2.79
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#include "xs_unicode.h"
12
13#include "snac.h"
14
15/* emoticons, people laughing and such */
16const char *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 **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 xs *list;
282 char *p;
283 const char *v;
284
285 /* work by lines */
286 list = xs_split(content, "\n");
287
288 p = list;
289 while (xs_list_iter(&p, &v)) {
290 xs *ss = NULL;
291
292 if (strcmp(v, "```") == 0) {
293 if (!in_pre)
294 s = xs_str_cat(s, "<pre>");
295 else
296 s = xs_str_cat(s, "</pre>");
297
298 in_pre = !in_pre;
299 continue;
300 }
301
302 if (in_pre) {
303 // Encode all HTML characters when we're in pre element until we are out.
304 ss = encode_html(v);
305
306 s = xs_str_cat(s, ss);
307 s = xs_str_cat(s, "<br>");
308 continue;
309 }
310
311 else
312 ss = xs_strip_i(format_line(v, attach));
313
314 if (xs_startswith(ss, "---")) {
315 /* delete the --- */
316 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
317 s = xs_str_cat(s, "<hr>");
318
319 s = xs_str_cat(s, ss);
320
321 continue;
322 }
323
324 //anzu - begin
325 // h1 reserved for snac?
326 if (xs_startswith(ss, "# ")) {
327 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
328 s = xs_str_cat(s, "<h2>");
329 s = xs_str_cat(s, ss);
330 s = xs_str_cat(s, "</h2>");
331 continue;
332 }
333 if (xs_startswith(ss, "## ")) {
334 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
335 s = xs_str_cat(s, "<h2>");
336 s = xs_str_cat(s, ss);
337 s = xs_str_cat(s, "</h2>");
338 continue;
339 }
340 if (xs_startswith(ss, "### ")) {
341 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
342 s = xs_str_cat(s, "<h3>");
343 s = xs_str_cat(s, ss);
344 s = xs_str_cat(s, "</h3>");
345 continue;
346 }
347 //anzu - end
348
349 if (xs_startswith(ss, ">")) {
350 /* delete the > and subsequent spaces */
351 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
352
353 if (!in_blq) {
354 s = xs_str_cat(s, "<blockquote>");
355 in_blq = 1;
356 }
357
358 s = xs_str_cat(s, ss);
359 s = xs_str_cat(s, "<br>");
360
361 continue;
362 }
363
364 if (in_blq) {
365 s = xs_str_cat(s, "</blockquote>");
366 in_blq = 0;
367 }
368
369 s = xs_str_cat(s, ss);
370 s = xs_str_cat(s, "<br>");
371 }
372
373 if (in_blq)
374 s = xs_str_cat(s, "</blockquote>");
375 if (in_pre)
376 s = xs_str_cat(s, "</pre>");
377
378 /* some beauty fixes */
379 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
380 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
381 s = xs_replace_i(s, "</pre><br>", "</pre>");
382 s = xs_replace_i(s, "</h2><br>", "</h2>"); //anzu ???
383 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
384
385 {
386 /* traditional emoticons */
387 xs *d = emojis();
388 int c = 0;
389 const char *k, *v;
390
391 while (xs_dict_next(d, &k, &v, &c)) {
392 const char *t = xs_mime_by_ext(v);
393
394 /* is it an URL to an image? */
395 if (xs_startswith(v, "https:/" "/") &&
396 (xs_startswith(t, "image/") || strcmp(t, "application/octet-stream") == 0)) {
397 if (tag && xs_str_in(s, k) != -1) {
398 /* add the emoji to the tag list */
399 xs *e = xs_dict_new();
400 xs *i = xs_dict_new();
401 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
402
403 e = xs_dict_append(e, "id", v);
404 e = xs_dict_append(e, "type", "Emoji");
405 e = xs_dict_append(e, "name", k);
406 e = xs_dict_append(e, "updated", u);
407
408 i = xs_dict_append(i, "type", "Image");
409 i = xs_dict_append(i, "mediaType", t);
410 i = xs_dict_append(i, "url", v);
411 e = xs_dict_append(e, "icon", i);
412
413 *tag = xs_list_append(*tag, e);
414 }
415 }
416 else
417 s = xs_replace_i(s, k, v);
418 }
419 }
420
421 return s;
422}
423
424
425const char *valid_tags[] = {
426 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
427 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi",
428 "h2","h3", //anzu
429 NULL
430};
431
432xs_str *sanitize(const char *content)
433/* cleans dangerous HTML output */
434{
435 xs_str *s = xs_str_new(NULL);
436 xs *sl;
437 int n = 0;
438 char *p;
439 const char *v;
440
441 sl = xs_regex_split(content, "</?[^>]+>");
442
443 p = sl;
444
445 n = 0;
446 while (xs_list_iter(&p, &v)) {
447 if (n & 0x1) {
448 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
449 xs *l1 = xs_split_n(s1, " ", 1);
450 xs *tag = xs_utf8_to_lower(xs_list_get(l1, 0));
451 xs *s2 = NULL;
452 int i;
453
454 /* check if it's one of the valid tags */
455 for (i = 0; valid_tags[i]; i++) {
456 if (strcmp(tag, valid_tags[i]) == 0)
457 break;
458 }
459
460 if (valid_tags[i]) {
461 /* accepted tag: rebuild it with only the accepted elements */
462 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=(\"[^\"]*\"|'[^']*')");
463 xs *s3 = xs_join(el, " ");
464
465 s2 = xs_fmt("<%s%s%s%s>",
466 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
467
468 s = xs_str_cat(s, s2);
469 } else {
470 /* treat end of divs as paragraph breaks */
471 if (strcmp(v, "</div>"))
472 s = xs_str_cat(s, "<p>");
473 }
474 }
475 else {
476 /* non-tag */
477 s = xs_str_cat(s, v);
478 }
479
480 n++;
481 }
482
483 return s;
484}
485
486
487xs_str *encode_html(const char *str)
488/* escapes html characters */
489{
490 xs_str *encoded = xs_html_encode((char *)str);
491
492 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
493 encoded = xs_replace_i(encoded, "<br>", "<br>");
494
495 return encoded;
496}
