2.78
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_strip_chars_i(
153 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
154 "![)");
155 xs *l = xs_split_n(w, "](", 1);
156
157 if (xs_list_len(l) == 2) {
158 const char *name = xs_list_get(l, 0);
159 const char *url = xs_list_get(l, 1);
160
161 xs *link = xs_fmt("<a href=\"%s\">%s</a>", url, name);
162
163 s = xs_str_cat(s, link);
164 }
165 else
166 s = xs_str_cat(s, v);
167 }
168 else
169 if (*v == '!') {
170 /* markdown-like images  */
171 xs *w = xs_strip_chars_i(
172 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
173 "![)");
174 xs *l = xs_split_n(w, "](", 1);
175
176 if (xs_list_len(l) == 2) {
177 const char *alt_text = xs_list_get(l, 0);
178 const char *img_url = xs_list_get(l, 1);
179 const char *mime = xs_mime_by_ext(img_url);
180
181 if (attach != NULL && xs_startswith(mime, "image/")) {
182 const xs_dict *ad;
183 int add = 1;
184
185 xs_list_foreach(*attach, ad) {
186 if (strcmp(xs_dict_get_def(ad, "url", ""), img_url) == 0) {
187 add = 0;
188 break;
189 }
190 }
191
192 if (add) {
193 xs *d = xs_dict_new();
194
195 d = xs_dict_append(d, "mediaType", mime);
196 d = xs_dict_append(d, "url", img_url);
197 d = xs_dict_append(d, "name", alt_text);
198 d = xs_dict_append(d, "type", "Image");
199
200 *attach = xs_list_append(*attach, d);
201 }
202 }
203 else {
204 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
205
206 s = xs_str_cat(s, link);
207 }
208 }
209 else
210 s = xs_str_cat(s, v);
211 }
212 else
213 if (xs_str_in(v, ":/" "/") != -1) {
214 /* direct URLs in the post body */
215 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
216
217 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
218
219 const char *mime = xs_mime_by_ext(v2);
220
221 if (attach != NULL && xs_startswith(mime, "image/")) {
222 /* if it's a link to an image, insert it as an attachment */
223 const xs_dict *ad;
224 int add = 1;
225
226 xs_list_foreach(*attach, ad) {
227 if (strcmp(xs_dict_get_def(ad, "url", ""), v2) == 0) {
228 add = 0;
229 break;
230 }
231 }
232
233 if (add) {
234 xs *d = xs_dict_new();
235
236 d = xs_dict_append(d, "mediaType", mime);
237 d = xs_dict_append(d, "url", v2);
238 d = xs_dict_append(d, "name", "");
239 d = xs_dict_append(d, "type", "Image");
240
241 *attach = xs_list_append(*attach, d);
242 }
243 }
244 else {
245 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
246 s = xs_str_cat(s, s1);
247 }
248 }
249 else
250 if (xs_match(v, "mailto*|xmpp*")) {
251 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
252
253 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
254
255 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
256 s = xs_str_cat(s, s1);
257 }
258 else
259 s = xs_str_cat(s, v);
260 }
261 else
262 /* surrounded text, copy directly */
263 s = xs_str_cat(s, v);
264
265 n++;
266 }
267
268 return s;
269}
270
271
272xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
273/* formats a content using some Markdown rules */
274{
275 xs_str *s = xs_str_new(NULL);
276 int in_pre = 0;
277 int in_blq = 0;
278 xs *list;
279 char *p;
280 const char *v;
281
282 /* work by lines */
283 list = xs_split(content, "\n");
284
285 p = list;
286 while (xs_list_iter(&p, &v)) {
287 xs *ss = NULL;
288
289 if (strcmp(v, "```") == 0) {
290 if (!in_pre)
291 s = xs_str_cat(s, "<pre>");
292 else
293 s = xs_str_cat(s, "</pre>");
294
295 in_pre = !in_pre;
296 continue;
297 }
298
299 if (in_pre) {
300 // Encode all HTML characters when we're in pre element until we are out.
301 ss = encode_html(v);
302
303 s = xs_str_cat(s, ss);
304 s = xs_str_cat(s, "<br>");
305 continue;
306 }
307
308 else
309 ss = xs_strip_i(format_line(v, attach));
310
311 if (xs_startswith(ss, "---")) {
312 /* delete the --- */
313 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
314 s = xs_str_cat(s, "<hr>");
315
316 s = xs_str_cat(s, ss);
317
318 continue;
319 }
320
321 //anzu - begin
322 // h1 reserved for snac?
323 if (xs_startswith(ss, "# ")) {
324 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
325 s = xs_str_cat(s, "<h2>");
326 s = xs_str_cat(s, ss);
327 s = xs_str_cat(s, "</h2>");
328 continue;
329 }
330 if (xs_startswith(ss, "## ")) {
331 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
332 s = xs_str_cat(s, "<h2>");
333 s = xs_str_cat(s, ss);
334 s = xs_str_cat(s, "</h2>");
335 continue;
336 }
337 if (xs_startswith(ss, "### ")) {
338 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
339 s = xs_str_cat(s, "<h3>");
340 s = xs_str_cat(s, ss);
341 s = xs_str_cat(s, "</h3>");
342 continue;
343 }
344 //anzu - end
345
346 if (xs_startswith(ss, ">")) {
347 /* delete the > and subsequent spaces */
348 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
349
350 if (!in_blq) {
351 s = xs_str_cat(s, "<blockquote>");
352 in_blq = 1;
353 }
354
355 s = xs_str_cat(s, ss);
356 s = xs_str_cat(s, "<br>");
357
358 continue;
359 }
360
361 if (in_blq) {
362 s = xs_str_cat(s, "</blockquote>");
363 in_blq = 0;
364 }
365
366 s = xs_str_cat(s, ss);
367 s = xs_str_cat(s, "<br>");
368 }
369
370 if (in_blq)
371 s = xs_str_cat(s, "</blockquote>");
372 if (in_pre)
373 s = xs_str_cat(s, "</pre>");
374
375 /* some beauty fixes */
376 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
377 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
378 s = xs_replace_i(s, "</pre><br>", "</pre>");
379 s = xs_replace_i(s, "</h2><br>", "</h2>"); //anzu ???
380 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
381
382 {
383 /* traditional emoticons */
384 xs *d = emojis();
385 int c = 0;
386 const char *k, *v;
387
388 while (xs_dict_next(d, &k, &v, &c)) {
389 const char *t = xs_mime_by_ext(v);
390
391 /* is it an URL to an image? */
392 if (xs_startswith(v, "https:/" "/") &&
393 (xs_startswith(t, "image/") || strcmp(t, "application/octet-stream") == 0)) {
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_utf8_to_lower(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}
