2.53
format.c
1/* snac - A simple, minimalistic ActivityPub instance */
2/* copyright (c) 2022 - 2024 grunfink et al. / MIT license */
3
4#include "xs.h"
5#include "xs_regex.h"
6#include "xs_mime.h"
7#include "xs_html.h"
8#include "xs_json.h"
9#include "xs_time.h"
10
11#include "snac.h"
12
13/* emoticons, people laughing and such */
14const char *smileys[] = {
15 ":-)", "🙂",
16 ":-D", "😀",
17 "X-D", "😆",
18 ";-)", "😉",
19 "B-)", "😎",
20 ">:-(", "😡",
21 ":-(", "😞",
22 ":-*", "😘",
23 ":-/", "😕",
24 "8-o", "😲",
25 "%-)", "🤪",
26 ":_(", "😢",
27 ":-|", "😐",
28 "<3", "❤️",
29 ":facepalm:", "🤦",
30 ":shrug:", "🤷",
31 ":shrug2:", "¯\\_(ツ)_/¯",
32 ":eyeroll:", "🙄",
33 ":beer:", "🍺",
34 ":beers:", "🍻",
35 ":munch:", "😱",
36 ":thumb:", "👍",
37 NULL, NULL
38};
39
40
41xs_dict *emojis(void)
42/* returns a dict with the emojis */
43{
44 xs *fn = xs_fmt("%s/emojis.json", srv_basedir);
45 FILE *f;
46
47 if (mtime(fn) == 0) {
48 /* file does not exist; create it with the defaults */
49 xs *d = xs_dict_new();
50 const char **emo = smileys;
51
52 while (*emo) {
53 d = xs_dict_append(d, emo[0], emo[1]);
54 emo += 2;
55 }
56
57 if ((f = fopen(fn, "w")) != NULL) {
58 xs_json_dump(d, 4, f);
59 fclose(f);
60 }
61 else
62 srv_log(xs_fmt("Error creating '%s'", fn));
63 }
64
65 xs_dict *d = NULL;
66
67 if ((f = fopen(fn, "r")) != NULL) {
68 d = xs_json_load(f);
69 fclose(f);
70
71 if (d == NULL)
72 srv_log(xs_fmt("JSON parse error in '%s'", fn));
73 }
74 else
75 srv_log(xs_fmt("Error opening '%s'", fn));
76
77 return d;
78}
79
80
81static xs_str *format_line(const char *line, xs_list **attach)
82/* formats a line */
83{
84 xs_str *s = xs_str_new(NULL);
85 char *p, *v;
86
87 /* split by markup */
88 xs *sm = xs_regex_split(line,
89 "(`[^`]+`|\\*\\*?[^\\*]+\\*?\\*|https?:/" "/[^[:space:]]+)");
90 int n = 0;
91
92 p = sm;
93 while (xs_list_iter(&p, &v)) {
94 if ((n & 0x1)) {
95 /* markup */
96 if (xs_startswith(v, "`")) {
97 xs *s1 = xs_crop_i(xs_dup(v), 1, -1);
98 xs *e1 = encode_html(s1);
99 xs *s2 = xs_fmt("<code>%s</code>", e1);
100 s = xs_str_cat(s, s2);
101 }
102 else
103 if (xs_startswith(v, "**")) {
104 xs *s1 = xs_crop_i(xs_dup(v), 2, -2);
105 xs *s2 = xs_fmt("<b>%s</b>", s1);
106 s = xs_str_cat(s, s2);
107 }
108 else
109 if (xs_startswith(v, "*")) {
110 xs *s1 = xs_crop_i(xs_dup(v), 1, -1);
111 xs *s2 = xs_fmt("<i>%s</i>", s1);
112 s = xs_str_cat(s, s2);
113 }
114 else
115 if (xs_startswith(v, "http")) {
116 xs *u = xs_replace(v, "#", "#");
117 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
118
119 const char *mime = xs_mime_by_ext(v2);
120
121 if (attach != NULL && xs_startswith(mime, "image/")) {
122 /* if it's a link to an image, insert it as an attachment */
123 xs *d = xs_dict_new();
124
125 d = xs_dict_append(d, "mediaType", mime);
126 d = xs_dict_append(d, "url", v2);
127 d = xs_dict_append(d, "name", "");
128 d = xs_dict_append(d, "type", "Image");
129
130 *attach = xs_list_append(*attach, d);
131 }
132 else {
133 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
134 s = xs_str_cat(s, s1);
135 }
136 }
137 else
138 s = xs_str_cat(s, v);
139 }
140 else
141 /* surrounded text, copy directly */
142 s = xs_str_cat(s, v);
143
144 n++;
145 }
146
147 return s;
148}
149
150
151xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
152/* formats a content using some Markdown rules */
153{
154 xs_str *s = xs_str_new(NULL);
155 int in_pre = 0;
156 int in_blq = 0;
157 xs *list;
158 char *p, *v;
159
160 /* work by lines */
161 list = xs_split(content, "\n");
162
163 p = list;
164 while (xs_list_iter(&p, &v)) {
165 xs *ss = NULL;
166
167 if (strcmp(v, "```") == 0) {
168 if (!in_pre)
169 s = xs_str_cat(s, "<pre>");
170 else
171 s = xs_str_cat(s, "</pre>");
172
173 in_pre = !in_pre;
174 continue;
175 }
176
177 if (in_pre) {
178 // Encode all HTML characters when we're in pre element until we are out.
179 ss = encode_html(v);
180
181 s = xs_str_cat(s, ss);
182 s = xs_str_cat(s, "<br>");
183 continue;
184 }
185
186 else
187 ss = xs_strip_i(format_line(v, attach));
188
189 if (xs_startswith(ss, "---")) {
190 /* delete the --- */
191 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
192 s = xs_str_cat(s, "<hr>");
193
194 s = xs_str_cat(s, ss);
195
196 continue;
197 }
198
199 if (xs_startswith(ss, ">")) {
200 /* delete the > and subsequent spaces */
201 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
202
203 if (!in_blq) {
204 s = xs_str_cat(s, "<blockquote>");
205 in_blq = 1;
206 }
207
208 s = xs_str_cat(s, ss);
209 s = xs_str_cat(s, "<br>");
210
211 continue;
212 }
213
214 if (in_blq) {
215 s = xs_str_cat(s, "</blockquote>");
216 in_blq = 0;
217 }
218
219 s = xs_str_cat(s, ss);
220 s = xs_str_cat(s, "<br>");
221 }
222
223 if (in_blq)
224 s = xs_str_cat(s, "</blockquote>");
225 if (in_pre)
226 s = xs_str_cat(s, "</pre>");
227
228 /* some beauty fixes */
229 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
230 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
231 s = xs_replace_i(s, "</pre><br>", "</pre>");
232
233 {
234 /* traditional emoticons */
235 xs *d = emojis();
236 int c = 0;
237 char *k, *v;
238
239 while (xs_dict_next(d, &k, &v, &c)) {
240 const char *t = NULL;
241
242 /* is it an URL to an image? */
243 if (xs_startswith(v, "https:/" "/") && xs_startswith((t = xs_mime_by_ext(v)), "image/")) {
244 if (tag && xs_str_in(s, k) != -1) {
245 /* add the emoji to the tag list */
246 xs *e = xs_dict_new();
247 xs *i = xs_dict_new();
248 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
249
250 e = xs_dict_append(e, "id", v);
251 e = xs_dict_append(e, "type", "Emoji");
252 e = xs_dict_append(e, "name", k);
253 e = xs_dict_append(e, "updated", u);
254
255 i = xs_dict_append(i, "type", "Image");
256 i = xs_dict_append(i, "mediaType", t);
257 i = xs_dict_append(i, "url", v);
258 e = xs_dict_append(e, "icon", i);
259
260 *tag = xs_list_append(*tag, e);
261 }
262 }
263 else
264 s = xs_replace_i(s, k, v);
265 }
266 }
267
268 return s;
269}
270
271
272const char *valid_tags[] = {
273 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
274 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi", NULL
275};
276
277xs_str *sanitize(const char *content)
278/* cleans dangerous HTML output */
279{
280 xs_str *s = xs_str_new(NULL);
281 xs *sl;
282 int n = 0;
283 char *p, *v;
284
285 sl = xs_regex_split(content, "</?[^>]+>");
286
287 p = sl;
288
289 n = 0;
290 while (xs_list_iter(&p, &v)) {
291 if (n & 0x1) {
292 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
293 xs *l1 = xs_split_n(s1, " ", 1);
294 xs *tag = xs_tolower_i(xs_dup(xs_list_get(l1, 0)));
295 xs *s2 = NULL;
296 int i;
297
298 /* check if it's one of the valid tags */
299 for (i = 0; valid_tags[i]; i++) {
300 if (strcmp(tag, valid_tags[i]) == 0)
301 break;
302 }
303
304 if (valid_tags[i]) {
305 /* accepted tag: rebuild it with only the accepted elements */
306 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=\"[^\"]*\"");
307 xs *s3 = xs_join(el, " ");
308
309 s2 = xs_fmt("<%s%s%s%s>",
310 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
311
312 s = xs_str_cat(s, s2);
313 } else {
314 /* treat end of divs as paragraph breaks */
315 if (strcmp(v, "</div>"))
316 s = xs_str_cat(s, "<p>");
317 }
318 }
319 else {
320 /* non-tag */
321 s = xs_str_cat(s, v);
322 }
323
324 n++;
325 }
326
327 return s;
328}
329
330
331xs_str *encode_html(const char *str)
332/* escapes html characters */
333{
334 xs_str *encoded = xs_html_encode((char *)str);
335
336 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
337 encoded = xs_replace_i(encoded, "<br>", "<br>");
338
339 return encoded;
340}
