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[0] != '%')
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 	    || valid_multibyte_utf8(p)) {
152 		/* normalize the host name. */
153 		if (*p->iri < 0x7F)
154 			*p->iri = tolower(*p->iri);
155 		p->iri++;
156 	}
157 
158 	if (p->err != NULL)
159 		return 0;
160 
161 	if (*p->iri == ':') {
162 		*p->iri = '\0';
163 		p->iri++;
164 		return parse_port(p);
165 	} else
166 		p->parsed->port_no = 1965;
167 
168 	if (*p->iri == '/') {
169 		*p->iri = '\0';
170 		p->iri++;
171 		return 1;
172 	}
173 
174 	if (*p->iri == '\0')
175 		return 1;
176 
177 	p->err = "illegal character in authority section";
178 	return 0;
179 }
180 
181 /* Routine for path_clean.  Elide the pointed .. with the preceding
182  * element.  Return 0 if it's not possible.  incr is the length of
183  * the increment, 3 for ../ and 2 for .. */
184 static int
185 path_elide_dotdot(char *path, char *i, int incr)
186 {
187 	char *j;
188 
189 	if (i == path)
190 		return 0;
191 	for (j = i-2; j != path && *j != '/'; j--)
192                 /* noop */ ;
193 	if (*j == '/')
194 		j++;
195 	i += incr;
196 	memmove(j, i, strlen(i)+1);
197 	return 1;
198 }
199 
200 /*
201  * Use an algorithm similar to the one implemented in go' path.Clean:
202  *
203  * 1. Replace multiple slashes with a single slash
204  * 2. Eliminate each . path name element
205  * 3. Eliminate each inner .. along with the non-.. element that precedes it
206  * 4. Eliminate trailing .. if possible or error (go would only discard)
207  *
208  * Unlike path.Clean, this function return the empty string if the
209  * original path is equivalent to "/".
210  */
211 static int
212 path_clean(char *path)
213 {
214 	char *i;
215 
216 	/* 1. replace multiple slashes with a single one */
217 	for (i = path; *i; ++i) {
218 		if (*i == '/' && *(i+1) == '/') {
219 			memmove(i, i+1, strlen(i)); /* move also the \0 */
220 			i--;
221 		}
222 	}
223 
224 	/* 2. eliminate each . path name element */
225 	for (i = path; *i; ++i) {
226 		if ((i == path || *i == '/') &&
227 		    *i != '.' && i[1] == '.' && i[2] == '/') {
228 			/* move also the \0 */
229 			memmove(i, i+2, strlen(i)-1);
230 			i--;
231 		}
232 	}
233 	if (!strcmp(path, ".") || !strcmp(path, "/.")) {
234 		*path = '\0';
235 		return 1;
236 	}
237 
238 	/* 3. eliminate each inner .. along with the preceding non-.. */
239 	for (i = strstr(path, "../"); i != NULL; i = strstr(path, ".."))
240 		if (!path_elide_dotdot(path, i, 3))
241 			return 0;
242 
243 	/* 4. eliminate trailing ..*/
244 	if ((i = strstr(path, "..")) != NULL)
245 		if (!path_elide_dotdot(path, i, 2))
246 			return 0;
247 
248 	return 1;
249 }
250 
251 static int
252 parse_query(struct parser *p)
253 {
254 	p->parsed->query = p->iri;
255 	if (*p->iri == '\0')
256 		return 1;
257 
258 	while (unreserved(*p->iri)
259 	    || sub_delimiters(*p->iri)
260 	    || *p->iri == '/'
261 	    || *p->iri == '?'
262 	    || parse_pct_encoded(p)
263 	    || valid_multibyte_utf8(p))
264 		p->iri++;
265 
266 	if (p->err != NULL)
267 		return 0;
268 
269 	if (*p->iri != '\0' && *p->iri != '#') {
270 		p->err = "illegal character in query";
271 		return 0;
272 	}
273 
274 	if (*p->iri != '\0') {
275 		*p->iri = '\0';
276 		p->iri++;
277 	}
278 
279 	return 1;
280 }
281 
282 /* don't even bother */
283 static int
284 parse_fragment(struct parser *p)
285 {
286 	p->parsed->fragment = p->iri;
287 	return 1;
288 }
289 
290 /* XXX: is it too broad? */
291 /* *(pchar / "/") */
292 static int
293 parse_path(struct parser *p)
294 {
295 	char c;
296 
297 	/* trim initial slashes */
298 	while (*p->iri == '/')
299 		p->iri++;
300 
301 	p->parsed->path = p->iri;
302 	if (*p->iri == '\0') {
303 		p->parsed->query = p->parsed->fragment = p->iri;
304 		return 1;
305 	}
306 
307 	while (unreserved(*p->iri)
308 	    || sub_delimiters(*p->iri)
309 	    || *p->iri == '/'
310 	    || parse_pct_encoded(p)
311 	    || valid_multibyte_utf8(p))
312 		p->iri++;
313 
314 	if (p->err != NULL)
315 		return 0;
316 
317 	if (*p->iri != '\0' && *p->iri != '?' && *p->iri != '#') {
318 		p->err = "illegal character in path";
319 		return 0;
320 	}
321 
322 	if (*p->iri != '\0') {
323 		c = *p->iri;
324 		*p->iri = '\0';
325 		p->iri++;
326 
327 		if (c == '#') {
328 			if (!parse_fragment(p))
329 				return 0;
330 		} else
331 			if (!parse_query(p) || !parse_fragment(p))
332 				return 0;
333 	}
334 
335 	if (!path_clean(p->parsed->path)) {
336 		p->err = "illegal path";
337 		return 0;
338 	}
339 
340 	return 1;
341 }
342 
343 int
344 parse_iri(char *iri, struct iri *ret, const char **err_ret)
345 {
346 	char *end;
347 	struct parser p = {iri, ret, NULL};
348 
349 	bzero(ret, sizeof(*ret));
350 
351 	/* initialize optional stuff to the empty string */
352 	end = iri + strlen(iri);
353 	p.parsed->host = end;
354 	p.parsed->port = end;
355 	p.parsed->path = end;
356 	p.parsed->query = end;
357 	p.parsed->fragment = end;
358 
359 	if (!parse_scheme(&p) || !parse_authority(&p) || !parse_path(&p)) {
360 		*err_ret = p.err;
361 		return 0;
362 	}
363 
364 	*err_ret = NULL;
365 	return 1;
366 }
367 
368 int
369 trim_req_iri(char *iri, const char **err)
370 {
371 	char *i;
372 
373 	if ((i = strstr(iri, "\r\n")) == NULL) {
374 		*err = "missing CRLF";
375 		return 0;
376 	}
377 	*i = '\0';
378 	return 1;
379 }