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 valid_pct_enc_string(char *s)
54 if (!isxdigit(s[1]) || !isxdigit(s[2]))
57 if (s[1] == '0' && s[2] == '0')
64 valid_pct_encoded(struct parser *p)
69 if (!valid_pct_enc_string(p->iri)) {
70 p->err = "illegal percent-encoding";
81 sscanf(s+1, "%2hhx", s);
82 memmove(s+1, s+3, strlen(s+3)+1);
86 parse_pct_encoded(struct parser *p)
91 if (!valid_pct_enc_string(p->iri)) {
92 p->err = "illegal percent-encoding";
97 if (*p->iri == '\0') {
98 p->err = "illegal percent-encoding";
105 /* ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) "://" */
107 parse_scheme(struct parser *p)
109 p->parsed->schema = p->iri;
111 if (!isalpha(*p->iri)) {
112 p->err = "illegal character in scheme";
118 * normalize the scheme (i.e. lowercase it)
120 * XXX: since we cannot have good things, tolower
121 * behaviour depends on the LC_CTYPE locale. The good
122 * news is that we're sure p->iri points to something
123 * that's in the ASCII range, so tolower can't
124 * mis-behave on some systems due to the locale.
126 *p->iri = tolower(*p->iri);
128 } while (isalnum(*p->iri)
133 if (*p->iri != ':') {
134 p->err = "illegal character in scheme";
139 if (p->iri[1] != '/' || p->iri[2] != '/') {
140 p->err = "invalid marker after scheme";
150 parse_port(struct parser *p)
154 p->parsed->port = p->iri;
156 for (; isdigit(*p->iri); p->iri++) {
157 i = i * 10 + *p->iri - '0';
158 if (i > UINT16_MAX) {
159 p->err = "port number too large";
164 if (*p->iri != '/' && *p->iri != '\0') {
165 p->err = "illegal character in port number";
169 p->parsed->port_no = i;
171 if (*p->iri != '\0') {
179 /* TODO: add support for ip-literal and ipv4addr ? */
180 /* *( unreserved / sub-delims / pct-encoded ) */
182 parse_authority(struct parser *p)
184 p->parsed->host = p->iri;
186 while (unreserved(*p->iri)
187 || sub_delimiters(*p->iri)
188 || parse_pct_encoded(p)
189 || valid_multibyte_utf8(p)) {
190 /* normalize the host name. */
192 *p->iri = tolower(*p->iri);
199 if (*p->iri == ':') {
202 return parse_port(p);
204 p->parsed->port_no = 1965;
206 if (*p->iri == '/') {
215 p->err = "illegal character in authority section";
220 * Routine for path_clean. Elide the pointed .. with the preceding
221 * element. Return 0 if it's not possible. incr is the length of
222 * the increment, 3 for ../ and 2 for ..
225 path_elide_dotdot(char *path, char *i, int incr)
231 for (j = i-2; j != path && *j != '/'; j--)
236 memmove(j, i, strlen(i)+1);
241 * Use an algorithm similar to the one implemented in go' path.Clean:
243 * 1. Replace multiple slashes with a single slash
244 * 2. Eliminate each . path name element
245 * 3. Eliminate each inner .. along with the non-.. element that precedes it
246 * 4. Eliminate trailing .. if possible or error (go would only discard)
248 * Unlike path.Clean, this function return the empty string if the
249 * original path is equivalent to "/".
252 path_clean(char *path)
256 /* 1. replace multiple slashes with a single one */
257 for (i = path; *i; ++i) {
258 if (*i == '/' && *(i+1) == '/') {
259 memmove(i, i+1, strlen(i)); /* move also the \0 */
264 /* 2. eliminate each . path name element */
265 for (i = path; *i; ++i) {
266 if ((i == path || *i == '/') &&
267 *i != '.' && i[1] == '.' && i[2] == '/') {
268 /* move also the \0 */
269 memmove(i, i+2, strlen(i)-1);
273 if (!strcmp(path, ".") || !strcmp(path, "/.")) {
278 /* 3. eliminate each inner .. along with the preceding non-.. */
279 for (i = strstr(path, "../"); i != NULL; i = strstr(path, "..")) {
280 /* break if we've found a trailing .. */
283 if (!path_elide_dotdot(path, i, 3))
287 /* 4. eliminate trailing ..*/
288 if ((i = strstr(path, "..")) != NULL)
289 if (!path_elide_dotdot(path, i, 2))
296 parse_query(struct parser *p)
298 p->parsed->query = p->iri;
302 while (unreserved(*p->iri)
303 || sub_delimiters(*p->iri)
308 || valid_pct_encoded(p)
309 || valid_multibyte_utf8(p))
315 if (*p->iri != '\0' && *p->iri != '#') {
316 p->err = "illegal character in query";
320 if (*p->iri != '\0') {
328 /* don't even bother */
330 parse_fragment(struct parser *p)
332 p->parsed->fragment = p->iri;
336 /* XXX: is it too broad? */
339 parse_path(struct parser *p)
343 /* trim initial slashes */
344 while (*p->iri == '/')
347 p->parsed->path = p->iri;
348 if (*p->iri == '\0') {
349 p->parsed->query = p->parsed->fragment = p->iri;
353 while (unreserved(*p->iri)
354 || sub_delimiters(*p->iri)
356 || parse_pct_encoded(p)
357 || valid_multibyte_utf8(p))
363 if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
364 p->err = "illegal character in path";
368 if (*p->iri != '\0') {
374 if (!parse_fragment(p))
377 if (!parse_query(p) || !parse_fragment(p))
381 if (!path_clean(p->parsed->path)) {
382 p->err = "illegal path";
390 parse_iri(char *iri, struct iri *ret, const char **err_ret)
399 memset(ret, 0, sizeof(*ret));
401 /* initialize optional stuff to the empty string */
402 end = iri + strlen(iri);
403 p.parsed->host = end;
404 p.parsed->port = end;
405 p.parsed->path = end;
406 p.parsed->query = end;
407 p.parsed->fragment = end;
409 if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
419 serialize_iri(struct iri *i, char *buf, size_t len)
423 /* in ex.c we receive empty "" strings as NULL */
424 if (i->schema == NULL || i->host == NULL) {
429 strlcpy(buf, i->schema, len);
430 strlcat(buf, "://", len);
431 strlcat(buf, i->host, len);
432 strlcat(buf, "/", len);
435 l = strlcat(buf, i->path, len);
437 if (i->query != NULL && *i->query != '\0') {
438 strlcat(buf, "?", len);
439 l = strlcat(buf, i->query, len);
446 pct_decode_str(char *s)
450 for (t = s; *t; ++t) {
451 if (*t == '%' && valid_pct_enc_string(t))