2.70
format.c
1/* snac - A simple, minimalistic ActivityPub instance */
2/* copyright (c) 2022 - 2025 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#include "xs_match.h"
11
12#include "snac.h"
13
14/* emoticons, people laughing and such */
15const char *smileys[] = {
16 ":-)", "🙂",
17 ":-D", "😀",
18 "X-D", "😆",
19 ";-)", "😉",
20 "B-)", "😎",
21 ">:-(", "😡",
22 ":-(", "😞",
23 ":-*", "😘",
24 ":-/", "😕",
25 "8-o", "😲",
26 "%-)", "🤪",
27 ":_(", "😢",
28 ":-|", "😐",
29 "<3", "❤️",
30 ":facepalm:", "🤦",
31 ":shrug:", "🤷",
32 ":shrug2:", "¯\\_(ツ)_/¯",
33 ":eyeroll:", "🙄",
34 ":beer:", "🍺",
35 ":beers:", "🍻",
36 ":munch:", "😱",
37 ":thumb:", "👍",
38 NULL, NULL
39};
40
41
42xs_dict *emojis(void)
43/* returns a dict with the emojis */
44{
45 xs *fn = xs_fmt("%s/emojis.json", srv_basedir);
46 FILE *f;
47
48 if (mtime(fn) == 0) {
49 /* file does not exist; create it with the defaults */
50 xs *d = xs_dict_new();
51 const char **emo = smileys;
52
53 while (*emo) {
54 d = xs_dict_append(d, emo[0], emo[1]);
55 emo += 2;
56 }
57
58 if ((f = fopen(fn, "w")) != NULL) {
59 xs_json_dump(d, 4, f);
60 fclose(f);
61 }
62 else
63 srv_log(xs_fmt("Error creating '%s'", fn));
64 }
65
66 xs_dict *d = NULL;
67
68 if ((f = fopen(fn, "r")) != NULL) {
69 d = xs_json_load(f);
70 fclose(f);
71
72 if (d == NULL)
73 srv_log(xs_fmt("JSON parse error in '%s'", fn));
74 }
75 else
76 srv_log(xs_fmt("Error opening '%s'", fn));
77
78 return d;
79}
80
81
82static xs_str *format_line(const char *line, xs_list **attach)
83/* formats a line */
84{
85 xs_str *s = xs_str_new(NULL);
86 char *p;
87 const char *v;
88
89 /* split by markup */
90 xs *sm = xs_regex_split(line,
91 "("
92 "`[^`]+`" "|"
93 "~~[^~]+~~" "|"
94 "\\*\\*?\\*?[^\\*]+\\*?\\*?\\*" "|"
95 "_[^_]+_" "|" //anzu
96 "__[^_]+__" "|" //anzu
97 "!\\[[^]]+\\]\\([^\\)]+\\)" "|"
98 "\\[[^]]+\\]\\([^\\)]+\\)" "|"
99 "[a-z]+:/" "/[^[:space:]]+" "|"
100 "(mailto|xmpp):[^@[:space:]]+@[^[:space:]]+"
101 ")");
102 int n = 0;
103
104 p = sm;
105 while (xs_list_iter(&p, &v)) {
106 if ((n & 0x1)) {
107 /* markup */
108 if (xs_startswith(v, "`")) {
109 xs *s1 = xs_strip_chars_i(xs_dup(v), "`");
110 xs *e1 = encode_html(s1);
111 xs *s2 = xs_fmt("<code>%s</code>", e1);
112 s = xs_str_cat(s, s2);
113 }
114 else
115 if (xs_startswith(v, "***")) {
116 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
117 xs *s2 = xs_fmt("<b><i>%s</i></b>", s1);
118 s = xs_str_cat(s, s2);
119 }
120 else
121 if (xs_startswith(v, "**")) {
122 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
123 xs *s2 = xs_fmt("<b>%s</b>", s1);
124 s = xs_str_cat(s, s2);
125 }
126 else
127 if (xs_startswith(v, "*")) {
128 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
129 xs *s2 = xs_fmt("<i>%s</i>", s1);
130 s = xs_str_cat(s, s2);
131 }
132 //anzu - begin
133 else
134 if (xs_startswith(v, "__")) {
135 xs *s1 = xs_strip_chars_i(xs_dup(v), "_");
136 xs *s2 = xs_fmt("<u>%s</u>", s1);
137 s = xs_str_cat(s, s2);
138 }
139 else
140 if (xs_startswith(v, "_")) {
141 xs *s1 = xs_strip_chars_i(xs_dup(v), "_");
142 xs *s2 = xs_fmt("<i>%s</i>", s1);
143 s = xs_str_cat(s, s2);
144 }
145 //anzu - end
146 else
147 if (xs_startswith(v, "~~")) {
148 xs *s1 = xs_strip_chars_i(xs_dup(v), "~");
149 xs *e1 = encode_html(s1);
150 xs *s2 = xs_fmt("<s>%s</s>", e1);
151 s = xs_str_cat(s, s2);
152 }
153 else
154 if (*v == '[') {
155 /* markdown-like links [label](url) */
156 xs *w = xs_strip_chars_i(
157 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
158 "![)");
159 xs *l = xs_split_n(w, "](", 1);
160
161 if (xs_list_len(l) == 2) {
162 xs *link = xs_fmt("<a href=\"%s\">%s</a>",
163 xs_list_get(l, 1), xs_list_get(l, 0));
164
165 s = xs_str_cat(s, link);
166 }
167 else
168 s = xs_str_cat(s, v);
169 }
170 else
171 if (*v == '!') {
172 /* markdown-like images  */
173 xs *w = xs_strip_chars_i(
174 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
175 "![)");
176 xs *l = xs_split_n(w, "](", 1);
177
178 if (xs_list_len(l) == 2) {
179 const char *alt_text = xs_list_get(l, 0);
180 const char *img_url = xs_list_get(l, 1);
181 const char *mime = xs_mime_by_ext(img_url);
182
183 if (attach != NULL && xs_startswith(mime, "image/")) {
184 const xs_dict *ad;
185 int add = 1;
186
187 xs_list_foreach(*attach, ad) {
188 if (strcmp(xs_dict_get_def(ad, "url", ""), img_url) == 0) {
189 add = 0;
190 break;
191 }
192 }
193
194 if (add) {
195 xs *d = xs_dict_new();
196
197 d = xs_dict_append(d, "mediaType", mime);
198 d = xs_dict_append(d, "url", img_url);
199 d = xs_dict_append(d, "name", alt_text);
200 d = xs_dict_append(d, "type", "Image");
201
202 *attach = xs_list_append(*attach, d);
203 }
204 }
205 else {
206 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
207
208 s = xs_str_cat(s, link);
209 }
210 }
211 else
212 s = xs_str_cat(s, v);
213 }
214 else
215 if (xs_str_in(v, ":/" "/") != -1) {
216 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
217
218 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
219
220 const char *mime = xs_mime_by_ext(v2);
221
222 if (attach != NULL && xs_startswith(mime, "image/")) {
223 /* if it's a link to an image, insert it as an attachment */
224 const xs_dict *ad;
225 int add = 1;
226
227 xs_list_foreach(*attach, ad) {
228 if (strcmp(xs_dict_get_def(ad, "url", ""), v2) == 0) {
229 add = 0;
230 break;
231 }
232 }
233
234 if (add) {
235 xs *d = xs_dict_new();
236
237 d = xs_dict_append(d, "mediaType", mime);
238 d = xs_dict_append(d, "url", v2);
239 d = xs_dict_append(d, "name", "");
240 d = xs_dict_append(d, "type", "Image");
241
242 *attach = xs_list_append(*attach, d);
243 }
244 }
245 else {
246 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
247 s = xs_str_cat(s, s1);
248 }
249 }
250 else
251 if (xs_match(v, "mailto*|xmpp*")) {
252 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
253
254 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
255
256 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
257 s = xs_str_cat(s, s1);
258 }
259 else
260 s = xs_str_cat(s, v);
261 }
262 else
263 /* surrounded text, copy directly */
264 s = xs_str_cat(s, v);
265
266 n++;
267 }
268
269 return s;
270}
271
272
273xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
274/* formats a content using some Markdown rules */
275{
276 xs_str *s = xs_str_new(NULL);
277 int in_pre = 0;
278 int in_blq = 0;
279 xs *list;
280 char *p;
281 const char *v;
282
283 /* work by lines */
284 list = xs_split(content, "\n");
285
286 p = list;
287 while (xs_list_iter(&p, &v)) {
288 xs *ss = NULL;
289
290 if (strcmp(v, "```") == 0) {
291 if (!in_pre)
292 s = xs_str_cat(s, "<pre>");
293 else
294 s = xs_str_cat(s, "</pre>");
295
296 in_pre = !in_pre;
297 continue;
298 }
299
300 if (in_pre) {
301 // Encode all HTML characters when we're in pre element until we are out.
302 ss = encode_html(v);
303
304 s = xs_str_cat(s, ss);
305 s = xs_str_cat(s, "<br>");
306 continue;
307 }
308
309 else
310 ss = xs_strip_i(format_line(v, attach));
311
312 if (xs_startswith(ss, "---")) {
313 /* delete the --- */
314 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
315 s = xs_str_cat(s, "<hr>");
316
317 s = xs_str_cat(s, ss);
318
319 continue;
320 }
321
322 //anzu - begin
323 // h1 reserved for snac?
324 if (xs_startswith(ss, "# ")) {
325 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
326 s = xs_str_cat(s, "<h2>");
327 s = xs_str_cat(s, ss);
328 s = xs_str_cat(s, "</h2>");
329 continue;
330 }
331 if (xs_startswith(ss, "## ")) {
332 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
333 s = xs_str_cat(s, "<h2>");
334 s = xs_str_cat(s, ss);
335 s = xs_str_cat(s, "</h2>");
336 continue;
337 }
338 if (xs_startswith(ss, "### ")) {
339 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
340 s = xs_str_cat(s, "<h3>");
341 s = xs_str_cat(s, ss);
342 s = xs_str_cat(s, "</h3>");
343 continue;
344 }
345 //anzu - end
346
347 if (xs_startswith(ss, ">")) {
348 /* delete the > and subsequent spaces */
349 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
350
351 if (!in_blq) {
352 s = xs_str_cat(s, "<blockquote>");
353 in_blq = 1;
354 }
355
356 s = xs_str_cat(s, ss);
357 s = xs_str_cat(s, "<br>");
358
359 continue;
360 }
361
362 if (in_blq) {
363 s = xs_str_cat(s, "</blockquote>");
364 in_blq = 0;
365 }
366
367 s = xs_str_cat(s, ss);
368 s = xs_str_cat(s, "<br>");
369 }
370
371 if (in_blq)
372 s = xs_str_cat(s, "</blockquote>");
373 if (in_pre)
374 s = xs_str_cat(s, "</pre>");
375
376 /* some beauty fixes */
377 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
378 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
379 s = xs_replace_i(s, "</pre><br>", "</pre>");
380 s = xs_replace_i(s, "</h2><br>", "</h2>"); //anzu ???
381 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
382
383 {
384 /* traditional emoticons */
385 xs *d = emojis();
386 int c = 0;
387 const char *k, *v;
388
389 while (xs_dict_next(d, &k, &v, &c)) {
390 const char *t = NULL;
391
392 /* is it an URL to an image? */
393 if (xs_startswith(v, "https:/" "/") && xs_startswith((t = xs_mime_by_ext(v)), "image/")) {
394 if (tag && xs_str_in(s, k) != -1) {
395 /* add the emoji to the tag list */
396 xs *e = xs_dict_new();
397 xs *i = xs_dict_new();
398 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
399
400 e = xs_dict_append(e, "id", v);
401 e = xs_dict_append(e, "type", "Emoji");
402 e = xs_dict_append(e, "name", k);
403 e = xs_dict_append(e, "updated", u);
404
405 i = xs_dict_append(i, "type", "Image");
406 i = xs_dict_append(i, "mediaType", t);
407 i = xs_dict_append(i, "url", v);
408 e = xs_dict_append(e, "icon", i);
409
410 *tag = xs_list_append(*tag, e);
411 }
412 }
413 else
414 s = xs_replace_i(s, k, v);
415 }
416 }
417
418 return s;
419}
420
421
422const char *valid_tags[] = {
423 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
424 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi",
425 "h2","h3", //anzu
426 NULL
427};
428
429xs_str *sanitize(const char *content)
430/* cleans dangerous HTML output */
431{
432 xs_str *s = xs_str_new(NULL);
433 xs *sl;
434 int n = 0;
435 char *p;
436 const char *v;
437
438 sl = xs_regex_split(content, "</?[^>]+>");
439
440 p = sl;
441
442 n = 0;
443 while (xs_list_iter(&p, &v)) {
444 if (n & 0x1) {
445 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
446 xs *l1 = xs_split_n(s1, " ", 1);
447 xs *tag = xs_tolower_i(xs_dup(xs_list_get(l1, 0)));
448 xs *s2 = NULL;
449 int i;
450
451 /* check if it's one of the valid tags */
452 for (i = 0; valid_tags[i]; i++) {
453 if (strcmp(tag, valid_tags[i]) == 0)
454 break;
455 }
456
457 if (valid_tags[i]) {
458 /* accepted tag: rebuild it with only the accepted elements */
459 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=\"[^\"]*\"");
460 xs *s3 = xs_join(el, " ");
461
462 s2 = xs_fmt("<%s%s%s%s>",
463 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
464
465 s = xs_str_cat(s, s2);
466 } else {
467 /* treat end of divs as paragraph breaks */
468 if (strcmp(v, "</div>"))
469 s = xs_str_cat(s, "<p>");
470 }
471 }
472 else {
473 /* non-tag */
474 s = xs_str_cat(s, v);
475 }
476
477 n++;
478 }
479
480 return s;
481}
482
483
484xs_str *encode_html(const char *str)
485/* escapes html characters */
486{
487 xs_str *encoded = xs_html_encode((char *)str);
488
489 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
490 encoded = xs_replace_i(encoded, "<br>", "<br>");
491
492 return encoded;
493}
