/*
- * Bring external RSS feeds into rooms.
+ * Bring external RSS and/or Atom feeds into rooms. This module implements a
+ * very loose parser that scrapes both kinds of feeds and is not picky about
+ * the standards compliance of the source data.
*
- * Copyright (c) 2007-2017 by the citadel.org team
+ * Copyright (c) 2007-2021 by the citadel.org team
*
- * This program is open source software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 3.
+ * This program is open source software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 3.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.See the
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*/
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
-
-#if TIME_WITH_SYS_TIME
-# include <sys/time.h>
-# include <time.h>
-#else
-# if HAVE_SYS_TIME_H
-#include <sys/time.h>
-# else
#include <time.h>
-# endif
-#endif
-
#include <ctype.h>
#include <string.h>
#include <errno.h>
#include "parsedate.h"
#include "database.h"
#include "citadel_dirs.h"
-#include "md5.h"
#include "context.h"
#include "internet_addressing.h"
};
time_t last_run = 0L;
-struct CitContext rss_CC;
struct rssurl *rsstodo = NULL;
// This handler is called whenever an XML tag opens.
//
-void rss_start_element(void *data, const char *el, const char **attribute)
-{
+void rss_start_element(void *data, const char *el, const char **attribute) {
struct rssparser *r = (struct rssparser *)data;
int i;
+ if (server_shutting_down) return; // shunt the whole operation if we're exiting
+
if (
(!strcasecmp(el, "entry"))
|| (!strcasecmp(el, "item"))
// This handler is called whenever an XML tag closes.
//
-void rss_end_element(void *data, const char *el)
-{
+void rss_end_element(void *data, const char *el) {
struct rssparser *r = (struct rssparser *)data;
+ StrBuf *encoded_field;
+
+ if (server_shutting_down) return; // shunt the whole operation if we're exiting
+
+ if (StrLength(r->CData) > 0) { // strip leading/trailing whitespace from field
+ StrBufTrim(r->CData);
+ }
if ( // end of a new item(rss) or entry(atom)
(!strcasecmp(el, "entry"))
|| (!strcasecmp(el, "item"))
) {
-
if (r->msg != NULL) { // Save the message to the rooms
// use the link as an item id if nothing else is available
// check the use table
StrBuf *u = NewStrBuf();
StrBufAppendPrintf(u, "rss/%s", r->item_id);
- time_t already_seen = CheckIfAlreadySeen(u, time(NULL), 0, eUpdate);
+ int already_seen = CheckIfAlreadySeen(u);
FreeStrBuf(&u);
if (already_seen == 0) {
long msgnum = (-1);
for (rr=r->rooms; rr!=NULL; rr=rr->next) {
if (rr == r->rooms) {
- msgnum = CtdlSubmitMsg(r->msg, NULL, rr->room, 0);
+ msgnum = CtdlSubmitMsg(r->msg, NULL, rr->room); // in first room, save msg
}
else {
- CtdlSaveMsgPointerInRoom(rr->room, msgnum, 0, NULL);
+ CtdlSaveMsgPointerInRoom(rr->room, msgnum, 0, NULL); // elsewhere, save a pointer
}
- syslog(LOG_DEBUG, "Saved message %ld to %s", msgnum, rr->room);
+ syslog(LOG_DEBUG, "rssclient: saved message %ld to %s", msgnum, rr->room);
}
}
else {
- syslog(LOG_DEBUG, "%s was already seen %ld seconds ago", r->item_id, already_seen);
+ syslog(LOG_DEBUG, "rssclient: already seen %s", r->item_id);
}
CM_Free(r->msg);
else if (!strcasecmp(el, "title")) { // item subject (rss and atom)
if ((r->msg != NULL) && (CM_IsEmpty(r->msg, eMsgSubject))) {
- CM_SetField(r->msg, eMsgSubject, ChrPtr(r->CData), StrLength(r->CData));
- striplt(r->msg->cm_fields[eMsgSubject]);
+ encoded_field = NewStrBuf();
+ StrBufRFC2047encode(&encoded_field, r->CData);
+ CM_SetAsFieldSB(r->msg, eMsgSubject, &encoded_field);
}
}
- else if (!strcasecmp(el, "author")) { // author of item (rss and maybe atom)
+ else if (!strcasecmp(el, "creator")) { // <creator> can be used if <author> is not present
if ((r->msg != NULL) && (CM_IsEmpty(r->msg, eAuthor))) {
- CM_SetField(r->msg, eAuthor, ChrPtr(r->CData), StrLength(r->CData));
- striplt(r->msg->cm_fields[eAuthor]);
+ encoded_field = NewStrBuf();
+ StrBufRFC2047encode(&encoded_field, r->CData);
+ CM_SetAsFieldSB(r->msg, eAuthor, &encoded_field);
+ }
+ }
+
+ else if (!strcasecmp(el, "author")) { // <author> supercedes <creator> if both are present
+ if (r->msg != NULL) {
+ encoded_field = NewStrBuf();
+ StrBufRFC2047encode(&encoded_field, r->CData);
+ CM_SetAsFieldSB(r->msg, eAuthor, &encoded_field);
}
}
r->link = NULL;
}
r->link = strdup(ChrPtr(r->CData));
- striplt(r->link);
}
else if (
r->item_id = NULL;
}
r->item_id = strdup(ChrPtr(r->CData));
- striplt(r->item_id);
}
else if (
r->description = NULL;
}
r->description = strdup(ChrPtr(r->CData));
- striplt(r->description);
}
if (r->CData != NULL) {
struct rssurl *thisone = NULL;
struct rssroom *newroom = NULL;
- syslog(LOG_DEBUG, "rssclient_push_todo(%s, %s)", rssurl, roomname);
+ syslog(LOG_DEBUG, "rssclient: will fetch %s to %s", rssurl, roomname);
for (r=rsstodo; r!=NULL; r=r->next) {
if (!strcasecmp(r->url, rssurl)) {
thisone = r;
}
}
+
if (thisone == NULL) {
thisone = malloc(sizeof(struct rssurl));
thisone->url = strdup(rssurl);
}
-// Callback function for curl
-//
-size_t rss_pof_write_data(void *buffer, size_t size, size_t nmemb, void *userp)
-{
- StrBuf *Downloaded = (StrBuf *)userp;
- size_t bytes = size * nmemb;
- StrBufAppendBufPlain(Downloaded, buffer, bytes, 0);
- return(bytes);
-}
-
-
// pull one feed (possibly multiple rooms)
//
void rss_pull_one_feed(struct rssurl *url)
CURLcode res;
StrBuf *Downloaded = NULL;
- syslog(LOG_DEBUG, "rss_pull_one_feed(%s)", url->url);
+ syslog(LOG_DEBUG, "rssclient: fetching %s", url->url);
curl = curl_easy_init();
if (!curl) {
Downloaded = NewStrBuf();
curl_easy_setopt(curl, CURLOPT_URL, url->url);
+ curl_easy_setopt(curl, CURLOPT_SSL_VERIFYPEER, 0L);
+ curl_easy_setopt(curl, CURLOPT_SSL_VERIFYHOST, 0L);
curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L); // Follow redirects
- curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, rss_pof_write_data); // What to do with downloaded data
+ curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, CurlFillStrBuf_callback); // What to do with downloaded data
curl_easy_setopt(curl, CURLOPT_WRITEDATA, Downloaded); // Give it our StrBuf to work with
curl_easy_setopt(curl, CURLOPT_TIMEOUT, 20L); // Time out after 20 seconds
res = curl_easy_perform(curl); // Perform the request
if (res != CURLE_OK) {
- syslog(LOG_WARNING, "Failed to load feed: %s", curl_easy_strerror(res));
+ syslog(LOG_WARNING, "rssclient: failed to load feed: %s", curl_easy_strerror(res));
}
curl_easy_cleanup(curl);
struct rssurl *r;
struct rssroom *rr;
- while (rsstodo != NULL) {
+ while ((rsstodo != NULL) && (!server_shutting_down)) {
rss_pull_one_feed(rsstodo);
r = rsstodo;
rsstodo = rsstodo->next;
char cfgline[SIZ];
int i = 0;
+ if (server_shutting_down) return;
+
serialized_config = LoadRoomNetConfigFile(qrbuf->QRnumber);
if (!serialized_config) {
return;
/* Run no more than once every 15 minutes. */
if ((now - last_run) < 900) {
syslog(LOG_DEBUG,
- "Client: polling interval not yet reached; last run was %ldm%lds ago",
+ "rssclient: polling interval not yet reached; last run was %ldm%lds ago",
((now - last_run) / 60),
((now - last_run) % 60)
);
return;
}
- become_session(&rss_CC);
- syslog(LOG_DEBUG, "rssclient started");
+ syslog(LOG_DEBUG, "rssclient: started");
CtdlForEachRoom(rssclient_scan_room, NULL);
rss_pull_feeds();
- syslog(LOG_DEBUG, "rssclient ended");
+ syslog(LOG_DEBUG, "rssclient: ended");
last_run = time(NULL);
return;
}
{
if (!threading)
{
- syslog(LOG_INFO, "%s", curl_version());
+ syslog(LOG_INFO, "rssclient: using %s", curl_version());
CtdlRegisterSessionHook(rssclient_scan, EVT_TIMER, PRIO_AGGR + 300);
}
- else
- {
- CtdlFillSystemContext(&rss_CC, "rssclient");
- }
return "rssclient";
}
-
projects / citadel.git / blobdiff