2.76
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/* Non-whitespace without trailing comma, period or closing paren */
82#define NOSPACE "([^[:space:],.)]+|[,.)]+[^[:space:],.)])+"
83
84static xs_str *format_line(const char *line, xs_list **attach)
85/* formats a line */
86{
87 xs_str *s = xs_str_new(NULL);
88 char *p;
89 const char *v;
90
91 /* split by markup */
92 xs *sm = xs_regex_split(line,
93 "("
94 "`[^`]+`" "|"
95 "~~[^~]+~~" "|"
96 "\\*\\*?\\*?[^\\*]+\\*?\\*?\\*" "|"
97 "__[^_]+__" "|" //anzu
98 "!\\[[^]]+\\]\\([^\\)]+\\)" "|"
99 "\\[[^]]+\\]\\([^\\)]+\\)" "|"
100 "[a-z]+:/" "/" NOSPACE "|"
101 "(mailto|xmpp):[^@[:space:]]+@" NOSPACE
102 ")");
103 int n = 0;
104
105 p = sm;
106 while (xs_list_iter(&p, &v)) {
107 if ((n & 0x1)) {
108 /* markup */
109 if (xs_startswith(v, "`")) {
110 xs *s1 = xs_strip_chars_i(xs_dup(v), "`");
111 xs *e1 = encode_html(s1);
112 xs *s2 = xs_fmt("<code>%s</code>", e1);
113 s = xs_str_cat(s, s2);
114 }
115 else
116 if (xs_startswith(v, "***")) {
117 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
118 xs *s2 = xs_fmt("<b><i>%s</i></b>", s1);
119 s = xs_str_cat(s, s2);
120 }
121 else
122 if (xs_startswith(v, "**")) {
123 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
124 xs *s2 = xs_fmt("<b>%s</b>", s1);
125 s = xs_str_cat(s, s2);
126 }
127 else
128 if (xs_startswith(v, "*")) {
129 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
130 xs *s2 = xs_fmt("<i>%s</i>", s1);
131 s = xs_str_cat(s, s2);
132 }
133 //anzu - begin
134 else
135 if (xs_startswith(v, "__")) {
136 xs *s1 = xs_strip_chars_i(xs_dup(v), "_");
137 xs *s2 = xs_fmt("<u>%s</u>", s1);
138 s = xs_str_cat(s, s2);
139 }
140 //anzu - end
141 else
142 if (xs_startswith(v, "~~")) {
143 xs *s1 = xs_strip_chars_i(xs_dup(v), "~");
144 xs *e1 = encode_html(s1);
145 xs *s2 = xs_fmt("<s>%s</s>", e1);
146 s = xs_str_cat(s, s2);
147 }
148 else
149 if (*v == '[') {
150 /* markdown-like links [label](url) */
151 xs *w = xs_strip_chars_i(
152 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
153 "![)");
154 xs *l = xs_split_n(w, "](", 1);
155
156 if (xs_list_len(l) == 2) {
157 const char *name = xs_list_get(l, 0);
158 const char *url = xs_list_get(l, 1);
159
160 xs *link = xs_fmt("<a href=\"%s\">%s</a>", url, name);
161
162 s = xs_str_cat(s, link);
163
164 if (attach) {
165 /* also add the link as an attachment */
166 xs *d = xs_dict_new();
167
168 d = xs_dict_append(d, "mediaType", "text/html");
169 d = xs_dict_append(d, "url", url);
170 d = xs_dict_append(d, "name", name);
171 d = xs_dict_append(d, "type", "Link");
172
173 *attach = xs_list_append(*attach, d);
174 }
175 }
176 else
177 s = xs_str_cat(s, v);
178 }
179 else
180 if (*v == '!') {
181 /* markdown-like images  */
182 xs *w = xs_strip_chars_i(
183 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
184 "![)");
185 xs *l = xs_split_n(w, "](", 1);
186
187 if (xs_list_len(l) == 2) {
188 const char *alt_text = xs_list_get(l, 0);
189 const char *img_url = xs_list_get(l, 1);
190 const char *mime = xs_mime_by_ext(img_url);
191
192 if (attach != NULL && xs_startswith(mime, "image/")) {
193 const xs_dict *ad;
194 int add = 1;
195
196 xs_list_foreach(*attach, ad) {
197 if (strcmp(xs_dict_get_def(ad, "url", ""), img_url) == 0) {
198 add = 0;
199 break;
200 }
201 }
202
203 if (add) {
204 xs *d = xs_dict_new();
205
206 d = xs_dict_append(d, "mediaType", mime);
207 d = xs_dict_append(d, "url", img_url);
208 d = xs_dict_append(d, "name", alt_text);
209 d = xs_dict_append(d, "type", "Image");
210
211 *attach = xs_list_append(*attach, d);
212 }
213 }
214 else {
215 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
216
217 s = xs_str_cat(s, link);
218 }
219 }
220 else
221 s = xs_str_cat(s, v);
222 }
223 else
224 if (xs_str_in(v, ":/" "/") != -1) {
225 /* direct URLs in the post body */
226 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
227
228 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
229
230 const char *mime = xs_mime_by_ext(v2);
231
232 if (attach != NULL && xs_startswith(mime, "image/")) {
233 /* if it's a link to an image, insert it as an attachment */
234 const xs_dict *ad;
235 int add = 1;
236
237 xs_list_foreach(*attach, ad) {
238 if (strcmp(xs_dict_get_def(ad, "url", ""), v2) == 0) {
239 add = 0;
240 break;
241 }
242 }
243
244 if (add) {
245 xs *d = xs_dict_new();
246
247 d = xs_dict_append(d, "mediaType", mime);
248 d = xs_dict_append(d, "url", v2);
249 d = xs_dict_append(d, "name", "");
250 d = xs_dict_append(d, "type", "Image");
251
252 *attach = xs_list_append(*attach, d);
253 }
254 }
255 else {
256 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
257 s = xs_str_cat(s, s1);
258
259 if (attach) {
260 /* also add the link as an attachment */
261 xs *d = xs_dict_new();
262
263 d = xs_dict_append(d, "mediaType", "text/html");
264 d = xs_dict_append(d, "url", v2);
265 d = xs_dict_append(d, "name", "");
266 d = xs_dict_append(d, "type", "Link");
267
268 *attach = xs_list_append(*attach, d);
269 }
270 }
271 }
272 else
273 if (xs_match(v, "mailto*|xmpp*")) {
274 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
275
276 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
277
278 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
279 s = xs_str_cat(s, s1);
280 }
281 else
282 s = xs_str_cat(s, v);
283 }
284 else
285 /* surrounded text, copy directly */
286 s = xs_str_cat(s, v);
287
288 n++;
289 }
290
291 return s;
292}
293
294
295xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
296/* formats a content using some Markdown rules */
297{
298 xs_str *s = xs_str_new(NULL);
299 int in_pre = 0;
300 int in_blq = 0;
301 xs *list;
302 char *p;
303 const char *v;
304
305 /* work by lines */
306 list = xs_split(content, "\n");
307
308 p = list;
309 while (xs_list_iter(&p, &v)) {
310 xs *ss = NULL;
311
312 if (strcmp(v, "```") == 0) {
313 if (!in_pre)
314 s = xs_str_cat(s, "<pre>");
315 else
316 s = xs_str_cat(s, "</pre>");
317
318 in_pre = !in_pre;
319 continue;
320 }
321
322 if (in_pre) {
323 // Encode all HTML characters when we're in pre element until we are out.
324 ss = encode_html(v);
325
326 s = xs_str_cat(s, ss);
327 s = xs_str_cat(s, "<br>");
328 continue;
329 }
330
331 else
332 ss = xs_strip_i(format_line(v, attach));
333
334 if (xs_startswith(ss, "---")) {
335 /* delete the --- */
336 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
337 s = xs_str_cat(s, "<hr>");
338
339 s = xs_str_cat(s, ss);
340
341 continue;
342 }
343
344 //anzu - begin
345 // h1 reserved for snac?
346 if (xs_startswith(ss, "# ")) {
347 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
348 s = xs_str_cat(s, "<h2>");
349 s = xs_str_cat(s, ss);
350 s = xs_str_cat(s, "</h2>");
351 continue;
352 }
353 if (xs_startswith(ss, "## ")) {
354 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
355 s = xs_str_cat(s, "<h2>");
356 s = xs_str_cat(s, ss);
357 s = xs_str_cat(s, "</h2>");
358 continue;
359 }
360 if (xs_startswith(ss, "### ")) {
361 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
362 s = xs_str_cat(s, "<h3>");
363 s = xs_str_cat(s, ss);
364 s = xs_str_cat(s, "</h3>");
365 continue;
366 }
367 //anzu - end
368
369 if (xs_startswith(ss, ">")) {
370 /* delete the > and subsequent spaces */
371 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
372
373 if (!in_blq) {
374 s = xs_str_cat(s, "<blockquote>");
375 in_blq = 1;
376 }
377
378 s = xs_str_cat(s, ss);
379 s = xs_str_cat(s, "<br>");
380
381 continue;
382 }
383
384 if (in_blq) {
385 s = xs_str_cat(s, "</blockquote>");
386 in_blq = 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, "</h2><br>", "</h2>"); //anzu ???
403 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
404
405 {
406 /* traditional emoticons */
407 xs *d = emojis();
408 int c = 0;
409 const char *k, *v;
410
411 while (xs_dict_next(d, &k, &v, &c)) {
412 const char *t = xs_mime_by_ext(v);
413
414 /* is it an URL to an image? */
415 if (xs_startswith(v, "https:/" "/") &&
416 (xs_startswith(t, "image/") || strcmp(t, "application/octet-stream") == 0)) {
417 if (tag && xs_str_in(s, k) != -1) {
418 /* add the emoji to the tag list */
419 xs *e = xs_dict_new();
420 xs *i = xs_dict_new();
421 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
422
423 e = xs_dict_append(e, "id", v);
424 e = xs_dict_append(e, "type", "Emoji");
425 e = xs_dict_append(e, "name", k);
426 e = xs_dict_append(e, "updated", u);
427
428 i = xs_dict_append(i, "type", "Image");
429 i = xs_dict_append(i, "mediaType", t);
430 i = xs_dict_append(i, "url", v);
431 e = xs_dict_append(e, "icon", i);
432
433 *tag = xs_list_append(*tag, e);
434 }
435 }
436 else
437 s = xs_replace_i(s, k, v);
438 }
439 }
440
441 return s;
442}
443
444
445const char *valid_tags[] = {
446 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
447 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi",
448 "h2","h3", //anzu
449 NULL
450};
451
452xs_str *sanitize(const char *content)
453/* cleans dangerous HTML output */
454{
455 xs_str *s = xs_str_new(NULL);
456 xs *sl;
457 int n = 0;
458 char *p;
459 const char *v;
460
461 sl = xs_regex_split(content, "</?[^>]+>");
462
463 p = sl;
464
465 n = 0;
466 while (xs_list_iter(&p, &v)) {
467 if (n & 0x1) {
468 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
469 xs *l1 = xs_split_n(s1, " ", 1);
470 xs *tag = xs_tolower_i(xs_dup(xs_list_get(l1, 0)));
471 xs *s2 = NULL;
472 int i;
473
474 /* check if it's one of the valid tags */
475 for (i = 0; valid_tags[i]; i++) {
476 if (strcmp(tag, valid_tags[i]) == 0)
477 break;
478 }
479
480 if (valid_tags[i]) {
481 /* accepted tag: rebuild it with only the accepted elements */
482 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=(\"[^\"]*\"|'[^']*')");
483 xs *s3 = xs_join(el, " ");
484
485 s2 = xs_fmt("<%s%s%s%s>",
486 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
487
488 s = xs_str_cat(s, s2);
489 } else {
490 /* treat end of divs as paragraph breaks */
491 if (strcmp(v, "</div>"))
492 s = xs_str_cat(s, "<p>");
493 }
494 }
495 else {
496 /* non-tag */
497 s = xs_str_cat(s, v);
498 }
499
500 n++;
501 }
502
503 return s;
504}
505
506
507xs_str *encode_html(const char *str)
508/* escapes html characters */
509{
510 xs_str *encoded = xs_html_encode((char *)str);
511
512 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
513 encoded = xs_replace_i(encoded, "<br>", "<br>");
514
515 return encoded;
516}
