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[1] != '/' || p->iri[2] != '/') {
101 		p->err = "invalid marker after scheme";
102 		return 0;
103 	}
104 
105 	p->iri += 3;
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 	} else
164 		p->parsed->port_no = 1965;
165 
166 	if (*p->iri == '/') {
167 		*p->iri = '\0';
168 		p->iri++;
169 		return 1;
170 	}
171 
172 	if (*p->iri == '\0')
173 		return 1;
174 
175 	p->err = "illegal character in authority section";
176 	return 0;
177 }
178 
179 /* Routine for path_clean.  Elide the pointed .. with the preceding
180  * element.  Return 0 if it's not possible.  incr is the length of
181  * the increment, 3 for ../ and 2 for .. */
182 static int
183 path_elide_dotdot(char *path, char *i, int incr)
184 {
185 	char *j;
186 
187 	if (i == path)
188 		return 0;
189 	for (j = i-2; j != path && *j != '/'; j--)
190                 /* noop */ ;
191 	if (*j == '/')
192 		j++;
193 	i += incr;
194 	memmove(j, i, strlen(i)+1);
195 	return 1;
196 }
197 
198 /*
199  * Use an algorithm similar to the one implemented in go' path.Clean:
200  *
201  * 1. Replace multiple slashes with a single slash
202  * 2. Eliminate each . path name element
203  * 3. Eliminate each inner .. along with the non-.. element that precedes it
204  * 4. Eliminate trailing .. if possible or error (go would only discard)
205  *
206  * Unlike path.Clean, this function return the empty string if the
207  * original path is equivalent to "/".
208  */
209 static int
210 path_clean(char *path)
211 {
212 	char *i;
213 
214 	/* 1. replace multiple slashes with a single one */
215 	for (i = path; *i; ++i) {
216 		if (*i == '/' && *(i+1) == '/') {
217 			memmove(i, i+1, strlen(i)); /* move also the \0 */
218 			i--;
219 		}
220 	}
221 
222 	/* 2. eliminate each . path name element */
223 	for (i = path; *i; ++i) {
224 		if ((i == path || *i == '/') &&
225 		    *i != '.' && i[1] == '.' && i[2] == '/') {
226 			/* move also the \0 */
227 			memmove(i, i+2, strlen(i)-1);
228 			i--;
229 		}
230 	}
231 	if (!strcmp(path, ".") || !strcmp(path, "/.")) {
232 		*path = '\0';
233 		return 1;
234 	}
235 
236 	/* 3. eliminate each inner .. along with the preceding non-.. */
237 	for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
238 		if (!path_elide_dotdot(path, i, 3))
239 			return 0;
240 
241 	/* 4. eliminate trailing ..*/
242 	if ((i = strstr(path, "..")) != NULL)
243 		if (!path_elide_dotdot(path, i, 2))
244 			return 0;
245 
246 	return 1;
247 }
248 
249 static int
250 parse_query(struct parser *p)
251 {
252 	p->parsed->query = p->iri;
253 	if (*p->iri == '\0')
254 		return 1;
255 
256 	while (unreserved(*p->iri)
257 	    || sub_delimiters(*p->iri)
258 	    || *p->iri == '/'
259 	    || *p->iri == '?'
260 	    || parse_pct_encoded(p)
261 	    || valid_multibyte_utf8(p))
262 		p->iri++;
263 
264 	if (p->err != NULL)
265 		return 0;
266 
267 	if (*p->iri != '\0' && *p->iri != '#') {
268 		p->err = "illegal character in query";
269 		return 0;
270 	}
271 
272 	if (*p->iri != '\0') {
273 		*p->iri = '\0';
274 		p->iri++;
275 	}
276 
277 	return 1;
278 }
279 
280 /* don't even bother */
281 static int
282 parse_fragment(struct parser *p)
283 {
284 	p->parsed->fragment = p->iri;
285 	return 1;
286 }
287 
288 /* XXX: is it too broad? */
289 /* *(pchar / "/") */
290 static int
291 parse_path(struct parser *p)
292 {
293 	char c;
294 
295 	p->parsed->path = p->iri;
296 	if (*p->iri == '\0') {
297 		p->parsed->query = p->parsed->fragment = p->iri;
298 		return 1;
299 	}
300 
301 	while (unreserved(*p->iri)
302 	    || sub_delimiters(*p->iri)
303 	    || *p->iri == '/'
304 	    || parse_pct_encoded(p)
305 	    || valid_multibyte_utf8(p))
306 		p->iri++;
307 
308 	if (p->err != NULL)
309 		return 0;
310 
311 	if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
312 		p->err = "illegal character in path";
313 		return 0;
314 	}
315 
316 	if (*p->iri != '\0') {
317 		c = *p->iri;
318 		*p->iri = '\0';
319 		p->iri++;
320 
321 		if (c == '#') {
322 			if (!parse_fragment(p))
323 				return 0;
324 		} else
325 			if (!parse_query(p) || !parse_fragment(p))
326 				return 0;
327 	}
328 
329 	if (!path_clean(p->parsed->path)) {
330 		p->err = "illegal path";
331 		return 0;
332 	}
333 
334 	return 1;
335 }
336 
337 int
338 parse_iri(char *iri, struct iri *ret, const char **err_ret)
339 {
340 	char *end;
341 	struct parser p = {iri, ret, NULL};
342 
343 	bzero(ret, sizeof(*ret));
344 
345 	/* initialize optional stuff to the empty string */
346 	end = iri + strlen(iri);
347 	p.parsed->port = end;
348 	p.parsed->path = end;
349 	p.parsed->query = end;
350 	p.parsed->fragment = end;
351 
352 	if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
353 		*err_ret = p.err;
354 		return 0;
355 	}
356 
357 	*err_ret = NULL;
358 	return 1;
359 }
360 
361 int
362 trim_req_iri(char *iri)
363 {
364 	char *i;
365 
366 	if ((i = strstr(iri, "\r\n")) == NULL)
367 		return 0;
368 	*i = '\0';
369 	return 1;
370 }