2.60
format.c
1/* snac - A simple, minimalistic ActivityPub instance */
2/* copyright (c) 2022 - 2024 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
11#include "snac.h"
12
13/* emoticons, people laughing and such */
14const char *smileys[] = {
15 ":-)", "🙂",
16 ":-D", "😀",
17 "X-D", "😆",
18 ";-)", "😉",
19 "B-)", "😎",
20 ">:-(", "😡",
21 ":-(", "😞",
22 ":-*", "😘",
23 ":-/", "😕",
24 "8-o", "😲",
25 "%-)", "🤪",
26 ":_(", "😢",
27 ":-|", "😐",
28 "<3", "❤️",
29 ":facepalm:", "🤦",
30 ":shrug:", "🤷",
31 ":shrug2:", "¯\\_(ツ)_/¯",
32 ":eyeroll:", "🙄",
33 ":beer:", "🍺",
34 ":beers:", "🍻",
35 ":munch:", "😱",
36 ":thumb:", "👍",
37 NULL, NULL
38};
39
40
41xs_dict *emojis(void)
42/* returns a dict with the emojis */
43{
44 xs *fn = xs_fmt("%s/emojis.json", srv_basedir);
45 FILE *f;
46
47 if (mtime(fn) == 0) {
48 /* file does not exist; create it with the defaults */
49 xs *d = xs_dict_new();
50 const char **emo = smileys;
51
52 while (*emo) {
53 d = xs_dict_append(d, emo[0], emo[1]);
54 emo += 2;
55 }
56
57 if ((f = fopen(fn, "w")) != NULL) {
58 xs_json_dump(d, 4, f);
59 fclose(f);
60 }
61 else
62 srv_log(xs_fmt("Error creating '%s'", fn));
63 }
64
65 xs_dict *d = NULL;
66
67 if ((f = fopen(fn, "r")) != NULL) {
68 d = xs_json_load(f);
69 fclose(f);
70
71 if (d == NULL)
72 srv_log(xs_fmt("JSON parse error in '%s'", fn));
73 }
74 else
75 srv_log(xs_fmt("Error opening '%s'", fn));
76
77 return d;
78}
79
80
81static xs_str *format_line(const char *line, xs_list **attach)
82/* formats a line */
83{
84 xs_str *s = xs_str_new(NULL);
85 char *p;
86 const char *v;
87
88 /* split by markup */
89 xs *sm = xs_regex_split(line,
90 "("
91 "`[^`]+`" "|"
92 "~~[^~]+~~" "|"
93 "\\*\\*?\\*?[^\\*]+\\*?\\*?\\*" "|"
94 "!\\[[^]]+\\]\\([^\\)]+\\)" "|"
95 "\\[[^]]+\\]\\([^\\)]+\\)" "|"
96 "https?:/" "/[^[:space:]]+"
97 ")");
98 int n = 0;
99
100 p = sm;
101 while (xs_list_iter(&p, &v)) {
102 if ((n & 0x1)) {
103 /* markup */
104 if (xs_startswith(v, "`")) {
105 xs *s1 = xs_strip_chars_i(xs_dup(v), "`");
106 xs *e1 = encode_html(s1);
107 xs *s2 = xs_fmt("<code>%s</code>", e1);
108 s = xs_str_cat(s, s2);
109 }
110 else
111 if (xs_startswith(v, "***")) {
112 xs *s1 = xs_strip_chars_i(xs_dup(v), "*");
113 xs *s2 = xs_fmt("<b><i>%s</i></b>", s1);
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>%s</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("<i>%s</i>", 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 *e1 = encode_html(s1);
132 xs *s2 = xs_fmt("<s>%s</s>", e1);
133 s = xs_str_cat(s, s2);
134 }
135 else
136 if (xs_startswith(v, "http")) {
137 xs *u = xs_replace(v, "#", "#");
138 xs *v2 = xs_strip_chars_i(xs_dup(u), ".,)");
139
140 const char *mime = xs_mime_by_ext(v2);
141
142 if (attach != NULL && xs_startswith(mime, "image/")) {
143 /* if it's a link to an image, insert it as an attachment */
144 xs *d = xs_dict_new();
145
146 d = xs_dict_append(d, "mediaType", mime);
147 d = xs_dict_append(d, "url", v2);
148 d = xs_dict_append(d, "name", "");
149 d = xs_dict_append(d, "type", "Image");
150
151 *attach = xs_list_append(*attach, d);
152 }
153 else {
154 xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v2, u);
155 s = xs_str_cat(s, s1);
156 }
157 }
158 else
159 if (*v == '[') {
160 /* markdown-like links [label](url) */
161 xs *w = xs_strip_chars_i(xs_replace(v, "#", "#"), "[)");
162 xs *l = xs_split_n(w, "](", 1);
163
164 if (xs_list_len(l) == 2) {
165 xs *link = xs_fmt("<a href=\"%s\">%s</a>",
166 xs_list_get(l, 1), xs_list_get(l, 0));
167
168 s = xs_str_cat(s, link);
169 }
170 else
171 s = xs_str_cat(s, v);
172 }
173 else
174 if (*v == '!') {
175 /* markdown-like images  */
176 xs *w = xs_strip_chars_i(xs_replace(v, "#", "#"), "![)");
177 xs *l = xs_split_n(w, "](", 1);
178
179 if (xs_list_len(l) == 2) {
180 const char *alt_text = xs_list_get(l, 0);
181 const char *img_url = xs_list_get(l, 1);
182 const char *mime = xs_mime_by_ext(img_url);
183
184 if (attach != NULL && xs_startswith(mime, "image/")) {
185 xs *d = xs_dict_new();
186
187 d = xs_dict_append(d, "mediaType", mime);
188 d = xs_dict_append(d, "url", img_url);
189 d = xs_dict_append(d, "name", alt_text);
190 d = xs_dict_append(d, "type", "Image");
191
192 *attach = xs_list_append(*attach, d);
193 }
194 else {
195 xs *link = xs_fmt("<a href=\"%s\">%s</a>", img_url, alt_text);
196
197 s = xs_str_cat(s, link);
198 }
199 }
200 else
201 s = xs_str_cat(s, v);
202 }
203 else
204 s = xs_str_cat(s, v);
205 }
206 else
207 /* surrounded text, copy directly */
208 s = xs_str_cat(s, v);
209
210 n++;
211 }
212
213 return s;
214}
215
216
217xs_str *not_really_markdown(const char *content, xs_list **attach, xs_list **tag)
218/* formats a content using some Markdown rules */
219{
220 xs_str *s = xs_str_new(NULL);
221 int in_pre = 0;
222 int in_blq = 0;
223 xs *list;
224 char *p;
225 const char *v;
226
227 /* work by lines */
228 list = xs_split(content, "\n");
229
230 p = list;
231 while (xs_list_iter(&p, &v)) {
232 xs *ss = NULL;
233
234 if (strcmp(v, "```") == 0) {
235 if (!in_pre)
236 s = xs_str_cat(s, "<pre>");
237 else
238 s = xs_str_cat(s, "</pre>");
239
240 in_pre = !in_pre;
241 continue;
242 }
243
244 if (in_pre) {
245 // Encode all HTML characters when we're in pre element until we are out.
246 ss = encode_html(v);
247
248 s = xs_str_cat(s, ss);
249 s = xs_str_cat(s, "<br>");
250 continue;
251 }
252
253 else
254 ss = xs_strip_i(format_line(v, attach));
255
256 if (xs_startswith(ss, "---")) {
257 /* delete the --- */
258 ss = xs_strip_i(xs_crop_i(ss, 3, 0));
259 s = xs_str_cat(s, "<hr>");
260
261 s = xs_str_cat(s, ss);
262
263 continue;
264 }
265
266 if (xs_startswith(ss, ">")) {
267 /* delete the > and subsequent spaces */
268 ss = xs_strip_i(xs_crop_i(ss, 1, 0));
269
270 if (!in_blq) {
271 s = xs_str_cat(s, "<blockquote>");
272 in_blq = 1;
273 }
274
275 s = xs_str_cat(s, ss);
276 s = xs_str_cat(s, "<br>");
277
278 continue;
279 }
280
281 if (in_blq) {
282 s = xs_str_cat(s, "</blockquote>");
283 in_blq = 0;
284 }
285
286 s = xs_str_cat(s, ss);
287 s = xs_str_cat(s, "<br>");
288 }
289
290 if (in_blq)
291 s = xs_str_cat(s, "</blockquote>");
292 if (in_pre)
293 s = xs_str_cat(s, "</pre>");
294
295 /* some beauty fixes */
296 s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
297 s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
298 s = xs_replace_i(s, "</pre><br>", "</pre>");
299
300 {
301 /* traditional emoticons */
302 xs *d = emojis();
303 int c = 0;
304 const char *k, *v;
305
306 while (xs_dict_next(d, &k, &v, &c)) {
307 const char *t = NULL;
308
309 /* is it an URL to an image? */
310 if (xs_startswith(v, "https:/" "/") && xs_startswith((t = xs_mime_by_ext(v)), "image/")) {
311 if (tag && xs_str_in(s, k) != -1) {
312 /* add the emoji to the tag list */
313 xs *e = xs_dict_new();
314 xs *i = xs_dict_new();
315 xs *u = xs_str_utctime(0, ISO_DATE_SPEC);
316
317 e = xs_dict_append(e, "id", v);
318 e = xs_dict_append(e, "type", "Emoji");
319 e = xs_dict_append(e, "name", k);
320 e = xs_dict_append(e, "updated", u);
321
322 i = xs_dict_append(i, "type", "Image");
323 i = xs_dict_append(i, "mediaType", t);
324 i = xs_dict_append(i, "url", v);
325 e = xs_dict_append(e, "icon", i);
326
327 *tag = xs_list_append(*tag, e);
328 }
329 }
330 else
331 s = xs_replace_i(s, k, v);
332 }
333 }
334
335 return s;
336}
337
338
339const char *valid_tags[] = {
340 "a", "p", "br", "br/", "blockquote", "ul", "ol", "li", "cite", "small",
341 "span", "i", "b", "u", "s", "pre", "code", "em", "strong", "hr", "img", "del", "bdi", NULL
342};
343
344xs_str *sanitize(const char *content)
345/* cleans dangerous HTML output */
346{
347 xs_str *s = xs_str_new(NULL);
348 xs *sl;
349 int n = 0;
350 char *p;
351 const char *v;
352
353 sl = xs_regex_split(content, "</?[^>]+>");
354
355 p = sl;
356
357 n = 0;
358 while (xs_list_iter(&p, &v)) {
359 if (n & 0x1) {
360 xs *s1 = xs_strip_i(xs_crop_i(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
361 xs *l1 = xs_split_n(s1, " ", 1);
362 xs *tag = xs_tolower_i(xs_dup(xs_list_get(l1, 0)));
363 xs *s2 = NULL;
364 int i;
365
366 /* check if it's one of the valid tags */
367 for (i = 0; valid_tags[i]; i++) {
368 if (strcmp(tag, valid_tags[i]) == 0)
369 break;
370 }
371
372 if (valid_tags[i]) {
373 /* accepted tag: rebuild it with only the accepted elements */
374 xs *el = xs_regex_select(v, "(src|href|rel|class|target)=\"[^\"]*\"");
375 xs *s3 = xs_join(el, " ");
376
377 s2 = xs_fmt("<%s%s%s%s>",
378 v[1] == '/' ? "/" : "", tag, xs_list_len(el) ? " " : "", s3);
379
380 s = xs_str_cat(s, s2);
381 } else {
382 /* treat end of divs as paragraph breaks */
383 if (strcmp(v, "</div>"))
384 s = xs_str_cat(s, "<p>");
385 }
386 }
387 else {
388 /* non-tag */
389 s = xs_str_cat(s, v);
390 }
391
392 n++;
393 }
394
395 return s;
396}
397
398
399xs_str *encode_html(const char *str)
400/* escapes html characters */
401{
402 xs_str *encoded = xs_html_encode((char *)str);
403
404 /* Restore only <br>. Probably safe. Let's hope nothing goes wrong with this. */
405 encoded = xs_replace_i(encoded, "<br>", "<br>");
406
407 return encoded;
408}
