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