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 		 * 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);
88 		p->iri++;
89 	} while (isalnum(*p->iri)
90 	    || *p->iri == '+'
91 	    || *p->iri == '-'
92 	    || *p->iri == '.');
93 
94 	if (*p->iri != ':') {
95 		p->err = "illegal character in scheme";
96 		return 0;
97 	}
98 
99 	*p->iri = '\0';
100 	if (*(++p->iri) != '/' || *(++p->iri) != '/') {
101 		p->err = "invalid marker after scheme";
102 		return 0;
103 	}
104 
105 	p->iri++;
106 	return 1;
107 }
108 
109 /* *DIGIT */
110 static int
111 parse_port(struct parser *p)
112 {
113 	uint32_t i = 0;
114 
115 	p->parsed->port = p->iri;
116 
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";
121 			return 0;
122 		}
123 	}
124 
125 	if (*p->iri != '/' && *p->iri != '\0') {
126 		p->err = "illegal character in port number";
127 		return 0;
128 	}
129 
130 	p->parsed->port_no = i;
131 
132 	if (*p->iri != '\0') {
133 		*p->iri = '\0';
134 		p->iri++;
135 	}
136 
137 	return 1;
138 }
139 
140 /* TODO: add support for ip-literal and ipv4addr ? */
141 /* *( unreserved / sub-delims / pct-encoded ) */
142 static int
143 parse_authority(struct parser *p)
144 {
145 	p->parsed->host = p->iri;
146 
147 	while (unreserved(*p->iri)
148 	    || sub_delimiters(*p->iri)
149 	    || parse_pct_encoded(p)) {
150 		/* normalize the host name. */
151 		if (*p->iri < 0x7F)
152 			*p->iri = tolower(*p->iri);
153 		p->iri++;
154 	}
155 
156 	if (p->err != NULL)
157 		return 0;
158 
159 	if (*p->iri == ':') {
160 		*p->iri = '\0';
161 		p->iri++;
162 		return parse_port(p);
163 	}
164 
165 	if (*p->iri == '/') {
166 		*p->iri = '\0';
167 		p->iri++;
168 		return 1;
169 	}
170 
171 	if (*p->iri == '\0')
172 		return 1;
173 
174 	p->err = "illegal character in authority section";
175 	return 0;
176 }
177 
178 /* Routine for path_clean.  Elide the pointed .. with the preceding
179  * element.  Return 0 if it's not possible.  incr is the length of
180  * the increment, 3 for ../ and 2 for .. */
181 static int
182 path_elide_dotdot(char *path, char *i, int incr)
183 {
184 	char *j;
185 
186 	if (i == path)
187 		return 0;
188 	for (j = i-2; j != path && *j != '/'; j--)
189                 /* noop */ ;
190 	if (*j == '/')
191 		j++;
192 	i += incr;
193 	memmove(j, i, strlen(i)+1);
194 	return 1;
195 }
196 
197 /*
198  * Use an algorithm similar to the one implemented in go' path.Clean:
199  *
200  * 1. Replace multiple slashes with a single slash
201  * 2. Eliminate each . path name element
202  * 3. Eliminate each inner .. along with the non-.. element that precedes it
203  * 4. Eliminate trailing .. if possible or error (go would only discard)
204  *
205  * Unlike path.Clean, this function return the empty string if the
206  * original path is equivalent to "/".
207  */
208 static int
209 path_clean(char *path)
210 {
211 	char *i;
212 
213 	/* 1. replace multiple slashes with a single one */
214 	for (i = path; *i; ++i) {
215 		if (*i == '/' && *(i+1) == '/') {
216 			memmove(i, i+1, strlen(i)); /* move also the \0 */
217 			i--;
218 		}
219 	}
220 
221 	/* 2. eliminate each . path name element */
222 	for (i = path; *i; ++i) {
223 		if ((i == path || *i == '/') &&
224 		    *i != '.' && i[1] == '.' && i[2] == '/') {
225 			/* move also the \0 */
226 			memmove(i, i+2, strlen(i)-1);
227 			i--;
228 		}
229 	}
230 	if (!strcmp(path, ".") || !strcmp(path, "/.")) {
231 		*path = '\0';
232 		return 1;
233 	}
234 
235 	/* 3. eliminate each inner .. along with the preceding non-.. */
236 	for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
237 		if (!path_elide_dotdot(path, i, 3))
238 			return 0;
239 
240 	/* 4. eliminate trailing ..*/
241 	if ((i = strstr(path, "..")) != NULL)
242 		if (!path_elide_dotdot(path, i, 2))
243 			return 0;
244 
245 	return 1;
246 }
247 
248 static int
249 parse_query(struct parser *p)
250 {
251 	p->parsed->query = p->iri;
252 	if (*p->iri == '\0')
253 		return 1;
254 
255 	while (unreserved(*p->iri)
256 	    || sub_delimiters(*p->iri)
257 	    || *p->iri == '/'
258 	    || *p->iri == '?'
259 	    || parse_pct_encoded(p)
260 	    || valid_multibyte_utf8(p))
261 		p->iri++;
262 
263 	if (p->err != NULL)
264 		return 0;
265 
266 	if (*p->iri != '\0' && *p->iri != '#') {
267 		p->err = "illegal character in query";
268 		return 0;
269 	}
270 
271 	if (*p->iri != '\0') {
272 		*p->iri = '\0';
273 		p->iri++;
274 	}
275 
276 	return 1;
277 }
278 
279 /* don't even bother */
280 static int
281 parse_fragment(struct parser *p)
282 {
283 	p->parsed->fragment = p->iri;
284 	return 1;
285 }
286 
287 /* XXX: is it too broad? */
288 /* *(pchar / "/") */
289 static int
290 parse_path(struct parser *p)
291 {
292 	char c;
293 
294 	p->parsed->path = p->iri;
295 	if (*p->iri == '\0') {
296 		p->parsed->query = p->parsed->fragment = p->iri;
297 		return 1;
298 	}
299 
300 	while (unreserved(*p->iri)
301 	    || sub_delimiters(*p->iri)
302 	    || *p->iri == '/'
303 	    || parse_pct_encoded(p)
304 	    || valid_multibyte_utf8(p))
305 		p->iri++;
306 
307 	if (p->err != NULL)
308 		return 0;
309 
310 	if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
311 		p->err = "illegal character in path";
312 		return 0;
313 	}
314 
315 	if (*p->iri != '\0') {
316 		c = *p->iri;
317 		*p->iri = '\0';
318 		p->iri++;
319 
320 		if (c == '#') {
321 			if (!parse_fragment(p))
322 				return 0;
323 		} else
324 			if (!parse_query(p) || !parse_fragment(p))
325 				return 0;
326 	}
327 
328 	if (!path_clean(p->parsed->path)) {
329 		p->err = "illegal path";
330 		return 0;
331 	}
332 
333 	return 1;
334 }
335 
336 int
337 parse_iri(char *iri, struct iri *ret, const char **err_ret)
338 {
339 	char *end;
340 	struct parser p = {iri, ret, NULL};
341 
342 	bzero(ret, sizeof(*ret));
343 
344 	/* initialize optional stuff to the empty string */
345 	end = iri + strlen(iri);
346 	p.parsed->port = end;
347 	p.parsed->path = end;
348 	p.parsed->query = end;
349 	p.parsed->fragment = end;
350 
351 	if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
352 		*err_ret = p.err;
353 		return 0;
354 	}
355 
356 	*err_ret = NULL;
357 	return 1;
358 }
359 
360 int
361 trim_req_iri(char *iri)
362 {
363 	char *i;
364 
365 	if ((i = strstr(iri, "\r\n")) == NULL)
366 		return 0;
367 	*i = '\0';
368 	return 1;
369 }