2 * Copyright (c) 2020 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.
49 parse_pct_encoded(struct parser *p)
54 if (!isxdigit(*(p->iri+1)) || !isxdigit(*(p->iri+2))) {
55 p->err = "illegal percent-encoding";
59 sscanf(p->iri+1, "%2hhx", p->iri);
60 memmove(p->iri+1, p->iri+3, strlen(p->iri+3)+1);
61 if (*p->iri == '\0') {
62 p->err = "illegal percent-encoding";
69 /* ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) "://" */
71 parse_scheme(struct parser *p)
73 p->parsed->schema = p->iri;
75 if (!isalpha(*p->iri)) {
76 p->err = "illegal character in scheme";
81 /* normalize the scheme (i.e. lowercase it)
83 * XXX: since we cannot have good things, tolower
84 * depends on the LC_CTYPE locale. The good things is
85 * that we're sure p->iri points to something in the
86 * ASCII range, so it shouldn't do weird stuff. */
87 *p->iri = tolower(*p->iri);
89 } while (isalnum(*p->iri)
95 p->err = "illegal character in scheme";
100 if (*(++p->iri) != '/' || *(++p->iri) != '/') {
101 p->err = "invalid marker after scheme";
111 parse_port(struct parser *p)
115 p->parsed->port = p->iri;
117 for (; isdigit(*p->iri); p->iri++) {
118 i = i * 10 + *p->iri - '0';
119 if (i > UINT16_MAX) {
120 p->err = "port number too large";
125 if (*p->iri != '/' && *p->iri != '\0') {
126 p->err = "illegal character in port number";
130 p->parsed->port_no = i;
132 if (*p->iri != '\0') {
140 /* TODO: add support for ip-literal and ipv4addr ? */
141 /* *( unreserved / sub-delims / pct-encoded ) */
143 parse_authority(struct parser *p)
145 p->parsed->host = p->iri;
147 while (unreserved(*p->iri)
148 || sub_delimiters(*p->iri)
149 || parse_pct_encoded(p))
155 if (*p->iri == ':') {
158 return parse_port(p);
161 if (*p->iri == '/') {
170 p->err = "illegal character in authority section";
174 /* Routine for path_clean. Elide the pointed .. with the preceding
175 * element. Return 0 if it's not possible. incr is the length of
176 * the increment, 3 for ../ and 2 for .. */
178 path_elide_dotdot(char *path, char *i, int incr)
184 for (j = i-2; j != path && *j != '/'; j--)
189 memmove(j, i, strlen(i)+1);
194 * Use an algorithm similar to the one implemented in go' path.Clean:
196 * 1. Replace multiple slashes with a single slash
197 * 2. Eliminate each . path name element
198 * 3. Eliminate each inner .. along with the non-.. element that precedes it
199 * 4. Eliminate trailing .. if possible or error (go would only discard)
201 * Unlike path.Clean, this function return the empty string if the
202 * original path is equivalent to "/".
205 path_clean(char *path)
209 /* 1. replace multiple slashes with a single one */
210 for (i = path; *i; ++i) {
211 if (*i == '/' && *(i+1) == '/') {
212 memmove(i, i+1, strlen(i)); /* move also the \0 */
217 /* 2. eliminate each . path name element */
218 for (i = path; *i; ++i) {
219 if ((i == path || *i == '/') &&
220 *i != '.' && i[1] == '.' && i[2] == '/') {
221 /* move also the \0 */
222 memmove(i, i+2, strlen(i)-1);
226 if (!strcmp(path, ".") || !strcmp(path, "/.")) {
231 /* 3. eliminate each inner .. along with the preceding non-.. */
232 for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
233 if (!path_elide_dotdot(path, i, 3))
236 /* 4. eliminate trailing ..*/
237 if ((i = strstr(path, "..")) != NULL)
238 if (!path_elide_dotdot(path, i, 2))
245 parse_query(struct parser *p)
247 p->parsed->query = p->iri;
251 while (unreserved(*p->iri)
252 || sub_delimiters(*p->iri)
255 || parse_pct_encoded(p)
256 || valid_multibyte_utf8(p))
262 if (*p->iri != '\0' && *p->iri != '#') {
263 p->err = "illegal character in query";
267 if (*p->iri != '\0') {
275 /* don't even bother */
277 parse_fragment(struct parser *p)
279 p->parsed->fragment = p->iri;
283 /* XXX: is it too broad? */
286 parse_path(struct parser *p)
290 p->parsed->path = p->iri;
291 if (*p->iri == '\0') {
292 p->parsed->query = p->parsed->fragment = p->iri;
296 while (unreserved(*p->iri)
297 || sub_delimiters(*p->iri)
299 || parse_pct_encoded(p)
300 || valid_multibyte_utf8(p))
306 if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
307 p->err = "illegal character in path";
311 if (*p->iri != '\0') {
317 if (!parse_fragment(p))
320 if (!parse_query(p) || !parse_fragment(p))
324 if (!path_clean(p->parsed->path)) {
325 p->err = "illegal path";
333 parse_iri(char *iri, struct iri *ret, const char **err_ret)
336 struct parser p = {iri, ret, NULL};
338 bzero(ret, sizeof(*ret));
340 /* initialize optional stuff to the empty string */
341 end = iri + strlen(iri);
342 p.parsed->port = end;
343 p.parsed->path = end;
344 p.parsed->query = end;
345 p.parsed->fragment = end;
347 if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
357 trim_req_iri(char *iri)
361 if ((i = strstr(iri, "\r\n")) == NULL)