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.
25 #include "telescope.h"
31 static int gemtext_parse(struct parser*, const char*, size_t);
32 static int gemtext_foreach_line(struct parser*, const char*, size_t);
33 static int gemtext_free(struct parser*);
35 static int parse_text(struct parser*, enum line_type, const char*, size_t);
36 static int parse_link(struct parser*, enum line_type, const char*, size_t);
37 static int parse_title(struct parser*, enum line_type, const char*, size_t);
38 static int parse_item(struct parser*, enum line_type, const char*, size_t);
39 static int parse_quote(struct parser*, enum line_type, const char*, size_t);
40 static int parse_pre_start(struct parser*, enum line_type, const char*, size_t);
41 static int parse_pre_cnt(struct parser*, enum line_type, const char*, size_t);
42 static int parse_pre_end(struct parser*, enum line_type, const char*, size_t);
43 static void search_title(struct parser*, enum line_type);
45 typedef int (parselinefn)(struct parser*, enum line_type, const char*, size_t);
47 static parselinefn *parsers[] = {
48 parse_text, /* LINE_TEXT */
49 parse_link, /* LINE_LINK */
50 parse_title, /* LINE_TITLE_1 */
51 parse_title, /* LINE_TITLE_2 */
52 parse_title, /* LINE_TITLE_3 */
53 parse_item, /* LINE_ITEM */
54 parse_quote, /* LINE_QUOTE */
55 parse_pre_start, /* LINE_PRE_START */
56 parse_pre_cnt, /* LINE_PRE_CONTENT */
57 parse_pre_end, /* LINE_PRE_END */
61 gemtext_initparser(struct parser *p)
63 memset(p, 0, sizeof(*p));
65 p->name = "text/gemini";
66 p->parse = &gemtext_parse;
67 p->free = &gemtext_free;
71 emit_line(struct parser *p, enum line_type type, char *line, char *alt)
75 if ((l = calloc(1, sizeof(*l))) == NULL)
88 case LINE_PRE_CONTENT:
97 if (TAILQ_EMPTY(&p->head))
98 TAILQ_INSERT_HEAD(&p->head, l, lines);
100 TAILQ_INSERT_TAIL(&p->head, l, lines);
106 parse_text(struct parser *p, enum line_type t, const char *buf, size_t len)
110 if ((l = calloc(1, len+1)) == NULL)
113 return emit_line(p, t, l, NULL);
117 parse_link(struct parser *p, enum line_type t, const char *buf, size_t len)
120 const char *url_start;
123 return emit_line(p, t, NULL, NULL);
127 while (len > 0 && isspace(buf[0])) {
133 return emit_line(p, t, NULL, NULL);
136 while (len > 0 && !isspace(buf[0])) {
141 if ((u = calloc(1, buf - url_start + 1)) == NULL)
143 memcpy(u, url_start, buf - url_start);
148 while (len > 0 && isspace(buf[0])) {
156 if ((l = calloc(1, len + 1)) == NULL)
160 return emit_line(p, t, l, u);
163 if ((l = strdup(u)) == NULL)
165 return emit_line(p, t, l, u);
169 parse_title(struct parser *p, enum line_type t, const char *buf, size_t len)
176 return emit_line(p, t, NULL, NULL);
182 return emit_line(p, t, NULL, NULL);
188 return emit_line(p, t, NULL, NULL);
197 while (len > 0 && isspace(buf[0])) {
203 return emit_line(p, t, NULL, NULL);
205 if (t == LINE_TITLE_1 && *p->title == '\0')
206 strncpy(p->title, buf, MIN(sizeof(p->title)-1, len));
208 if ((l = calloc(1, len+1)) == NULL)
211 return emit_line(p, t, l, NULL);
215 parse_item(struct parser *p, enum line_type t, const char *buf, size_t len)
220 return emit_line(p, t, NULL, NULL);
225 while (len > 0 && isspace(buf[0])) {
231 return emit_line(p, t, NULL, NULL);
233 if ((l = calloc(1, len+1)) == NULL)
236 return emit_line(p, t, l, NULL);
240 parse_quote(struct parser *p, enum line_type t, const char *buf, size_t len)
245 return emit_line(p, t, NULL, NULL);
250 while (len > 0 && isspace(buf[0])) {
256 return emit_line(p, t, NULL, NULL);
258 if ((l = calloc(1, len+1)) == NULL)
261 return emit_line(p, t, l, NULL);
265 parse_pre_start(struct parser *p, enum line_type t, const char *buf, size_t len)
270 return emit_line(p, t, NULL, NULL);
275 while (len > 0 && isspace(buf[0])) {
281 return emit_line(p, t, NULL, NULL);
283 if ((l = calloc(1, len+1)) == NULL)
287 return emit_line(p, t, l, NULL);
291 parse_pre_cnt(struct parser *p, enum line_type t, const char *buf, size_t len)
296 return emit_line(p, t, NULL, NULL);
298 if ((l = calloc(1, len+1)) == NULL)
301 return emit_line(p, t, l, NULL);
305 parse_pre_end(struct parser *p, enum line_type t, const char *buf, size_t len)
307 return emit_line(p, t, NULL, NULL);
310 static inline enum line_type
311 detect_line_type(const char *buf, size_t len, int in_pre)
315 buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
318 return LINE_PRE_CONTENT;
325 case '*': return LINE_ITEM;
326 case '>': return LINE_QUOTE;
328 if (len >= 1 && buf[1] == '>')
346 if (buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
347 return LINE_PRE_START;
355 gemtext_parse(struct parser *p, const char *buf, size_t size)
357 return parser_foreach_line(p, buf, size, gemtext_foreach_line);
361 gemtext_foreach_line(struct parser *p, const char *line, size_t linelen)
365 t = detect_line_type(line, linelen, p->flags & PARSER_IN_PRE);
366 if (t == LINE_PRE_START)
367 p->flags ^= PARSER_IN_PRE;
368 if (t == LINE_PRE_END)
369 p->flags ^= PARSER_IN_PRE;
370 return parsers[t](p, t, line, linelen);
374 gemtext_free(struct parser *p)
378 /* flush the buffer */
380 t = detect_line_type(p->buf, p->len, p->flags & PARSER_IN_PRE);
381 if (!parsers[t](p, t, p->buf, p->len))
383 if ((p->flags & PARSER_IN_PRE) &&
384 !emit_line(p, LINE_PRE_END, NULL, NULL))
391 * use the first level 2 or 3 header as page title if none
394 if (*p->title == '\0')
395 search_title(p, LINE_TITLE_2);
396 if (*p->title == '\0')
397 search_title(p, LINE_TITLE_3);
403 search_title(struct parser *p, enum line_type level)
407 TAILQ_FOREACH(l, &p->head, lines) {
408 if (l->type == level) {
411 strlcpy(p->title, l->line, sizeof(p->title));