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 	p->iri++;
81 	while (isalnum(*p->iri)
82 	    || *p->iri == '+'
83 	    || *p->iri == '-'
84 	    || *p->iri == '.')
85 		p->iri++;
86 
87 	if (*p->iri != ':') {
88 		p->err = "illegal character in scheme";
89 		return 0;
90 	}
91 
92 	*p->iri = '\0';
93 	if (*(++p->iri) != '/' || *(++p->iri) != '/') {
94 		p->err = "invalid marker after scheme";
95 		return 0;
96 	}
97 
98 	p->iri++;
99 	return 1;
100 }
101 
102 /* *DIGIT */
103 static int
104 parse_port(struct parser *p)
105 {
106 	uint32_t i = 0;
107 
108 	p->parsed->port = p->iri;
109 
110 	for (; isdigit(*p->iri); p->iri++) {
111 		i = i * 10 + *p->iri - '0';
112 		if (i > UINT16_MAX) {
113 			p->err = "port number too large";
114 			return 0;
115 		}
116 	}
117 
118 	if (*p->iri != '/' && *p->iri != '\0') {
119 		p->err = "illegal character in port number";
120 		return 0;
121 	}
122 
123 	p->parsed->port_no = i;
124 
125 	if (*p->iri != '\0') {
126 		*p->iri = '\0';
127 		p->iri++;
128 	}
129 
130 	return 1;
131 }
132 
133 /* TODO: add support for ip-literal and ipv4addr ? */
134 /* *( unreserved / sub-delims / pct-encoded ) */
135 static int
136 parse_authority(struct parser *p)
137 {
138 	p->parsed->host = p->iri;
139 
140 	while (unreserved(*p->iri)
141 	    || sub_delimiters(*p->iri)
142 	    || parse_pct_encoded(p))
143 		p->iri++;
144 
145 	if (p->err != NULL)
146 		return 0;
147 
148 	if (*p->iri == ':') {
149 		*p->iri = '\0';
150 		p->iri++;
151 		return parse_port(p);
152 	}
153 
154 	if (*p->iri == '/') {
155 		*p->iri = '\0';
156 		p->iri++;
157 		return 1;
158 	}
159 
160 	if (*p->iri == '\0')
161 		return 1;
162 
163 	p->err = "illegal character in authority section";
164 	return 0;
165 }
166 
167 /* Routine for path_clean.  Elide the pointed .. with the preceding
168  * element.  Return 0 if it's not possible.  incr is the length of
169  * the increment, 3 for ../ and 2 for .. */
170 static int
171 path_elide_dotdot(char *path, char *i, int incr)
172 {
173 	char *j;
174 
175 	if (i == path)
176 		return 0;
177 	for (j = i-2; j != path && *j != '/'; j--)
178                 /* noop */ ;
179 	if (*j == '/')
180 		j++;
181 	i += incr;
182 	memmove(j, i, strlen(i)+1);
183 	return 1;
184 }
185 
186 /*
187  * Use an algorithm similar to the one implemented in go' path.Clean:
188  *
189  * 1. Replace multiple slashes with a single slash
190  * 2. Eliminate each . path name element
191  * 3. Eliminate each inner .. along with the non-.. element that precedes it
192  * 4. Eliminate trailing .. if possible or error (go would only discard)
193  *
194  * Unlike path.Clean, this function return the empty string if the
195  * original path is equivalent to "/".
196  */
197 static int
198 path_clean(char *path)
199 {
200 	char *i;
201 
202 	/* 1. replace multiple slashes with a single one */
203 	for (i = path; *i; ++i) {
204 		if (*i == '/' && *(i+1) == '/') {
205 			memmove(i, i+1, strlen(i)); /* move also the \0 */
206 			i--;
207 		}
208 	}
209 
210 	/* 2. eliminate each . path name element */
211 	for (i = path; *i; ++i) {
212 		if ((i == path || *i == '/') && *(i+1) == '.' &&
213 		    *(i+2) == '/') {
214 			/* move also the \0 */
215 			memmove(i, i+2, strlen(i)-1);
216 			i--;
217 		}
218 	}
219 	if (!strcmp(path, ".") || !strcmp(path, "/.")) {
220 		*path = '\0';
221 		return 1;
222 	}
223 
224 	/* 3. eliminate each inner .. along with the preceding non-.. */
225 	for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
226 		if (!path_elide_dotdot(path, i, 3))
227 			return 0;
228 
229 	/* 4. eliminate trailing ..*/
230 	if ((i = strstr(path, "..")) != NULL)
231 		if (!path_elide_dotdot(path, i, 2))
232 			return 0;
233 
234 	return 1;
235 }
236 
237 static int
238 parse_query(struct parser *p)
239 {
240 	p->parsed->query = p->iri;
241 	if (*p->iri == '\0')
242 		return 1;
243 
244 	while (unreserved(*p->iri)
245 	    || sub_delimiters(*p->iri)
246 	    || *p->iri == '/'
247 	    || *p->iri == '?'
248 	    || parse_pct_encoded(p)
249 	    || valid_multibyte_utf8(p))
250 		p->iri++;
251 
252 	if (p->err != NULL)
253 		return 0;
254 
255 	if (*p->iri != '\0' && *p->iri != '#') {
256 		p->err = "illegal character in query";
257 		return 0;
258 	}
259 
260 	if (*p->iri != '\0') {
261 		*p->iri = '\0';
262 		p->iri++;
263 	}
264 
265 	return 1;
266 }
267 
268 /* don't even bother */
269 static int
270 parse_fragment(struct parser *p)
271 {
272 	p->parsed->fragment = p->iri;
273 	return 1;
274 }
275 
276 /* XXX: is it too broad? */
277 /* *(pchar / "/") */
278 static int
279 parse_path(struct parser *p)
280 {
281 	char c;
282 
283 	p->parsed->path = p->iri;
284 	if (*p->iri == '\0') {
285 		p->parsed->query = p->parsed->fragment = p->iri;
286 		return 1;
287 	}
288 
289 	while (unreserved(*p->iri)
290 	    || sub_delimiters(*p->iri)
291 	    || *p->iri == '/'
292 	    || parse_pct_encoded(p)
293 	    || valid_multibyte_utf8(p))
294 		p->iri++;
295 
296 	if (p->err != NULL)
297 		return 0;
298 
299 	if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
300 		p->err = "illegal character in path";
301 		return 0;
302 	}
303 
304 	if (*p->iri != '\0') {
305 		c = *p->iri;
306 		*p->iri = '\0';
307 		p->iri++;
308 
309 		if (c == '#') {
310 			if (!parse_fragment(p))
311 				return 0;
312 		} else
313 			if (!parse_query(p) || !parse_fragment(p))
314 				return 0;
315 	}
316 
317 	if (!path_clean(p->parsed->path)) {
318 		p->err = "illegal path";
319 		return 0;
320 	}
321 
322 	return 1;
323 }
324 
325 int
326 parse_iri(char *iri, struct iri *ret, const char **err_ret)
327 {
328 	char *end;
329 	struct parser p = {iri, ret, NULL};
330 
331 	bzero(ret, sizeof(*ret));
332 
333 	/* initialize optional stuff to the empty string */
334 	end = iri + strlen(iri);
335 	p.parsed->port = end;
336 	p.parsed->path = end;
337 	p.parsed->query = end;
338 	p.parsed->fragment = end;
339 
340 	if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
341 		*err_ret = p.err;
342 		return 0;
343 	}
344 
345 	*err_ret = NULL;
346 	return 1;
347 }
348 
349 int
350 trim_req_iri(char *iri)
351 {
352 	char *i;
353 
354 	if ((i = strstr(iri, "\r\n")) == NULL)
355 		return 0;
356 	*i = '\0';
357 	return 1;
358 }