op public repos

Blob

Date:: Tue Mar 2 13:35:07 2021 UTC
Message:: initial commit
Actions:: History | Blame | Raw File
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 /*
18  * A streaming gemtext parser.
19  *
20  * TODO:
21  *  - handle NULs
22  *  - UTF8
23  */
24 
25 #include <telescope.h>
26 
27 #include <ctype.h>
28 #include <string.h>
29 #include <stdlib.h>
30 
31 static int	gemtext_parse(struct parser*, const char*, size_t);
32 static void	gemtext_free(struct parser*);
33 
34 static int	parse_text(struct parser*, enum line_type, const char*, size_t);
35 static int	parse_link(struct parser*, enum line_type, const char*, size_t);
36 static int	parse_title(struct parser*, enum line_type, const char*, size_t);
37 static int	parse_item(struct parser*, enum line_type, const char*, size_t);
38 static int	parse_quote(struct parser*, enum line_type, const char*, size_t);
39 static int	parse_pre_start(struct parser*, enum line_type, const char*, size_t);
40 static int	parse_pre_cnt(struct parser*, enum line_type, const char*, size_t);
41 static int	parse_pre_end(struct parser*, enum line_type, const char*, size_t);
42 
43 typedef int (parselinefn)(struct parser*, enum line_type, const char*, size_t);
44 
45 static parselinefn *parsers[] = {
46 	parse_text,		/* LINE_TEXT */
47 	parse_link,		/* LINE_LINK */
48 	parse_title,		/* LINE_TITLE_1 */
49 	parse_title,		/* LINE_TITLE_2 */
50 	parse_title,		/* LINE_TITLE_3 */
51 	parse_item,		/* LINE_ITEM */
52 	parse_quote,		/* LINE_QUOTE */
53 	parse_pre_start,	/* LINE_PRE_START */
54 	parse_pre_cnt,		/* LINE_PRE_CONTENT */
55 	parse_pre_end,		/* LINE_PRE_END */
56 };
57 
58 void
59 gemtext_initparser(struct parser *p)
60 {
61 	memset(p, 0, sizeof(*p));
62 
63 	p->parse = &gemtext_parse;
64 	p->free  = &gemtext_free;
65 }
66 
67 static inline int
68 emit_line(struct parser *p, enum line_type type, char *line, char *alt)
69 {
70 	struct line *l;
71 
72 	if ((l = calloc(1, sizeof(*l))) == NULL)
73 		return 0;
74 
75 	l->type = type;
76 	l->line = line;
77 	l->alt = alt;
78 
79 	if (TAILQ_EMPTY(&p->head))
80 		TAILQ_INSERT_HEAD(&p->head, l, lines);
81 	else
82 		TAILQ_INSERT_TAIL(&p->head, l, lines);
83 
84 	return 1;
85 }
86 
87 static int
88 parse_text(struct parser *p, enum line_type t, const char *buf, size_t len)
89 {
90 	char *l;
91 
92 	if ((l = calloc(1, len+1)) == NULL)
93 		return 0;
94 	memcpy(l, buf, len);
95 	return emit_line(p, t, l, NULL);
96 }
97 
98 static int
99 parse_link(struct parser *p, enum line_type t, const char *buf, size_t len)
100 {
101 	char *l, *u;
102 	const char *url_start;
103 
104 	if (len <= 2)
105 		return emit_line(p, t, NULL, NULL);
106 	buf += 2;
107 	len -= 2;
108 
109 	while (len > 0 && isspace(buf[0])) {
110 		buf++;
111 		len--;
112 	}
113 
114 	if (len == 0)
115 		return emit_line(p, t, NULL, NULL);
116 
117 	url_start = buf;
118 	while (len > 0 && !isspace(buf[0])) {
119 		buf++;
120 		len--;
121 	}
122 
123 	if ((u = calloc(1, buf - url_start + 1)) == NULL)
124 		return 0;
125 	memcpy(u, url_start, buf - url_start);
126 
127 	if (len == 0)
128 		return emit_line(p, t, u, NULL);
129 
130 	while (len > 0) {
131 		buf++;
132 		len--;
133 	}
134 
135 	if (len == 0)
136 		return emit_line(p, t, u, NULL);
137 
138 	if ((l = calloc(1, len + 1)) == NULL)
139 		return 0;
140 
141 	memcpy(l, buf, len);
142 	return emit_line(p, t, u, l);
143 }
144 
145 static int
146 parse_title(struct parser *p, enum line_type t, const char *buf, size_t len)
147 {
148 	char *l;
149 
150 	switch (t) {
151 	case LINE_TITLE_1:
152 		if (len <= 1)
153 			return emit_line(p, t, NULL, NULL);
154 		buf++;
155 		len--;
156 		break;
157 	case LINE_TITLE_2:
158 		if (len <= 2)
159 			return emit_line(p, t, NULL, NULL);
160 		buf += 2;
161 		len -= 2;
162 		break;
163 	case LINE_TITLE_3:
164 		if (len <= 3)
165 			return emit_line(p, t, NULL, NULL);
166 		buf += 3;
167 		len -= 3;
168 		break;
169 	default:
170 		/* unreachable */
171 		abort();
172 	}
173 
174 	while (len > 0 && isspace(buf[0])) {
175 		buf++;
176 		len--;
177 	}
178 
179 	if (len == 0)
180 		return emit_line(p, t, NULL, NULL);
181 
182 	if ((l = calloc(1, len+1)) == NULL)
183 		return 0;
184 	memcpy(l, buf, len);
185 	return emit_line(p, t, l, NULL);
186 }
187 
188 static int
189 parse_item(struct parser *p, enum line_type t, const char *buf, size_t len)
190 {
191 	char *l;
192 
193 	if (len == 1)
194 		return emit_line(p, t, NULL, NULL);
195 
196 	buf++;
197 	len--;
198 
199 	while (len > 0 && isspace(buf[0])) {
200 		buf++;
201 		len--;
202 	}
203 
204 	if (len == 0)
205 		return emit_line(p, t, NULL, NULL);
206 
207 	if ((l = calloc(1, len+1)) == NULL)
208 		return 0;
209 	memcpy(l, buf, len);
210 	return emit_line(p, t, l, NULL);
211 }
212 
213 static int
214 parse_quote(struct parser *p, enum line_type t, const char *buf, size_t len)
215 {
216 	char *l;
217 
218 	if (len == 1)
219 		return emit_line(p, t, NULL, NULL);
220 
221 	buf++;
222 	len--;
223 
224 	while (len > 0 && isspace(buf[0])) {
225 		buf++;
226 		len--;
227 	}
228 
229 	if (len == 0)
230 		return emit_line(p, t, NULL, NULL);
231 
232 	if ((l = calloc(1, len+1)) == NULL)
233 		return 0;
234 	memcpy(l, buf, len);
235 	return emit_line(p, t, l, NULL);
236 }
237 
238 static int
239 parse_pre_start(struct parser *p, enum line_type t, const char *buf, size_t len)
240 {
241 	char *l;
242 
243 	if (len <= 3)
244 		return emit_line(p, t, NULL, NULL);
245 
246 	buf += 3;
247 	len += 3;
248 
249 	while (len > 0 && isspace(buf[0])) {
250 		buf++;
251 		len--;
252 	}
253 
254 	if (len == 0)
255 		return emit_line(p, t, NULL, NULL);
256 
257 	if ((l = calloc(1, len+1)) == NULL)
258 		return 0;
259 
260 	memcpy(l, buf, len);
261 	return emit_line(p, t, NULL, l);
262 }
263 
264 static int
265 parse_pre_cnt(struct parser *p, enum line_type t, const char *buf, size_t len)
266 {
267 	char *l;
268 
269 	if (len == 0)
270 		return emit_line(p, t, NULL, NULL);
271 
272 	if ((l = calloc(1, len+1)) == NULL)
273 		return 0;
274 	memcpy(l, buf, len);
275 	return emit_line(p, t, l, NULL);
276 }
277 
278 static int
279 parse_pre_end(struct parser *p, enum line_type t, const char *buf, size_t len)
280 {
281 	return emit_line(p, t, NULL, NULL);
282 }
283 
284 static inline enum line_type
285 detect_line_type(const char *buf, size_t len, int in_pre)
286 {
287 	size_t i;
288 
289 	if (len == 0)
290 		return LINE_TEXT;
291 
292 	if (in_pre) {
293 		if (len >= 3 &&
294 		    buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
295 			return LINE_PRE_END;
296 		else
297 			return LINE_PRE_CONTENT;
298 	}
299 
300 	switch (*buf) {
301 	case '*': return LINE_ITEM;
302 	case '>': return LINE_QUOTE;
303 	case '=':
304 		if (len >= 1 && buf[1] == '>')
305 			return LINE_LINK;
306 		break;
307 	case '#':
308 		if (len == 1)
309 			return LINE_TEXT;
310 		if (buf[1] != '#')
311 			return LINE_TITLE_1;
312 		if (len == 2)
313 			return LINE_TEXT;
314 		if (buf[2] != '#')
315 			return LINE_TITLE_2;
316 		if (len == 3)
317 			return LINE_TEXT;
318 		return LINE_TITLE_3;
319 	case '`':
320 		if (len < 3)
321 			return LINE_TEXT;
322 		if (buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
323 			return LINE_PRE_START;
324 		break;
325 	}
326 
327 	return LINE_TEXT;
328 }
329 
330 static inline int
331 append(struct parser *p, const char *buf, size_t len)
332 {
333 	size_t newlen;
334 	char *t;
335 
336 	newlen = len + p->len;
337 	if ((t = calloc(1, newlen)) == NULL)
338 		return 0;
339 	free(p->buf);
340 	p->buf = t;
341 	p->len = newlen;
342 	return 1;
343 }
344 
345 static inline int
346 set_buf(struct parser *p, const char *buf, size_t len)
347 {
348 	free(p->buf);
349 	p->buf = NULL;
350 
351 	if (len == 0)
352 		return 1;
353 
354 	if ((p->buf = calloc(1, len)) == NULL)
355 		return 0;
356 	memcpy(p->buf, buf, len);
357 	return 1;
358 }
359 
360 static int
361 gemtext_parse(struct parser *p, const char *buf, size_t size)
362 {
363 	const char	*b, *e;
364 	enum line_type	 t;
365 	size_t		 len, l;
366 
367 	if (p->len == 0) {
368 		b = buf;
369 		len = size;
370 	} else {
371 		if (!append(p, buf, size))
372 			return 0;
373 		b = p->buf;
374 		len = p->len;
375 	}
376 
377 	while (len > 0) {
378 		if ((e = telescope_strnchr((char*)b, '\n', len)) == NULL)
379 			break;
380 		l = e - b;
381 		t = detect_line_type(b, l, p->flags);
382 		if (t == LINE_PRE_START)
383 			p->flags = 1;
384 		if (t == LINE_PRE_END)
385 			p->flags = 0;
386 		if (!parsers[t](p, t, b, l))
387 			return 0;
388 
389 		len -= l;
390 		b += l;
391 
392 		if (len > 0) {
393 			/* skip \n */
394 			len--;
395 			b++;
396 		}
397 	}
398 
399 	return set_buf(p, b, len);
400 }
401 
402 static void
403 gemtext_free(struct parser *p)
404 {
405 	free(p->buf);
406 }