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[1] != '/' || p->iri[2] != '/') {
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)) {
150 /* normalize the host name. */
152 *p->iri = tolower(*p->iri);
159 if (*p->iri == ':') {
162 return parse_port(p);
164 p->parsed->port_no = 1965;
166 if (*p->iri == '/') {
175 p->err = "illegal character in authority section";
179 /* Routine for path_clean. Elide the pointed .. with the preceding
180 * element. Return 0 if it's not possible. incr is the length of
181 * the increment, 3 for ../ and 2 for .. */
183 path_elide_dotdot(char *path, char *i, int incr)
189 for (j = i-2; j != path && *j != '/'; j--)
194 memmove(j, i, strlen(i)+1);
199 * Use an algorithm similar to the one implemented in go' path.Clean:
201 * 1. Replace multiple slashes with a single slash
202 * 2. Eliminate each . path name element
203 * 3. Eliminate each inner .. along with the non-.. element that precedes it
204 * 4. Eliminate trailing .. if possible or error (go would only discard)
206 * Unlike path.Clean, this function return the empty string if the
207 * original path is equivalent to "/".
210 path_clean(char *path)
214 /* 1. replace multiple slashes with a single one */
215 for (i = path; *i; ++i) {
216 if (*i == '/' && *(i+1) == '/') {
217 memmove(i, i+1, strlen(i)); /* move also the \0 */
222 /* 2. eliminate each . path name element */
223 for (i = path; *i; ++i) {
224 if ((i == path || *i == '/') &&
225 *i != '.' && i[1] == '.' && i[2] == '/') {
226 /* move also the \0 */
227 memmove(i, i+2, strlen(i)-1);
231 if (!strcmp(path, ".") || !strcmp(path, "/.")) {
236 /* 3. eliminate each inner .. along with the preceding non-.. */
237 for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
238 if (!path_elide_dotdot(path, i, 3))
241 /* 4. eliminate trailing ..*/
242 if ((i = strstr(path, "..")) != NULL)
243 if (!path_elide_dotdot(path, i, 2))
250 parse_query(struct parser *p)
252 p->parsed->query = p->iri;
256 while (unreserved(*p->iri)
257 || sub_delimiters(*p->iri)
260 || parse_pct_encoded(p)
261 || valid_multibyte_utf8(p))
267 if (*p->iri != '\0' && *p->iri != '#') {
268 p->err = "illegal character in query";
272 if (*p->iri != '\0') {
280 /* don't even bother */
282 parse_fragment(struct parser *p)
284 p->parsed->fragment = p->iri;
288 /* XXX: is it too broad? */
291 parse_path(struct parser *p)
295 p->parsed->path = p->iri;
296 if (*p->iri == '\0') {
297 p->parsed->query = p->parsed->fragment = p->iri;
301 while (unreserved(*p->iri)
302 || sub_delimiters(*p->iri)
304 || parse_pct_encoded(p)
305 || valid_multibyte_utf8(p))
311 if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
312 p->err = "illegal character in path";
316 if (*p->iri != '\0') {
322 if (!parse_fragment(p))
325 if (!parse_query(p) || !parse_fragment(p))
329 if (!path_clean(p->parsed->path)) {
330 p->err = "illegal path";
338 parse_iri(char *iri, struct iri *ret, const char **err_ret)
341 struct parser p = {iri, ret, NULL};
343 bzero(ret, sizeof(*ret));
345 /* initialize optional stuff to the empty string */
346 end = iri + strlen(iri);
347 p.parsed->port = end;
348 p.parsed->path = end;
349 p.parsed->query = end;
350 p.parsed->fragment = end;
352 if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
362 trim_req_iri(char *iri)
366 if ((i = strstr(iri, "\r\n")) == NULL)