2 * Bring external RSS feeds into rooms.
4 * Copyright (c) 2007-2010 by the citadel.org team
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 3 of the License, or
9 * (at your option) any later version.
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.
16 * You should have received a copy of the GNU General Public License
17 * along with this program; if not, write to the Free Software
18 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
25 #if TIME_WITH_SYS_TIME
26 # include <sys/time.h>
30 # include <sys/time.h>
39 #include <sys/types.h>
42 #include <curl/curl.h>
43 #include <libcitadel.h>
46 #include "citserver.h"
50 #include "ctdl_module.h"
52 #include "parsedate.h"
54 #include "citadel_dirs.h"
57 #include "event_client.h"
58 #include "rss_atom_parser.h"
61 #define TMP_MSGDATA 0xFF
62 #define TMP_SHORTER_URL_OFFSET 0xFE
63 #define TMP_SHORTER_URLS 0xFD
65 citthread_mutex_t RSSQueueMutex; /* locks the access to the following vars: */
66 HashList *RSSQueueRooms = NULL; /* rss_room_counter */
67 HashList *RSSFetchUrls = NULL; /* -> rss_aggregator; ->RefCount access to be locked too. */
69 eNextState RSSAggregatorTerminate(AsyncIO *IO);
72 struct rssnetcfg *rnclist = NULL;
73 void AppendLink(StrBuf *Message, StrBuf *link, StrBuf *LinkTitle, const char *Title)
75 if (StrLength(link) > 0)
77 StrBufAppendBufPlain(Message, HKEY("<a href=\""), 0);
78 StrBufAppendBuf(Message, link, 0);
79 StrBufAppendBufPlain(Message, HKEY("\">"), 0);
80 if (StrLength(LinkTitle) > 0)
81 StrBufAppendBuf(Message, LinkTitle, 0);
82 else if ((Title != NULL) && !IsEmptyStr(Title))
83 StrBufAppendBufPlain(Message, Title, -1, 0);
85 StrBufAppendBuf(Message, link, 0);
86 StrBufAppendBufPlain(Message, HKEY("</a><br>\n"), 0);
89 typedef struct __networker_save_message {
91 struct CtdlMessage *Msg;
92 struct recptypes *recp;
97 } networker_save_message;
100 void DeleteRoomReference(long QRnumber)
106 rss_room_counter *pRoomC;
108 At = GetNewHashPos(RSSQueueRooms, 0);
110 GetHashPosFromKey(RSSQueueRooms, LKEY(QRnumber), At);
111 GetHashPos(RSSQueueRooms, At, &HKLen, &HK, &vData);
114 pRoomC = (rss_room_counter *) vData;
116 if (pRoomC->count == 0)
117 DeleteEntryFromHash(RSSQueueRooms, At);
122 void UnlinkRooms(rss_aggregator *Cfg)
125 DeleteRoomReference(Cfg->QRnumber);
126 if (Cfg->OtherQRnumbers != NULL)
133 At = GetNewHashPos(Cfg->OtherQRnumbers, 0);
134 while (GetNextHashPos(Cfg->OtherQRnumbers, At, &HKLen, &HK, &vData) &&
137 long *lData = (long*) vData;
138 DeleteRoomReference(*lData);
146 void UnlinkRSSAggregator(rss_aggregator *Cfg)
152 At = GetNewHashPos(RSSFetchUrls, 0);
153 if (GetHashPosFromKey(RSSFetchUrls, SKEY(Cfg->Url), At) == 0)
155 DeleteEntryFromHash(RSSFetchUrls, At);
160 eNextState FreeNetworkSaveMessage (AsyncIO *IO)
162 networker_save_message *Ctx = (networker_save_message *) IO->Data;
164 citthread_mutex_lock(&RSSQueueMutex);
165 Ctx->Cfg->RefCount --;
167 if (Ctx->Cfg->RefCount == 0)
169 UnlinkRSSAggregator(Ctx->Cfg);
172 citthread_mutex_unlock(&RSSQueueMutex);
174 CtdlFreeMessage(Ctx->Msg);
175 free_recipients(Ctx->recp);
176 FreeStrBuf(&Ctx->Message);
177 FreeStrBuf(&Ctx->MsgGUID);
182 eNextState AbortNetworkSaveMessage (AsyncIO *IO)
184 return eAbort; ///TODO
187 eNextState RSSSaveMessage(AsyncIO *IO)
189 networker_save_message *Ctx = (networker_save_message *) IO->Data;
191 Ctx->Msg->cm_fields['M'] = SmashStrBuf(&Ctx->Message);
193 CtdlSubmitMsg(Ctx->Msg, Ctx->recp, NULL, 0);
195 /* write the uidl to the use table so we don't store this item again */
196 cdb_store(CDB_USETABLE, SKEY(Ctx->MsgGUID), &Ctx->ut, sizeof(struct UseTable) );
198 return eTerminateConnection;
201 // TODO: relink me: ExpandShortUrls(ri->description);
203 eNextState RSS_FetchNetworkUsetableEntry(AsyncIO *IO)
205 struct cdbdata *cdbut;
206 networker_save_message *Ctx = (networker_save_message *) IO->Data;
208 /* Find out if we've already seen this item */
209 strcpy(Ctx->ut.ut_msgid, ChrPtr(Ctx->MsgGUID)); /// TODO
210 Ctx->ut.ut_timestamp = time(NULL);
212 cdbut = cdb_fetch(CDB_USETABLE, SKEY(Ctx->MsgGUID));
215 /* Item has already been seen */
216 CtdlLogPrintf(CTDL_DEBUG, "%s has already been seen\n", ChrPtr(Ctx->MsgGUID));
219 /* rewrite the record anyway, to update the timestamp */
220 cdb_store(CDB_USETABLE,
222 &Ctx->ut, sizeof(struct UseTable) );
228 NextDBOperation(IO, RSSSaveMessage);
232 void RSSQueueSaveMessage(struct CtdlMessage *Msg, struct recptypes *recp, StrBuf *MsgGUID, StrBuf *MessageBody, rss_aggregator *Cfg)
234 networker_save_message *Ctx;
236 Ctx = (networker_save_message *) malloc(sizeof(networker_save_message));
237 memset(Ctx, 0, sizeof(networker_save_message));
239 Ctx->MsgGUID = MsgGUID;
240 Ctx->Message = MessageBody;
245 Ctx->IO.CitContext = CloneContext(CC);
246 Ctx->IO.Terminate = FreeNetworkSaveMessage;
247 Ctx->IO.ShutdownAbort = AbortNetworkSaveMessage;
248 QueueDBOperation(&Ctx->IO, RSS_FetchNetworkUsetableEntry);
253 * Commit a fetched and parsed RSS item to disk
255 void rss_save_item(rss_item *ri, rss_aggregator *Cfg)
258 struct MD5Context md5context;
259 u_char rawdigest[MD5_DIGEST_LEN];
260 struct CtdlMessage *msg;
261 struct recptypes *recp = NULL;
267 recp = (struct recptypes *) malloc(sizeof(struct recptypes));
268 if (recp == NULL) return;
269 memset(recp, 0, sizeof(struct recptypes));
270 Buf = NewStrBufDup(Cfg->rooms);
271 recp->recp_room = SmashStrBuf(&Buf);
272 recp->num_room = Cfg->roomlist_parts;
273 recp->recptypes_magic = RECPTYPES_MAGIC;
276 /* Construct a GUID to use in the S_USETABLE table.
277 * If one is not present in the item itself, make one up.
279 if (ri->guid != NULL) {
280 StrBufSpaceToBlank(ri->guid);
281 StrBufTrim(ri->guid);
282 guid = NewStrBufPlain(HKEY("rss/"));
283 StrBufAppendBuf(guid, ri->guid, 0);
286 MD5Init(&md5context);
287 if (ri->title != NULL) {
288 MD5Update(&md5context, (const unsigned char*)ChrPtr(ri->title), StrLength(ri->title));
290 if (ri->link != NULL) {
291 MD5Update(&md5context, (const unsigned char*)ChrPtr(ri->link), StrLength(ri->link));
293 MD5Final(rawdigest, &md5context);
294 guid = NewStrBufPlain(NULL, MD5_DIGEST_LEN * 2 + 12 /* _rss2ctdl*/);
295 StrBufHexEscAppend(guid, NULL, rawdigest, MD5_DIGEST_LEN);
296 StrBufAppendBufPlain(guid, HKEY("_rss2ctdl"), 0);
299 /* translate Item into message. */
300 CtdlLogPrintf(CTDL_DEBUG, "RSS: translating item...\n");
301 if (ri->description == NULL) ri->description = NewStrBufPlain(HKEY(""));
302 StrBufSpaceToBlank(ri->description);
303 msg = malloc(sizeof(struct CtdlMessage));
304 memset(msg, 0, sizeof(struct CtdlMessage));
305 msg->cm_magic = CTDLMESSAGE_MAGIC;
306 msg->cm_anon_type = MES_NORMAL;
307 msg->cm_format_type = FMT_RFC822;
309 if (ri->guid != NULL) {
310 msg->cm_fields['E'] = strdup(ChrPtr(ri->guid));
313 if (ri->author_or_creator != NULL) {
315 StrBuf *Encoded = NULL;
318 From = html_to_ascii(ChrPtr(ri->author_or_creator),
319 StrLength(ri->author_or_creator),
321 StrBufPlain(ri->author_or_creator, From, -1);
322 StrBufTrim(ri->author_or_creator);
325 FromAt = strchr(ChrPtr(ri->author_or_creator), '@') != NULL;
326 if (!FromAt && StrLength (ri->author_email) > 0)
328 StrBufRFC2047encode(&Encoded, ri->author_or_creator);
329 msg->cm_fields['A'] = SmashStrBuf(&Encoded);
330 msg->cm_fields['P'] = SmashStrBuf(&ri->author_email);
335 msg->cm_fields['P'] = SmashStrBuf(&ri->author_or_creator);
338 StrBufRFC2047encode(&Encoded, ri->author_or_creator);
339 msg->cm_fields['A'] = SmashStrBuf(&Encoded);
340 msg->cm_fields['P'] = strdup("rss@localhost");
345 msg->cm_fields['A'] = strdup("rss");
348 msg->cm_fields['N'] = strdup(NODENAME);
349 if (ri->title != NULL) {
352 StrBuf *Encoded, *QPEncoded;
355 StrBufSpaceToBlank(ri->title);
356 len = StrLength(ri->title);
357 Sbj = html_to_ascii(ChrPtr(ri->title), len, 512, 0);
359 if (Sbj[len - 1] == '\n')
364 Encoded = NewStrBufPlain(Sbj, len);
368 StrBufRFC2047encode(&QPEncoded, Encoded);
370 msg->cm_fields['U'] = SmashStrBuf(&QPEncoded);
371 FreeStrBuf(&Encoded);
373 msg->cm_fields['T'] = malloc(64);
374 snprintf(msg->cm_fields['T'], 64, "%ld", ri->pubdate);
375 if (ri->channel_title != NULL) {
376 if (StrLength(ri->channel_title) > 0) {
377 msg->cm_fields['O'] = strdup(ChrPtr(ri->channel_title));
380 if (ri->link == NULL)
381 ri->link = NewStrBufPlain(HKEY(""));
383 #if 0 /* temporarily disable shorter urls. */
384 msg->cm_fields[TMP_SHORTER_URLS] = GetShorterUrls(ri->description);
387 msglen += 1024 + StrLength(ri->link) + StrLength(ri->description) ;
389 Message = NewStrBufPlain(NULL, StrLength(ri->description));
391 StrBufPlain(Message, HKEY(
392 "Content-type: text/html; charset=\"UTF-8\"\r\n\r\n"
394 #if 0 /* disable shorter url for now. */
395 msg->cm_fields[TMP_SHORTER_URL_OFFSET] = StrLength(Message);
397 StrBufAppendBuf(Message, ri->description, 0);
398 StrBufAppendBufPlain(Message, HKEY("<br><br>\n"), 0);
400 AppendLink(Message, ri->link, ri->linkTitle, NULL);
401 AppendLink(Message, ri->reLink, ri->reLinkTitle, "Reply to this");
402 StrBufAppendBufPlain(Message, HKEY("</body></html>\n"), 0);
404 RSSQueueSaveMessage(msg, recp, guid, Message, Cfg);
412 int rss_do_fetching(rss_aggregator *Cfg)
421 if ((Cfg->next_poll != 0) && (now < Cfg->next_poll))
425 ri = (rss_item*) malloc(sizeof(rss_item));
426 memset(ri, 0, sizeof(rss_item));
429 IO->CitContext = CloneContext(CC);
433 CtdlLogPrintf(CTDL_DEBUG, "Fetching RSS feed <%s>\n", ChrPtr(Cfg->Url));
434 ParseURL(&IO->ConnectMe, Cfg->Url, 80);
435 CurlPrepareURL(IO->ConnectMe);
437 if (! evcurl_init(IO,
440 "Citadel RSS Client",
442 RSSAggregatorTerminate))
444 CtdlLogPrintf(CTDL_ALERT, "Unable to initialize libcurl.\n");
448 evcurl_handle_start(IO);
454 void DeleteRssCfg(void *vptr)
456 rss_aggregator *rncptr = (rss_aggregator *)vptr;
458 FreeStrBuf(&rncptr->Url);
459 FreeStrBuf(&rncptr->rooms);
460 FreeStrBuf(&rncptr->CData);
461 FreeStrBuf(&rncptr->Key);
463 DeleteHash(&rncptr->OtherQRnumbers);
465 if (rncptr->Item != NULL)
467 FreeStrBuf(&rncptr->Item->guid);
468 FreeStrBuf(&rncptr->Item->title);
469 FreeStrBuf(&rncptr->Item->link);
470 FreeStrBuf(&rncptr->Item->linkTitle);
471 FreeStrBuf(&rncptr->Item->reLink);
472 FreeStrBuf(&rncptr->Item->reLinkTitle);
473 FreeStrBuf(&rncptr->Item->description);
474 FreeStrBuf(&rncptr->Item->channel_title);
475 FreeStrBuf(&rncptr->Item->author_or_creator);
476 FreeStrBuf(&rncptr->Item->author_url);
477 FreeStrBuf(&rncptr->Item->author_email);
484 eNextState RSSAggregatorTerminate(AsyncIO *IO)
486 rss_aggregator *rncptr = (rss_aggregator *)IO->Data;
492 citthread_mutex_lock(&RSSQueueMutex);
494 if (rncptr->RefCount == 0)
496 UnlinkRSSAggregator(rncptr);
499 citthread_mutex_unlock(&RSSQueueMutex);
501 At = GetNewHashPos(RSSFetchUrls, 0);
503 citthread_mutex_lock(&RSSQueueMutex);
504 GetHashPosFromKey(RSSFetchUrls, SKEY(rncptr->Url), At);
505 GetHashPos(RSSFetchUrls, At, &HKLen, &HK, &vData);
506 DeleteEntryFromHash(RSSFetchUrls, At);
507 citthread_mutex_unlock(&RSSQueueMutex);
515 * Scan a room's netconfig to determine whether it is requesting any RSS feeds
517 void rssclient_scan_room(struct ctdlroom *qrbuf, void *data)
522 rss_room_counter *Count = NULL;
524 char filename[PATH_MAX];
527 rss_aggregator *rncptr = NULL;
528 rss_aggregator *use_this_rncptr = NULL;
530 const char *CfgPtr, *lPtr;
533 citthread_mutex_lock(&RSSQueueMutex);
534 if (GetHash(RSSQueueRooms, LKEY(qrbuf->QRnumber), &vptr))
536 CtdlLogPrintf(CTDL_DEBUG,
537 "rssclient: [%ld] %s already in progress.\n",
540 citthread_mutex_unlock(&RSSQueueMutex);
543 citthread_mutex_unlock(&RSSQueueMutex);
545 assoc_file_name(filename, sizeof filename, qrbuf, ctdl_netcfg_dir);
547 if (CtdlThreadCheckStop())
550 /* Only do net processing for rooms that have netconfigs */
551 fd = open(filename, 0);
553 //CtdlLogPrintf(CTDL_DEBUG, "rssclient: %s no config.\n", qrbuf->QRname);
556 if (CtdlThreadCheckStop())
558 if (fstat(fd, &statbuf) == -1) {
559 CtdlLogPrintf(CTDL_DEBUG, "ERROR: could not stat configfile '%s' - %s\n",
560 filename, strerror(errno));
563 if (CtdlThreadCheckStop())
565 CfgData = NewStrBufPlain(NULL, statbuf.st_size + 1);
566 if (StrBufReadBLOB(CfgData, &fd, 1, statbuf.st_size, &Err) < 0) {
568 FreeStrBuf(&CfgData);
569 CtdlLogPrintf(CTDL_DEBUG, "ERROR: reading config '%s' - %s<br>\n",
570 filename, strerror(errno));
574 if (CtdlThreadCheckStop())
578 CfgType = NewStrBuf();
579 Line = NewStrBufPlain(NULL, StrLength(CfgData));
583 Done = StrBufSipLine(Line, CfgData, &CfgPtr) == 0;
584 if (StrLength(Line) > 0)
587 StrBufExtract_NextToken(CfgType, Line, &lPtr, '|');
588 if (!strcasecmp("rssclient", ChrPtr(CfgType)))
592 Count = malloc(sizeof(rss_room_counter));
596 rncptr = (rss_aggregator *) malloc(sizeof(rss_aggregator));
597 memset (rncptr, 0, sizeof(rss_aggregator));
598 rncptr->roomlist_parts = 1;
599 rncptr->Url = NewStrBuf();
600 StrBufExtract_NextToken(rncptr->Url, Line, &lPtr, '|');
602 citthread_mutex_lock(&RSSQueueMutex);
603 GetHash(RSSFetchUrls, SKEY(rncptr->Url), &vptr);
604 use_this_rncptr = (rss_aggregator *)vptr;
605 if (use_this_rncptr != NULL)
607 /* mustn't attach to an active session */
608 if (use_this_rncptr->RefCount > 0)
610 DeleteRssCfg(rncptr);
616 StrBufAppendBufPlain(use_this_rncptr->rooms,
619 if (use_this_rncptr->roomlist_parts == 1)
621 use_this_rncptr->OtherQRnumbers = NewHash(1, lFlathash);
623 QRnumber = (long*)malloc(sizeof(long));
624 *QRnumber = qrbuf->QRnumber;
625 Put(use_this_rncptr->OtherQRnumbers, LKEY(qrbuf->QRnumber), QRnumber, NULL);
626 use_this_rncptr->roomlist_parts++;
628 citthread_mutex_unlock(&RSSQueueMutex);
631 citthread_mutex_unlock(&RSSQueueMutex);
633 rncptr->ItemType = RSS_UNSET;
635 rncptr->rooms = NewStrBufPlain(qrbuf->QRname, -1);
637 citthread_mutex_lock(&RSSQueueMutex);
638 Put(RSSFetchUrls, SKEY(rncptr->Url), rncptr, DeleteRssCfg);
639 citthread_mutex_unlock(&RSSQueueMutex);
645 Count->QRnumber = qrbuf->QRnumber;
646 citthread_mutex_lock(&RSSQueueMutex);
647 CtdlLogPrintf(CTDL_DEBUG, "rssclient: [%ld] %s now starting.\n",
648 qrbuf->QRnumber, qrbuf->QRname);
649 Put(RSSQueueRooms, LKEY(qrbuf->QRnumber), Count, NULL);
650 citthread_mutex_unlock(&RSSQueueMutex);
652 FreeStrBuf(&CfgData);
653 FreeStrBuf(&CfgType);
658 * Scan for rooms that have RSS client requests configured
660 void rssclient_scan(void) {
661 static int doing_rssclient = 0;
662 rss_aggregator *rptr = NULL;
669 * This is a simple concurrency check to make sure only one rssclient run
670 * is done at a time. We could do this with a mutex, but since we
671 * don't really require extremely fine granularity here, we'll do it
672 * with a static variable instead.
674 if (doing_rssclient) return;
677 CtdlLogPrintf(CTDL_DEBUG, "rssclient started\n");
678 CtdlForEachRoom(rssclient_scan_room, NULL);
680 citthread_mutex_lock(&RSSQueueMutex);
682 it = GetNewHashPos(RSSFetchUrls, 0);
683 while (GetNextHashPos(RSSFetchUrls, it, &len, &Key, &vrptr) &&
685 rptr = (rss_aggregator *)vrptr;
686 if (rptr->RefCount == 0)
687 if (!rss_do_fetching(rptr))
688 UnlinkRSSAggregator(rptr);
691 citthread_mutex_unlock(&RSSQueueMutex);
693 CtdlLogPrintf(CTDL_DEBUG, "rssclientscheduler ended\n");
698 void rss_cleanup(void)
700 citthread_mutex_destroy(&RSSQueueMutex);
701 DeleteHash(&RSSFetchUrls);
702 DeleteHash(&RSSQueueRooms);
706 CTDL_MODULE_INIT(rssclient)
710 citthread_mutex_init(&RSSQueueMutex, NULL);
711 RSSQueueRooms = NewHash(1, lFlathash);
712 RSSFetchUrls = NewHash(1, NULL);
713 CtdlLogPrintf(CTDL_INFO, "%s\n", curl_version());
714 CtdlRegisterSessionHook(rssclient_scan, EVT_TIMER);
715 CtdlRegisterCleanupHook(rss_cleanup);