1 /*
2  * Copyright (c) 2021, 2022 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 "compat.h"
26 
27 #include <ctype.h>
28 #include <string.h>
29 #include <stdlib.h>
30 
31 #include "defaults.h"
32 #include "parser.h"
33 #include "telescope.h"
34 #include "utf8.h"
35 #include "xwrapper.h"
36 
37 static int	gemtext_parse_line(struct buffer *, const char *, size_t);
38 static int	gemtext_free(struct buffer *);
39 static int	gemtext_serialize(struct buffer *, FILE *);
40 
41 static int	parse_link(struct buffer *, const char*, size_t);
42 static int	parse_title(struct buffer *, const char*, size_t);
43 static void	search_title(struct buffer *, enum line_type);
44 
45 const struct parser gemtext_parser = {
46 	.name = "text/gemini",
47 	.parseline = &gemtext_parse_line,
48 	.free = &gemtext_free,
49 	.serialize = &gemtext_serialize,
50 };
51 
52 static inline int
53 emit_line(struct buffer *b, enum line_type type, char *line, char *alt)
54 {
55 	struct line *l;
56 
57 	l = xcalloc(1, sizeof(*l));
58 
59 	l->type = type;
60 	l->line = line;
61 	l->alt = alt;
62 
63 	switch (l->type) {
64 	case LINE_PRE_START:
65 	case LINE_PRE_END:
66 		if (hide_pre_context)
67 			l->flags = L_HIDDEN;
68 		if (l->type == LINE_PRE_END &&
69 		    hide_pre_closing_line)
70 			l->flags = L_HIDDEN;
71 		break;
72 	case LINE_PRE_CONTENT:
73 		if (hide_pre_blocks)
74 			l->flags = L_HIDDEN;
75 		break;
76 	case LINE_LINK:
77 		if (emojify_link &&
78 		    !emojied_line(line, (const char **)&l->data))
79 			l->data = NULL;
80 		break;
81 	default:
82 		break;
83 	}
84 
85 	if (dont_apply_styling)
86 		l->flags &= ~L_HIDDEN;
87 
88 	TAILQ_INSERT_TAIL(&b->head, l, lines);
89 
90 	return 1;
91 }
92 
93 static int
94 parse_link(struct buffer *b, const char *line, size_t len)
95 {
96 	char *label, *url;
97 	const char *start;
98 
99 	if (len <= 2)
100 		return emit_line(b, LINE_TEXT, NULL, NULL);
101 
102 	line += 2, len -= 2;
103 	while (len > 0 && isspace((unsigned char)line[0]))
104 		line++, len--;
105 
106 	if (len == 0)
107 		return emit_line(b, LINE_TEXT, NULL, NULL);
108 
109 	start = line;
110 	while (len > 0 && !isspace((unsigned char)line[0]))
111 		line++, len--;
112 
113 	url = xstrndup(start, line - start);
114 
115 	while (len > 0 && isspace(line[0]))
116 		line++, len--;
117 
118 	if (len == 0) {
119 		label = xstrdup(url);
120 	} else {
121 		label = xstrndup(line, len);
122 	}
123 
124 	return emit_line(b, LINE_LINK, label, url);
125 }
126 
127 static int
128 parse_title(struct buffer *b, const char *line, size_t len)
129 {
130 	enum line_type t = LINE_TITLE_1;
131 	char *l;
132 
133 	line++, len--;
134 	while (len > 0 && *line == '#') {
135 		line++, len--;
136 		t++;
137 		if (t == LINE_TITLE_3)
138 			break;
139 	}
140 
141 	while (len > 0 && isspace((unsigned char)*line))
142 		line++, len--;
143 
144 	if (len == 0)
145 		return emit_line(b, t, NULL, NULL);
146 
147 	if (t == LINE_TITLE_1 && *b->title == '\0')
148 		strncpy(b->title, line, MIN(sizeof(b->title)-1, len));
149 
150 	l = xstrndup(line, len);
151 	return emit_line(b, t, l, NULL);
152 }
153 
154 static int
155 gemtext_parse_line(struct buffer *b, const char *line, size_t len)
156 {
157 	char *l;
158 
159 	if (b->parser_flags & PARSER_IN_PRE) {
160 		if (len >= 3 && !strncmp(line, "```", 3)) {
161 			b->parser_flags ^= PARSER_IN_PRE;
162 			return emit_line(b, LINE_PRE_END, NULL, NULL);
163 		}
164 
165 		if (len == 0)
166 			return emit_line(b, LINE_PRE_CONTENT, NULL, NULL);
167 		l = xstrndup(line, len);
168 		return emit_line(b, LINE_PRE_CONTENT, l, NULL);
169 	}
170 
171 	if (len == 0)
172 		return emit_line(b, LINE_TEXT, NULL, NULL);
173 
174 	switch (*line) {
175 	case '*':
176 		if (len < 1 || line[1] != ' ')
177 			break;
178 
179 		line += 2, len -= 2;
180 		while (len > 0 && isspace((unsigned char)*line))
181 			line++, len--;
182 		if (len == 0)
183 			return emit_line(b, LINE_ITEM, NULL, NULL);
184 		l = xstrndup(line, len);
185 		return emit_line(b, LINE_ITEM, l, NULL);
186 
187 	case '>':
188 		line++, len--;
189 		while (len > 0 && isspace((unsigned char)*line))
190 			line++, len--;
191 		if (len == 0)
192 			return emit_line(b, LINE_QUOTE, NULL, NULL);
193 		l = xstrndup(line, len);
194 		return emit_line(b, LINE_QUOTE, l, NULL);
195 
196 	case '=':
197 		if (len > 1 && line[1] == '>')
198 			return parse_link(b, line, len);
199 		break;
200 
201 	case '#':
202 		return parse_title(b, line, len);
203 
204 	case '`':
205 		if (len < 3 || strncmp(line, "```", 3) != 0)
206 			break;
207 
208 		b->parser_flags |= PARSER_IN_PRE;
209 		line += 3, len -= 3;
210 		while (len > 0 && isspace((unsigned char)*line))
211 			line++, len--;
212 		if (len == 0)
213 			return emit_line(b, LINE_PRE_START,
214 			    NULL, NULL);
215 		l = xstrndup(line, len);
216 		return emit_line(b, LINE_PRE_START, l, NULL);
217 	}
218 
219 	l = xstrndup(line, len);
220 	return emit_line(b, LINE_TEXT, l, NULL);
221 }
222 
223 static int
224 gemtext_free(struct buffer *b)
225 {
226 	/* flush the buffer */
227 	if (b->len != 0) {
228 		if (!gemtext_parse_line(b, b->buf, b->len))
229 			return 0;
230 		if ((b->parser_flags & PARSER_IN_PRE) &&
231 		    !emit_line(b, LINE_PRE_END, NULL, NULL))
232 			return 0;
233 	}
234 
235 	/*
236 	 * use the first level 2 or 3 header as page title if none
237 	 * found yet.
238 	 */
239 	if (*b->title == '\0')
240 		search_title(b, LINE_TITLE_2);
241 	if (*b->title == '\0')
242 		search_title(b, LINE_TITLE_3);
243 
244 	return 1;
245 }
246 
247 static void
248 search_title(struct buffer *b, enum line_type level)
249 {
250 	struct line *l;
251 
252 	TAILQ_FOREACH(l, &b->head, lines) {
253 		if (l->type == level) {
254 			if (l->line == NULL)
255 				continue;
256 			strlcpy(b->title, l->line, sizeof(b->title));
257 			break;
258 		}
259 	}
260 }
261 
262 static const char *gemtext_prefixes[] = {
263 	[LINE_TEXT] = "",
264 	[LINE_TITLE_1] = "# ",
265 	[LINE_TITLE_2] = "## ",
266 	[LINE_TITLE_3] = "### ",
267 	[LINE_ITEM] = "* ",
268 	[LINE_QUOTE] = "> ",
269 	[LINE_PRE_START] = "``` ",
270 	[LINE_PRE_CONTENT] = "",
271 	[LINE_PRE_END] = "```",
272 };
273 
274 static int
275 gemtext_serialize(struct buffer *b, FILE *fp)
276 {
277 	struct line	*line;
278 	const char	*text;
279 	const char	*alt;
280 	int		 r;
281 
282 	TAILQ_FOREACH(line, &b->head, lines) {
283 		if ((text = line->line) == NULL)
284 			text = "";
285 
286 		if ((alt = line->alt) == NULL)
287 			alt = "";
288 
289 		switch (line->type) {
290 		case LINE_TEXT:
291 		case LINE_TITLE_1:
292 		case LINE_TITLE_2:
293 		case LINE_TITLE_3:
294 		case LINE_ITEM:
295 		case LINE_QUOTE:
296 		case LINE_PRE_START:
297 		case LINE_PRE_CONTENT:
298 		case LINE_PRE_END:
299 			r = fprintf(fp, "%s%s\n", gemtext_prefixes[line->type],
300 			    text);
301 			break;
302 
303 		case LINE_LINK:
304 			r = fprintf(fp, "=> %s %s\n", alt, text);
305 			break;
306 
307 		default:
308 			/* not reached */
309 			abort();
310 		}
311 
312 		if (r == -1)
313 			return 0;
314 	}
315 
316 	return 1;
317 }