2 * Copyright (c) 2022 Omar Polo <op@omarpolo.com>
4 * Permission to use, copy, modify, and distribute this software for any
5 * purpose with or without fee is hereby granted, provided that the above
6 * copyright notice and this permission notice appear in all copies.
8 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
26 /* TODO: URI -> IRI. accept IRI but emit always URI */
29 cpstr(const char *start, const char *till, char *buf, size_t len)
31 size_t slen = till - start;
35 memcpy(buf, start, slen);
43 return (isalnum((unsigned char)c) ||
55 return (t[0] == '%' &&
56 isxdigit((unsigned char)t[1]) &&
57 isxdigit((unsigned char)t[2]));
63 return (c == '!' || c == '$' || c == '&' || c == '\'' ||
64 c == '(' || c == ')' || c == '*' || c == '+' || c == ',' ||
65 c == ';' || c == '=');
68 static inline const char *
69 advance_pchar(const char *s)
71 if (unreserved(*s) || sub_delims(*s) || *s == ':' || *s == '@')
78 static inline const char *
79 advance_segment(const char *s)
83 while ((t = advance_pchar(s)) != NULL)
88 static inline const char *
89 advance_segment_nz(const char *s)
93 if ((t = advance_pchar(s)) == NULL)
95 return (advance_segment(t));
98 static inline const char *
99 advance_segment_nz_nc(const char *s)
104 if (unreserved(*t) || sub_delims(*t) || *t == '@')
112 return (t != s ? t : NULL);
116 parse_scheme(const char *s, struct iri *iri)
120 if (!isalpha((unsigned char)*t))
123 while (isalnum((unsigned char)*t) ||
129 if (cpstr(s, t, iri->iri_scheme, sizeof(iri->iri_scheme)) == -1)
132 iri->iri_flags |= IH_SCHEME;
136 /* userinfo is always optional */
138 parse_uinfo(const char *s, struct iri *iri)
143 if (unreserved(*t) || sub_delims(*t) || *t == ':')
154 if (cpstr(s, t, iri->iri_uinfo, sizeof(iri->iri_uinfo)) == -1)
156 iri->iri_flags |= IH_UINFO;
161 parse_host(const char *s, struct iri *iri)
166 * cheating a bit by relaxing and merging the rule for
167 * IPv6address and IPvFuture and by merging IPv4address and
172 while (*t && *t != ']')
177 if (cpstr(s, t, iri->iri_host, sizeof(iri->iri_host)) == -1)
179 iri->iri_flags |= IH_HOST;
184 if (unreserved(*t) || sub_delims(*t))
192 if (cpstr(s, t, iri->iri_host, sizeof(iri->iri_host)) == -1)
194 iri->iri_flags |= IH_HOST;
199 parse_port(const char *s, struct iri *iri)
204 while (isdigit((unsigned char)*t))
206 if (cpstr(s, t, iri->iri_portstr, sizeof(iri->iri_portstr)) == -1)
208 iri->iri_port = strtonum(iri->iri_portstr, 1, UINT16_MAX, &errstr);
211 iri->iri_flags |= IH_PORT;
216 parse_authority(const char *s, struct iri *iri)
220 if ((t = parse_uinfo(s, iri)) == NULL)
223 if ((t = parse_host(t, iri)) == NULL)
227 return (parse_port(t, iri));
233 parse_path_abempty(const char *s, struct iri *iri)
238 t = advance_segment(t + 1);
240 if (cpstr(s, t, iri->iri_path, sizeof(iri->iri_path)) == -1)
242 iri->iri_flags |= IH_PATH;
247 parse_path_absolute(const char *s, struct iri *iri)
254 if ((t = advance_segment_nz(s + 1)) == NULL)
258 t = advance_segment(t + 1);
260 if (cpstr(s, t, iri->iri_path, sizeof(iri->iri_path)) == -1)
262 iri->iri_flags |= IH_PATH;
267 parse_path_rootless(const char *s, struct iri *iri)
271 if ((t = advance_segment_nz(s)) == NULL)
275 t = advance_segment(t + 1);
277 if (cpstr(s, t, iri->iri_path, sizeof(iri->iri_path)) == -1)
279 iri->iri_flags |= IH_PATH;
284 parse_path_noscheme(const char *s, struct iri *iri)
288 if ((t = advance_segment_nz_nc(s)) == NULL)
292 t = advance_segment(t + 1);
294 if (cpstr(s, t, iri->iri_path, sizeof(iri->iri_path)) == -1)
296 iri->iri_flags |= IH_PATH;
301 parse_path_empty(const char *s, struct iri *iri)
303 iri->iri_path[0] = '\0';
304 iri->iri_flags |= IH_PATH;
309 parse_hier(const char *s, struct iri *iri)
313 if (!strncmp(s, "//", 2)) {
314 if ((t = parse_authority(s + 2, iri)) == NULL)
316 return (parse_path_abempty(t, iri));
319 if ((t = parse_path_absolute(s, iri)) != NULL)
322 if ((t = parse_path_rootless(s, iri)) != NULL)
325 return (parse_path_empty(s, iri));
329 parse_relative(const char *s, struct iri *iri)
333 if (!strncmp(s, "//", 2)) {
334 if ((t = parse_authority(s + 2, iri)) == NULL)
336 return (parse_path_abempty(t, iri));
339 if ((t = parse_path_absolute(s, iri)) != NULL)
342 if ((t = parse_path_noscheme(s, iri)) != NULL)
345 return (parse_path_empty(s, iri));
349 parse_query(const char *s, struct iri *iri)
351 const char *n, *t = s;
354 if ((n = advance_pchar(t)) != NULL)
356 else if (*t == '/' || *t == '?')
362 if (cpstr(s, t, iri->iri_query, sizeof(iri->iri_query)) == -1)
364 iri->iri_flags |= IH_QUERY;
369 parse_uri(const char *s, struct iri *iri)
371 if ((s = parse_scheme(s, iri)) == NULL)
377 if ((s = parse_hier(s + 1, iri)) == NULL)
380 if (*s == '?' && (s = parse_query(s + 1, iri)) == NULL)
384 if (*s == '#' || *s == '\0')
391 parse_relative_ref(const char *s, struct iri *iri)
393 if ((s = parse_relative(s, iri)) == NULL)
396 if (*s == '?' && (s = parse_query(s + 1, iri)) == NULL)
400 if (*s == '#' || *s == '\0')
407 parse(const char *s, struct iri *iri)
414 if (parse_uri(s, iri) == -1) {
416 if (parse_relative_ref(s, iri) == -1)
427 *s = tolower((unsigned char)*s);
431 cpfields(struct iri *dest, const struct iri *src, int flags)
433 if (flags & IH_SCHEME) {
434 dest->iri_flags |= IH_SCHEME;
435 if (src->iri_flags & IH_SCHEME)
436 memcpy(dest->iri_scheme, src->iri_scheme,
437 sizeof(dest->iri_scheme));
438 lowerify(dest->iri_scheme);
440 if (flags & IH_UINFO) {
441 if (src->iri_flags & IH_UINFO) {
442 memcpy(dest->iri_uinfo, src->iri_uinfo,
443 sizeof(dest->iri_uinfo));
444 dest->iri_flags |= IH_UINFO;
447 if (flags & IH_HOST) {
448 dest->iri_flags |= IH_HOST;
449 if (src->iri_flags & IH_HOST)
450 memcpy(dest->iri_host, src->iri_host,
451 sizeof(dest->iri_host));
452 lowerify(dest->iri_host);
454 if (flags & IH_PORT) {
455 if (src->iri_flags & IH_PORT) {
456 dest->iri_port = src->iri_port;
457 dest->iri_flags |= IH_PORT;
460 if (flags & IH_PATH) {
461 dest->iri_flags |= IH_PATH;
462 if (src->iri_flags & IH_PATH)
463 memcpy(dest->iri_path, src->iri_path,
464 sizeof(dest->iri_path));
466 if (flags & IH_QUERY) {
467 if (src->iri_flags & IH_QUERY) {
468 dest->iri_flags |= IH_QUERY;
469 memcpy(dest->iri_query, src->iri_query,
470 sizeof(dest->iri_query));
476 remove_dot_segments(struct iri *iri)
478 char *p, *q, *buf, *s;
481 buf = p = q = iri->iri_path;
482 bufsize = sizeof(iri->iri_path);
484 while (*p && (q - buf < bufsize)) {
486 if (!strncmp(p, "../", 3)) {
490 if (!strncmp(p, "./", 3)) {
495 if (!strncmp(p, "/./", 3)) {
500 if (!strcmp(p, "/.")) {
505 if (p[0] == '/' && p[1] == '.' && p[2] == '.' &&
506 (p[3] == '/' || p[3] == '\0')) {
508 while (q != buf && *--q != '/')
513 if (!strcmp(p, ".")) {
517 if (!strcmp(p, "..")) {
522 s = strchr(p + 1, '/');
523 while (*p && p != s && (q - buf < bufsize))
527 if (*p == '\0' && (q - buf < bufsize)) {
532 errno = ENAMETOOLONG;
537 mergepath(struct iri *i, struct iri *base, struct iri *r)
539 const char *bpath, *rpath, *s;
541 bpath = (base->iri_flags & IH_PATH) ? base->iri_path : "/";
542 rpath = (r->iri_flags & IH_PATH) ? r->iri_path : "/";
544 i->iri_flags |= IH_PATH;
545 i->iri_path[0] = '\0';
547 if ((base->iri_flags & IH_AUTHORITY) &&
548 (*bpath == '\0' || !strcmp(bpath, "/"))) {
551 strlcpy(i->iri_path, "/", sizeof(i->iri_path));
552 strlcat(i->iri_path, rpath, sizeof(i->iri_path));
556 if ((s = strrchr(bpath, '/')) != NULL) {
557 cpstr(bpath, s + 1, i->iri_path, sizeof(i->iri_path));
561 if (strlcat(i->iri_path, rpath, sizeof(i->iri_path)) >=
562 sizeof(i->iri_path)) {
563 errno = ENAMETOOLONG;
571 iri_parse(const char *base, const char *str, struct iri *iri)
573 static struct iri ibase, iparsed;
575 memset(iri, 0, sizeof(*iri));
579 if (parse_uri(str, &iparsed) == -1) {
584 if (parse_uri(base, &ibase) == -1 ||
585 parse(str, &iparsed) == -1) {
591 if (iparsed.iri_flags & IH_SCHEME) {
592 cpfields(iri, &iparsed, iparsed.iri_flags);
593 remove_dot_segments(iri);
597 /* if fragments are supported, copy iparsed fragment to iri */
599 cpfields(iri, &ibase, IH_SCHEME);
601 if (iparsed.iri_flags & IH_HOST) {
602 cpfields(iri, &iparsed, IH_AUTHORITY|IH_PATH|IH_QUERY);
603 remove_dot_segments(iri);
607 cpfields(iri, &ibase, IH_AUTHORITY);
609 if ((iparsed.iri_flags & IH_PATH) && *iparsed.iri_path == '\0') {
610 cpfields(iri, &ibase, IH_PATH);
611 if (iparsed.iri_flags & IH_QUERY)
612 cpfields(iri, &iparsed, IH_QUERY);
614 cpfields(iri, &ibase, IH_QUERY);
618 cpfields(iri, &iparsed, IH_QUERY);
619 if ((iparsed.iri_flags & IH_PATH) && !strcmp(iparsed.iri_path, "/"))
620 cpfields(iri, &iparsed, IH_PATH);
622 if (!(ibase.iri_flags & IH_PATH))
623 ibase.iri_path[0] = '\0';
624 if (!(iparsed.iri_flags & IH_PATH))
625 iparsed.iri_path[0] = '\0';
626 if (mergepath(iri, &ibase, &iparsed) == -1)
629 if (remove_dot_segments(iri) == -1)
635 iri_unparse(const struct iri *i, char *buf, size_t buflen)
640 /* TODO: should %enc octets if needed */
644 if (i->iri_flags & IH_SCHEME) {
645 if (strlcat(buf, i->iri_scheme, buflen) >= buflen ||
646 strlcat(buf, ":", buflen) >= buflen)
650 if (i->iri_flags & IH_AUTHORITY) {
651 if (strlcat(buf, "//", buflen) >= buflen)
655 if (i->iri_flags & IH_UINFO) {
656 if (strlcat(buf, i->iri_uinfo, buflen) >= buflen ||
657 strlcat(buf, "@", buflen) >= buflen)
660 if (i->iri_flags & IH_HOST) {
661 if (strlcat(buf, i->iri_host, buflen) >= buflen)
664 if (i->iri_flags & IH_PORT) {
665 if (strlcat(buf, ":", buflen) >= buflen ||
666 strlcat(buf, i->iri_portstr, buflen) >= buflen)
670 if (i->iri_flags & IH_PATH) {
671 if (!(i->iri_flags & IH_AUTHORITY) &&
672 i->iri_path[0] != '/' &&
673 strlcat(buf, "/", buflen) >= buflen)
675 if (strlcat(buf, i->iri_path, buflen) >= buflen)
679 if (i->iri_flags & IH_QUERY) {
680 if (strlcat(buf, "?", buflen) >= buflen ||
681 strlcat(buf, i->iri_query, buflen) >= buflen)
693 iri_human(const struct iri *iri, char *buf, size_t buflen)
695 memset(buf, 0, buflen);
700 iri_setquery(struct iri *iri, const char *text)