2.95
rss.c
1/* snac - A simple, minimalistic ActivityPub instance */
2/* copyright (c) 2025 - 2026 grunfink et al. / MIT license */
3
4#include "xs.h"
5#include "xs_html.h"
6#include "xs_regex.h"
7#include "xs_time.h"
8#include "xs_match.h"
9#include "xs_curl.h"
10#include "xs_openssl.h"
11#include "xs_json.h"
12#include "xs_http.h"
13#include "xs_unicode.h"
14
15#include "snac.h"
16
17xs_str *rss_from_timeline(snac *user, const xs_list *timeline,
18 const char *title, const char *link, const char *desc)
19/* converts a timeline to rss */
20{
21 xs_html *rss = xs_html_tag("rss",
22 xs_html_attr("xmlns:content", "http:/" "/purl.org/rss/1.0/modules/content/"),
23 xs_html_attr("version", "2.0"),
24 xs_html_attr("xmlns:atom", "http:/" "/www.w3.org/2005/Atom"));
25
26 xs_html *channel = xs_html_tag("channel",
27 xs_html_tag("title",
28 xs_html_text(title)),
29 xs_html_tag("language",
30 xs_html_text("en")),
31 xs_html_tag("link",
32 xs_html_text(link)),
33 xs_html_sctag("atom:link",
34 xs_html_attr("href", link),
35 xs_html_attr("rel", "self"),
36 xs_html_attr("type", "application/rss+xml")),
37 xs_html_tag("generator",
38 xs_html_text(USER_AGENT)),
39 xs_html_tag("description",
40 xs_html_text(desc)));
41
42 xs_html_add(rss, channel);
43
44 int cnt = 0;
45 const char *v;
46
47 int show_unlisted = user ? xs_is_true(xs_dict_get(user->config, "show_unlisted")) : 0;
48
49 xs_list_foreach(timeline, v) {
50 xs *msg = NULL;
51
52 if (user) {
53 if (!valid_status(timeline_get_by_md5(user, v, &msg)))
54 continue;
55 }
56 else {
57 if (!valid_status(object_get_by_md5(v, &msg)))
58 continue;
59 }
60
61 const char *id = xs_dict_get(msg, "id");
62 const char *content = xs_dict_get(msg, "content");
63 const char *published = xs_dict_get(msg, "published");
64
65 if (user && !is_msg_mine(user, id))
66 continue;
67
68 if (!id || !content || !published)
69 continue;
70
71 int scope = get_msg_visibility(msg);
72
73 if (scope == SCOPE_MENTIONED || scope == SCOPE_FOLLOWERS)
74 continue;
75
76 if (scope == SCOPE_UNLISTED && !show_unlisted)
77 continue;
78
79 /* create a title with the first line of the content */
80 xs *title = xs_replace(content, "<br>", "\n");
81 title = xs_regex_replace_i(title, "<[^>]+>", " ");
82 title = xs_regex_replace_i(title, "&[^;]+;", " ");
83 int i;
84
85 for (i = 0; title[i] && title[i] != '\n' && i < 50; ) {
86 const char *p = &title[i];
87 unsigned int cp = xs_utf8_dec(&p);
88 int n = p - title;
89 if (cp == 0xfffd || n > 50)
90 break;
91 i = n;
92 }
93
94 if (title[i] != '\0') {
95 title[i] = '\0';
96 title = xs_str_cat(title, "...");
97 }
98
99 title = xs_strip_i(title);
100
101 /* convert the date */
102 time_t t = xs_parse_iso_date(published, 0);
103 xs *rss_date = xs_str_utctime(t, "%a, %d %b %Y %T +0000");
104
105 /* if it's the first one, add it to the header */
106 if (cnt == 0)
107 xs_html_add(channel,
108 xs_html_tag("lastBuildDate",
109 xs_html_text(rss_date)));
110
111 xs *att_content = xs_dup(content);
112 xs *attach_list = get_attachments(msg);
113
114 /* add possible attachments */
115 if (xs_list_len(attach_list)) {
116 const xs_val *a;
117 xs_html *content_attachments = xs_html_tag("div",
118 xs_html_attr("class", "snac-content-attachments"));
119
120 xs_list_foreach(attach_list, a) {
121 const char *type = xs_dict_get(a, "type");
122 const char *o_href = xs_dict_get(a, "href");
123 const char *name = xs_dict_get_def(a, "name", "");
124
125 if (!xs_is_string(type) || !xs_is_string(o_href))
126 continue;
127
128 if (xs_startswith(type, "image/")) {
129 xs_html_add(content_attachments,
130 xs_html_tag("figure",
131 xs_html_sctag("img",
132 xs_html_attr("src", o_href)),
133 xs_html_tag("figcaption",
134 xs_html_text(name))));
135 }
136 else {
137 xs_html_add(content_attachments,
138 xs_html_tag("a",
139 xs_html_attr("href", o_href),
140 xs_html_text(name),
141 xs_html_text(" "),
142 xs_html_text(o_href),
143 xs_html_sctag("br", NULL)));
144 }
145 }
146
147 xs *s1 = xs_html_render(content_attachments);
148 att_content = xs_str_cat(att_content, s1);
149 }
150
151 xs_html_add(channel,
152 xs_html_tag("item",
153 xs_html_tag("title",
154 xs_html_text(title)),
155 xs_html_tag("link",
156 xs_html_text(id)),
157 xs_html_tag("guid",
158 xs_html_text(id)),
159 xs_html_tag("pubDate",
160 xs_html_text(rss_date)),
161 xs_html_tag("description",
162 xs_html_text(att_content))));
163
164 cnt++;
165 }
166
167 return xs_html_render_s(rss, "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n");
168}
169
170
171void rss_to_timeline(snac *user, const char *url)
172/* reads an RSS and inserts all ActivityPub posts into the user's timeline */
173{
174 if (!xs_startswith(url, "https:/") && !xs_startswith(url, "http:/"))
175 return;
176
177 xs *hdrs = xs_dict_new();
178 hdrs = xs_dict_set(hdrs, "accept", "application/rss+xml");
179 hdrs = xs_dict_set(hdrs, "user-agent", USER_AGENT);
180
181 /* get the RSS metadata */
182 xs *md5 = xs_md5_hex(url, strlen(url));
183 xs *rss_md_fn = xs_fmt("%s/rss", user->basedir);
184 mkdirx(rss_md_fn);
185 rss_md_fn = xs_str_cat(rss_md_fn, "/", md5, ".json");
186
187 xs *rss_md = NULL;
188 const char *etag = NULL;
189
190 FILE *f;
191 if ((f = fopen(rss_md_fn, "r")) != NULL) {
192 rss_md = xs_json_load(f);
193 fclose(f);
194
195 etag = xs_dict_get(rss_md, "etag");
196
197 if (xs_is_string(etag))
198 hdrs = xs_dict_set(hdrs, "if-none-match", etag);
199 }
200
201 if (rss_md == NULL)
202 rss_md = xs_dict_new();
203
204 xs *payload = NULL;
205 int status;
206 int p_size;
207
208 xs *rsp = xs_http_request("GET", url, hdrs, NULL, 0, &status, &payload, &p_size, 0);
209
210 snac_log(user, xs_fmt("parsing RSS %s %d", url, status));
211
212 if (!valid_status(status) || !xs_is_string(payload))
213 return;
214
215 /* not an RSS? done */
216 const char *ctype = xs_dict_get(rsp, "content-type");
217 if (!xs_is_string(ctype) || xs_str_in(ctype, "application/rss+xml") == -1)
218 return;
219
220 /* yes, parsing is done with regexes (now I have two problems blah blah blah) */
221 xs *links = xs_regex_select(payload, "<link>[^<]+</link>");
222 const char *link;
223
224 xs_list_foreach(links, link) {
225 xs *l = xs_replace(link, "<link>", "");
226 char *p = strchr(l, '<');
227
228 if (p == NULL)
229 continue;
230 *p = '\0';
231
232 /* skip this same URL */
233 if (strcmp(l, url) == 0)
234 continue;
235
236 /* skip crap */
237 if (!xs_startswith(l, "https:/") && !xs_startswith(l, "http:/"))
238 continue;
239
240 snac_debug(user, 1, xs_fmt("RSS link: %s", l));
241
242 if (timeline_here(user, l)) {
243 snac_debug(user, 1, xs_fmt("RSS entry already in timeline %s", l));
244 continue;
245 }
246
247 /* special trick for Mastodon: convert from the alternate format */
248 if (strchr(l, '@') != NULL) {
249 xs *l2 = xs_split(l, "/");
250
251 if (xs_list_len(l2) == 5) {
252 const char *uid = xs_list_get(l2, 3);
253 if (*uid == '@') {
254 xs *guessed_id = xs_fmt("https:/" "/%s/users/%s/statuses/%s",
255 xs_list_get(l2, 2), uid + 1, xs_list_get(l2, -1));
256
257 if (timeline_here(user, guessed_id)) {
258 snac_debug(user, 1, xs_fmt("RSS entry already in timeline (alt) %s", guessed_id));
259 continue;
260 }
261 }
262 }
263 }
264
265 xs *obj = NULL;
266
267 if (!valid_status(object_get(l, &obj))) {
268 /* object is not here: bring it */
269 if (!valid_status(activitypub_request(user, l, &obj)))
270 continue;
271 }
272
273 if (xs_is_dict(obj)) {
274 const char *id = xs_dict_get(obj, "id");
275 const char *type = xs_dict_get(obj, "type");
276 const char *attr_to = get_atto(obj);
277
278 if (!xs_is_string(id) || !xs_is_string(type) || !xs_is_string(attr_to))
279 continue;
280
281 if (!xs_match(type, POSTLIKE_OBJECT_TYPE))
282 continue;
283
284 /* the instance is [back] online */
285 instance_failure(id, OP_DEL);
286
287 if (timeline_here(user, id)) {
288 snac_debug(user, 1, xs_fmt("RSS entry already in timeline (id) %s", id));
289 continue;
290 }
291
292 enqueue_actor_refresh(user, attr_to, 0);
293
294 timeline_add(user, id, obj);
295
296 snac_log(user, xs_fmt("new '%s' (RSS) %s %s", type, attr_to, id));
297 }
298 }
299
300 /* update the RSS metadata */
301 etag = xs_dict_get(rsp, "etag");
302
303 if (xs_is_string(etag)) {
304 rss_md = xs_dict_set(rss_md, "etag", etag);
305 rss_md = xs_dict_set(rss_md, "url", url);
306 if ((f = fopen(rss_md_fn, "w")) != NULL) {
307 xs_json_dump(rss_md, 4, f);
308 fclose(f);
309 }
310 }
311}
312
313
314void rss_poll_hashtags(void)
315/* parses all RSS from all users */
316{
317 xs *list = user_list();
318 const char *uid;
319
320 xs_list_foreach(list, uid) {
321 snac user;
322
323 if (user_open(&user, uid)) {
324 const xs_list *rss = xs_dict_get(user.config, "followed_hashtags");
325
326 if (xs_is_list(rss)) {
327 const char *url;
328
329 xs_list_foreach(rss, url)
330 rss_to_timeline(&user, url);
331 }
332
333 user_free(&user);
334 }
335 }
336}
