2.77
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 else
165 s = xs_str_cat(s, v);
166 }
167 else
168 if (*v == '!') {
169 /* markdown-like images  */
170 xs *w = xs_strip_chars_i(
171 xs_replace_i(xs_replace(v, "#", "#"), "@", "@"),
172 "![)");
173 xs *l = xs_split_n(w, "](", 1);
174
175 if (xs_list_len(l) == 2) {
176 const char *alt_text = xs_list_get(l, 0);
177 const char *img_url = xs_list_get(l, 1);
178 const char *mime = xs_mime_by_ext(img_url);
179
180 if (attach != NULL && xs_startswith(mime, "image/")) {
181 const xs_dict *ad;
182 int add = 1;
183
184 xs_list_foreach(*attach, ad) {
185 if (strcmp(xs_dict_get_def(ad, "url", ""), img_url) == 0) {
186 add = 0;
187 break;
188 }
189 }
190
191 if (add) {
192 xs *d = xs_dict_new();
193
194 d = xs_dict_append(d, "mediaType", mime);
195 d = xs_dict_append(d, "url", img_url);
196 d = xs_dict_append(d, "name", alt_text);
197 d = xs_dict_append(d, "type", "Image");
198
199 *attach = xs_list_append(*attach, d);
200 }
201 }
202 else {
203 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
204
205 s = xs_str_cat(s, link);
206 }
207 }
208 else
209 s = xs_str_cat(s, v);
210 }
211 else
212 if (xs_str_in(v, ":/" "/") != -1) {
213 /* direct URLs in the post body */
214 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
215
216 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
217
218 const char *mime = xs_mime_by_ext(v2);
219
220 if (attach != NULL && xs_startswith(mime, "image/")) {
221 /* if it's a link to an image, insert it as an attachment */
222 const xs_dict *ad;
223 int add = 1;
224
225 xs_list_foreach(*attach, ad) {
226 if (strcmp(xs_dict_get_def(ad, "url", ""), v2) == 0) {
227 add = 0;
228 break;
229 }
230 }
231
232 if (add) {
233 xs *d = xs_dict_new();
234
235 d = xs_dict_append(d, "mediaType", mime);
236 d = xs_dict_append(d, "url", v2);
237 d = xs_dict_append(d, "name", "");
238 d = xs_dict_append(d, "type", "Image");
239
240 *attach = xs_list_append(*attach, d);
241 }
242 }
243 else {
244 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
245 s = xs_str_cat(s, s1);
246 }
247 }
248 else
249 if (xs_match(v, "mailto*|xmpp*")) {
250 xs *u = xs_replace_i(xs_replace(v, "#", "#"), "@", "@");
251
252 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
253
254 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
255 s = xs_str_cat(s, s1);
256 }
257 else
258 s = xs_str_cat(s, v);
259 }
260 else
261 /* surrounded text, copy directly */
262 s = xs_str_cat(s, v);
263
264 n++;
265 }
266
267 return s;
268}
269
270
271xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
272/* formats a content using some Markdown rules */
273{
274 xs_str *s = xs_str_new(NULL);
275 int in_pre = 0;
276 int in_blq = 0;
277 xs *list;
278 char *p;
279 const char *v;
280
281 /* work by lines */
282 list = xs_split(content, "\n");
283
284 p = list;
285 while (xs_list_iter(&p, &v)) {
286 xs *ss = NULL;
287
288 if (strcmp(v, "```") == 0) {
289 if (!in_pre)
290 s = xs_str_cat(s, "<pre>");
291 else
292 s = xs_str_cat(s, "</pre>");
293
294 in_pre = !in_pre;
295 continue;
296 }
297
298 if (in_pre) {
299 // Encode all HTML characters when we're in pre element until we are out.
300 ss = encode_html(v);
301
302 s = xs_str_cat(s, ss);
303 s = xs_str_cat(s, "<br>");
304 continue;
305 }
306
307 else
308 ss = xs_strip_i(format_line(v, attach));
309
310 if (xs_startswith(ss, "---")) {
311 /* delete the --- */
312 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
313 s = xs_str_cat(s, "<hr>");
314
315 s = xs_str_cat(s, ss);
316
317 continue;
318 }
319
320 //anzu - begin
321 // h1 reserved for snac?
322 if (xs_startswith(ss, "# ")) {
323 ss = xs_strip_i(xs_crop_i(ss, 2, 0));
324 s = xs_str_cat(s, "<h2>");
325 s = xs_str_cat(s, ss);
326 s = xs_str_cat(s, "</h2>");
327 continue;
328 }
329 if (xs_startswith(ss, "## ")) {
330 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
331 s = xs_str_cat(s, "<h2>");
332 s = xs_str_cat(s, ss);
333 s = xs_str_cat(s, "</h2>");
334 continue;
335 }
336 if (xs_startswith(ss, "### ")) {
337 ss = xs_strip_i(xs_crop_i(ss, 4, 0));
338 s = xs_str_cat(s, "<h3>");
339 s = xs_str_cat(s, ss);
340 s = xs_str_cat(s, "</h3>");
341 continue;
342 }
343 //anzu - end
344
345 if (xs_startswith(ss, ">")) {
346 /* delete the > and subsequent spaces */
347 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
348
349 if (!in_blq) {
350 s = xs_str_cat(s, "<blockquote>");
351 in_blq = 1;
352 }
353
354 s = xs_str_cat(s, ss);
355 s = xs_str_cat(s, "<br>");
356
357 continue;
358 }
359
360 if (in_blq) {
361 s = xs_str_cat(s, "</blockquote>");
362 in_blq = 0;
363 }
364
365 s = xs_str_cat(s, ss);
366 s = xs_str_cat(s, "<br>");
367 }
368
369 if (in_blq)
370 s = xs_str_cat(s, "</blockquote>");
371 if (in_pre)
372 s = xs_str_cat(s, "</pre>");
373
374 /* some beauty fixes */
375 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
376 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
377 s = xs_replace_i(s, "</pre><br>", "</pre>");
378 s = xs_replace_i(s, "</h2><br>", "</h2>"); //anzu ???
379 s = xs_replace_i(s, "</h3><br>", "</h3>"); //anzu ???
380
381 {
382 /* traditional emoticons */
383 xs *d = emojis();
384 int c = 0;
385 const char *k, *v;
386
387 while (xs_dict_next(d, &k, &v, &c)) {
388 const char *t = xs_mime_by_ext(v);
389
390 /* is it an URL to an image? */
391 if (xs_startswith(v, "https:/" "/") &&
392 (xs_startswith(t, "image/") || strcmp(t, "application/octet-stream") == 0)) {
393 if (tag && xs_str_in(s, k) != -1) {
394 /* add the emoji to the tag list */
395 xs *e = xs_dict_new();
396 xs *i = xs_dict_new();
397 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
398
399 e = xs_dict_append(e, "id", v);
400 e = xs_dict_append(e, "type", "Emoji");
401 e = xs_dict_append(e, "name", k);
402 e = xs_dict_append(e, "updated", u);
403
404 i = xs_dict_append(i, "type", "Image");
405 i = xs_dict_append(i, "mediaType", t);
406 i = xs_dict_append(i, "url", v);
407 e = xs_dict_append(e, "icon", i);
408
409 *tag = xs_list_append(*tag, e);
410 }
411 }
412 else
413 s = xs_replace_i(s, k, v);
414 }
415 }
416
417 return s;
418}
419
420
421const char *valid_tags[] = {
422 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
423 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi",
424 "h2","h3", //anzu
425 NULL
426};
427
428xs_str *sanitize(const char *content)
429/* cleans dangerous HTML output */
430{
431 xs_str *s = xs_str_new(NULL);
432 xs *sl;
433 int n = 0;
434 char *p;
435 const char *v;
436
437 sl = xs_regex_split(content, "</?[^>]+>");
438
439 p = sl;
440
441 n = 0;
442 while (xs_list_iter(&p, &v)) {
443 if (n & 0x1) {
444 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
445 xs *l1 = xs_split_n(s1, " ", 1);
446 xs *tag = xs_tolower_i(xs_dup(xs_list_get(l1, 0)));
447 xs *s2 = NULL;
448 int i;
449
450 /* check if it's one of the valid tags */
451 for (i = 0; valid_tags[i]; i++) {
452 if (strcmp(tag, valid_tags[i]) == 0)
453 break;
454 }
455
456 if (valid_tags[i]) {
457 /* accepted tag: rebuild it with only the accepted elements */
458 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=(\"[^\"]*\"|'[^']*')");
459 xs *s3 = xs_join(el, " ");
460
461 s2 = xs_fmt("<%s%s%s%s>",
462 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
463
464 s = xs_str_cat(s, s2);
465 } else {
466 /* treat end of divs as paragraph breaks */
467 if (strcmp(v, "</div>"))
468 s = xs_str_cat(s, "<p>");
469 }
470 }
471 else {
472 /* non-tag */
473 s = xs_str_cat(s, v);
474 }
475
476 n++;
477 }
478
479 return s;
480}
481
482
483xs_str *encode_html(const char *str)
484/* escapes html characters */
485{
486 xs_str *encoded = xs_html_encode((char *)str);
487
488 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
489 encoded = xs_replace_i(encoded, "<br>", "<br>");
490
491 return encoded;
492}
