grep/icase: avoid kwsset on literal non-ascii strings
[git.git] / strbuf.c
blobbab316dda8f4946efb9a908f17c98b4db3d21e98
1 #include "cache.h"
2 #include "refs.h"
3 #include "utf8.h"
5 int starts_with(const char *str, const char *prefix)
7 for (; ; str++, prefix++)
8 if (!*prefix)
9 return 1;
10 else if (*str != *prefix)
11 return 0;
15 * Used as the default ->buf value, so that people can always assume
16 * buf is non NULL and ->buf is NUL terminated even for a freshly
17 * initialized strbuf.
19 char strbuf_slopbuf[1];
21 void strbuf_init(struct strbuf *sb, size_t hint)
23 sb->alloc = sb->len = 0;
24 sb->buf = strbuf_slopbuf;
25 if (hint)
26 strbuf_grow(sb, hint);
29 void strbuf_release(struct strbuf *sb)
31 if (sb->alloc) {
32 free(sb->buf);
33 strbuf_init(sb, 0);
37 char *strbuf_detach(struct strbuf *sb, size_t *sz)
39 char *res;
40 strbuf_grow(sb, 0);
41 res = sb->buf;
42 if (sz)
43 *sz = sb->len;
44 strbuf_init(sb, 0);
45 return res;
48 void strbuf_attach(struct strbuf *sb, void *buf, size_t len, size_t alloc)
50 strbuf_release(sb);
51 sb->buf = buf;
52 sb->len = len;
53 sb->alloc = alloc;
54 strbuf_grow(sb, 0);
55 sb->buf[sb->len] = '\0';
58 void strbuf_grow(struct strbuf *sb, size_t extra)
60 int new_buf = !sb->alloc;
61 if (unsigned_add_overflows(extra, 1) ||
62 unsigned_add_overflows(sb->len, extra + 1))
63 die("you want to use way too much memory");
64 if (new_buf)
65 sb->buf = NULL;
66 ALLOC_GROW(sb->buf, sb->len + extra + 1, sb->alloc);
67 if (new_buf)
68 sb->buf[0] = '\0';
71 void strbuf_trim(struct strbuf *sb)
73 strbuf_rtrim(sb);
74 strbuf_ltrim(sb);
76 void strbuf_rtrim(struct strbuf *sb)
78 while (sb->len > 0 && isspace((unsigned char)sb->buf[sb->len - 1]))
79 sb->len--;
80 sb->buf[sb->len] = '\0';
83 void strbuf_ltrim(struct strbuf *sb)
85 char *b = sb->buf;
86 while (sb->len > 0 && isspace(*b)) {
87 b++;
88 sb->len--;
90 memmove(sb->buf, b, sb->len);
91 sb->buf[sb->len] = '\0';
94 int strbuf_reencode(struct strbuf *sb, const char *from, const char *to)
96 char *out;
97 int len;
99 if (same_encoding(from, to))
100 return 0;
102 out = reencode_string_len(sb->buf, sb->len, to, from, &len);
103 if (!out)
104 return -1;
106 strbuf_attach(sb, out, len, len);
107 return 0;
110 void strbuf_tolower(struct strbuf *sb)
112 char *p = sb->buf, *end = sb->buf + sb->len;
113 for (; p < end; p++)
114 *p = tolower(*p);
117 struct strbuf **strbuf_split_buf(const char *str, size_t slen,
118 int terminator, int max)
120 struct strbuf **ret = NULL;
121 size_t nr = 0, alloc = 0;
122 struct strbuf *t;
124 while (slen) {
125 int len = slen;
126 if (max <= 0 || nr + 1 < max) {
127 const char *end = memchr(str, terminator, slen);
128 if (end)
129 len = end - str + 1;
131 t = xmalloc(sizeof(struct strbuf));
132 strbuf_init(t, len);
133 strbuf_add(t, str, len);
134 ALLOC_GROW(ret, nr + 2, alloc);
135 ret[nr++] = t;
136 str += len;
137 slen -= len;
139 ALLOC_GROW(ret, nr + 1, alloc); /* In case string was empty */
140 ret[nr] = NULL;
141 return ret;
144 void strbuf_list_free(struct strbuf **sbs)
146 struct strbuf **s = sbs;
148 while (*s) {
149 strbuf_release(*s);
150 free(*s++);
152 free(sbs);
155 int strbuf_cmp(const struct strbuf *a, const struct strbuf *b)
157 int len = a->len < b->len ? a->len: b->len;
158 int cmp = memcmp(a->buf, b->buf, len);
159 if (cmp)
160 return cmp;
161 return a->len < b->len ? -1: a->len != b->len;
164 void strbuf_splice(struct strbuf *sb, size_t pos, size_t len,
165 const void *data, size_t dlen)
167 if (unsigned_add_overflows(pos, len))
168 die("you want to use way too much memory");
169 if (pos > sb->len)
170 die("`pos' is too far after the end of the buffer");
171 if (pos + len > sb->len)
172 die("`pos + len' is too far after the end of the buffer");
174 if (dlen >= len)
175 strbuf_grow(sb, dlen - len);
176 memmove(sb->buf + pos + dlen,
177 sb->buf + pos + len,
178 sb->len - pos - len);
179 memcpy(sb->buf + pos, data, dlen);
180 strbuf_setlen(sb, sb->len + dlen - len);
183 void strbuf_insert(struct strbuf *sb, size_t pos, const void *data, size_t len)
185 strbuf_splice(sb, pos, 0, data, len);
188 void strbuf_remove(struct strbuf *sb, size_t pos, size_t len)
190 strbuf_splice(sb, pos, len, NULL, 0);
193 void strbuf_add(struct strbuf *sb, const void *data, size_t len)
195 strbuf_grow(sb, len);
196 memcpy(sb->buf + sb->len, data, len);
197 strbuf_setlen(sb, sb->len + len);
200 void strbuf_adddup(struct strbuf *sb, size_t pos, size_t len)
202 strbuf_grow(sb, len);
203 memcpy(sb->buf + sb->len, sb->buf + pos, len);
204 strbuf_setlen(sb, sb->len + len);
207 void strbuf_addchars(struct strbuf *sb, int c, size_t n)
209 strbuf_grow(sb, n);
210 memset(sb->buf + sb->len, c, n);
211 strbuf_setlen(sb, sb->len + n);
214 void strbuf_addf(struct strbuf *sb, const char *fmt, ...)
216 va_list ap;
217 va_start(ap, fmt);
218 strbuf_vaddf(sb, fmt, ap);
219 va_end(ap);
222 static void add_lines(struct strbuf *out,
223 const char *prefix1,
224 const char *prefix2,
225 const char *buf, size_t size)
227 while (size) {
228 const char *prefix;
229 const char *next = memchr(buf, '\n', size);
230 next = next ? (next + 1) : (buf + size);
232 prefix = ((prefix2 && (buf[0] == '\n' || buf[0] == '\t'))
233 ? prefix2 : prefix1);
234 strbuf_addstr(out, prefix);
235 strbuf_add(out, buf, next - buf);
236 size -= next - buf;
237 buf = next;
239 strbuf_complete_line(out);
242 void strbuf_add_commented_lines(struct strbuf *out, const char *buf, size_t size)
244 static char prefix1[3];
245 static char prefix2[2];
247 if (prefix1[0] != comment_line_char) {
248 xsnprintf(prefix1, sizeof(prefix1), "%c ", comment_line_char);
249 xsnprintf(prefix2, sizeof(prefix2), "%c", comment_line_char);
251 add_lines(out, prefix1, prefix2, buf, size);
254 void strbuf_commented_addf(struct strbuf *sb, const char *fmt, ...)
256 va_list params;
257 struct strbuf buf = STRBUF_INIT;
258 int incomplete_line = sb->len && sb->buf[sb->len - 1] != '\n';
260 va_start(params, fmt);
261 strbuf_vaddf(&buf, fmt, params);
262 va_end(params);
264 strbuf_add_commented_lines(sb, buf.buf, buf.len);
265 if (incomplete_line)
266 sb->buf[--sb->len] = '\0';
268 strbuf_release(&buf);
271 void strbuf_vaddf(struct strbuf *sb, const char *fmt, va_list ap)
273 int len;
274 va_list cp;
276 if (!strbuf_avail(sb))
277 strbuf_grow(sb, 64);
278 va_copy(cp, ap);
279 len = vsnprintf(sb->buf + sb->len, sb->alloc - sb->len, fmt, cp);
280 va_end(cp);
281 if (len < 0)
282 die("BUG: your vsnprintf is broken (returned %d)", len);
283 if (len > strbuf_avail(sb)) {
284 strbuf_grow(sb, len);
285 len = vsnprintf(sb->buf + sb->len, sb->alloc - sb->len, fmt, ap);
286 if (len > strbuf_avail(sb))
287 die("BUG: your vsnprintf is broken (insatiable)");
289 strbuf_setlen(sb, sb->len + len);
292 void strbuf_expand(struct strbuf *sb, const char *format, expand_fn_t fn,
293 void *context)
295 for (;;) {
296 const char *percent;
297 size_t consumed;
299 percent = strchrnul(format, '%');
300 strbuf_add(sb, format, percent - format);
301 if (!*percent)
302 break;
303 format = percent + 1;
305 if (*format == '%') {
306 strbuf_addch(sb, '%');
307 format++;
308 continue;
311 consumed = fn(sb, format, context);
312 if (consumed)
313 format += consumed;
314 else
315 strbuf_addch(sb, '%');
319 size_t strbuf_expand_dict_cb(struct strbuf *sb, const char *placeholder,
320 void *context)
322 struct strbuf_expand_dict_entry *e = context;
323 size_t len;
325 for (; e->placeholder && (len = strlen(e->placeholder)); e++) {
326 if (!strncmp(placeholder, e->placeholder, len)) {
327 if (e->value)
328 strbuf_addstr(sb, e->value);
329 return len;
332 return 0;
335 void strbuf_addbuf_percentquote(struct strbuf *dst, const struct strbuf *src)
337 int i, len = src->len;
339 for (i = 0; i < len; i++) {
340 if (src->buf[i] == '%')
341 strbuf_addch(dst, '%');
342 strbuf_addch(dst, src->buf[i]);
346 size_t strbuf_fread(struct strbuf *sb, size_t size, FILE *f)
348 size_t res;
349 size_t oldalloc = sb->alloc;
351 strbuf_grow(sb, size);
352 res = fread(sb->buf + sb->len, 1, size, f);
353 if (res > 0)
354 strbuf_setlen(sb, sb->len + res);
355 else if (oldalloc == 0)
356 strbuf_release(sb);
357 return res;
360 ssize_t strbuf_read(struct strbuf *sb, int fd, size_t hint)
362 size_t oldlen = sb->len;
363 size_t oldalloc = sb->alloc;
365 strbuf_grow(sb, hint ? hint : 8192);
366 for (;;) {
367 ssize_t want = sb->alloc - sb->len - 1;
368 ssize_t got = read_in_full(fd, sb->buf + sb->len, want);
370 if (got < 0) {
371 if (oldalloc == 0)
372 strbuf_release(sb);
373 else
374 strbuf_setlen(sb, oldlen);
375 return -1;
377 sb->len += got;
378 if (got < want)
379 break;
380 strbuf_grow(sb, 8192);
383 sb->buf[sb->len] = '\0';
384 return sb->len - oldlen;
387 ssize_t strbuf_read_once(struct strbuf *sb, int fd, size_t hint)
389 ssize_t cnt;
391 strbuf_grow(sb, hint ? hint : 8192);
392 cnt = xread(fd, sb->buf + sb->len, sb->alloc - sb->len - 1);
393 if (cnt > 0)
394 strbuf_setlen(sb, sb->len + cnt);
395 return cnt;
398 #define STRBUF_MAXLINK (2*PATH_MAX)
400 int strbuf_readlink(struct strbuf *sb, const char *path, size_t hint)
402 size_t oldalloc = sb->alloc;
404 if (hint < 32)
405 hint = 32;
407 while (hint < STRBUF_MAXLINK) {
408 int len;
410 strbuf_grow(sb, hint);
411 len = readlink(path, sb->buf, hint);
412 if (len < 0) {
413 if (errno != ERANGE)
414 break;
415 } else if (len < hint) {
416 strbuf_setlen(sb, len);
417 return 0;
420 /* .. the buffer was too small - try again */
421 hint *= 2;
423 if (oldalloc == 0)
424 strbuf_release(sb);
425 return -1;
428 int strbuf_getcwd(struct strbuf *sb)
430 size_t oldalloc = sb->alloc;
431 size_t guessed_len = 128;
433 for (;; guessed_len *= 2) {
434 strbuf_grow(sb, guessed_len);
435 if (getcwd(sb->buf, sb->alloc)) {
436 strbuf_setlen(sb, strlen(sb->buf));
437 return 0;
439 if (errno != ERANGE)
440 break;
442 if (oldalloc == 0)
443 strbuf_release(sb);
444 else
445 strbuf_reset(sb);
446 return -1;
449 #ifdef HAVE_GETDELIM
450 int strbuf_getwholeline(struct strbuf *sb, FILE *fp, int term)
452 ssize_t r;
454 if (feof(fp))
455 return EOF;
457 strbuf_reset(sb);
459 /* Translate slopbuf to NULL, as we cannot call realloc on it */
460 if (!sb->alloc)
461 sb->buf = NULL;
462 r = getdelim(&sb->buf, &sb->alloc, term, fp);
464 if (r > 0) {
465 sb->len = r;
466 return 0;
468 assert(r == -1);
471 * Normally we would have called xrealloc, which will try to free
472 * memory and recover. But we have no way to tell getdelim() to do so.
473 * Worse, we cannot try to recover ENOMEM ourselves, because we have
474 * no idea how many bytes were read by getdelim.
476 * Dying here is reasonable. It mirrors what xrealloc would do on
477 * catastrophic memory failure. We skip the opportunity to free pack
478 * memory and retry, but that's unlikely to help for a malloc small
479 * enough to hold a single line of input, anyway.
481 if (errno == ENOMEM)
482 die("Out of memory, getdelim failed");
484 /* Restore slopbuf that we moved out of the way before */
485 if (!sb->buf)
486 strbuf_init(sb, 0);
487 return EOF;
489 #else
490 int strbuf_getwholeline(struct strbuf *sb, FILE *fp, int term)
492 int ch;
494 if (feof(fp))
495 return EOF;
497 strbuf_reset(sb);
498 flockfile(fp);
499 while ((ch = getc_unlocked(fp)) != EOF) {
500 if (!strbuf_avail(sb))
501 strbuf_grow(sb, 1);
502 sb->buf[sb->len++] = ch;
503 if (ch == term)
504 break;
506 funlockfile(fp);
507 if (ch == EOF && sb->len == 0)
508 return EOF;
510 sb->buf[sb->len] = '\0';
511 return 0;
513 #endif
515 static int strbuf_getdelim(struct strbuf *sb, FILE *fp, int term)
517 if (strbuf_getwholeline(sb, fp, term))
518 return EOF;
519 if (sb->buf[sb->len - 1] == term)
520 strbuf_setlen(sb, sb->len - 1);
521 return 0;
524 int strbuf_getline(struct strbuf *sb, FILE *fp)
526 if (strbuf_getwholeline(sb, fp, '\n'))
527 return EOF;
528 if (sb->buf[sb->len - 1] == '\n') {
529 strbuf_setlen(sb, sb->len - 1);
530 if (sb->len && sb->buf[sb->len - 1] == '\r')
531 strbuf_setlen(sb, sb->len - 1);
533 return 0;
536 int strbuf_getline_lf(struct strbuf *sb, FILE *fp)
538 return strbuf_getdelim(sb, fp, '\n');
541 int strbuf_getline_nul(struct strbuf *sb, FILE *fp)
543 return strbuf_getdelim(sb, fp, '\0');
546 int strbuf_getwholeline_fd(struct strbuf *sb, int fd, int term)
548 strbuf_reset(sb);
550 while (1) {
551 char ch;
552 ssize_t len = xread(fd, &ch, 1);
553 if (len <= 0)
554 return EOF;
555 strbuf_addch(sb, ch);
556 if (ch == term)
557 break;
559 return 0;
562 ssize_t strbuf_read_file(struct strbuf *sb, const char *path, size_t hint)
564 int fd;
565 ssize_t len;
567 fd = open(path, O_RDONLY);
568 if (fd < 0)
569 return -1;
570 len = strbuf_read(sb, fd, hint);
571 close(fd);
572 if (len < 0)
573 return -1;
575 return len;
578 void strbuf_add_lines(struct strbuf *out, const char *prefix,
579 const char *buf, size_t size)
581 add_lines(out, prefix, NULL, buf, size);
584 void strbuf_addstr_xml_quoted(struct strbuf *buf, const char *s)
586 while (*s) {
587 size_t len = strcspn(s, "\"<>&");
588 strbuf_add(buf, s, len);
589 s += len;
590 switch (*s) {
591 case '"':
592 strbuf_addstr(buf, "&quot;");
593 break;
594 case '<':
595 strbuf_addstr(buf, "&lt;");
596 break;
597 case '>':
598 strbuf_addstr(buf, "&gt;");
599 break;
600 case '&':
601 strbuf_addstr(buf, "&amp;");
602 break;
603 case 0:
604 return;
606 s++;
610 static int is_rfc3986_reserved(char ch)
612 switch (ch) {
613 case '!': case '*': case '\'': case '(': case ')': case ';':
614 case ':': case '@': case '&': case '=': case '+': case '$':
615 case ',': case '/': case '?': case '#': case '[': case ']':
616 return 1;
618 return 0;
621 static int is_rfc3986_unreserved(char ch)
623 return isalnum(ch) ||
624 ch == '-' || ch == '_' || ch == '.' || ch == '~';
627 static void strbuf_add_urlencode(struct strbuf *sb, const char *s, size_t len,
628 int reserved)
630 strbuf_grow(sb, len);
631 while (len--) {
632 char ch = *s++;
633 if (is_rfc3986_unreserved(ch) ||
634 (!reserved && is_rfc3986_reserved(ch)))
635 strbuf_addch(sb, ch);
636 else
637 strbuf_addf(sb, "%%%02x", ch);
641 void strbuf_addstr_urlencode(struct strbuf *sb, const char *s,
642 int reserved)
644 strbuf_add_urlencode(sb, s, strlen(s), reserved);
647 void strbuf_humanise_bytes(struct strbuf *buf, off_t bytes)
649 if (bytes > 1 << 30) {
650 strbuf_addf(buf, "%u.%2.2u GiB",
651 (int)(bytes >> 30),
652 (int)(bytes & ((1 << 30) - 1)) / 10737419);
653 } else if (bytes > 1 << 20) {
654 int x = bytes + 5243; /* for rounding */
655 strbuf_addf(buf, "%u.%2.2u MiB",
656 x >> 20, ((x & ((1 << 20) - 1)) * 100) >> 20);
657 } else if (bytes > 1 << 10) {
658 int x = bytes + 5; /* for rounding */
659 strbuf_addf(buf, "%u.%2.2u KiB",
660 x >> 10, ((x & ((1 << 10) - 1)) * 100) >> 10);
661 } else {
662 strbuf_addf(buf, "%u bytes", (int)bytes);
666 void strbuf_add_absolute_path(struct strbuf *sb, const char *path)
668 if (!*path)
669 die("The empty string is not a valid path");
670 if (!is_absolute_path(path)) {
671 struct stat cwd_stat, pwd_stat;
672 size_t orig_len = sb->len;
673 char *cwd = xgetcwd();
674 char *pwd = getenv("PWD");
675 if (pwd && strcmp(pwd, cwd) &&
676 !stat(cwd, &cwd_stat) &&
677 (cwd_stat.st_dev || cwd_stat.st_ino) &&
678 !stat(pwd, &pwd_stat) &&
679 pwd_stat.st_dev == cwd_stat.st_dev &&
680 pwd_stat.st_ino == cwd_stat.st_ino)
681 strbuf_addstr(sb, pwd);
682 else
683 strbuf_addstr(sb, cwd);
684 if (sb->len > orig_len && !is_dir_sep(sb->buf[sb->len - 1]))
685 strbuf_addch(sb, '/');
686 free(cwd);
688 strbuf_addstr(sb, path);
691 int printf_ln(const char *fmt, ...)
693 int ret;
694 va_list ap;
695 va_start(ap, fmt);
696 ret = vprintf(fmt, ap);
697 va_end(ap);
698 if (ret < 0 || putchar('\n') == EOF)
699 return -1;
700 return ret + 1;
703 int fprintf_ln(FILE *fp, const char *fmt, ...)
705 int ret;
706 va_list ap;
707 va_start(ap, fmt);
708 ret = vfprintf(fp, fmt, ap);
709 va_end(ap);
710 if (ret < 0 || putc('\n', fp) == EOF)
711 return -1;
712 return ret + 1;
715 char *xstrdup_tolower(const char *string)
717 char *result;
718 size_t len, i;
720 len = strlen(string);
721 result = xmalloc(len + 1);
722 for (i = 0; i < len; i++)
723 result[i] = tolower(string[i]);
724 result[i] = '\0';
725 return result;
728 char *xstrvfmt(const char *fmt, va_list ap)
730 struct strbuf buf = STRBUF_INIT;
731 strbuf_vaddf(&buf, fmt, ap);
732 return strbuf_detach(&buf, NULL);
735 char *xstrfmt(const char *fmt, ...)
737 va_list ap;
738 char *ret;
740 va_start(ap, fmt);
741 ret = xstrvfmt(fmt, ap);
742 va_end(ap);
744 return ret;
747 void strbuf_addftime(struct strbuf *sb, const char *fmt, const struct tm *tm)
749 size_t hint = 128;
750 size_t len;
752 if (!*fmt)
753 return;
755 strbuf_grow(sb, hint);
756 len = strftime(sb->buf + sb->len, sb->alloc - sb->len, fmt, tm);
758 if (!len) {
760 * strftime reports "0" if it could not fit the result in the buffer.
761 * Unfortunately, it also reports "0" if the requested time string
762 * takes 0 bytes. So our strategy is to munge the format so that the
763 * output contains at least one character, and then drop the extra
764 * character before returning.
766 struct strbuf munged_fmt = STRBUF_INIT;
767 strbuf_addf(&munged_fmt, "%s ", fmt);
768 while (!len) {
769 hint *= 2;
770 strbuf_grow(sb, hint);
771 len = strftime(sb->buf + sb->len, sb->alloc - sb->len,
772 munged_fmt.buf, tm);
774 strbuf_release(&munged_fmt);
775 len--; /* drop munged space */
777 strbuf_setlen(sb, sb->len + len);
780 void strbuf_add_unique_abbrev(struct strbuf *sb, const unsigned char *sha1,
781 int abbrev_len)
783 int r;
784 strbuf_grow(sb, GIT_SHA1_HEXSZ + 1);
785 r = find_unique_abbrev_r(sb->buf + sb->len, sha1, abbrev_len);
786 strbuf_setlen(sb, sb->len + r);
790 * Returns the length of a line, without trailing spaces.
792 * If the line ends with newline, it will be removed too.
794 static size_t cleanup(char *line, size_t len)
796 while (len) {
797 unsigned char c = line[len - 1];
798 if (!isspace(c))
799 break;
800 len--;
803 return len;
807 * Remove empty lines from the beginning and end
808 * and also trailing spaces from every line.
810 * Turn multiple consecutive empty lines between paragraphs
811 * into just one empty line.
813 * If the input has only empty lines and spaces,
814 * no output will be produced.
816 * If last line does not have a newline at the end, one is added.
818 * Enable skip_comments to skip every line starting with comment
819 * character.
821 void strbuf_stripspace(struct strbuf *sb, int skip_comments)
823 int empties = 0;
824 size_t i, j, len, newlen;
825 char *eol;
827 /* We may have to add a newline. */
828 strbuf_grow(sb, 1);
830 for (i = j = 0; i < sb->len; i += len, j += newlen) {
831 eol = memchr(sb->buf + i, '\n', sb->len - i);
832 len = eol ? eol - (sb->buf + i) + 1 : sb->len - i;
834 if (skip_comments && len && sb->buf[i] == comment_line_char) {
835 newlen = 0;
836 continue;
838 newlen = cleanup(sb->buf + i, len);
840 /* Not just an empty line? */
841 if (newlen) {
842 if (empties > 0 && j > 0)
843 sb->buf[j++] = '\n';
844 empties = 0;
845 memmove(sb->buf + j, sb->buf + i, newlen);
846 sb->buf[newlen + j++] = '\n';
847 } else {
848 empties++;
852 strbuf_setlen(sb, j);