1 /*
2  * Copyright (c) 2021 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 "compat.h"
18 
19 #include <stdio.h>
20 #include <stdlib.h>
21 #include <string.h>
22 
23 #include "parser.h"
24 #include "utils.h"
25 
26 #ifndef LINE_MAX
27 #define LINE_MAX 2048
28 #endif
29 
30 struct gm_selector {
31 	char		 type;
32 	const char	*ds;
33 	const char	*selector;
34 	const char	*addr;
35 	const char	*port;
36 };
37 
38 static void	gm_parse_selector(char *, struct gm_selector *);
39 
40 static int	gm_parse(struct parser *, const char *, size_t);
41 static int	gm_foreach_line(struct parser *, const char *, size_t);
42 static int	gm_free(struct parser *);
43 static int	gm_serialize(struct parser *, FILE *);
44 
45 void
46 gophermap_initparser(struct parser *p)
47 {
48 	memset(p, 0, sizeof(*p));
49 
50 	p->name = "gophermap";
51 	p->parse = &gm_parse;
52 	p->free = &gm_free;
53 	p->serialize = &gm_serialize;
54 
55 	TAILQ_INIT(&p->head);
56 }
57 
58 static void
59 gm_parse_selector(char *line, struct gm_selector *s)
60 {
61 	s->type = *line++;
62 	s->ds = line;
63 	s->selector = "";
64 	s->addr = "";
65 	s->port = "";
66 
67 	if ((line = strchr(line, '\t')) == NULL)
68 		return;
69 	*line++ = '\0';
70 	s->selector = line;
71 
72 	if ((line = strchr(line, '\t')) == NULL)
73 		return;
74 	*line++ = '\0';
75 	s->addr = line;
76 
77 	if ((line = strchr(line, '\t')) == NULL)
78 		return;
79 	*line++ = '\0';
80 	s->port = line;
81 }
82 
83 static int
84 gm_parse(struct parser *p, const char *buf, size_t size)
85 {
86 	return parser_foreach_line(p, buf, size, gm_foreach_line);
87 }
88 
89 static inline int
90 emit_line(struct parser *p, enum line_type type, struct gm_selector *s)
91 {
92 	struct line *l;
93 	char buf[LINE_MAX], b[2] = {0};
94 
95 	if ((l = calloc(1, sizeof(*l))) == NULL)
96 		goto err;
97 
98 	if ((l->line = strdup(s->ds)) == NULL)
99 		goto err;
100 
101 	switch (l->type = type) {
102 	case LINE_LINK:
103 		if (s->type == 'h' && !strncmp(s->selector, "URL:", 4)) {
104 			strlcpy(buf, s->selector+4, sizeof(buf));
105 		} else {
106 			strlcpy(buf, "gopher://", sizeof(buf));
107 			strlcat(buf, s->addr, sizeof(buf));
108 			strlcat(buf, ":", sizeof(buf));
109 			strlcat(buf, s->port, sizeof(buf));
110 			strlcat(buf, "/", sizeof(buf));
111 			b[0] = s->type;
112 			strlcat(buf, b, sizeof(buf));
113 			if (*s->selector != '/')
114 				strlcat(buf, "/", sizeof(buf));
115 			strlcat(buf, s->selector, sizeof(buf));
116 		}
117 
118 		if ((l->alt = strdup(buf)) == NULL)
119 			goto err;
120 		break;
121 
122 	default:
123 		break;
124 	}
125 
126 	TAILQ_INSERT_TAIL(&p->head, l, lines);
127 
128 	return 1;
129 
130 err:
131 	if (l != NULL) {
132 		free(l->line);
133 		free(l->alt);
134 		free(l);
135 	}
136 	return 0;
137 }
138 
139 static int
140 gm_foreach_line(struct parser *p, const char *line, size_t linelen)
141 {
142 	char buf[LINE_MAX] = {0};
143 	struct gm_selector s = {0};
144 
145 	memcpy(buf, line, MIN(sizeof(buf)-1, linelen));
146 	gm_parse_selector(buf, &s);
147 
148 	switch (s.type) {
149 	case '0':	/* text file */
150 	case '1':	/* gopher submenu */
151 	case '2':	/* CCSO nameserver */
152 	case '4':	/* binhex-encoded file */
153 	case '5':	/* DOS file */
154 	case '6':	/* uuencoded file */
155 	case '7':	/* full-text search */
156 	case '8':	/* telnet */
157 	case '9':	/* binary file */
158 	case '+':	/* mirror or alternate server */
159 	case 'g':	/* gif */
160 	case 'I':	/* image */
161 	case 'T':	/* telnet 3270 */
162 	case ':':	/* gopher+: bitmap image */
163 	case ';':	/* gopher+: movie file */
164 	case 'd':	/* non-canonical: doc */
165 	case 'h':	/* non-canonical: html file */
166 	case 's':	/* non-canonical: sound file */
167 		if (!emit_line(p, LINE_LINK, &s))
168 			return 0;
169 		break;
170 
171 	case 'i':	/* non-canonical: message */
172 		if (!emit_line(p, LINE_TEXT, &s))
173 			return 0;
174 		break;
175 
176 	case '3':	/* error code */
177 		if (!emit_line(p, LINE_QUOTE, &s))
178 			return 0;
179 		break;
180 	}
181 
182 	return 1;
183 }
184 
185 static int
186 gm_free(struct parser *p)
187 {
188 	/* flush the buffer */
189 	if (p->len != 0)
190 		gm_foreach_line(p, p->buf, p->len);
191 
192 	free(p->buf);
193 
194 	return 1;
195 }
196 
197 static inline const char *
198 gopher_skip_selector(const char *path, int *ret_type)
199 {
200 	*ret_type = 0;
201 
202 	if (!strcmp(path, "/") || *path == '\0') {
203 		*ret_type = '1';
204 		return path;
205 	}
206 
207 	if (*path != '/')
208 		return path;
209 	path++;
210 
211 	switch (*ret_type = *path) {
212 	case '0':
213 	case '1':
214 	case '7':
215 		break;
216 
217 	default:
218 		*ret_type = 0;
219 		path -= 1;
220 		return path;
221 	}
222 
223 	return ++path;
224 }
225 
226 static int
227 serialize_link(struct line *line, const char *text, FILE *fp)
228 {
229 	size_t		 portlen = 0;
230 	int		 type;
231 	const char	*uri, *endhost, *port, *path, *colon;
232 
233 	if ((uri = line->alt) == NULL)
234 		return -1;
235 
236 	if (strncmp(uri, "gopher://", 9) != 0)
237 		return fprintf(fp, "h%s\tURL:%s\terror.host\t1\n",
238 		    text, line->alt);
239 
240 	uri += 9; /* skip gopher:// */
241 
242 	path = strchr(uri, '/');
243 	colon = strchr(uri, ':');
244 
245 	if (path != NULL && colon > path)
246 		colon = NULL;
247 
248 	if ((endhost = colon) == NULL &&
249 	    (endhost = path) == NULL)
250 		endhost = strchr(uri, '\0');
251 
252 	if (colon != NULL) {
253 		for (port = colon+1; *port && *port != '/'; ++port)
254 			++portlen;
255 		port = colon+1;
256 	} else {
257 		port = "70";
258 		portlen = 2;
259 	}
260 
261 	if (path == NULL) {
262 		type = '1';
263 		path = "";
264 	} else
265 		path = gopher_skip_selector(path, &type);
266 
267 	return fprintf(fp, "%c%s\t%s\t%.*s\t%.*s\n", type, text,
268 	    path, (int)(endhost - uri), uri, (int)portlen, port);
269 }
270 
271 static int
272 gm_serialize(struct parser *p, FILE *fp)
273 {
274 	struct line	*line;
275 	const char	*text;
276 	int		 r;
277 
278 	TAILQ_FOREACH(line, &p->head, lines) {
279 		if ((text = line->line) == NULL)
280 			text = "";
281 
282 		switch (line->type) {
283 		case LINE_LINK:
284 			r = serialize_link(line, text, fp);
285 			break;
286 
287 		case LINE_TEXT:
288 			r = fprintf(fp, "i%s\t\terror.host\t1\n", text);
289 			break;
290 
291 		case LINE_QUOTE:
292 			r = fprintf(fp, "3%s\t\terror.host\t1\n", text);
293 			break;
294 
295 		default:
296 			/* unreachable */
297 			abort();
298 		}
299 
300 		if (r == -1)
301 			return 0;
302 	}
303 
304 	return 1;
305 }