. */ | |
93 | ||
94 | sl = strlen(s); | |
95 | wl = strlen(word); | |
96 | ||
97 | if (sl < wl) | |
98 | return NULL; | |
99 | ||
100 | if (wl == 0) | |
101 | return (char*) s; | |
102 | ||
103 | if (memcmp(s, word, wl) != 0) | |
104 | return NULL; | |
105 | ||
106 | p = s + wl; | |
107 | if (*p == 0) | |
108 | return (char*) p; | |
109 | ||
110 | if (!strchr(WHITESPACE, *p)) | |
111 | return NULL; | |
112 | ||
113 | p += strspn(p, WHITESPACE); | |
114 | return (char*) p; | |
115 | } | |
116 | ||
117 | static size_t strcspn_escaped(const char *s, const char *reject) { | |
118 | bool escaped = false; | |
119 | int n; | |
120 | ||
121 | for (n=0; s[n]; n++) { | |
122 | if (escaped) | |
123 | escaped = false; | |
124 | else if (s[n] == '\\') | |
125 | escaped = true; | |
126 | else if (strchr(reject, s[n])) | |
127 | break; | |
128 | } | |
129 | ||
130 | /* if s ends in \, return index of previous char */ | |
131 | return n - escaped; | |
132 | } | |
133 | ||
134 | /* Split a string into words. */ | |
135 | const char* split(const char **state, size_t *l, const char *separator, bool quoted) { | |
136 | const char *current; | |
137 | ||
138 | current = *state; | |
139 | ||
140 | if (!*current) { | |
141 | assert(**state == '\0'); | |
142 | return NULL; | |
143 | } | |
144 | ||
145 | current += strspn(current, separator); | |
146 | if (!*current) { | |
147 | *state = current; | |
148 | return NULL; | |
149 | } | |
150 | ||
151 | if (quoted && strchr("\'\"", *current)) { | |
152 | char quotechars[2] = {*current, '\0'}; | |
153 | ||
154 | *l = strcspn_escaped(current + 1, quotechars); | |
155 | if (current[*l + 1] == '\0' || current[*l + 1] != quotechars[0] || | |
156 | (current[*l + 2] && !strchr(separator, current[*l + 2]))) { | |
157 | /* right quote missing or garbage at the end */ | |
158 | *state = current; | |
159 | return NULL; | |
160 | } | |
161 | *state = current++ + *l + 2; | |
162 | } else if (quoted) { | |
163 | *l = strcspn_escaped(current, separator); | |
164 | if (current[*l] && !strchr(separator, current[*l])) { | |
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) { | |
271 | char *e; | |
272 | ||
7546145e LP |
273 | if (!s) |
274 | return NULL; | |
275 | ||
07630cea LP |
276 | /* Drops trailing whitespace. Modifies the string in |
277 | * place. Returns pointer to first non-space character */ | |
278 | ||
279 | s += strspn(s, WHITESPACE); | |
280 | ||
281 | for (e = strchr(s, 0); e > s; e --) | |
282 | if (!strchr(WHITESPACE, e[-1])) | |
283 | break; | |
284 | ||
285 | *e = 0; | |
286 | ||
287 | return s; | |
288 | } | |
289 | ||
290 | char *delete_chars(char *s, const char *bad) { | |
291 | char *f, *t; | |
292 | ||
7546145e LP |
293 | /* Drops all specified bad characters, regardless where in the string */ |
294 | ||
295 | if (!s) | |
296 | return NULL; | |
297 | ||
298 | if (!bad) | |
299 | bad = WHITESPACE; | |
07630cea LP |
300 | |
301 | for (f = s, t = s; *f; f++) { | |
302 | if (strchr(bad, *f)) | |
303 | continue; | |
304 | ||
305 | *(t++) = *f; | |
306 | } | |
307 | ||
308 | *t = 0; | |
309 | ||
310 | return s; | |
311 | } | |
312 | ||
7546145e LP |
313 | char *delete_trailing_chars(char *s, const char *bad) { |
314 | char *p, *c = s; | |
315 | ||
316 | /* Drops all specified bad characters, at the end of the string */ | |
317 | ||
318 | if (!s) | |
319 | return NULL; | |
320 | ||
321 | if (!bad) | |
322 | bad = WHITESPACE; | |
323 | ||
324 | for (p = s; *p; p++) | |
325 | if (!strchr(bad, *p)) | |
326 | c = p + 1; | |
327 | ||
328 | *c = 0; | |
329 | ||
330 | return s; | |
331 | } | |
332 | ||
07630cea LP |
333 | char *truncate_nl(char *s) { |
334 | assert(s); | |
335 | ||
336 | s[strcspn(s, NEWLINE)] = 0; | |
337 | return s; | |
338 | } | |
339 | ||
b577e3d5 LP |
340 | char ascii_tolower(char x) { |
341 | ||
342 | if (x >= 'A' && x <= 'Z') | |
343 | return x - 'A' + 'a'; | |
344 | ||
345 | return x; | |
346 | } | |
347 | ||
846b8fc3 LP |
348 | char ascii_toupper(char x) { |
349 | ||
350 | if (x >= 'a' && x <= 'z') | |
351 | return x - 'a' + 'A'; | |
352 | ||
353 | return x; | |
354 | } | |
355 | ||
07630cea LP |
356 | char *ascii_strlower(char *t) { |
357 | char *p; | |
358 | ||
359 | assert(t); | |
360 | ||
361 | for (p = t; *p; p++) | |
b577e3d5 LP |
362 | *p = ascii_tolower(*p); |
363 | ||
364 | return t; | |
365 | } | |
366 | ||
846b8fc3 LP |
367 | char *ascii_strupper(char *t) { |
368 | char *p; | |
369 | ||
370 | assert(t); | |
371 | ||
372 | for (p = t; *p; p++) | |
373 | *p = ascii_toupper(*p); | |
374 | ||
375 | return t; | |
376 | } | |
377 | ||
b577e3d5 LP |
378 | char *ascii_strlower_n(char *t, size_t n) { |
379 | size_t i; | |
380 | ||
381 | if (n <= 0) | |
382 | return t; | |
383 | ||
384 | for (i = 0; i < n; i++) | |
385 | t[i] = ascii_tolower(t[i]); | |
07630cea LP |
386 | |
387 | return t; | |
388 | } | |
522d85ae LP |
389 | |
390 | int ascii_strcasecmp_n(const char *a, const char *b, size_t n) { | |
391 | ||
392 | for (; n > 0; a++, b++, n--) { | |
393 | int x, y; | |
394 | ||
395 | x = (int) (uint8_t) ascii_tolower(*a); | |
396 | y = (int) (uint8_t) ascii_tolower(*b); | |
397 | ||
398 | if (x != y) | |
399 | return x - y; | |
400 | } | |
401 | ||
402 | return 0; | |
403 | } | |
c1749834 LP |
404 | |
405 | int ascii_strcasecmp_nn(const char *a, size_t n, const char *b, size_t m) { | |
406 | int r; | |
407 | ||
408 | r = ascii_strcasecmp_n(a, b, MIN(n, m)); | |
409 | if (r != 0) | |
410 | return r; | |
411 | ||
412 | if (n < m) | |
413 | return -1; | |
414 | else if (n > m) | |
415 | return 1; | |
416 | else | |
417 | return 0; | |
418 | } | |
07630cea LP |
419 | |
420 | bool chars_intersect(const char *a, const char *b) { | |
421 | const char *p; | |
422 | ||
423 | /* Returns true if any of the chars in a are in b. */ | |
424 | for (p = a; *p; p++) | |
425 | if (strchr(b, *p)) | |
426 | return true; | |
427 | ||
428 | return false; | |
429 | } | |
430 | ||
431 | bool string_has_cc(const char *p, const char *ok) { | |
432 | const char *t; | |
433 | ||
434 | assert(p); | |
435 | ||
436 | /* | |
437 | * Check if a string contains control characters. If 'ok' is | |
438 | * non-NULL it may be a string containing additional CCs to be | |
439 | * considered OK. | |
440 | */ | |
441 | ||
442 | for (t = p; *t; t++) { | |
443 | if (ok && strchr(ok, *t)) | |
444 | continue; | |
445 | ||
446 | if (*t > 0 && *t < ' ') | |
447 | return true; | |
448 | ||
449 | if (*t == 127) | |
450 | return true; | |
451 | } | |
452 | ||
453 | return false; | |
454 | } | |
455 | ||
456 | static char *ascii_ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 457 | size_t x, need_space; |
07630cea LP |
458 | char *r; |
459 | ||
460 | assert(s); | |
461 | assert(percent <= 100); | |
c30a49b2 | 462 | assert(new_length != (size_t) -1); |
07630cea | 463 | |
c30a49b2 | 464 | if (old_length <= new_length) |
07630cea LP |
465 | return strndup(s, old_length); |
466 | ||
c30a49b2 LP |
467 | /* Special case short ellipsations */ |
468 | switch (new_length) { | |
469 | ||
470 | case 0: | |
471 | return strdup(""); | |
472 | ||
473 | case 1: | |
474 | if (is_locale_utf8()) | |
475 | return strdup("…"); | |
476 | else | |
477 | return strdup("."); | |
478 | ||
479 | case 2: | |
480 | if (!is_locale_utf8()) | |
481 | return strdup(".."); | |
482 | ||
483 | break; | |
484 | ||
485 | default: | |
486 | break; | |
487 | } | |
488 | ||
489 | /* Calculate how much space the ellipsis will take up. If we are in UTF-8 mode we only need space for one | |
490 | * character ("…"), otherwise for three characters ("..."). Note that in both cases we need 3 bytes of storage, | |
491 | * either for the UTF-8 encoded character or for three ASCII characters. */ | |
492 | need_space = is_locale_utf8() ? 1 : 3; | |
493 | ||
494 | r = new(char, new_length+3); | |
07630cea LP |
495 | if (!r) |
496 | return NULL; | |
497 | ||
c30a49b2 | 498 | assert(new_length >= need_space); |
07630cea | 499 | |
c30a49b2 LP |
500 | x = ((new_length - need_space) * percent + 50) / 100; |
501 | assert(x <= new_length - need_space); | |
07630cea LP |
502 | |
503 | memcpy(r, s, x); | |
c30a49b2 LP |
504 | |
505 | if (is_locale_utf8()) { | |
506 | r[x+0] = 0xe2; /* tri-dot ellipsis: … */ | |
507 | r[x+1] = 0x80; | |
508 | r[x+2] = 0xa6; | |
509 | } else { | |
510 | r[x+0] = '.'; | |
511 | r[x+1] = '.'; | |
512 | r[x+2] = '.'; | |
513 | } | |
514 | ||
07630cea | 515 | memcpy(r + x + 3, |
c30a49b2 LP |
516 | s + old_length - (new_length - x - need_space), |
517 | new_length - x - need_space + 1); | |
07630cea LP |
518 | |
519 | return r; | |
520 | } | |
521 | ||
522 | char *ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 523 | size_t x, k, len, len2; |
07630cea | 524 | const char *i, *j; |
c30a49b2 | 525 | char *e; |
c932fb71 | 526 | int r; |
07630cea | 527 | |
c30a49b2 LP |
528 | /* Note that 'old_length' refers to bytes in the string, while 'new_length' refers to character cells taken up |
529 | * on screen. This distinction doesn't matter for ASCII strings, but it does matter for non-ASCII UTF-8 | |
530 | * strings. | |
531 | * | |
532 | * Ellipsation is done in a locale-dependent way: | |
533 | * 1. If the string passed in is fully ASCII and the current locale is not UTF-8, three dots are used ("...") | |
534 | * 2. Otherwise, a unicode ellipsis is used ("…") | |
535 | * | |
536 | * In other words: you'll get a unicode ellipsis as soon as either the string contains non-ASCII characters or | |
537 | * the current locale is UTF-8. | |
538 | */ | |
539 | ||
07630cea LP |
540 | assert(s); |
541 | assert(percent <= 100); | |
ddbc9319 LP |
542 | |
543 | if (new_length == (size_t) -1) | |
544 | return strndup(s, old_length); | |
545 | ||
c30a49b2 LP |
546 | if (new_length == 0) |
547 | return strdup(""); | |
07630cea | 548 | |
c30a49b2 | 549 | /* If no multibyte characters use ascii_ellipsize_mem for speed */ |
07630cea LP |
550 | if (ascii_is_valid(s)) |
551 | return ascii_ellipsize_mem(s, old_length, new_length, percent); | |
552 | ||
c30a49b2 LP |
553 | x = ((new_length - 1) * percent) / 100; |
554 | assert(x <= new_length - 1); | |
07630cea LP |
555 | |
556 | k = 0; | |
557 | for (i = s; k < x && i < s + old_length; i = utf8_next_char(i)) { | |
c932fb71 | 558 | char32_t c; |
07630cea | 559 | |
c932fb71 SL |
560 | r = utf8_encoded_to_unichar(i, &c); |
561 | if (r < 0) | |
07630cea LP |
562 | return NULL; |
563 | k += unichar_iswide(c) ? 2 : 1; | |
564 | } | |
565 | ||
566 | if (k > x) /* last character was wide and went over quota */ | |
313cefa1 | 567 | x++; |
07630cea LP |
568 | |
569 | for (j = s + old_length; k < new_length && j > i; ) { | |
c932fb71 | 570 | char32_t c; |
07630cea LP |
571 | |
572 | j = utf8_prev_char(j); | |
c932fb71 SL |
573 | r = utf8_encoded_to_unichar(j, &c); |
574 | if (r < 0) | |
07630cea LP |
575 | return NULL; |
576 | k += unichar_iswide(c) ? 2 : 1; | |
577 | } | |
578 | assert(i <= j); | |
579 | ||
580 | /* we don't actually need to ellipsize */ | |
581 | if (i == j) | |
582 | return memdup(s, old_length + 1); | |
583 | ||
584 | /* make space for ellipsis */ | |
585 | j = utf8_next_char(j); | |
586 | ||
587 | len = i - s; | |
588 | len2 = s + old_length - j; | |
589 | e = new(char, len + 3 + len2 + 1); | |
590 | if (!e) | |
591 | return NULL; | |
592 | ||
593 | /* | |
594 | printf("old_length=%zu new_length=%zu x=%zu len=%u len2=%u k=%u\n", | |
595 | old_length, new_length, x, len, len2, k); | |
596 | */ | |
597 | ||
598 | memcpy(e, s, len); | |
c30a49b2 | 599 | e[len + 0] = 0xe2; /* tri-dot ellipsis: … */ |
07630cea LP |
600 | e[len + 1] = 0x80; |
601 | e[len + 2] = 0xa6; | |
602 | ||
603 | memcpy(e + len + 3, j, len2 + 1); | |
604 | ||
605 | return e; | |
606 | } | |
607 | ||
608 | char *ellipsize(const char *s, size_t length, unsigned percent) { | |
ddbc9319 LP |
609 | |
610 | if (length == (size_t) -1) | |
611 | return strdup(s); | |
612 | ||
07630cea LP |
613 | return ellipsize_mem(s, strlen(s), length, percent); |
614 | } | |
615 | ||
2d5dece8 | 616 | bool nulstr_contains(const char *nulstr, const char *needle) { |
07630cea LP |
617 | const char *i; |
618 | ||
619 | if (!nulstr) | |
620 | return false; | |
621 | ||
622 | NULSTR_FOREACH(i, nulstr) | |
623 | if (streq(i, needle)) | |
624 | return true; | |
625 | ||
626 | return false; | |
627 | } | |
628 | ||
629 | char* strshorten(char *s, size_t l) { | |
630 | assert(s); | |
631 | ||
47b33c7d | 632 | if (strnlen(s, l+1) > l) |
07630cea LP |
633 | s[l] = 0; |
634 | ||
635 | return s; | |
636 | } | |
637 | ||
638 | char *strreplace(const char *text, const char *old_string, const char *new_string) { | |
9d73565a LP |
639 | size_t l, old_len, new_len, allocated = 0; |
640 | char *t, *ret = NULL; | |
07630cea | 641 | const char *f; |
07630cea | 642 | |
07630cea LP |
643 | assert(old_string); |
644 | assert(new_string); | |
645 | ||
9d73565a LP |
646 | if (!text) |
647 | return NULL; | |
648 | ||
07630cea LP |
649 | old_len = strlen(old_string); |
650 | new_len = strlen(new_string); | |
651 | ||
652 | l = strlen(text); | |
9d73565a | 653 | if (!GREEDY_REALLOC(ret, allocated, l+1)) |
07630cea LP |
654 | return NULL; |
655 | ||
656 | f = text; | |
9d73565a | 657 | t = ret; |
07630cea | 658 | while (*f) { |
07630cea LP |
659 | size_t d, nl; |
660 | ||
661 | if (!startswith(f, old_string)) { | |
662 | *(t++) = *(f++); | |
663 | continue; | |
664 | } | |
665 | ||
9d73565a | 666 | d = t - ret; |
07630cea | 667 | nl = l - old_len + new_len; |
9d73565a LP |
668 | |
669 | if (!GREEDY_REALLOC(ret, allocated, nl + 1)) | |
670 | return mfree(ret); | |
07630cea LP |
671 | |
672 | l = nl; | |
9d73565a | 673 | t = ret + d; |
07630cea LP |
674 | |
675 | t = stpcpy(t, new_string); | |
676 | f += old_len; | |
677 | } | |
678 | ||
679 | *t = 0; | |
9d73565a | 680 | return ret; |
07630cea LP |
681 | } |
682 | ||
b4766d5f ZJS |
683 | static void advance_offsets(ssize_t diff, size_t offsets[2], size_t shift[2], size_t size) { |
684 | if (!offsets) | |
685 | return; | |
686 | ||
687 | if ((size_t) diff < offsets[0]) | |
688 | shift[0] += size; | |
689 | if ((size_t) diff < offsets[1]) | |
690 | shift[1] += size; | |
691 | } | |
692 | ||
693 | char *strip_tab_ansi(char **ibuf, size_t *_isz, size_t highlight[2]) { | |
07630cea LP |
694 | const char *i, *begin = NULL; |
695 | enum { | |
696 | STATE_OTHER, | |
697 | STATE_ESCAPE, | |
695a944c LP |
698 | STATE_CSI, |
699 | STATE_CSO, | |
07630cea LP |
700 | } state = STATE_OTHER; |
701 | char *obuf = NULL; | |
b4766d5f | 702 | size_t osz = 0, isz, shift[2] = {}; |
07630cea LP |
703 | FILE *f; |
704 | ||
705 | assert(ibuf); | |
706 | assert(*ibuf); | |
707 | ||
695a944c LP |
708 | /* This does three things: |
709 | * | |
710 | * 1. Replaces TABs by 8 spaces | |
711 | * 2. Strips ANSI color sequences (a subset of CSI), i.e. ESC '[' … 'm' sequences | |
712 | * 3. Strips ANSI operating system sequences (CSO), i.e. ESC ']' … BEL sequences | |
713 | * | |
714 | * Everything else will be left as it is. In particular other ANSI sequences are left as they are, as are any | |
715 | * other special characters. Truncated ANSI sequences are left-as is too. This call is supposed to suppress the | |
716 | * most basic formatting noise, but nothing else. | |
717 | * | |
718 | * Why care for CSO sequences? Well, to undo what terminal_urlify() and friends generate. */ | |
07630cea LP |
719 | |
720 | isz = _isz ? *_isz : strlen(*ibuf); | |
721 | ||
722 | f = open_memstream(&obuf, &osz); | |
723 | if (!f) | |
724 | return NULL; | |
725 | ||
0d536673 LP |
726 | /* Note we turn off internal locking on f for performance reasons. It's safe to do so since we created f here |
727 | * and it doesn't leave our scope. */ | |
728 | ||
729 | (void) __fsetlocking(f, FSETLOCKING_BYCALLER); | |
db3f45e2 | 730 | |
07630cea LP |
731 | for (i = *ibuf; i < *ibuf + isz + 1; i++) { |
732 | ||
733 | switch (state) { | |
734 | ||
735 | case STATE_OTHER: | |
736 | if (i >= *ibuf + isz) /* EOT */ | |
737 | break; | |
738 | else if (*i == '\x1B') | |
739 | state = STATE_ESCAPE; | |
b4766d5f | 740 | else if (*i == '\t') { |
0d536673 | 741 | fputs(" ", f); |
b4766d5f ZJS |
742 | advance_offsets(i - *ibuf, highlight, shift, 7); |
743 | } else | |
0d536673 | 744 | fputc(*i, f); |
b4766d5f | 745 | |
07630cea LP |
746 | break; |
747 | ||
748 | case STATE_ESCAPE: | |
749 | if (i >= *ibuf + isz) { /* EOT */ | |
0d536673 | 750 | fputc('\x1B', f); |
b4766d5f | 751 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea | 752 | break; |
695a944c LP |
753 | } else if (*i == '[') { /* ANSI CSI */ |
754 | state = STATE_CSI; | |
755 | begin = i + 1; | |
756 | } else if (*i == ']') { /* ANSI CSO */ | |
757 | state = STATE_CSO; | |
07630cea LP |
758 | begin = i + 1; |
759 | } else { | |
0d536673 LP |
760 | fputc('\x1B', f); |
761 | fputc(*i, f); | |
b4766d5f | 762 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea LP |
763 | state = STATE_OTHER; |
764 | } | |
765 | ||
766 | break; | |
767 | ||
695a944c | 768 | case STATE_CSI: |
07630cea | 769 | |
695a944c LP |
770 | if (i >= *ibuf + isz || /* EOT … */ |
771 | !strchr("01234567890;m", *i)) { /* … or invalid chars in sequence */ | |
0d536673 LP |
772 | fputc('\x1B', f); |
773 | fputc('[', f); | |
b4766d5f | 774 | advance_offsets(i - *ibuf, highlight, shift, 2); |
07630cea LP |
775 | state = STATE_OTHER; |
776 | i = begin-1; | |
777 | } else if (*i == 'm') | |
778 | state = STATE_OTHER; | |
695a944c LP |
779 | |
780 | break; | |
781 | ||
782 | case STATE_CSO: | |
783 | ||
784 | if (i >= *ibuf + isz || /* EOT … */ | |
785 | (*i != '\a' && (uint8_t) *i < 32U) || (uint8_t) *i > 126U) { /* … or invalid chars in sequence */ | |
786 | fputc('\x1B', f); | |
787 | fputc(']', f); | |
788 | advance_offsets(i - *ibuf, highlight, shift, 2); | |
789 | state = STATE_OTHER; | |
790 | i = begin-1; | |
791 | } else if (*i == '\a') | |
792 | state = STATE_OTHER; | |
793 | ||
07630cea LP |
794 | break; |
795 | } | |
796 | } | |
797 | ||
c7e03d2e | 798 | if (fflush_and_check(f) < 0) { |
07630cea | 799 | fclose(f); |
6b430fdb | 800 | return mfree(obuf); |
07630cea LP |
801 | } |
802 | ||
803 | fclose(f); | |
804 | ||
805 | free(*ibuf); | |
806 | *ibuf = obuf; | |
807 | ||
808 | if (_isz) | |
809 | *_isz = osz; | |
810 | ||
b4766d5f ZJS |
811 | if (highlight) { |
812 | highlight[0] += shift[0]; | |
813 | highlight[1] += shift[1]; | |
814 | } | |
815 | ||
07630cea LP |
816 | return obuf; |
817 | } | |
818 | ||
bb8ad9ea LP |
819 | char *strextend_with_separator(char **x, const char *separator, ...) { |
820 | bool need_separator; | |
821 | size_t f, l, l_separator; | |
07630cea | 822 | char *r, *p; |
bb8ad9ea | 823 | va_list ap; |
07630cea LP |
824 | |
825 | assert(x); | |
826 | ||
7bf7ce28 | 827 | l = f = strlen_ptr(*x); |
07630cea | 828 | |
bb8ad9ea LP |
829 | need_separator = !isempty(*x); |
830 | l_separator = strlen_ptr(separator); | |
831 | ||
832 | va_start(ap, separator); | |
07630cea LP |
833 | for (;;) { |
834 | const char *t; | |
835 | size_t n; | |
836 | ||
837 | t = va_arg(ap, const char *); | |
838 | if (!t) | |
839 | break; | |
840 | ||
841 | n = strlen(t); | |
bb8ad9ea LP |
842 | |
843 | if (need_separator) | |
844 | n += l_separator; | |
845 | ||
07630cea LP |
846 | if (n > ((size_t) -1) - l) { |
847 | va_end(ap); | |
848 | return NULL; | |
849 | } | |
850 | ||
851 | l += n; | |
bb8ad9ea | 852 | need_separator = true; |
07630cea LP |
853 | } |
854 | va_end(ap); | |
855 | ||
bb8ad9ea LP |
856 | need_separator = !isempty(*x); |
857 | ||
07630cea LP |
858 | r = realloc(*x, l+1); |
859 | if (!r) | |
860 | return NULL; | |
861 | ||
862 | p = r + f; | |
863 | ||
bb8ad9ea | 864 | va_start(ap, separator); |
07630cea LP |
865 | for (;;) { |
866 | const char *t; | |
867 | ||
868 | t = va_arg(ap, const char *); | |
869 | if (!t) | |
870 | break; | |
871 | ||
bb8ad9ea LP |
872 | if (need_separator && separator) |
873 | p = stpcpy(p, separator); | |
874 | ||
07630cea | 875 | p = stpcpy(p, t); |
bb8ad9ea LP |
876 | |
877 | need_separator = true; | |
07630cea LP |
878 | } |
879 | va_end(ap); | |
880 | ||
bb8ad9ea LP |
881 | assert(p == r + l); |
882 | ||
07630cea LP |
883 | *p = 0; |
884 | *x = r; | |
885 | ||
886 | return r + l; | |
887 | } | |
888 | ||
889 | char *strrep(const char *s, unsigned n) { | |
890 | size_t l; | |
891 | char *r, *p; | |
892 | unsigned i; | |
893 | ||
894 | assert(s); | |
895 | ||
896 | l = strlen(s); | |
897 | p = r = malloc(l * n + 1); | |
898 | if (!r) | |
899 | return NULL; | |
900 | ||
901 | for (i = 0; i < n; i++) | |
902 | p = stpcpy(p, s); | |
903 | ||
904 | *p = 0; | |
905 | return r; | |
906 | } | |
907 | ||
908 | int split_pair(const char *s, const char *sep, char **l, char **r) { | |
909 | char *x, *a, *b; | |
910 | ||
911 | assert(s); | |
912 | assert(sep); | |
913 | assert(l); | |
914 | assert(r); | |
915 | ||
916 | if (isempty(sep)) | |
917 | return -EINVAL; | |
918 | ||
919 | x = strstr(s, sep); | |
920 | if (!x) | |
921 | return -EINVAL; | |
922 | ||
923 | a = strndup(s, x - s); | |
924 | if (!a) | |
925 | return -ENOMEM; | |
926 | ||
927 | b = strdup(x + strlen(sep)); | |
928 | if (!b) { | |
929 | free(a); | |
930 | return -ENOMEM; | |
931 | } | |
932 | ||
933 | *l = a; | |
934 | *r = b; | |
935 | ||
936 | return 0; | |
937 | } | |
938 | ||
939 | int free_and_strdup(char **p, const char *s) { | |
940 | char *t; | |
941 | ||
942 | assert(p); | |
943 | ||
944 | /* Replaces a string pointer with an strdup()ed new string, | |
945 | * possibly freeing the old one. */ | |
946 | ||
947 | if (streq_ptr(*p, s)) | |
948 | return 0; | |
949 | ||
950 | if (s) { | |
951 | t = strdup(s); | |
952 | if (!t) | |
953 | return -ENOMEM; | |
954 | } else | |
955 | t = NULL; | |
956 | ||
957 | free(*p); | |
958 | *p = t; | |
959 | ||
960 | return 1; | |
961 | } | |
962 | ||
4b9545f1 | 963 | #if !HAVE_EXPLICIT_BZERO |
b6b609db MB |
964 | /* |
965 | * Pointer to memset is volatile so that compiler must de-reference | |
966 | * the pointer and can't assume that it points to any function in | |
967 | * particular (such as memset, which it then might further "optimize") | |
968 | * This approach is inspired by openssl's crypto/mem_clr.c. | |
969 | */ | |
970 | typedef void *(*memset_t)(void *,int,size_t); | |
9fe4ea21 | 971 | |
b6b609db | 972 | static volatile memset_t memset_func = memset; |
9fe4ea21 | 973 | |
2d26d8e0 ZJS |
974 | void explicit_bzero(void *p, size_t l) { |
975 | memset_func(p, '\0', l); | |
9fe4ea21 | 976 | } |
2d26d8e0 | 977 | #endif |
9fe4ea21 | 978 | |
9fe4ea21 | 979 | char* string_erase(char *x) { |
07630cea | 980 | if (!x) |
9fe4ea21 | 981 | return NULL; |
07630cea LP |
982 | |
983 | /* A delicious drop of snake-oil! To be called on memory where | |
984 | * we stored passphrases or so, after we used them. */ | |
2d26d8e0 ZJS |
985 | explicit_bzero(x, strlen(x)); |
986 | return x; | |
07630cea LP |
987 | } |
988 | ||
989 | char *string_free_erase(char *s) { | |
9fe4ea21 | 990 | return mfree(string_erase(s)); |
07630cea | 991 | } |
f3e2e81d LP |
992 | |
993 | bool string_is_safe(const char *p) { | |
994 | const char *t; | |
995 | ||
996 | if (!p) | |
997 | return false; | |
998 | ||
999 | for (t = p; *t; t++) { | |
1000 | if (*t > 0 && *t < ' ') /* no control characters */ | |
1001 | return false; | |
1002 | ||
1003 | if (strchr(QUOTES "\\\x7f", *t)) | |
1004 | return false; | |
1005 | } | |
1006 | ||
1007 | return true; | |
1008 | } |