2 * Copyright (c) 2021 Omar Polo <op@omarpolo.com>
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.
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.
18 * A streaming gemtext parser.
35 static int gemtext_parse(struct parser*, const char*, size_t);
36 static int gemtext_foreach_line(struct parser*, const char*, size_t);
37 static int gemtext_free(struct parser*);
39 static int parse_text(struct parser*, enum line_type, const char*, size_t);
40 static int parse_link(struct parser*, enum line_type, const char*, size_t);
41 static int parse_title(struct parser*, enum line_type, const char*, size_t);
42 static int parse_item(struct parser*, enum line_type, const char*, size_t);
43 static int parse_quote(struct parser*, enum line_type, const char*, size_t);
44 static int parse_pre_start(struct parser*, enum line_type, const char*, size_t);
45 static int parse_pre_cnt(struct parser*, enum line_type, const char*, size_t);
46 static int parse_pre_end(struct parser*, enum line_type, const char*, size_t);
47 static void search_title(struct parser*, enum line_type);
49 typedef int (parselinefn)(struct parser*, enum line_type, const char*, size_t);
51 static parselinefn *parsers[] = {
52 [LINE_TEXT] = parse_text,
53 [LINE_LINK] = parse_link,
54 [LINE_TITLE_1] = parse_title,
55 [LINE_TITLE_2] = parse_title,
56 [LINE_TITLE_3] = parse_title,
57 [LINE_ITEM] = parse_item,
58 [LINE_QUOTE] = parse_quote,
59 [LINE_PRE_START] = parse_pre_start,
60 [LINE_PRE_CONTENT] = parse_pre_cnt,
61 [LINE_PRE_END] = parse_pre_end,
65 gemtext_initparser(struct parser *p)
67 memset(p, 0, sizeof(*p));
69 p->name = "text/gemini";
70 p->parse = &gemtext_parse;
71 p->free = &gemtext_free;
75 emit_line(struct parser *p, enum line_type type, char *line, char *alt)
79 if ((l = calloc(1, sizeof(*l))) == NULL)
91 if (l->type == LINE_PRE_END &&
92 hide_pre_closing_line)
95 case LINE_PRE_CONTENT:
101 !emojied_line(line, (const char **)&l->data))
108 if (TAILQ_EMPTY(&p->head))
109 TAILQ_INSERT_HEAD(&p->head, l, lines);
111 TAILQ_INSERT_TAIL(&p->head, l, lines);
117 parse_text(struct parser *p, enum line_type t, const char *buf, size_t len)
121 if ((l = calloc(1, len+1)) == NULL)
124 return emit_line(p, t, l, NULL);
128 parse_link(struct parser *p, enum line_type t, const char *buf, size_t len)
131 const char *url_start;
134 return emit_line(p, t, NULL, NULL);
138 while (len > 0 && isspace(buf[0])) {
144 return emit_line(p, t, NULL, NULL);
147 while (len > 0 && !isspace(buf[0])) {
152 if ((u = calloc(1, buf - url_start + 1)) == NULL)
154 memcpy(u, url_start, buf - url_start);
159 while (len > 0 && isspace(buf[0])) {
167 if ((l = calloc(1, len + 1)) == NULL)
171 return emit_line(p, t, l, u);
174 if ((l = strdup(u)) == NULL)
176 return emit_line(p, t, l, u);
180 parse_title(struct parser *p, enum line_type t, const char *buf, size_t len)
187 return emit_line(p, t, NULL, NULL);
193 return emit_line(p, t, NULL, NULL);
199 return emit_line(p, t, NULL, NULL);
208 while (len > 0 && isspace(buf[0])) {
214 return emit_line(p, t, NULL, NULL);
216 if (t == LINE_TITLE_1 && *p->title == '\0')
217 strncpy(p->title, buf, MIN(sizeof(p->title)-1, len));
219 if ((l = calloc(1, len+1)) == NULL)
222 return emit_line(p, t, l, NULL);
226 parse_item(struct parser *p, enum line_type t, const char *buf, size_t len)
231 return emit_line(p, t, NULL, NULL);
236 while (len > 0 && isspace(buf[0])) {
242 return emit_line(p, t, NULL, NULL);
244 if ((l = calloc(1, len+1)) == NULL)
247 return emit_line(p, t, l, NULL);
251 parse_quote(struct parser *p, enum line_type t, const char *buf, size_t len)
256 return emit_line(p, t, NULL, NULL);
261 while (len > 0 && isspace(buf[0])) {
267 return emit_line(p, t, NULL, NULL);
269 if ((l = calloc(1, len+1)) == NULL)
272 return emit_line(p, t, l, NULL);
276 parse_pre_start(struct parser *p, enum line_type t, const char *buf, size_t len)
281 return emit_line(p, t, NULL, NULL);
286 while (len > 0 && isspace(buf[0])) {
292 return emit_line(p, t, NULL, NULL);
294 if ((l = calloc(1, len+1)) == NULL)
298 return emit_line(p, t, l, NULL);
302 parse_pre_cnt(struct parser *p, enum line_type t, const char *buf, size_t len)
307 return emit_line(p, t, NULL, NULL);
309 if ((l = calloc(1, len+1)) == NULL)
312 return emit_line(p, t, l, NULL);
316 parse_pre_end(struct parser *p, enum line_type t, const char *buf, size_t len)
318 return emit_line(p, t, NULL, NULL);
321 static inline enum line_type
322 detect_line_type(const char *buf, size_t len, int in_pre)
326 buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
329 return LINE_PRE_CONTENT;
336 case '*': return LINE_ITEM;
337 case '>': return LINE_QUOTE;
339 if (len >= 1 && buf[1] == '>')
357 if (buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
358 return LINE_PRE_START;
366 gemtext_parse(struct parser *p, const char *buf, size_t size)
368 return parser_foreach_line(p, buf, size, gemtext_foreach_line);
372 gemtext_foreach_line(struct parser *p, const char *line, size_t linelen)
376 t = detect_line_type(line, linelen, p->flags & PARSER_IN_PRE);
377 if (t == LINE_PRE_START)
378 p->flags ^= PARSER_IN_PRE;
379 if (t == LINE_PRE_END)
380 p->flags ^= PARSER_IN_PRE;
381 return parsers[t](p, t, line, linelen);
385 gemtext_free(struct parser *p)
389 /* flush the buffer */
391 t = detect_line_type(p->buf, p->len, p->flags & PARSER_IN_PRE);
392 if (!parsers[t](p, t, p->buf, p->len))
394 if ((p->flags & PARSER_IN_PRE) &&
395 !emit_line(p, LINE_PRE_END, NULL, NULL))
402 * use the first level 2 or 3 header as page title if none
405 if (*p->title == '\0')
406 search_title(p, LINE_TITLE_2);
407 if (*p->title == '\0')
408 search_title(p, LINE_TITLE_3);
414 search_title(struct parser *p, enum line_type level)
418 TAILQ_FOREACH(l, &p->head, lines) {
419 if (l->type == level) {
422 strlcpy(p->title, l->line, sizeof(p->title));