2 * Bring external RSS and/or Atom feeds into rooms. This module implements a
3 * very loose parser that scrapes both kinds of feeds and is not picky about
4 * the standards compliance of the source data.
6 * Copyright (c) 2007-2018 by the citadel.org team
8 * This program is open source software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License version 3.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.See the
14 * GNU General Public License for more details.
21 #if TIME_WITH_SYS_TIME
22 # include <sys/time.h>
35 #include <sys/types.h>
38 #include <curl/curl.h>
39 #include <libcitadel.h>
42 #include "citserver.h"
46 #include "ctdl_module.h"
48 #include "parsedate.h"
50 #include "citadel_dirs.h"
53 #include "internet_addressing.h"
63 struct rssroom *rooms;
68 struct CtdlMessage *msg;
72 struct rssroom *rooms;
76 struct CitContext rss_CC;
77 struct rssurl *rsstodo = NULL;
80 // This handler is called whenever an XML tag opens.
82 void rss_start_element(void *data, const char *el, const char **attribute)
84 struct rssparser *r = (struct rssparser *)data;
88 (!strcasecmp(el, "entry"))
89 || (!strcasecmp(el, "item"))
91 // this is the start of a new item(rss) or entry(atom)
96 r->msg = malloc(sizeof(struct CtdlMessage));
97 memset(r->msg, 0, sizeof(struct CtdlMessage));
98 r->msg->cm_magic = CTDLMESSAGE_MAGIC;
99 r->msg->cm_anon_type = MES_NORMAL;
100 r->msg->cm_format_type = FMT_RFC822;
103 else if (!strcasecmp(el, "link")) { // atom feeds have the link as an attribute
104 for(i = 0; attribute[i]; i += 2) {
105 if (!strcasecmp(attribute[i], "href")) {
106 if (r->link != NULL) {
110 r->link = strdup(attribute[i+1]);
118 // This handler is called whenever an XML tag closes.
120 void rss_end_element(void *data, const char *el)
122 struct rssparser *r = (struct rssparser *)data;
124 if ( // end of a new item(rss) or entry(atom)
125 (!strcasecmp(el, "entry"))
126 || (!strcasecmp(el, "item"))
129 if (r->msg != NULL) { // Save the message to the rooms
131 // use the link as an item id if nothing else is available
132 if ((r->item_id == NULL) && (r->link != NULL)) {
133 r->item_id = strdup(r->link);
136 // check the use table
137 StrBuf *u = NewStrBuf();
138 StrBufAppendPrintf(u, "rss/%s", r->item_id);
139 int already_seen = CheckIfAlreadySeen(u);
142 if (already_seen == 0) {
144 // Compose the message text
145 StrBuf *TheMessage = NewStrBuf();
146 StrBufAppendPrintf(TheMessage,
147 "Content-type: text/html\n\n"
149 "<html><head></head><body>"
152 if (r->description != NULL) {
153 StrBufAppendPrintf(TheMessage, "%s<br><br>\r\n", r->description);
154 free(r->description);
155 r->description = NULL;
158 if (r->link != NULL) {
159 StrBufAppendPrintf(TheMessage, "<a href=\"%s\">%s</a>\r\n", r->link, r->link);
164 StrBufAppendPrintf(TheMessage, "</body></html>\r\n");
165 CM_SetField(r->msg, eMesageText, ChrPtr(TheMessage), StrLength(TheMessage));
166 FreeStrBuf(&TheMessage);
168 if (CM_IsEmpty(r->msg, eAuthor)) {
169 CM_SetField(r->msg, eAuthor, HKEY("rss"));
172 if (CM_IsEmpty(r->msg, eTimestamp)) {
173 CM_SetFieldLONG(r->msg, eTimestamp, time(NULL));
176 // Save it to the room(s)
177 struct rssroom *rr = NULL;
179 for (rr=r->rooms; rr!=NULL; rr=rr->next) {
180 if (rr == r->rooms) {
181 msgnum = CtdlSubmitMsg(r->msg, NULL, rr->room, 0);
184 CtdlSaveMsgPointerInRoom(rr->room, msgnum, 0, NULL);
186 syslog(LOG_DEBUG, "rssclient: saved message %ld to %s", msgnum, rr->room);
190 syslog(LOG_DEBUG, "rssclient: already seen %s", r->item_id);
197 if (r->item_id != NULL) {
203 else if (!strcasecmp(el, "title")) { // item subject (rss and atom)
204 if ((r->msg != NULL) && (CM_IsEmpty(r->msg, eMsgSubject))) {
205 CM_SetField(r->msg, eMsgSubject, ChrPtr(r->CData), StrLength(r->CData));
206 striplt(r->msg->cm_fields[eMsgSubject]);
210 else if (!strcasecmp(el, "creator")) { // <creator> can be used if <author> is not present
211 if ((r->msg != NULL) && (CM_IsEmpty(r->msg, eAuthor))) {
212 CM_SetField(r->msg, eAuthor, ChrPtr(r->CData), StrLength(r->CData));
213 striplt(r->msg->cm_fields[eAuthor]);
217 else if (!strcasecmp(el, "author")) { // <author> supercedes <creator> if both are present
218 if (r->msg != NULL) {
219 CM_SetField(r->msg, eAuthor, ChrPtr(r->CData), StrLength(r->CData)); // CM_SetField will free() the previous value
220 striplt(r->msg->cm_fields[eAuthor]);
224 else if (!strcasecmp(el, "pubdate")) { // date/time stamp (rss) Sat, 25 Feb 2017 14:28:01 EST
225 if ((r->msg)&&(r->msg->cm_fields[eTimestamp]==NULL)) {
226 CM_SetFieldLONG(r->msg, eTimestamp, parsedate(ChrPtr(r->CData)));
230 else if (!strcasecmp(el, "updated")) { // date/time stamp (atom) 2003-12-13T18:30:02Z
231 if ((r->msg)&&(r->msg->cm_fields[eTimestamp]==NULL)) {
234 memset(&t, 0, sizeof t);
235 sscanf(ChrPtr(r->CData), "%d-%d-%dT%d:%d:%d%c", &t.tm_year, &t.tm_mon, &t.tm_mday, &t.tm_hour, &t.tm_min, &t.tm_sec, &zulu);
238 CM_SetFieldLONG(r->msg, eTimestamp, mktime(&t));
242 else if (!strcasecmp(el, "link")) { // link to story (rss)
243 if (r->link != NULL) {
247 r->link = strdup(ChrPtr(r->CData));
252 (!strcasecmp(el, "guid")) // unique item id (rss)
253 || (!strcasecmp(el, "id")) // unique item id (atom)
255 if (r->item_id != NULL) {
259 r->item_id = strdup(ChrPtr(r->CData));
264 (!strcasecmp(el, "description")) // message text (rss)
265 || (!strcasecmp(el, "summary")) // message text (atom)
266 || (!strcasecmp(el, "content")) // message text (atom)
268 if (r->description != NULL) {
269 free(r->description);
270 r->description = NULL;
272 r->description = strdup(ChrPtr(r->CData));
273 striplt(r->description);
276 if (r->CData != NULL) {
277 FreeStrBuf(&r->CData);
283 // This handler is called whenever data appears between opening and closing tags.
285 void rss_handle_data(void *data, const char *content, int length)
287 struct rssparser *r = (struct rssparser *)data;
289 if (r->CData == NULL) {
290 r->CData = NewStrBuf();
293 StrBufAppendBufPlain(r->CData, content, length, 0);
297 // Feed has been downloaded, now parse it.
299 void rss_parse_feed(StrBuf *Feed, struct rssroom *rooms)
303 memset(&r, 0, sizeof r);
305 XML_Parser p = XML_ParserCreate("UTF-8");
306 XML_SetElementHandler(p, rss_start_element, rss_end_element);
307 XML_SetCharacterDataHandler(p, rss_handle_data);
308 XML_SetUserData(p, (void *)&r);
309 XML_Parse(p, ChrPtr(Feed), StrLength(Feed), XML_TRUE);
314 // Add a feed/room pair into the todo list
316 void rssclient_push_todo(char *rssurl, char *roomname)
318 struct rssurl *r = NULL;
319 struct rssurl *thisone = NULL;
320 struct rssroom *newroom = NULL;
322 syslog(LOG_DEBUG, "rssclient: will fetch %s to %s", rssurl, roomname);
324 for (r=rsstodo; r!=NULL; r=r->next) {
325 if (!strcasecmp(r->url, rssurl)) {
329 if (thisone == NULL) {
330 thisone = malloc(sizeof(struct rssurl));
331 thisone->url = strdup(rssurl);
332 thisone->rooms = NULL;
333 thisone->next = rsstodo;
337 newroom = malloc(sizeof(struct rssroom));
338 newroom->room = strdup(roomname);
339 newroom->next = thisone->rooms;
340 thisone->rooms = newroom;
344 // pull one feed (possibly multiple rooms)
346 void rss_pull_one_feed(struct rssurl *url)
350 StrBuf *Downloaded = NULL;
352 syslog(LOG_DEBUG, "rssclient: fetching %s", url->url);
354 curl = curl_easy_init();
359 Downloaded = NewStrBuf();
361 curl_easy_setopt(curl, CURLOPT_URL, url->url);
362 curl_easy_setopt(curl, CURLOPT_SSL_VERIFYPEER, 0L);
363 curl_easy_setopt(curl, CURLOPT_SSL_VERIFYHOST, 0L);
364 curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L); // Follow redirects
365 curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, CurlFillStrBuf_callback); // What to do with downloaded data
366 curl_easy_setopt(curl, CURLOPT_WRITEDATA, Downloaded); // Give it our StrBuf to work with
367 curl_easy_setopt(curl, CURLOPT_TIMEOUT, 20L); // Time out after 20 seconds
368 res = curl_easy_perform(curl); // Perform the request
369 if (res != CURLE_OK) {
370 syslog(LOG_WARNING, "rssclient: failed to load feed: %s", curl_easy_strerror(res));
372 curl_easy_cleanup(curl);
374 rss_parse_feed(Downloaded, url->rooms); // parse the feed
375 FreeStrBuf(&Downloaded); // free the downloaded feed data
379 // We have a list, now download the feeds
381 void rss_pull_feeds(void)
386 while (rsstodo != NULL) {
387 rss_pull_one_feed(rsstodo);
389 rsstodo = rsstodo->next;
390 while (r->rooms != NULL) {
392 r->rooms = r->rooms->next;
402 // Scan a room's netconfig looking for RSS feed parsing requests
404 void rssclient_scan_room(struct ctdlroom *qrbuf, void *data)
406 char *serialized_config = NULL;
411 serialized_config = LoadRoomNetConfigFile(qrbuf->QRnumber);
412 if (!serialized_config) {
416 num_configs = num_tokens(serialized_config, '\n');
417 for (i=0; i<num_configs; ++i) {
418 extract_token(cfgline, serialized_config, i, '\n', sizeof cfgline);
419 if (!strncasecmp(cfgline, HKEY("rssclient|"))) {
420 strcpy(cfgline, &cfgline[10]);
421 char *vbar = strchr(cfgline, '|');
425 rssclient_push_todo(cfgline, qrbuf->QRname);
429 free(serialized_config);
434 * Scan for rooms that have RSS client requests configured
436 void rssclient_scan(void) {
437 time_t now = time(NULL);
439 /* Run no more than once every 15 minutes. */
440 if ((now - last_run) < 900) {
442 "rssclient: polling interval not yet reached; last run was %ldm%lds ago",
443 ((now - last_run) / 60),
444 ((now - last_run) % 60)
449 become_session(&rss_CC);
450 syslog(LOG_DEBUG, "rssclient: started");
451 CtdlForEachRoom(rssclient_scan_room, NULL);
453 syslog(LOG_DEBUG, "rssclient: ended");
454 last_run = time(NULL);
459 CTDL_MODULE_INIT(rssclient)
463 syslog(LOG_INFO, "rssclient: using %s", curl_version());
464 CtdlRegisterSessionHook(rssclient_scan, EVT_TIMER, PRIO_AGGR + 300);
468 CtdlFillSystemContext(&rss_CC, "rssclient");