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