1 /*
2  * Copyright (c) 2020, 2022 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 <err.h>
18 #include <stdio.h>
19 #include <string.h>
20 
21 #include "../gmid.h"
22 
23 #define ENCTEST(buf, len, raw, exp)				\
24 	if (encode_path(buf, len, raw) == -1) {			\
25 		fprintf(stderr, "%s:%d: failed to encode: %s\n", \
26 		    __FILE__, __LINE__, raw);			\
27 		exit(1);					\
28 	}							\
29 	if (strcmp(buf, exp) != 0) {				\
30 		fprintf(stderr, "%s:%d: error: "		\
31 		    "unexpected encoding: got %s, want %s\n",	\
32 		    __FILE__, __LINE__, buf, exp);		\
33 		exit(1);					\
34 	}
35 
36 #define TEST(iri, fail, exp, descr)				\
37 	if (!run_test(iri, fail, exp)) {			\
38 		fprintf(stderr, "%s:%d: error: %s\n",		\
39 		    __FILE__, __LINE__, descr);			\
40 		exit(1);					\
41 	}
42 
43 #define IRI(schema, host, port, path, query, frag)		\
44 	((struct iri){(char*)schema, (char*)host, (char*)port,	\
45 		 0, (char*)path, (char*)query,			\
46 		 (char*)frag})
47 
48 #define DIFF(wanted, got, field)					\
49 	if (wanted->field == NULL || got->field == NULL ||		\
50 	    strcmp(wanted->field, got->field)) {			\
51 		fprintf(stderr, #field ":\n\tgot: %s\n\twanted: %s\n",	\
52 		    got->field, wanted->field);				\
53 		return 0;						\
54 	}
55 
56 #define PASS 0
57 #define FAIL 1
58 
59 int	diff_iri(struct iri*, struct iri*);
60 int	run_test(const char*, int, struct iri);
61 
62 int
63 diff_iri(struct iri *p, struct iri *exp)
64 {
65         DIFF(p, exp, schema);
66         DIFF(p, exp, host);
67         DIFF(p, exp, port);
68         DIFF(p, exp, path);
69         DIFF(p, exp, query);
70         DIFF(p, exp, fragment);
71 	return 1;
72 }
73 
74 int
75 run_test(const char *iri, int should_fail, struct iri expected)
76 {
77 	int failed, ok = 1;
78 	char *iri_copy;
79 	struct iri parsed;
80 	const char *error;
81 
82 	if ((iri_copy = strdup(iri)) == NULL)
83 		err(1, "strdup");
84 
85 	fprintf(stderr, "=> %s\n", iri);
86 	failed = !parse_iri(iri_copy, &parsed, &error);
87 
88 	if (failed && should_fail)
89 		goto done;
90 
91 	if (error != NULL)
92 		fprintf(stderr, "> %s\n", error);
93 
94 	ok = !failed && !should_fail;
95 	if (ok)
96 		ok = diff_iri(&expected, &parsed);
97 
98 done:
99 	free(iri_copy);
100 	return ok;
101 }
102 
103 int
104 main(void)
105 {
106 	char buf[32];
107 	struct iri empty = IRI("", "", "", "", "", "");
108 
109 	ENCTEST(buf, sizeof(buf), "hello world", "hello%20world");
110 	ENCTEST(buf, sizeof(buf), "hello\nworld", "hello%0Aworld");
111 	ENCTEST(buf, sizeof(buf), "hello\r\nworld", "hello%0D%0Aworld");
112 
113 	TEST("http://omarpolo.com",
114 	    PASS,
115 	    IRI("http", "omarpolo.com", "", "", "", ""),
116 	    "can parse iri with empty path");
117 
118 	/* schema */
119 	TEST("omarpolo.com", FAIL, empty, "FAIL when the schema is missing");
120 	TEST("gemini:/omarpolo.com", FAIL, empty, "FAIL with invalid marker");
121 	TEST("gemini//omarpolo.com", FAIL, empty, "FAIL with invalid marker");
122 	TEST("h!!p://omarpolo.com", FAIL, empty, "FAIL with invalid schema");
123 	TEST("GEMINI://omarpolo.com",
124 	    PASS,
125 	    IRI("gemini", "omarpolo.com", "", "", "", ""),
126 	    "Schemas are case insensitive.");
127 
128 	/* authority */
129 	TEST("gemini://omarpolo.com",
130 	    PASS,
131 	    IRI("gemini", "omarpolo.com", "", "", "", ""),
132 	    "can parse authority with empty path");
133 	TEST("gemini://omarpolo.com/",
134 	    PASS,
135 	    IRI("gemini", "omarpolo.com", "", "", "", ""),
136 	    "can parse authority with empty path (alt)")
137 	TEST("gemini://omarpolo.com:1965",
138 	    PASS,
139 	    IRI("gemini", "omarpolo.com", "1965", "", "", ""),
140 	    "can parse with port and empty path");
141 	TEST("gemini://omarpolo.com:1965/",
142 	    PASS,
143 	    IRI("gemini", "omarpolo.com", "1965", "", "", ""),
144 	    "can parse with port and empty path")
145 	TEST("gemini://omarpolo.com:196s",
146 	    FAIL,
147 	    empty,
148 	    "FAIL with invalid port number");
149 	TEST("gemini://OmArPoLo.CoM",
150 	    PASS,
151 	    IRI("gemini", "omarpolo.com", "", "", "", ""),
152 	    "host is case-insensitive");
153 	TEST("gemini://xn--nave-6pa.omarpolo.com",
154 	    PASS,
155 	    IRI("gemini", "xn--nave-6pa.omarpolo.com", "", "", "", ""),
156 	    "Can parse punycode-encoded hostnames");
157 	TEST("gemini://naïve.omarpolo.com",
158 	    PASS,
159 	    IRI("gemini", "naïve.omarpolo.com", "", "", "", ""),
160 	    "Accept non punycode-encoded hostnames");
161 	TEST("gemini://na%c3%afve.omarpolo.com",
162 	    PASS,
163 	    IRI("gemini", "naïve.omarpolo.com", "", "", "", ""),
164 	    "Can percent decode hostnames");
165 
166 	/* path */
167 	TEST("gemini://omarpolo.com/foo/bar/baz",
168 	    PASS,
169 	    IRI("gemini", "omarpolo.com", "", "foo/bar/baz", "", ""),
170 	    "parse simple paths");
171 	TEST("gemini://omarpolo.com/foo//bar///baz",
172 	    PASS,
173 	    IRI("gemini", "omarpolo.com", "", "foo/bar/baz", "", ""),
174 	    "parse paths with multiple slashes");
175 	TEST("gemini://omarpolo.com/foo/./bar/./././baz",
176 	    PASS,
177 	    IRI("gemini", "omarpolo.com", "", "foo/bar/baz", "", ""),
178 	    "parse paths with . elements");
179 	TEST("gemini://omarpolo.com/foo/bar/../bar/baz",
180 	    PASS,
181 	    IRI("gemini", "omarpolo.com", "", "foo/bar/baz", "", ""),
182 	    "parse paths with .. elements");
183 	TEST("gemini://omarpolo.com/foo/../foo/bar/../bar/baz/../baz",
184 	    PASS,
185 	    IRI("gemini", "omarpolo.com", "", "foo/bar/baz", "", ""),
186 	    "parse paths with multiple .. elements");
187 	TEST("gemini://omarpolo.com/foo/..",
188 	    PASS,
189             IRI("gemini", "omarpolo.com", "", "", "", ""),
190 	    "parse paths with a trailing ..");
191 	TEST("gemini://omarpolo.com/foo/../",
192 	    PASS,
193             IRI("gemini", "omarpolo.com", "", "", "", ""),
194 	    "parse paths with a trailing ..");
195 	TEST("gemini://omarpolo.com/foo/../..",
196 	    FAIL,
197             empty,
198 	    "reject paths that would escape the root");
199 	TEST("gemini://omarpolo.com/foo/../../",
200 	    FAIL,
201             empty,
202 	    "reject paths that would escape the root")
203 	TEST("gemini://omarpolo.com/foo/../foo/../././/bar/baz/.././.././/",
204 	    PASS,
205             IRI("gemini", "omarpolo.com", "", "", "", ""),
206 	    "parse path with lots of cleaning available");
207 	TEST("gemini://omarpolo.com//foo",
208 	    PASS,
209             IRI("gemini", "omarpolo.com", "", "foo", "", ""),
210 	    "Trim initial slashes");
211 	TEST("gemini://omarpolo.com/////foo",
212 	    PASS,
213             IRI("gemini", "omarpolo.com", "", "foo", "", ""),
214 	    "Trim initial slashes (pt. 2)");
215 	TEST("http://a/b/c/../..",
216 	    PASS,
217 	    IRI("http", "a", "", "", "", ""),
218 	    "avoid infinite loops (see v1.6.1)");
219 	TEST("gemini://example.com/@f:b!(z$&)/baz",
220 	    PASS,
221 	    IRI("gemini", "example.com", "", "@f:b!(z$&)/baz", "", ""),
222 	    "allow @, :, !, (), $ and & in paths");
223 
224 	/* query */
225 	TEST("foo://example.com/foo/?gne",
226 	    PASS,
227 	    IRI("foo", "example.com", "", "foo/", "gne", ""),
228 	    "parse query strings");
229 	TEST("foo://example.com/foo/?gne&foo",
230 	    PASS,
231 	    IRI("foo", "example.com", "", "foo/", "gne&foo", ""),
232 	    "parse query strings");
233 	/* TEST("foo://example.com/foo/?gne%2F", */
234 	/*     PASS, */
235 	/*     IRI("foo", "example.com", "", "foo/", "gne/", ""), */
236 	/*     "parse query strings"); */
237 	TEST("foo://ex.com/robots.txt?name=foobar&url=https://foo.com",
238 	    PASS,
239 	    IRI("foo", "ex.com", "", "robots.txt", "name=foobar&url=https://foo.com", ""),
240 	    "Accepts : in queries");
241 	TEST("foo://ex.com/foo?email=foo@bar.com#quuz",
242 	    PASS,
243 	    IRI("foo", "ex.com", "", "foo", "email=foo@bar.com", "quuz"),
244 	    "Accepts @ in queries");
245 
246 	/* fragment */
247 	TEST("foo://bar.co/#foo",
248 	    PASS,
249 	    IRI("foo", "bar.co", "", "", "", "foo"),
250 	    "can recognize fragments");
251 
252 	/* percent encoding */
253 	TEST("foo://bar.com/caf%C3%A8.gmi",
254 	    PASS,
255 	    IRI("foo", "bar.com", "", "cafè.gmi", "", ""),
256 	    "can decode");
257 	TEST("foo://bar.com/caff%C3%A8%20macchiato.gmi",
258 	    PASS,
259 	    IRI("foo", "bar.com", "", "caffè macchiato.gmi", "", ""),
260 	    "can decode");
261 	TEST("foo://bar.com/caff%C3%A8+macchiato.gmi",
262 	    PASS,
263 	    IRI("foo", "bar.com", "", "caffè+macchiato.gmi", "", ""),
264 	    "can decode");
265 	TEST("foo://bar.com/foo%2F..%2F..",
266 	    FAIL,
267 	    empty,
268 	    "conversion and checking are done in the correct order");
269 	TEST("foo://bar.com/foo%00?baz",
270 	    FAIL,
271 	    empty,
272 	    "rejects %00");
273 
274 	/* IRI */
275         TEST("foo://bar.com/cafè.gmi",
276 	    PASS,
277 	    IRI("foo", "bar.com", "", "cafè.gmi", "" , ""),
278 	    "decode IRI (with a 2-byte utf8 seq)");
279 	TEST("foo://bar.com/世界.gmi",
280 	    PASS,
281 	    IRI("foo", "bar.com", "", "世界.gmi", "" , ""),
282 	    "decode IRI");
283 	TEST("foo://bar.com/😼.gmi",
284 	    PASS,
285 	    IRI("foo", "bar.com", "", "😼.gmi", "" , ""),
286 	    "decode IRI (with a 3-byte utf8 seq)");
287 	TEST("foo://bar.com/😼/𤭢.gmi",
288 	    PASS,
289 	    IRI("foo", "bar.com", "", "😼/𤭢.gmi", "" , ""),
290 	    "decode IRI (with a 3-byte and a 4-byte utf8 seq)");
291 	TEST("foo://bar.com/世界/\xC0\x80",
292 	    FAIL,
293 	    empty,
294 	    "reject invalid sequence (overlong NUL)");
295 
296 	return 0;
297 }