]>
Commit | Line | Data |
---|---|---|
db9ecf05 | 1 | /* SPDX-License-Identifier: LGPL-2.1-or-later */ |
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 | 22 | |
07630cea LP |
23 | char* first_word(const char *s, const char *word) { |
24 | size_t sl, wl; | |
25 | const char *p; | |
26 | ||
27 | assert(s); | |
28 | assert(word); | |
29 | ||
30 | /* Checks if the string starts with the specified word, either | |
31 | * followed by NUL or by whitespace. Returns a pointer to the | |
32 | * NUL or the first character after the whitespace. */ | |
33 | ||
34 | sl = strlen(s); | |
35 | wl = strlen(word); | |
36 | ||
37 | if (sl < wl) | |
38 | return NULL; | |
39 | ||
40 | if (wl == 0) | |
41 | return (char*) s; | |
42 | ||
43 | if (memcmp(s, word, wl) != 0) | |
44 | return NULL; | |
45 | ||
46 | p = s + wl; | |
47 | if (*p == 0) | |
48 | return (char*) p; | |
49 | ||
50 | if (!strchr(WHITESPACE, *p)) | |
51 | return NULL; | |
52 | ||
53 | p += strspn(p, WHITESPACE); | |
54 | return (char*) p; | |
55 | } | |
56 | ||
07630cea LP |
57 | char *strnappend(const char *s, const char *suffix, size_t b) { |
58 | size_t a; | |
59 | char *r; | |
60 | ||
61 | if (!s && !suffix) | |
62 | return strdup(""); | |
63 | ||
64 | if (!s) | |
65 | return strndup(suffix, b); | |
66 | ||
67 | if (!suffix) | |
68 | return strdup(s); | |
69 | ||
70 | assert(s); | |
71 | assert(suffix); | |
72 | ||
73 | a = strlen(s); | |
fd4e991d | 74 | if (b > SIZE_MAX - a) |
07630cea LP |
75 | return NULL; |
76 | ||
77 | r = new(char, a+b+1); | |
78 | if (!r) | |
79 | return NULL; | |
80 | ||
81 | memcpy(r, s, a); | |
82 | memcpy(r+a, suffix, b); | |
83 | r[a+b] = 0; | |
84 | ||
85 | return r; | |
86 | } | |
87 | ||
605405c6 | 88 | char *strjoin_real(const char *x, ...) { |
07630cea | 89 | va_list ap; |
6ced0770 | 90 | size_t l = 1; |
07630cea LP |
91 | char *r, *p; |
92 | ||
93 | va_start(ap, x); | |
020003f2 RV |
94 | for (const char *t = x; t; t = va_arg(ap, const char *)) { |
95 | size_t n; | |
07630cea | 96 | |
020003f2 RV |
97 | n = strlen(t); |
98 | if (n > SIZE_MAX - l) { | |
99 | va_end(ap); | |
100 | return NULL; | |
07630cea | 101 | } |
020003f2 RV |
102 | l += n; |
103 | } | |
07630cea LP |
104 | va_end(ap); |
105 | ||
6ced0770 | 106 | p = r = new(char, l); |
07630cea LP |
107 | if (!r) |
108 | return NULL; | |
109 | ||
020003f2 RV |
110 | va_start(ap, x); |
111 | for (const char *t = x; t; t = va_arg(ap, const char *)) | |
112 | p = stpcpy(p, t); | |
113 | va_end(ap); | |
07630cea | 114 | |
020003f2 | 115 | *p = 0; |
07630cea LP |
116 | |
117 | return r; | |
118 | } | |
119 | ||
120 | char *strstrip(char *s) { | |
7546145e LP |
121 | if (!s) |
122 | return NULL; | |
123 | ||
0a6ffc5c | 124 | /* Drops trailing whitespace. Modifies the string in place. Returns pointer to first non-space character */ |
07630cea | 125 | |
0a6ffc5c | 126 | return delete_trailing_chars(skip_leading_chars(s, WHITESPACE), WHITESPACE); |
07630cea LP |
127 | } |
128 | ||
129 | char *delete_chars(char *s, const char *bad) { | |
130 | char *f, *t; | |
131 | ||
7546145e LP |
132 | /* Drops all specified bad characters, regardless where in the string */ |
133 | ||
134 | if (!s) | |
135 | return NULL; | |
136 | ||
137 | if (!bad) | |
138 | bad = WHITESPACE; | |
07630cea LP |
139 | |
140 | for (f = s, t = s; *f; f++) { | |
141 | if (strchr(bad, *f)) | |
142 | continue; | |
143 | ||
144 | *(t++) = *f; | |
145 | } | |
146 | ||
147 | *t = 0; | |
148 | ||
149 | return s; | |
150 | } | |
151 | ||
7546145e LP |
152 | char *delete_trailing_chars(char *s, const char *bad) { |
153 | char *p, *c = s; | |
154 | ||
155 | /* Drops all specified bad characters, at the end of the string */ | |
156 | ||
157 | if (!s) | |
158 | return NULL; | |
159 | ||
160 | if (!bad) | |
161 | bad = WHITESPACE; | |
162 | ||
163 | for (p = s; *p; p++) | |
164 | if (!strchr(bad, *p)) | |
165 | c = p + 1; | |
166 | ||
167 | *c = 0; | |
168 | ||
169 | return s; | |
170 | } | |
171 | ||
07630cea LP |
172 | char *truncate_nl(char *s) { |
173 | assert(s); | |
174 | ||
175 | s[strcspn(s, NEWLINE)] = 0; | |
176 | return s; | |
177 | } | |
178 | ||
b577e3d5 LP |
179 | char ascii_tolower(char x) { |
180 | ||
181 | if (x >= 'A' && x <= 'Z') | |
182 | return x - 'A' + 'a'; | |
183 | ||
184 | return x; | |
185 | } | |
186 | ||
846b8fc3 LP |
187 | char ascii_toupper(char x) { |
188 | ||
189 | if (x >= 'a' && x <= 'z') | |
190 | return x - 'a' + 'A'; | |
191 | ||
192 | return x; | |
193 | } | |
194 | ||
07630cea LP |
195 | char *ascii_strlower(char *t) { |
196 | char *p; | |
197 | ||
198 | assert(t); | |
199 | ||
200 | for (p = t; *p; p++) | |
b577e3d5 LP |
201 | *p = ascii_tolower(*p); |
202 | ||
203 | return t; | |
204 | } | |
205 | ||
846b8fc3 LP |
206 | char *ascii_strupper(char *t) { |
207 | char *p; | |
208 | ||
209 | assert(t); | |
210 | ||
211 | for (p = t; *p; p++) | |
212 | *p = ascii_toupper(*p); | |
213 | ||
214 | return t; | |
215 | } | |
216 | ||
b577e3d5 LP |
217 | char *ascii_strlower_n(char *t, size_t n) { |
218 | size_t i; | |
219 | ||
220 | if (n <= 0) | |
221 | return t; | |
222 | ||
223 | for (i = 0; i < n; i++) | |
224 | t[i] = ascii_tolower(t[i]); | |
07630cea LP |
225 | |
226 | return t; | |
227 | } | |
522d85ae LP |
228 | |
229 | int ascii_strcasecmp_n(const char *a, const char *b, size_t n) { | |
230 | ||
231 | for (; n > 0; a++, b++, n--) { | |
232 | int x, y; | |
233 | ||
234 | x = (int) (uint8_t) ascii_tolower(*a); | |
235 | y = (int) (uint8_t) ascii_tolower(*b); | |
236 | ||
237 | if (x != y) | |
238 | return x - y; | |
239 | } | |
240 | ||
241 | return 0; | |
242 | } | |
c1749834 LP |
243 | |
244 | int ascii_strcasecmp_nn(const char *a, size_t n, const char *b, size_t m) { | |
245 | int r; | |
246 | ||
247 | r = ascii_strcasecmp_n(a, b, MIN(n, m)); | |
248 | if (r != 0) | |
249 | return r; | |
250 | ||
6dd91b36 | 251 | return CMP(n, m); |
c1749834 | 252 | } |
07630cea LP |
253 | |
254 | bool chars_intersect(const char *a, const char *b) { | |
255 | const char *p; | |
256 | ||
257 | /* Returns true if any of the chars in a are in b. */ | |
258 | for (p = a; *p; p++) | |
259 | if (strchr(b, *p)) | |
260 | return true; | |
261 | ||
262 | return false; | |
263 | } | |
264 | ||
265 | bool string_has_cc(const char *p, const char *ok) { | |
266 | const char *t; | |
267 | ||
268 | assert(p); | |
269 | ||
270 | /* | |
271 | * Check if a string contains control characters. If 'ok' is | |
272 | * non-NULL it may be a string containing additional CCs to be | |
273 | * considered OK. | |
274 | */ | |
275 | ||
276 | for (t = p; *t; t++) { | |
277 | if (ok && strchr(ok, *t)) | |
278 | continue; | |
279 | ||
280 | if (*t > 0 && *t < ' ') | |
281 | return true; | |
282 | ||
283 | if (*t == 127) | |
284 | return true; | |
285 | } | |
286 | ||
287 | return false; | |
288 | } | |
289 | ||
8409f688 ZJS |
290 | static int write_ellipsis(char *buf, bool unicode) { |
291 | if (unicode || is_locale_utf8()) { | |
292 | buf[0] = 0xe2; /* tri-dot ellipsis: … */ | |
293 | buf[1] = 0x80; | |
294 | buf[2] = 0xa6; | |
295 | } else { | |
296 | buf[0] = '.'; | |
297 | buf[1] = '.'; | |
298 | buf[2] = '.'; | |
299 | } | |
300 | ||
301 | return 3; | |
302 | } | |
303 | ||
07630cea | 304 | static char *ascii_ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { |
9924aef6 ZJS |
305 | size_t x, need_space, suffix_len; |
306 | char *t; | |
07630cea LP |
307 | |
308 | assert(s); | |
309 | assert(percent <= 100); | |
f5fbe71d | 310 | assert(new_length != SIZE_MAX); |
07630cea | 311 | |
c30a49b2 | 312 | if (old_length <= new_length) |
07630cea LP |
313 | return strndup(s, old_length); |
314 | ||
c30a49b2 LP |
315 | /* Special case short ellipsations */ |
316 | switch (new_length) { | |
317 | ||
318 | case 0: | |
319 | return strdup(""); | |
320 | ||
321 | case 1: | |
322 | if (is_locale_utf8()) | |
323 | return strdup("…"); | |
324 | else | |
325 | return strdup("."); | |
326 | ||
327 | case 2: | |
328 | if (!is_locale_utf8()) | |
329 | return strdup(".."); | |
330 | ||
331 | break; | |
332 | ||
333 | default: | |
334 | break; | |
335 | } | |
336 | ||
337 | /* Calculate how much space the ellipsis will take up. If we are in UTF-8 mode we only need space for one | |
338 | * character ("…"), otherwise for three characters ("..."). Note that in both cases we need 3 bytes of storage, | |
339 | * either for the UTF-8 encoded character or for three ASCII characters. */ | |
340 | need_space = is_locale_utf8() ? 1 : 3; | |
341 | ||
9924aef6 ZJS |
342 | t = new(char, new_length+3); |
343 | if (!t) | |
07630cea LP |
344 | return NULL; |
345 | ||
c30a49b2 | 346 | assert(new_length >= need_space); |
07630cea | 347 | |
c30a49b2 LP |
348 | x = ((new_length - need_space) * percent + 50) / 100; |
349 | assert(x <= new_length - need_space); | |
07630cea | 350 | |
9924aef6 ZJS |
351 | memcpy(t, s, x); |
352 | write_ellipsis(t + x, false); | |
353 | suffix_len = new_length - x - need_space; | |
354 | memcpy(t + x + 3, s + old_length - suffix_len, suffix_len); | |
355 | *(t + x + 3 + suffix_len) = '\0'; | |
07630cea | 356 | |
9924aef6 | 357 | return t; |
07630cea LP |
358 | } |
359 | ||
360 | char *ellipsize_mem(const char *s, size_t old_length, size_t new_length, unsigned percent) { | |
c30a49b2 | 361 | size_t x, k, len, len2; |
07630cea | 362 | const char *i, *j; |
c30a49b2 | 363 | char *e; |
c932fb71 | 364 | int r; |
07630cea | 365 | |
c30a49b2 LP |
366 | /* Note that 'old_length' refers to bytes in the string, while 'new_length' refers to character cells taken up |
367 | * on screen. This distinction doesn't matter for ASCII strings, but it does matter for non-ASCII UTF-8 | |
368 | * strings. | |
369 | * | |
370 | * Ellipsation is done in a locale-dependent way: | |
371 | * 1. If the string passed in is fully ASCII and the current locale is not UTF-8, three dots are used ("...") | |
372 | * 2. Otherwise, a unicode ellipsis is used ("…") | |
373 | * | |
374 | * In other words: you'll get a unicode ellipsis as soon as either the string contains non-ASCII characters or | |
375 | * the current locale is UTF-8. | |
376 | */ | |
377 | ||
07630cea LP |
378 | assert(s); |
379 | assert(percent <= 100); | |
ddbc9319 | 380 | |
f5fbe71d | 381 | if (new_length == SIZE_MAX) |
ddbc9319 LP |
382 | return strndup(s, old_length); |
383 | ||
c30a49b2 LP |
384 | if (new_length == 0) |
385 | return strdup(""); | |
07630cea | 386 | |
c30a49b2 | 387 | /* If no multibyte characters use ascii_ellipsize_mem for speed */ |
21e4e3e0 | 388 | if (ascii_is_valid_n(s, old_length)) |
07630cea LP |
389 | return ascii_ellipsize_mem(s, old_length, new_length, percent); |
390 | ||
c30a49b2 LP |
391 | x = ((new_length - 1) * percent) / 100; |
392 | assert(x <= new_length - 1); | |
07630cea LP |
393 | |
394 | k = 0; | |
9924aef6 | 395 | for (i = s; i < s + old_length; i = utf8_next_char(i)) { |
c932fb71 | 396 | char32_t c; |
9924aef6 | 397 | int w; |
07630cea | 398 | |
c932fb71 SL |
399 | r = utf8_encoded_to_unichar(i, &c); |
400 | if (r < 0) | |
07630cea | 401 | return NULL; |
07630cea | 402 | |
9924aef6 ZJS |
403 | w = unichar_iswide(c) ? 2 : 1; |
404 | if (k + w <= x) | |
405 | k += w; | |
406 | else | |
407 | break; | |
408 | } | |
07630cea | 409 | |
9924aef6 | 410 | for (j = s + old_length; j > i; ) { |
c932fb71 | 411 | char32_t c; |
9924aef6 ZJS |
412 | int w; |
413 | const char *jj; | |
07630cea | 414 | |
9924aef6 ZJS |
415 | jj = utf8_prev_char(j); |
416 | r = utf8_encoded_to_unichar(jj, &c); | |
c932fb71 | 417 | if (r < 0) |
07630cea | 418 | return NULL; |
9924aef6 ZJS |
419 | |
420 | w = unichar_iswide(c) ? 2 : 1; | |
421 | if (k + w <= new_length) { | |
422 | k += w; | |
423 | j = jj; | |
424 | } else | |
425 | break; | |
07630cea LP |
426 | } |
427 | assert(i <= j); | |
428 | ||
429 | /* we don't actually need to ellipsize */ | |
430 | if (i == j) | |
9924aef6 | 431 | return memdup_suffix0(s, old_length); |
07630cea | 432 | |
9924aef6 ZJS |
433 | /* make space for ellipsis, if possible */ |
434 | if (j < s + old_length) | |
435 | j = utf8_next_char(j); | |
436 | else if (i > s) | |
437 | i = utf8_prev_char(i); | |
07630cea LP |
438 | |
439 | len = i - s; | |
440 | len2 = s + old_length - j; | |
441 | e = new(char, len + 3 + len2 + 1); | |
442 | if (!e) | |
443 | return NULL; | |
444 | ||
445 | /* | |
446 | printf("old_length=%zu new_length=%zu x=%zu len=%u len2=%u k=%u\n", | |
447 | old_length, new_length, x, len, len2, k); | |
448 | */ | |
449 | ||
450 | memcpy(e, s, len); | |
8409f688 | 451 | write_ellipsis(e + len, true); |
9924aef6 ZJS |
452 | memcpy(e + len + 3, j, len2); |
453 | *(e + len + 3 + len2) = '\0'; | |
07630cea LP |
454 | |
455 | return e; | |
456 | } | |
457 | ||
8409f688 ZJS |
458 | char *cellescape(char *buf, size_t len, const char *s) { |
459 | /* Escape and ellipsize s into buffer buf of size len. Only non-control ASCII | |
460 | * characters are copied as they are, everything else is escaped. The result | |
461 | * is different then if escaping and ellipsization was performed in two | |
462 | * separate steps, because each sequence is either stored in full or skipped. | |
463 | * | |
464 | * This function should be used for logging about strings which expected to | |
465 | * be plain ASCII in a safe way. | |
466 | * | |
467 | * An ellipsis will be used if s is too long. It was always placed at the | |
468 | * very end. | |
469 | */ | |
470 | ||
61f6e276 LP |
471 | size_t i = 0, last_char_width[4] = {}, k = 0, j; |
472 | ||
473 | assert(len > 0); /* at least a terminating NUL */ | |
8409f688 | 474 | |
61f6e276 LP |
475 | for (;;) { |
476 | char four[4]; | |
477 | int w; | |
8409f688 | 478 | |
61f6e276 | 479 | if (*s == 0) /* terminating NUL detected? then we are done! */ |
8409f688 | 480 | goto done; |
61f6e276 LP |
481 | |
482 | w = cescape_char(*s, four); | |
483 | if (i + w + 1 > len) /* This character doesn't fit into the buffer anymore? In that case let's | |
484 | * ellipsize at the previous location */ | |
485 | break; | |
486 | ||
487 | /* OK, there was space, let's add this escaped character to the buffer */ | |
488 | memcpy(buf + i, four, w); | |
489 | i += w; | |
490 | ||
491 | /* And remember its width in the ring buffer */ | |
492 | last_char_width[k] = w; | |
493 | k = (k + 1) % 4; | |
494 | ||
495 | s++; | |
8409f688 ZJS |
496 | } |
497 | ||
61f6e276 LP |
498 | /* Ellipsation is necessary. This means we might need to truncate the string again to make space for 4 |
499 | * characters ideally, but the buffer is shorter than that in the first place take what we can get */ | |
500 | for (j = 0; j < ELEMENTSOF(last_char_width); j++) { | |
501 | ||
502 | if (i + 4 <= len) /* nice, we reached our space goal */ | |
503 | break; | |
504 | ||
505 | k = k == 0 ? 3 : k - 1; | |
506 | if (last_char_width[k] == 0) /* bummer, we reached the beginning of the strings */ | |
507 | break; | |
508 | ||
509 | assert(i >= last_char_width[k]); | |
510 | i -= last_char_width[k]; | |
8409f688 ZJS |
511 | } |
512 | ||
61f6e276 LP |
513 | if (i + 4 <= len) /* yay, enough space */ |
514 | i += write_ellipsis(buf + i, false); | |
515 | else if (i + 3 <= len) { /* only space for ".." */ | |
516 | buf[i++] = '.'; | |
517 | buf[i++] = '.'; | |
518 | } else if (i + 2 <= len) /* only space for a single "." */ | |
519 | buf[i++] = '.'; | |
520 | else | |
521 | assert(i + 1 <= len); | |
522 | ||
8409f688 ZJS |
523 | done: |
524 | buf[i] = '\0'; | |
525 | return buf; | |
526 | } | |
527 | ||
07630cea LP |
528 | char* strshorten(char *s, size_t l) { |
529 | assert(s); | |
530 | ||
47b33c7d | 531 | if (strnlen(s, l+1) > l) |
07630cea LP |
532 | s[l] = 0; |
533 | ||
534 | return s; | |
535 | } | |
536 | ||
537 | char *strreplace(const char *text, const char *old_string, const char *new_string) { | |
9d73565a LP |
538 | size_t l, old_len, new_len, allocated = 0; |
539 | char *t, *ret = NULL; | |
07630cea | 540 | const char *f; |
07630cea | 541 | |
07630cea LP |
542 | assert(old_string); |
543 | assert(new_string); | |
544 | ||
9d73565a LP |
545 | if (!text) |
546 | return NULL; | |
547 | ||
07630cea LP |
548 | old_len = strlen(old_string); |
549 | new_len = strlen(new_string); | |
550 | ||
551 | l = strlen(text); | |
9d73565a | 552 | if (!GREEDY_REALLOC(ret, allocated, l+1)) |
07630cea LP |
553 | return NULL; |
554 | ||
555 | f = text; | |
9d73565a | 556 | t = ret; |
07630cea | 557 | while (*f) { |
07630cea LP |
558 | size_t d, nl; |
559 | ||
560 | if (!startswith(f, old_string)) { | |
561 | *(t++) = *(f++); | |
562 | continue; | |
563 | } | |
564 | ||
9d73565a | 565 | d = t - ret; |
07630cea | 566 | nl = l - old_len + new_len; |
9d73565a LP |
567 | |
568 | if (!GREEDY_REALLOC(ret, allocated, nl + 1)) | |
569 | return mfree(ret); | |
07630cea LP |
570 | |
571 | l = nl; | |
9d73565a | 572 | t = ret + d; |
07630cea LP |
573 | |
574 | t = stpcpy(t, new_string); | |
575 | f += old_len; | |
576 | } | |
577 | ||
578 | *t = 0; | |
9d73565a | 579 | return ret; |
07630cea LP |
580 | } |
581 | ||
6fb05690 LP |
582 | static void advance_offsets( |
583 | ssize_t diff, | |
584 | size_t offsets[2], /* note: we can't use [static 2] here, since this may be NULL */ | |
585 | size_t shift[static 2], | |
586 | size_t size) { | |
587 | ||
b4766d5f ZJS |
588 | if (!offsets) |
589 | return; | |
590 | ||
6fb05690 LP |
591 | assert(shift); |
592 | ||
b4766d5f ZJS |
593 | if ((size_t) diff < offsets[0]) |
594 | shift[0] += size; | |
595 | if ((size_t) diff < offsets[1]) | |
596 | shift[1] += size; | |
597 | } | |
598 | ||
599 | char *strip_tab_ansi(char **ibuf, size_t *_isz, size_t highlight[2]) { | |
62a3fc6d | 600 | const char *begin = NULL; |
07630cea LP |
601 | enum { |
602 | STATE_OTHER, | |
603 | STATE_ESCAPE, | |
695a944c LP |
604 | STATE_CSI, |
605 | STATE_CSO, | |
07630cea LP |
606 | } state = STATE_OTHER; |
607 | char *obuf = NULL; | |
62a3fc6d | 608 | size_t osz = 0, isz, shift[2] = {}, n_carriage_returns = 0; |
07630cea LP |
609 | FILE *f; |
610 | ||
611 | assert(ibuf); | |
612 | assert(*ibuf); | |
613 | ||
695a944c LP |
614 | /* This does three things: |
615 | * | |
616 | * 1. Replaces TABs by 8 spaces | |
617 | * 2. Strips ANSI color sequences (a subset of CSI), i.e. ESC '[' … 'm' sequences | |
618 | * 3. Strips ANSI operating system sequences (CSO), i.e. ESC ']' … BEL sequences | |
62a3fc6d ZJS |
619 | * 4. Strip trailing \r characters (since they would "move the cursor", but have no |
620 | * other effect). | |
695a944c | 621 | * |
2fe21124 ZJS |
622 | * Everything else will be left as it is. In particular other ANSI sequences are left as they are, as |
623 | * are any other special characters. Truncated ANSI sequences are left-as is too. This call is | |
624 | * supposed to suppress the most basic formatting noise, but nothing else. | |
695a944c LP |
625 | * |
626 | * Why care for CSO sequences? Well, to undo what terminal_urlify() and friends generate. */ | |
07630cea LP |
627 | |
628 | isz = _isz ? *_isz : strlen(*ibuf); | |
629 | ||
2fe21124 ZJS |
630 | /* Note we turn off internal locking on f for performance reasons. It's safe to do so since we |
631 | * created f here and it doesn't leave our scope. */ | |
632 | f = open_memstream_unlocked(&obuf, &osz); | |
07630cea LP |
633 | if (!f) |
634 | return NULL; | |
635 | ||
62a3fc6d | 636 | for (const char *i = *ibuf; i < *ibuf + isz + 1; i++) { |
07630cea LP |
637 | |
638 | switch (state) { | |
639 | ||
640 | case STATE_OTHER: | |
641 | if (i >= *ibuf + isz) /* EOT */ | |
642 | break; | |
62a3fc6d ZJS |
643 | |
644 | if (*i == '\r') { | |
645 | n_carriage_returns++; | |
646 | break; | |
647 | } else if (*i == '\n') | |
648 | /* Ignore carriage returns before new line */ | |
649 | n_carriage_returns = 0; | |
650 | for (; n_carriage_returns > 0; n_carriage_returns--) | |
651 | fputc('\r', f); | |
652 | ||
653 | if (*i == '\x1B') | |
07630cea | 654 | state = STATE_ESCAPE; |
b4766d5f | 655 | else if (*i == '\t') { |
0d536673 | 656 | fputs(" ", f); |
b4766d5f ZJS |
657 | advance_offsets(i - *ibuf, highlight, shift, 7); |
658 | } else | |
0d536673 | 659 | fputc(*i, f); |
b4766d5f | 660 | |
07630cea LP |
661 | break; |
662 | ||
663 | case STATE_ESCAPE: | |
62a3fc6d ZJS |
664 | assert(n_carriage_returns == 0); |
665 | ||
07630cea | 666 | if (i >= *ibuf + isz) { /* EOT */ |
0d536673 | 667 | fputc('\x1B', f); |
b4766d5f | 668 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea | 669 | break; |
695a944c LP |
670 | } else if (*i == '[') { /* ANSI CSI */ |
671 | state = STATE_CSI; | |
672 | begin = i + 1; | |
673 | } else if (*i == ']') { /* ANSI CSO */ | |
674 | state = STATE_CSO; | |
07630cea LP |
675 | begin = i + 1; |
676 | } else { | |
0d536673 LP |
677 | fputc('\x1B', f); |
678 | fputc(*i, f); | |
b4766d5f | 679 | advance_offsets(i - *ibuf, highlight, shift, 1); |
07630cea LP |
680 | state = STATE_OTHER; |
681 | } | |
682 | ||
683 | break; | |
684 | ||
695a944c | 685 | case STATE_CSI: |
62a3fc6d | 686 | assert(n_carriage_returns == 0); |
07630cea | 687 | |
695a944c LP |
688 | if (i >= *ibuf + isz || /* EOT … */ |
689 | !strchr("01234567890;m", *i)) { /* … or invalid chars in sequence */ | |
0d536673 LP |
690 | fputc('\x1B', f); |
691 | fputc('[', f); | |
b4766d5f | 692 | advance_offsets(i - *ibuf, highlight, shift, 2); |
07630cea LP |
693 | state = STATE_OTHER; |
694 | i = begin-1; | |
695 | } else if (*i == 'm') | |
696 | state = STATE_OTHER; | |
695a944c LP |
697 | |
698 | break; | |
699 | ||
700 | case STATE_CSO: | |
62a3fc6d | 701 | assert(n_carriage_returns == 0); |
695a944c LP |
702 | |
703 | if (i >= *ibuf + isz || /* EOT … */ | |
704 | (*i != '\a' && (uint8_t) *i < 32U) || (uint8_t) *i > 126U) { /* … or invalid chars in sequence */ | |
705 | fputc('\x1B', f); | |
706 | fputc(']', f); | |
707 | advance_offsets(i - *ibuf, highlight, shift, 2); | |
708 | state = STATE_OTHER; | |
709 | i = begin-1; | |
710 | } else if (*i == '\a') | |
711 | state = STATE_OTHER; | |
712 | ||
07630cea LP |
713 | break; |
714 | } | |
715 | } | |
716 | ||
c7e03d2e | 717 | if (fflush_and_check(f) < 0) { |
07630cea | 718 | fclose(f); |
6b430fdb | 719 | return mfree(obuf); |
07630cea | 720 | } |
07630cea LP |
721 | fclose(f); |
722 | ||
6fb05690 | 723 | free_and_replace(*ibuf, obuf); |
07630cea LP |
724 | |
725 | if (_isz) | |
726 | *_isz = osz; | |
727 | ||
b4766d5f ZJS |
728 | if (highlight) { |
729 | highlight[0] += shift[0]; | |
730 | highlight[1] += shift[1]; | |
731 | } | |
732 | ||
6fb05690 | 733 | return *ibuf; |
07630cea LP |
734 | } |
735 | ||
c2bc710b | 736 | char *strextend_with_separator_internal(char **x, const char *separator, ...) { |
bb8ad9ea | 737 | size_t f, l, l_separator; |
c2bc710b LP |
738 | bool need_separator; |
739 | char *nr, *p; | |
bb8ad9ea | 740 | va_list ap; |
07630cea LP |
741 | |
742 | assert(x); | |
743 | ||
7bf7ce28 | 744 | l = f = strlen_ptr(*x); |
07630cea | 745 | |
bb8ad9ea LP |
746 | need_separator = !isempty(*x); |
747 | l_separator = strlen_ptr(separator); | |
748 | ||
749 | va_start(ap, separator); | |
07630cea LP |
750 | for (;;) { |
751 | const char *t; | |
752 | size_t n; | |
753 | ||
754 | t = va_arg(ap, const char *); | |
755 | if (!t) | |
756 | break; | |
757 | ||
758 | n = strlen(t); | |
bb8ad9ea LP |
759 | |
760 | if (need_separator) | |
761 | n += l_separator; | |
762 | ||
c2bc710b | 763 | if (n >= SIZE_MAX - l) { |
07630cea LP |
764 | va_end(ap); |
765 | return NULL; | |
766 | } | |
767 | ||
768 | l += n; | |
bb8ad9ea | 769 | need_separator = true; |
07630cea LP |
770 | } |
771 | va_end(ap); | |
772 | ||
bb8ad9ea LP |
773 | need_separator = !isempty(*x); |
774 | ||
2a4e1fd0 | 775 | nr = realloc(*x, GREEDY_ALLOC_ROUND_UP(l+1)); |
c2bc710b | 776 | if (!nr) |
07630cea LP |
777 | return NULL; |
778 | ||
c2bc710b LP |
779 | *x = nr; |
780 | p = nr + f; | |
07630cea | 781 | |
bb8ad9ea | 782 | va_start(ap, separator); |
07630cea LP |
783 | for (;;) { |
784 | const char *t; | |
785 | ||
786 | t = va_arg(ap, const char *); | |
787 | if (!t) | |
788 | break; | |
789 | ||
bb8ad9ea LP |
790 | if (need_separator && separator) |
791 | p = stpcpy(p, separator); | |
792 | ||
07630cea | 793 | p = stpcpy(p, t); |
bb8ad9ea LP |
794 | |
795 | need_separator = true; | |
07630cea LP |
796 | } |
797 | va_end(ap); | |
798 | ||
c2bc710b | 799 | assert(p == nr + l); |
bb8ad9ea | 800 | |
07630cea | 801 | *p = 0; |
07630cea | 802 | |
c2bc710b | 803 | return p; |
07630cea LP |
804 | } |
805 | ||
806 | char *strrep(const char *s, unsigned n) { | |
07630cea | 807 | char *r, *p; |
fe96c0f8 | 808 | size_t l; |
07630cea LP |
809 | |
810 | assert(s); | |
811 | ||
812 | l = strlen(s); | |
813 | p = r = malloc(l * n + 1); | |
814 | if (!r) | |
815 | return NULL; | |
816 | ||
fe96c0f8 | 817 | for (unsigned i = 0; i < n; i++) |
07630cea LP |
818 | p = stpcpy(p, s); |
819 | ||
820 | *p = 0; | |
821 | return r; | |
822 | } | |
823 | ||
824 | int split_pair(const char *s, const char *sep, char **l, char **r) { | |
825 | char *x, *a, *b; | |
826 | ||
827 | assert(s); | |
828 | assert(sep); | |
829 | assert(l); | |
830 | assert(r); | |
831 | ||
832 | if (isempty(sep)) | |
833 | return -EINVAL; | |
834 | ||
835 | x = strstr(s, sep); | |
836 | if (!x) | |
837 | return -EINVAL; | |
838 | ||
839 | a = strndup(s, x - s); | |
840 | if (!a) | |
841 | return -ENOMEM; | |
842 | ||
843 | b = strdup(x + strlen(sep)); | |
844 | if (!b) { | |
845 | free(a); | |
846 | return -ENOMEM; | |
847 | } | |
848 | ||
849 | *l = a; | |
850 | *r = b; | |
851 | ||
852 | return 0; | |
853 | } | |
854 | ||
855 | int free_and_strdup(char **p, const char *s) { | |
856 | char *t; | |
857 | ||
858 | assert(p); | |
859 | ||
7f546026 | 860 | /* Replaces a string pointer with a strdup()ed new string, |
07630cea LP |
861 | * possibly freeing the old one. */ |
862 | ||
863 | if (streq_ptr(*p, s)) | |
864 | return 0; | |
865 | ||
866 | if (s) { | |
867 | t = strdup(s); | |
868 | if (!t) | |
869 | return -ENOMEM; | |
870 | } else | |
871 | t = NULL; | |
872 | ||
873 | free(*p); | |
874 | *p = t; | |
875 | ||
876 | return 1; | |
877 | } | |
878 | ||
7f546026 ZJS |
879 | int free_and_strndup(char **p, const char *s, size_t l) { |
880 | char *t; | |
881 | ||
882 | assert(p); | |
883 | assert(s || l == 0); | |
884 | ||
885 | /* Replaces a string pointer with a strndup()ed new string, | |
886 | * freeing the old one. */ | |
887 | ||
888 | if (!*p && !s) | |
889 | return 0; | |
890 | ||
891 | if (*p && s && strneq(*p, s, l) && (l > strlen(*p) || (*p)[l] == '\0')) | |
892 | return 0; | |
893 | ||
894 | if (s) { | |
895 | t = strndup(s, l); | |
896 | if (!t) | |
897 | return -ENOMEM; | |
898 | } else | |
899 | t = NULL; | |
900 | ||
901 | free_and_replace(*p, t); | |
902 | return 1; | |
903 | } | |
904 | ||
f3e2e81d LP |
905 | bool string_is_safe(const char *p) { |
906 | const char *t; | |
907 | ||
908 | if (!p) | |
909 | return false; | |
910 | ||
839d1b20 LP |
911 | /* Checks if the specified string contains no quotes or control characters */ |
912 | ||
f3e2e81d LP |
913 | for (t = p; *t; t++) { |
914 | if (*t > 0 && *t < ' ') /* no control characters */ | |
915 | return false; | |
916 | ||
917 | if (strchr(QUOTES "\\\x7f", *t)) | |
918 | return false; | |
919 | } | |
920 | ||
921 | return true; | |
922 | } | |
53caaffd LP |
923 | |
924 | char* string_erase(char *x) { | |
925 | if (!x) | |
926 | return NULL; | |
927 | ||
928 | /* A delicious drop of snake-oil! To be called on memory where we stored passphrases or so, after we | |
929 | * used them. */ | |
930 | explicit_bzero_safe(x, strlen(x)); | |
931 | return x; | |
932 | } | |
8dd6491e LP |
933 | |
934 | int string_truncate_lines(const char *s, size_t n_lines, char **ret) { | |
935 | const char *p = s, *e = s; | |
936 | bool truncation_applied = false; | |
937 | char *copy; | |
938 | size_t n = 0; | |
939 | ||
940 | assert(s); | |
941 | ||
942 | /* Truncate after the specified number of lines. Returns > 0 if a truncation was applied or == 0 if | |
943 | * there were fewer lines in the string anyway. Trailing newlines on input are ignored, and not | |
944 | * generated either. */ | |
945 | ||
946 | for (;;) { | |
947 | size_t k; | |
948 | ||
949 | k = strcspn(p, "\n"); | |
950 | ||
951 | if (p[k] == 0) { | |
952 | if (k == 0) /* final empty line */ | |
953 | break; | |
954 | ||
955 | if (n >= n_lines) /* above threshold */ | |
956 | break; | |
957 | ||
958 | e = p + k; /* last line to include */ | |
959 | break; | |
960 | } | |
961 | ||
962 | assert(p[k] == '\n'); | |
963 | ||
964 | if (n >= n_lines) | |
965 | break; | |
966 | ||
967 | if (k > 0) | |
968 | e = p + k; | |
969 | ||
970 | p += k + 1; | |
971 | n++; | |
972 | } | |
973 | ||
974 | /* e points after the last character we want to keep */ | |
975 | if (isempty(e)) | |
976 | copy = strdup(s); | |
977 | else { | |
978 | if (!in_charset(e, "\n")) /* We only consider things truncated if we remove something that | |
979 | * isn't a new-line or a series of them */ | |
980 | truncation_applied = true; | |
981 | ||
982 | copy = strndup(s, e - s); | |
983 | } | |
984 | if (!copy) | |
985 | return -ENOMEM; | |
986 | ||
987 | *ret = copy; | |
988 | return truncation_applied; | |
989 | } | |
f6857fa6 LP |
990 | |
991 | int string_extract_line(const char *s, size_t i, char **ret) { | |
992 | const char *p = s; | |
993 | size_t c = 0; | |
994 | ||
995 | /* Extract the i'nth line from the specified string. Returns > 0 if there are more lines after that, | |
996 | * and == 0 if we are looking at the last line or already beyond the last line. As special | |
997 | * optimization, if the first line is requested and the string only consists of one line we return | |
998 | * NULL, indicating the input string should be used as is, and avoid a memory allocation for a very | |
999 | * common case. */ | |
1000 | ||
1001 | for (;;) { | |
1002 | const char *q; | |
1003 | ||
1004 | q = strchr(p, '\n'); | |
1005 | if (i == c) { | |
1006 | /* The line we are looking for! */ | |
1007 | ||
1008 | if (q) { | |
1009 | char *m; | |
1010 | ||
1011 | m = strndup(p, q - p); | |
1012 | if (!m) | |
1013 | return -ENOMEM; | |
1014 | ||
1015 | *ret = m; | |
1016 | return !isempty(q + 1); /* more coming? */ | |
1017 | } else { | |
1018 | if (p == s) | |
1019 | *ret = NULL; /* Just use the input string */ | |
1020 | else { | |
1021 | char *m; | |
1022 | ||
1023 | m = strdup(p); | |
1024 | if (!m) | |
1025 | return -ENOMEM; | |
1026 | ||
1027 | *ret = m; | |
1028 | } | |
1029 | ||
1030 | return 0; /* The end */ | |
1031 | } | |
1032 | } | |
1033 | ||
1034 | if (!q) { | |
1035 | char *m; | |
1036 | ||
1037 | /* No more lines, return empty line */ | |
1038 | ||
1039 | m = strdup(""); | |
1040 | if (!m) | |
1041 | return -ENOMEM; | |
1042 | ||
1043 | *ret = m; | |
1044 | return 0; /* The end */ | |
1045 | } | |
1046 | ||
1047 | p = q + 1; | |
1048 | c++; | |
1049 | } | |
1050 | } | |
53cd7f33 | 1051 | |
46bf625a | 1052 | int string_contains_word_strv(const char *string, const char *separators, char **words, const char **ret_word) { |
53cd7f33 ZJS |
1053 | /* In the default mode with no separators specified, we split on whitespace and |
1054 | * don't coalesce separators. */ | |
1055 | const ExtractFlags flags = separators ? EXTRACT_DONT_COALESCE_SEPARATORS : 0; | |
1056 | ||
46bf625a ZJS |
1057 | const char *found = NULL; |
1058 | ||
53cd7f33 ZJS |
1059 | for (const char *p = string;;) { |
1060 | _cleanup_free_ char *w = NULL; | |
1061 | int r; | |
1062 | ||
1063 | r = extract_first_word(&p, &w, separators, flags); | |
1064 | if (r < 0) | |
1065 | return r; | |
1066 | if (r == 0) | |
46bf625a ZJS |
1067 | break; |
1068 | ||
1069 | found = strv_find(words, w); | |
1070 | if (found) | |
1071 | break; | |
53cd7f33 | 1072 | } |
46bf625a ZJS |
1073 | |
1074 | if (ret_word) | |
1075 | *ret_word = found; | |
1076 | return !!found; | |
53cd7f33 | 1077 | } |