2.87
format.c
1/* snac - A simple, minimalistic ActivityPub instance */
2/* copyright (c) 2022 - 2026 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#include "xs_unicode.h"
12
13#include "snac.h"
14
15/* emoticons, people laughing and such */
16const char * const smileys[] = {
17 ":-)", "🙂",
18 ":-D", "😀",
19 "X-D", "😆",
20 ";-)", "😉",
21 "B-)", "😎",
22 ">:-(", "😡",
23 ":-(", "😞",
24 ":-*", "😘",
25 ":-/", "😕",
26 "8-o", "😲",
27 "%-)", "🤪",
28 ":_(", "😢",
29 ":-|", "😐",
30 "<3", "❤️",
31 ":facepalm:", "🤦",
32 ":shrug:", "🤷",
33 ":shrug2:", "¯\\_(ツ)_/¯",
34 ":eyeroll:", "🙄",
35 ":beer:", "🍺",
36 ":beers:", "🍻",
37 ":munch:", "😱",
38 ":thumb:", "👍",
39 NULL, NULL
40};
41
42
43xs_dict *emojis(void)
44/* returns a dict with the emojis */
45{
46 xs *fn = xs_fmt("%s/emojis.json", srv_basedir);
47 FILE *f;
48
49 if (mtime(fn) == 0) {
50 /* file does not exist; create it with the defaults */
51 xs *d = xs_dict_new();
52 const char * const *emo = smileys;
53
54 while (*emo) {
55 d = xs_dict_append(d, emo[0], emo[1]);
56 emo += 2;
57 }
58
59 if ((f = fopen(fn, "w")) != NULL) {
60 xs_json_dump(d, 4, f);
61 fclose(f);
62 }
63 else
64 srv_log(xs_fmt("Error creating '%s'", fn));
65 }
66
67 xs_dict *d = NULL;
68
69 if ((f = fopen(fn, "r")) != NULL) {
70 d = xs_json_load(f);
71 fclose(f);
72
73 if (d == NULL)
74 srv_log(xs_fmt("JSON parse error in '%s'", fn));
75 }
76 else
77 srv_log(xs_fmt("Error opening '%s'", fn));
78
79 return d;
80}
81
82/* Non-whitespace without trailing comma, period or closing paren */
83#define NOSPACE "([^[:space:],.)]+|[,.)]+[^[:space:],.)])+"
84
85static xs_str *format_line(const char *line, xs_list **attach)
86/* formats a line */
87{
88 xs_str *s = xs_str_new(NULL);
89 char *p;
90 const char *v;
91
92 /* split by markup */
93 xs *sm = xs_regex_split(line,
94 "("
95 "`[^`]+`" "|"
96 "~~[^~]+~~" "|"
97 "\\*\\*?\\*?[^\\*]+\\*?\\*?\\*" "|"
98 "__[^_]+__" "|" //anzu
99 "!\\[[^]]+\\]\\([^\\)]+\\)\\)?" "|"
100 "\\[[^]]+\\]\\([^\\)]+\\)\\)?" "|"
101 "[a-z]+:/" "/" NOSPACE "|"
102 "(mailto|xmpp):[^@[:space:]]+@" NOSPACE
103 ")");
104 int n = 0;
105
106 p = sm;
107 while (xs_list_iter(&p, &v)) {
108 if ((n & 0x1)) {
109 /* markup */
110 if (xs_startswith(v, "`")) {
111 xs *s1 = xs_strip_chars_i(xs_dup(v), "`");
112 xs *e1 = encode_html(s1);
113 xs *s2 = xs_fmt("<code>%s</code>", e1);
114 s = xs_str_cat(s, s2);
115 }
116 else
117 if (xs_startswith(v, "***")) {
118 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
119 xs *s2 = xs_fmt("<b><i>%s</i></b>", s1);
120 s = xs_str_cat(s, s2);
121 }
122 else
123 if (xs_startswith(v, "**")) {
124 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
125 xs *s2 = xs_fmt("<b>%s</b>", s1);
126 s = xs_str_cat(s, s2);
127 }
128 else
129 if (xs_startswith(v, "*")) {
130 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
131 xs *s2 = xs_fmt("<i>%s</i>", s1);
132 s = xs_str_cat(s, s2);
133 }
134 //anzu - begin
135 else
136 if (xs_startswith(v, "__")) {
137 xs *s1 = xs_strip_chars_i(xs_dup(v), "_");
138 xs *s2 = xs_fmt("<u>%s</u>", s1);
139 s = xs_str_cat(s, s2);
140 }
141 //anzu - end
142 else
143 if (xs_startswith(v, "~~")) {
144 xs *s1 = xs_strip_chars_i(xs_dup(v), "~");
145 xs *e1 = encode_html(s1);
146 xs *s2 = xs_fmt("<s>%s</s>", e1);
147 s = xs_str_cat(s, s2);
148 }
149 else
150 if (*v == '[') {
151 /* markdown-like links [label](url) */
152 xs *w = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
153 xs *l = xs_split_n(w, "](", 1);
154
155 if (xs_list_len(l) == 2) {
156 xs *name = xs_dup(xs_list_get(l, 0));
157 xs *url = xs_dup(xs_list_get(l, 1));
158
159 name = xs_crop_i(name, 1, 0);
160 url = xs_crop_i(url, 0, -1);
161
162 xs *link = xs_fmt("<a href=\"%s\">%s</a>", url, name);
163
164 s = xs_str_cat(s, link);
165 }
166 else
167 s = xs_str_cat(s, v);
168 }
169 else
170 if (*v == '!') {
171 /* markdown-like images  */
172 xs *w = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
173 xs *l = xs_split_n(w, "](", 1);
174
175 if (xs_list_len(l) == 2) {
176 xs *alt_text = xs_dup(xs_list_get(l, 0));
177 xs *img_url = xs_dup(xs_list_get(l, 1));
178
179 alt_text = xs_crop_i(alt_text, 2, 0);
180 img_url = xs_crop_i(img_url, 0, -1);
181
182 const char *mime = xs_mime_by_ext(img_url);
183
184 if (attach != NULL && xs_startswith(mime, "image/")) {
185 const xs_dict *ad;
186 int add = 1;
187
188 xs_list_foreach(*attach, ad) {
189 if (strcmp(xs_dict_get_def(ad, "url", ""), img_url) == 0) {
190 add = 0;
191 break;
192 }
193 }
194
195 if (add) {
196 xs *d = xs_dict_new();
197
198 d = xs_dict_append(d, "mediaType", mime);
199 d = xs_dict_append(d, "url", img_url);
200 d = xs_dict_append(d, "name", alt_text);
201 d = xs_dict_append(d, "type", "Image");
202
203 *attach = xs_list_append(*attach, d);
204 }
205 }
206 else {
207 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
208
209 s = xs_str_cat(s, link);
210 }
211 }
212 else
213 s = xs_str_cat(s, v);
214 }
215 else
216 if (xs_str_in(v, ":/" "/") != -1) {
217 /* direct URLs in the post body */
218 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
219
220 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
221
222 const char *mime = xs_mime_by_ext(v2);
223
224 if (attach != NULL && xs_startswith(mime, "image/")) {
225 /* if it's a link to an image, insert it as an attachment */
226 const xs_dict *ad;
227 int add = 1;
228
229 xs_list_foreach(*attach, ad) {
230 if (strcmp(xs_dict_get_def(ad, "url", ""), v2) == 0) {
231 add = 0;
232 break;
233 }
234 }
235
236 if (add) {
237 xs *d = xs_dict_new();
238
239 d = xs_dict_append(d, "mediaType", mime);
240 d = xs_dict_append(d, "url", v2);
241 d = xs_dict_append(d, "name", "");
242 d = xs_dict_append(d, "type", "Image");
243
244 *attach = xs_list_append(*attach, d);
245 }
246 }
247 else {
248 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
249 s = xs_str_cat(s, s1);
250 }
251 }
252 else
253 if (xs_match(v, "mailto*|xmpp*")) {
254 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
255
256 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
257
258 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
259 s = xs_str_cat(s, s1);
260 }
261 else
262 s = xs_str_cat(s, v);
263 }
264 else
265 /* surrounded text, copy directly */
266 s = xs_str_cat(s, v);
267
268 n++;
269 }
270
271 return s;
272}
273
274
275xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
276/* formats a content using some Markdown rules */
277{
278 xs_str *s = xs_str_new(NULL);
279 int in_pre = 0;
280 int in_blq = 0;
281 int in_ul = 0;
282 xs *list;
283 char *p;
284 const char *v;
285
286 /* work by lines */
287 list = xs_split(content, "\n");
288
289 p = list;
290 while (xs_list_iter(&p, &v)) {
291 xs *ss = NULL;
292
293 if (strcmp(v, "```") == 0) {
294 if (!in_pre)
295 s = xs_str_cat(s, "<pre>");
296 else
297 s = xs_str_cat(s, "</pre>");
298
299 in_pre = !in_pre;
300 continue;
301 }
302
303 if (in_pre) {
304 // Encode all HTML characters when we're in pre element until we are out.
305 ss = encode_html(v);
306
307 s = xs_str_cat(s, ss);
308 s = xs_str_cat(s, "<br>");
309 continue;
310 }
311
312 else
313 ss = xs_strip_i(format_line(v, attach));
314
315 if (xs_startswith(ss, "---")) {
316 /* delete the --- */
317 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
318 s = xs_str_cat(s, "<hr>");
319
320 s = xs_str_cat(s, ss);
321
322 continue;
323 }
324
325 //anzu - begin
326 // h1 reserved for snac?
327 if (xs_startswith(ss, "# ")) {
328 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
329 s = xs_str_cat(s, "<h2>");
330 s = xs_str_cat(s, ss);
331 s = xs_str_cat(s, "</h2>");
332 continue;
333 }
334 if (xs_startswith(ss, "## ")) {
335 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
336 s = xs_str_cat(s, "<h2>");
337 s = xs_str_cat(s, ss);
338 s = xs_str_cat(s, "</h2>");
339 continue;
340 }
341 if (xs_startswith(ss, "### ")) {
342 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
343 s = xs_str_cat(s, "<h3>");
344 s = xs_str_cat(s, ss);
345 s = xs_str_cat(s, "</h3>");
346 continue;
347 }
348 //anzu - end
349
350 if (xs_startswith(ss, ">")) {
351 /* delete the > and subsequent spaces */
352 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
353
354 if (!in_blq) {
355 s = xs_str_cat(s, "<blockquote>");
356 in_blq = 1;
357 }
358
359 s = xs_str_cat(s, ss);
360 s = xs_str_cat(s, "<br>");
361
362 continue;
363 }
364
365 if (xs_startswith(ss, "* ") || xs_startswith(ss, "- ")) {
366 /* unsorted list */
367 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
368
369 if (!in_ul) {
370 s = xs_str_cat(s, "<ul>");
371 in_ul = 1;
372 }
373
374 s = xs_str_cat(s, "<li>", ss);
375
376 continue;
377 }
378
379 if (in_blq) {
380 s = xs_str_cat(s, "</blockquote>");
381 in_blq = 0;
382 }
383
384 if (in_ul) {
385 s = xs_str_cat(s, "</ul>");
386 in_ul = 0;
387 }
388
389 s = xs_str_cat(s, ss);
390 s = xs_str_cat(s, "<br>");
391 }
392
393 if (in_blq)
394 s = xs_str_cat(s, "</blockquote>");
395 if (in_pre)
396 s = xs_str_cat(s, "</pre>");
397
398 /* some beauty fixes */
399 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
400 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
401 s = xs_replace_i(s, "</pre><br>", "</pre>");
402 s = xs_replace_i(s, "</ul><br>", "</ul>");
403 s = xs_replace_i(s, "</h2><br>", "</h2>"); //anzu ???
404 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
405
406 {
407 /* traditional emoticons */
408 xs *d = emojis();
409 int c = 0;
410 const char *k, *v;
411
412 while (xs_dict_next(d, &k, &v, &c)) {
413 const char *t = xs_mime_by_ext(v);
414
415 /* is it an URL to an image? */
416 if (xs_startswith(v, "https:/" "/") &&
417 (xs_startswith(t, "image/") || strcmp(t, "application/octet-stream") == 0)) {
418 if (tag && xs_str_in(s, k) != -1) {
419 /* add the emoji to the tag list */
420 xs *e = xs_dict_new();
421 xs *i = xs_dict_new();
422 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
423
424 e = xs_dict_append(e, "id", v);
425 e = xs_dict_append(e, "type", "Emoji");
426 e = xs_dict_append(e, "name", k);
427 e = xs_dict_append(e, "updated", u);
428
429 i = xs_dict_append(i, "type", "Image");
430 i = xs_dict_append(i, "mediaType", t);
431 i = xs_dict_append(i, "url", v);
432 e = xs_dict_append(e, "icon", i);
433
434 *tag = xs_list_append(*tag, e);
435 }
436 }
437 else
438 s = xs_replace_i(s, k, v);
439 }
440 }
441
442 return s;
443}
444
445
446const char * const valid_tags[] = {
447 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
448 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi",
449 "h2","h3", //anzu
450 NULL
451};
452
453xs_str *sanitize(const char *content)
454/* cleans dangerous HTML output */
455{
456 xs_str *s = xs_str_new(NULL);
457 xs *sl;
458 int n = 0;
459 char *p;
460 const char *v;
461
462 if (!content)
463 return NULL;
464
465 sl = xs_regex_split(content, "</?[^>]+>");
466
467 p = sl;
468
469 n = 0;
470 while (xs_list_iter(&p, &v)) {
471 if (n & 0x1) {
472 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
473 xs *l1 = xs_split_n(s1, " ", 1);
474 xs *tag = xs_utf8_to_lower(xs_list_get(l1, 0));
475 xs *s2 = NULL;
476 int i;
477
478 /* check if it's one of the valid tags */
479 for (i = 0; valid_tags[i]; i++) {
480 if (strcmp(tag, valid_tags[i]) == 0)
481 break;
482 }
483
484 if (valid_tags[i]) {
485 /* accepted tag: rebuild it with only the accepted elements */
486 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=(\"[^\"]*\"|'[^']*')");
487 xs *s3 = xs_join(el, " ");
488
489 s2 = xs_fmt("<%s%s%s%s>",
490 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
491
492 s = xs_str_cat(s, s2);
493 } else {
494 /* treat end of divs as paragraph breaks */
495 if (strcmp(v, "</div>"))
496 s = xs_str_cat(s, "<p>");
497 }
498 }
499 else {
500 /* non-tag */
501 s = xs_str_cat(s, v);
502 }
503
504 n++;
505 }
506
507 return s;
508}
509
510
511xs_str *encode_html(const char *str)
512/* escapes html characters */
513{
514 xs_str *encoded = xs_html_encode((char *)str);
515
516 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
517 encoded = xs_replace_i(encoded, "<br>", "<br>");
518
519 return encoded;
520}
