1 /*
2  * Copyright (c) 2020 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  */
16 
17 #include <ctype.h>
18 #include <string.h>
19 
20 #include "gmid.h"
21 
22 static inline int
23 unreserved(int p)
24 {
25 	return isalnum(p)
26 		|| p == '-'
27 		|| p == '.'
28 		|| p == '_'
29 		|| p == '~';
30 }
31 
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 }
47 
48 static int
49 parse_pct_encoded(struct parser *p)
50 {
51 	if (*p->iri != '%')
52 		return 0;
53 
54 	if (!isxdigit(*(p->iri+1)) || !isxdigit(*(p->iri+2))) {
55 		p->err = "illegal percent-encoding";
56 		return 0;
57 	}
58 
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";
63 		return 0;
64 	}
65 
66 	return 1;
67 }
68 
69 /* ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) "://" */
70 static int
71 parse_scheme(struct parser *p)
72 {
73 	p->parsed->schema = p->iri;
74 
75 	if (!isalpha(*p->iri)) {
76 		p->err = "illegal character in scheme";
77 		return 0;
78 	}
79 
80 	do {
81 		/* normalize the scheme (i.e. lowercase it)
82 		 *
83 		 * XXX: since we cannot have good things, tolower
84 		 * behaviour depends on the LC_CTYPE locale.  The good
85 		 * news is that we're sure p->iri points to something
86 		 * that's in the ASCII range, so tolower can't
87 		 * mis-behave on some systems due to the locale. */
88 		*p->iri = tolower(*p->iri);
89 		p->iri++;
90 	} while (isalnum(*p->iri)
91 	    || *p->iri == '+'
92 	    || *p->iri == '-'
93 	    || *p->iri == '.');
94 
95 	if (*p->iri != ':') {
96 		p->err = "illegal character in scheme";
97 		return 0;
98 	}
99 
100 	*p->iri = '\0';
101 	if (p->iri[1] != '/' || p->iri[2] != '/') {
102 		p->err = "invalid marker after scheme";
103 		return 0;
104 	}
105 
106 	p->iri += 3;
107 	return 1;
108 }
109 
110 /* *DIGIT */
111 static int
112 parse_port(struct parser *p)
113 {
114 	uint32_t i = 0;
115 
116 	p->parsed->port = p->iri;
117 
118 	for (; isdigit(*p->iri); p->iri++) {
119 		i = i * 10 + *p->iri - '0';
120 		if (i > UINT16_MAX) {
121 			p->err = "port number too large";
122 			return 0;
123 		}
124 	}
125 
126 	if (*p->iri != '/' && *p->iri != '\0') {
127 		p->err = "illegal character in port number";
128 		return 0;
129 	}
130 
131 	p->parsed->port_no = i;
132 
133 	if (*p->iri != '\0') {
134 		*p->iri = '\0';
135 		p->iri++;
136 	}
137 
138 	return 1;
139 }
140 
141 /* TODO: add support for ip-literal and ipv4addr ? */
142 /* *( unreserved / sub-delims / pct-encoded ) */
143 static int
144 parse_authority(struct parser *p)
145 {
146 	p->parsed->host = p->iri;
147 
148 	while (unreserved(*p->iri)
149 	    || sub_delimiters(*p->iri)
150 	    || parse_pct_encoded(p)) {
151 		/* normalize the host name. */
152 		if (*p->iri < 0x7F)
153 			*p->iri = tolower(*p->iri);
154 		p->iri++;
155 	}
156 
157 	if (p->err != NULL)
158 		return 0;
159 
160 	if (*p->iri == ':') {
161 		*p->iri = '\0';
162 		p->iri++;
163 		return parse_port(p);
164 	} else
165 		p->parsed->port_no = 1965;
166 
167 	if (*p->iri == '/') {
168 		*p->iri = '\0';
169 		p->iri++;
170 		return 1;
171 	}
172 
173 	if (*p->iri == '\0')
174 		return 1;
175 
176 	p->err = "illegal character in authority section";
177 	return 0;
178 }
179 
180 /* Routine for path_clean.  Elide the pointed .. with the preceding
181  * element.  Return 0 if it's not possible.  incr is the length of
182  * the increment, 3 for ../ and 2 for .. */
183 static int
184 path_elide_dotdot(char *path, char *i, int incr)
185 {
186 	char *j;
187 
188 	if (i == path)
189 		return 0;
190 	for (j = i-2; j != path && *j != '/'; j--)
191                 /* noop */ ;
192 	if (*j == '/')
193 		j++;
194 	i += incr;
195 	memmove(j, i, strlen(i)+1);
196 	return 1;
197 }
198 
199 /*
200  * Use an algorithm similar to the one implemented in go' path.Clean:
201  *
202  * 1. Replace multiple slashes with a single slash
203  * 2. Eliminate each . path name element
204  * 3. Eliminate each inner .. along with the non-.. element that precedes it
205  * 4. Eliminate trailing .. if possible or error (go would only discard)
206  *
207  * Unlike path.Clean, this function return the empty string if the
208  * original path is equivalent to "/".
209  */
210 static int
211 path_clean(char *path)
212 {
213 	char *i;
214 
215 	/* 1. replace multiple slashes with a single one */
216 	for (i = path; *i; ++i) {
217 		if (*i == '/' && *(i+1) == '/') {
218 			memmove(i, i+1, strlen(i)); /* move also the \0 */
219 			i--;
220 		}
221 	}
222 
223 	/* 2. eliminate each . path name element */
224 	for (i = path; *i; ++i) {
225 		if ((i == path || *i == '/') &&
226 		    *i != '.' && i[1] == '.' && i[2] == '/') {
227 			/* move also the \0 */
228 			memmove(i, i+2, strlen(i)-1);
229 			i--;
230 		}
231 	}
232 	if (!strcmp(path, ".") || !strcmp(path, "/.")) {
233 		*path = '\0';
234 		return 1;
235 	}
236 
237 	/* 3. eliminate each inner .. along with the preceding non-.. */
238 	for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
239 		if (!path_elide_dotdot(path, i, 3))
240 			return 0;
241 
242 	/* 4. eliminate trailing ..*/
243 	if ((i = strstr(path, "..")) != NULL)
244 		if (!path_elide_dotdot(path, i, 2))
245 			return 0;
246 
247 	return 1;
248 }
249 
250 static int
251 parse_query(struct parser *p)
252 {
253 	p->parsed->query = p->iri;
254 	if (*p->iri == '\0')
255 		return 1;
256 
257 	while (unreserved(*p->iri)
258 	    || sub_delimiters(*p->iri)
259 	    || *p->iri == '/'
260 	    || *p->iri == '?'
261 	    || parse_pct_encoded(p)
262 	    || valid_multibyte_utf8(p))
263 		p->iri++;
264 
265 	if (p->err != NULL)
266 		return 0;
267 
268 	if (*p->iri != '\0' && *p->iri != '#') {
269 		p->err = "illegal character in query";
270 		return 0;
271 	}
272 
273 	if (*p->iri != '\0') {
274 		*p->iri = '\0';
275 		p->iri++;
276 	}
277 
278 	return 1;
279 }
280 
281 /* don't even bother */
282 static int
283 parse_fragment(struct parser *p)
284 {
285 	p->parsed->fragment = p->iri;
286 	return 1;
287 }
288 
289 /* XXX: is it too broad? */
290 /* *(pchar / "/") */
291 static int
292 parse_path(struct parser *p)
293 {
294 	char c;
295 
296 	/* trim initial slashes */
297 	while (*p->iri == '/')
298 		p->iri++;
299 
300 	p->parsed->path = p->iri;
301 	if (*p->iri == '\0') {
302 		p->parsed->query = p->parsed->fragment = p->iri;
303 		return 1;
304 	}
305 
306 	while (unreserved(*p->iri)
307 	    || sub_delimiters(*p->iri)
308 	    || *p->iri == '/'
309 	    || parse_pct_encoded(p)
310 	    || valid_multibyte_utf8(p))
311 		p->iri++;
312 
313 	if (p->err != NULL)
314 		return 0;
315 
316 	if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
317 		p->err = "illegal character in path";
318 		return 0;
319 	}
320 
321 	if (*p->iri != '\0') {
322 		c = *p->iri;
323 		*p->iri = '\0';
324 		p->iri++;
325 
326 		if (c == '#') {
327 			if (!parse_fragment(p))
328 				return 0;
329 		} else
330 			if (!parse_query(p) || !parse_fragment(p))
331 				return 0;
332 	}
333 
334 	if (!path_clean(p->parsed->path)) {
335 		p->err = "illegal path";
336 		return 0;
337 	}
338 
339 	return 1;
340 }
341 
342 int
343 parse_iri(char *iri, struct iri *ret, const char **err_ret)
344 {
345 	char *end;
346 	struct parser p = {iri, ret, NULL};
347 
348 	bzero(ret, sizeof(*ret));
349 
350 	/* initialize optional stuff to the empty string */
351 	end = iri + strlen(iri);
352 	p.parsed->port = end;
353 	p.parsed->path = end;
354 	p.parsed->query = end;
355 	p.parsed->fragment = end;
356 
357 	if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
358 		*err_ret = p.err;
359 		return 0;
360 	}
361 
362 	*err_ret = NULL;
363 	return 1;
364 }
365 
366 int
367 trim_req_iri(char *iri)
368 {
369 	char *i;
370 
371 	if ((i = strstr(iri, "\r\n")) == NULL)
372 		return 0;
373 	*i = '\0';
374 	return 1;
375 }