Blob


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 */
17 /*
18 * A streaming gemtext parser.
19 *
20 * TODO:
21 * - handle NULs
22 * - UTF8
23 */
25 #include "compat.h"
27 #include <ctype.h>
28 #include <string.h>
29 #include <stdlib.h>
31 #include "defaults.h"
32 #include "parser.h"
33 #include "utf8.h"
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,
62 };
64 void
65 gemtext_initparser(struct parser *p)
66 {
67 memset(p, 0, sizeof(*p));
69 p->name = "text/gemini";
70 p->parse = &gemtext_parse;
71 p->free = &gemtext_free;
73 TAILQ_INIT(&p->head);
74 }
76 static inline int
77 emit_line(struct parser *p, enum line_type type, char *line, char *alt)
78 {
79 struct line *l;
81 if ((l = calloc(1, sizeof(*l))) == NULL)
82 return 0;
84 l->type = type;
85 l->line = line;
86 l->alt = alt;
88 switch (l->type) {
89 case LINE_PRE_START:
90 case LINE_PRE_END:
91 if (hide_pre_context)
92 l->flags = L_HIDDEN;
93 if (l->type == LINE_PRE_END &&
94 hide_pre_closing_line)
95 l->flags = L_HIDDEN;
96 break;
97 case LINE_PRE_CONTENT:
98 if (hide_pre_blocks)
99 l->flags = L_HIDDEN;
100 break;
101 case LINE_LINK:
102 if (emojify_link &&
103 !emojied_line(line, (const char **)&l->data))
104 l->data = NULL;
105 break;
106 default:
107 break;
110 TAILQ_INSERT_TAIL(&p->head, l, lines);
112 return 1;
115 static int
116 parse_text(struct parser *p, enum line_type t, const char *buf, size_t len)
118 char *l;
120 if ((l = calloc(1, len+1)) == NULL)
121 return 0;
122 memcpy(l, buf, len);
123 return emit_line(p, t, l, NULL);
126 static int
127 parse_link(struct parser *p, enum line_type t, const char *buf, size_t len)
129 char *l, *u;
130 const char *url_start;
132 if (len <= 2)
133 return emit_line(p, t, NULL, NULL);
134 buf += 2;
135 len -= 2;
137 while (len > 0 && isspace(buf[0])) {
138 buf++;
139 len--;
142 if (len == 0)
143 return emit_line(p, t, NULL, NULL);
145 url_start = buf;
146 while (len > 0 && !isspace(buf[0])) {
147 buf++;
148 len--;
151 if ((u = calloc(1, buf - url_start + 1)) == NULL)
152 return 0;
153 memcpy(u, url_start, buf - url_start);
155 if (len == 0)
156 goto nolabel;
158 while (len > 0 && isspace(buf[0])) {
159 buf++;
160 len--;
163 if (len == 0)
164 goto nolabel;
166 if ((l = calloc(1, len + 1)) == NULL)
167 return 0;
169 memcpy(l, buf, len);
170 return emit_line(p, t, l, u);
172 nolabel:
173 if ((l = strdup(u)) == NULL)
174 return 0;
175 return emit_line(p, t, l, u);
178 static int
179 parse_title(struct parser *p, enum line_type t, const char *buf, size_t len)
181 char *l;
183 switch (t) {
184 case LINE_TITLE_1:
185 if (len <= 1)
186 return emit_line(p, t, NULL, NULL);
187 buf++;
188 len--;
189 break;
190 case LINE_TITLE_2:
191 if (len <= 2)
192 return emit_line(p, t, NULL, NULL);
193 buf += 2;
194 len -= 2;
195 break;
196 case LINE_TITLE_3:
197 if (len <= 3)
198 return emit_line(p, t, NULL, NULL);
199 buf += 3;
200 len -= 3;
201 break;
202 default:
203 /* unreachable */
204 abort();
207 while (len > 0 && isspace(buf[0])) {
208 buf++;
209 len--;
212 if (len == 0)
213 return emit_line(p, t, NULL, NULL);
215 if (t == LINE_TITLE_1 && *p->title == '\0')
216 strncpy(p->title, buf, MIN(sizeof(p->title)-1, len));
218 if ((l = calloc(1, len+1)) == NULL)
219 return 0;
220 memcpy(l, buf, len);
221 return emit_line(p, t, l, NULL);
224 static int
225 parse_item(struct parser *p, enum line_type t, const char *buf, size_t len)
227 char *l;
229 if (len == 1)
230 return emit_line(p, t, NULL, NULL);
232 buf++;
233 len--;
235 while (len > 0 && isspace(buf[0])) {
236 buf++;
237 len--;
240 if (len == 0)
241 return emit_line(p, t, NULL, NULL);
243 if ((l = calloc(1, len+1)) == NULL)
244 return 0;
245 memcpy(l, buf, len);
246 return emit_line(p, t, l, NULL);
249 static int
250 parse_quote(struct parser *p, enum line_type t, const char *buf, size_t len)
252 char *l;
254 if (len == 1)
255 return emit_line(p, t, NULL, NULL);
257 buf++;
258 len--;
260 while (len > 0 && isspace(buf[0])) {
261 buf++;
262 len--;
265 if (len == 0)
266 return emit_line(p, t, NULL, NULL);
268 if ((l = calloc(1, len+1)) == NULL)
269 return 0;
270 memcpy(l, buf, len);
271 return emit_line(p, t, l, NULL);
274 static int
275 parse_pre_start(struct parser *p, enum line_type t, const char *buf, size_t len)
277 char *l;
279 if (len <= 3)
280 return emit_line(p, t, NULL, NULL);
282 buf += 3;
283 len -= 3;
285 while (len > 0 && isspace(buf[0])) {
286 buf++;
287 len--;
290 if (len == 0)
291 return emit_line(p, t, NULL, NULL);
293 if ((l = calloc(1, len+1)) == NULL)
294 return 0;
296 memcpy(l, buf, len);
297 return emit_line(p, t, l, NULL);
300 static int
301 parse_pre_cnt(struct parser *p, enum line_type t, const char *buf, size_t len)
303 char *l;
305 if (len == 0)
306 return emit_line(p, t, NULL, NULL);
308 if ((l = calloc(1, len+1)) == NULL)
309 return 0;
310 memcpy(l, buf, len);
311 return emit_line(p, t, l, NULL);
314 static int
315 parse_pre_end(struct parser *p, enum line_type t, const char *buf, size_t len)
317 return emit_line(p, t, NULL, NULL);
320 static inline enum line_type
321 detect_line_type(const char *buf, size_t len, int in_pre)
323 if (in_pre) {
324 if (len >= 3 &&
325 buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
326 return LINE_PRE_END;
327 else
328 return LINE_PRE_CONTENT;
331 if (len == 0)
332 return LINE_TEXT;
334 switch (*buf) {
335 case '*': return LINE_ITEM;
336 case '>': return LINE_QUOTE;
337 case '=':
338 if (len >= 1 && buf[1] == '>')
339 return LINE_LINK;
340 break;
341 case '#':
342 if (len == 1)
343 return LINE_TEXT;
344 if (buf[1] != '#')
345 return LINE_TITLE_1;
346 if (len == 2)
347 return LINE_TEXT;
348 if (buf[2] != '#')
349 return LINE_TITLE_2;
350 if (len == 3)
351 return LINE_TEXT;
352 return LINE_TITLE_3;
353 case '`':
354 if (len < 3)
355 return LINE_TEXT;
356 if (buf[0] == '`' && buf[1] == '`' && buf[2] == '`')
357 return LINE_PRE_START;
358 break;
361 return LINE_TEXT;
364 static int
365 gemtext_parse(struct parser *p, const char *buf, size_t size)
367 return parser_foreach_line(p, buf, size, gemtext_foreach_line);
370 static int
371 gemtext_foreach_line(struct parser *p, const char *line, size_t linelen)
373 enum line_type t;
375 t = detect_line_type(line, linelen, p->flags & PARSER_IN_PRE);
376 if (t == LINE_PRE_START)
377 p->flags ^= PARSER_IN_PRE;
378 if (t == LINE_PRE_END)
379 p->flags ^= PARSER_IN_PRE;
380 return parsers[t](p, t, line, linelen);
383 static int
384 gemtext_free(struct parser *p)
386 enum line_type t;
388 /* flush the buffer */
389 if (p->len != 0) {
390 t = detect_line_type(p->buf, p->len, p->flags & PARSER_IN_PRE);
391 if (!parsers[t](p, t, p->buf, p->len))
392 return 0;
393 if ((p->flags & PARSER_IN_PRE) &&
394 !emit_line(p, LINE_PRE_END, NULL, NULL))
395 return 0;
398 free(p->buf);
400 /*
401 * use the first level 2 or 3 header as page title if none
402 * found yet.
403 */
404 if (*p->title == '\0')
405 search_title(p, LINE_TITLE_2);
406 if (*p->title == '\0')
407 search_title(p, LINE_TITLE_3);
409 return 1;
412 static void
413 search_title(struct parser *p, enum line_type level)
415 struct line *l;
417 TAILQ_FOREACH(l, &p->head, lines) {
418 if (l->type == level) {
419 if (l->line == NULL)
420 continue;
421 strlcpy(p->title, l->line, sizeof(p->title));
422 break;