Blob


1 /*
2 * Copyright (c) 2020, 2022 Omar Polo <op@omarpolo.com>
3 *
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.
7 *
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.
15 */
17 #include "gmid.h"
19 #include <ctype.h>
20 #include <string.h>
22 static inline int
23 unreserved(int p)
24 {
25 return isalnum(p)
26 || p == '-'
27 || p == '.'
28 || p == '_'
29 || p == '~';
30 }
32 static inline int
33 sub_delimiters(int p)
34 {
35 return p == '!'
36 || p == '$'
37 || p == '&'
38 || p == '\''
39 || p == '('
40 || p == ')'
41 || p == '*'
42 || p == '+'
43 || p == ','
44 || p == ';'
45 || p == '=';
46 }
48 static int
49 valid_pct_enc_string(char *s)
50 {
51 if (*s != '%')
52 return 1;
54 if (!isxdigit(s[1]) || !isxdigit(s[2]))
55 return 0;
57 if (s[1] == '0' && s[2] == '0')
58 return 0;
60 return 1;
61 }
63 static int
64 valid_pct_encoded(struct parser *p)
65 {
66 if (p->iri[0] != '%')
67 return 0;
69 if (!valid_pct_enc_string(p->iri)) {
70 p->err = "illegal percent-encoding";
71 return 0;
72 }
74 p->iri += 2;
75 return 1;
76 }
78 static void
79 pct_decode(char *s)
80 {
81 sscanf(s+1, "%2hhx", s);
82 memmove(s+1, s+3, strlen(s+3)+1);
83 }
85 static int
86 parse_pct_encoded(struct parser *p)
87 {
88 if (p->iri[0] != '%')
89 return 0;
91 if (!valid_pct_enc_string(p->iri)) {
92 p->err = "illegal percent-encoding";
93 return 0;
94 }
96 pct_decode(p->iri);
97 if (*p->iri == '\0') {
98 p->err = "illegal percent-encoding";
99 return 0;
102 return 1;
105 /* ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) "://" */
106 static int
107 parse_scheme(struct parser *p)
109 p->parsed->schema = p->iri;
111 if (!isalpha(*p->iri)) {
112 p->err = "illegal character in scheme";
113 return 0;
116 do {
117 /*
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.
125 */
126 *p->iri = tolower(*p->iri);
127 p->iri++;
128 } while (isalnum(*p->iri)
129 || *p->iri == '+'
130 || *p->iri == '-'
131 || *p->iri == '.');
133 if (*p->iri != ':') {
134 p->err = "illegal character in scheme";
135 return 0;
138 *p->iri = '\0';
139 if (p->iri[1] != '/' || p->iri[2] != '/') {
140 p->err = "invalid marker after scheme";
141 return 0;
144 p->iri += 3;
145 return 1;
148 /* *DIGIT */
149 static int
150 parse_port(struct parser *p)
152 uint32_t i = 0;
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";
160 return 0;
164 if (*p->iri != '/' && *p->iri != '\0') {
165 p->err = "illegal character in port number";
166 return 0;
169 p->parsed->port_no = i;
171 if (*p->iri != '\0') {
172 *p->iri = '\0';
173 p->iri++;
176 return 1;
179 /* TODO: add support for ip-literal and ipv4addr ? */
180 /* *( unreserved / sub-delims / pct-encoded ) */
181 static int
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. */
191 if (*p->iri < 0x7F)
192 *p->iri = tolower(*p->iri);
193 p->iri++;
196 if (p->err != NULL)
197 return 0;
199 if (*p->iri == ':') {
200 *p->iri = '\0';
201 p->iri++;
202 return parse_port(p);
203 } else
204 p->parsed->port_no = 1965;
206 if (*p->iri == '/') {
207 *p->iri = '\0';
208 p->iri++;
209 return 1;
212 if (*p->iri == '\0')
213 return 1;
215 p->err = "illegal character in authority section";
216 return 0;
219 /*
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 ..
223 */
224 static int
225 path_elide_dotdot(char *path, char *i, int incr)
227 char *j;
229 if (i == path)
230 return 0;
231 for (j = i-2; j != path && *j != '/'; j--)
232 /* noop */ ;
233 if (*j == '/')
234 j++;
235 i += incr;
236 memmove(j, i, strlen(i)+1);
237 return 1;
240 /*
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 "/".
250 */
251 static int
252 path_clean(char *path)
254 char *i;
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 */
260 i--;
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);
270 i--;
273 if (!strcmp(path, ".") || !strcmp(path, "/.")) {
274 *path = '\0';
275 return 1;
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 .. */
281 if (i[2] == '\0')
282 break;
283 if (!path_elide_dotdot(path, i, 3))
284 return 0;
287 /* 4. eliminate trailing ..*/
288 if ((i = strstr(path, "..")) != NULL)
289 if (!path_elide_dotdot(path, i, 2))
290 return 0;
292 return 1;
295 static int
296 parse_query(struct parser *p)
298 p->parsed->query = p->iri;
299 if (*p->iri == '\0')
300 return 1;
302 while (unreserved(*p->iri)
303 || sub_delimiters(*p->iri)
304 || *p->iri == '/'
305 || *p->iri == '?'
306 || *p->iri == ':'
307 || *p->iri == '@'
308 || valid_pct_encoded(p)
309 || valid_multibyte_utf8(p))
310 p->iri++;
312 if (p->err != NULL)
313 return 0;
315 if (*p->iri != '\0' && *p->iri != '#') {
316 p->err = "illegal character in query";
317 return 0;
320 if (*p->iri != '\0') {
321 *p->iri = '\0';
322 p->iri++;
325 return 1;
328 /* don't even bother */
329 static int
330 parse_fragment(struct parser *p)
332 p->parsed->fragment = p->iri;
333 return 1;
336 /* XXX: is it too broad? */
337 /* *(pchar / "/") */
338 static int
339 parse_path(struct parser *p)
341 char c;
343 /* trim initial slashes */
344 while (*p->iri == '/')
345 p->iri++;
347 p->parsed->path = p->iri;
348 if (*p->iri == '\0') {
349 p->parsed->query = p->parsed->fragment = p->iri;
350 return 1;
353 while (unreserved(*p->iri)
354 || sub_delimiters(*p->iri)
355 || *p->iri == '@'
356 || *p->iri == ':'
357 || *p->iri == '/'
358 || parse_pct_encoded(p)
359 || valid_multibyte_utf8(p))
360 p->iri++;
362 if (p->err != NULL)
363 return 0;
365 if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
366 p->err = "illegal character in path";
367 return 0;
370 if (*p->iri != '\0') {
371 c = *p->iri;
372 *p->iri = '\0';
373 p->iri++;
375 if (c == '#') {
376 if (!parse_fragment(p))
377 return 0;
378 } else
379 if (!parse_query(p) || !parse_fragment(p))
380 return 0;
383 if (!path_clean(p->parsed->path)) {
384 p->err = "illegal path";
385 return 0;
388 return 1;
391 int
392 parse_iri(char *iri, struct iri *ret, const char **err_ret)
394 char *end;
395 struct parser p = {
396 .iri = iri,
397 .parsed = ret,
398 .err = NULL,
399 };
401 memset(ret, 0, sizeof(*ret));
403 /* initialize optional stuff to the empty string */
404 end = iri + strlen(iri);
405 p.parsed->host = end;
406 p.parsed->port = end;
407 p.parsed->path = end;
408 p.parsed->query = end;
409 p.parsed->fragment = end;
411 if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
412 *err_ret = p.err;
413 return 0;
416 *err_ret = NULL;
417 return 1;
420 int
421 serialize_iri(struct iri *i, char *buf, size_t len)
423 size_t l = 0;
425 /* in ex.c we receive empty "" strings as NULL */
426 if (i->schema == NULL || i->host == NULL) {
427 memset(buf, 0, len);
428 return 0;
431 strlcpy(buf, i->schema, len);
432 strlcat(buf, "://", len);
433 strlcat(buf, i->host, len);
434 strlcat(buf, "/", len);
436 if (i->path != NULL)
437 l = strlcat(buf, i->path, len);
439 if (i->query != NULL && *i->query != '\0') {
440 strlcat(buf, "?", len);
441 l = strlcat(buf, i->query, len);
444 return l < len;
447 int
448 encode_path(char *buf, size_t len, const char *path)
450 char *p = buf;
451 int a, b;
453 memset(buf, 0, len);
454 while (*path != '\0') {
455 if (len == 1) /* NUL */
456 return -1;
458 if (unreserved(*path) ||
459 sub_delimiters(*path) ||
460 *path == '@' ||
461 *path == ':' ||
462 *path == '/') {
463 *p++ = *path++;
464 len--;
465 } else if (len < 4)
466 return -1;
467 else {
468 a = (*path & 0xF0) >> 4;
469 b = (*path & 0x0F);
471 p[0] = '%';
472 p[1] = a <= 9 ? ('0' + a) : ('7' + a);
473 p[2] = b <= 9 ? ('0' + b) : ('7' + b);
475 path++;
476 p += 3;
477 len -= 3;
481 return 0;
484 char *
485 pct_decode_str(char *s)
487 char *t;
489 for (t = s; *t; ++t) {
490 if (*t == '%' && valid_pct_enc_string(t))
491 pct_decode(t);
494 return s;