Friendica/friendica-github
1
0
Fork 0
mirror of https://github.com/friendica/friendica synced 2024-11-10 12:22:53 +00:00
Code Issues Projects Releases Packages Wiki Activity

"Scrape" now respects the new url formats with "index.php"

Browse source
This commit is contained in:
Michael Vogel 2016-03-30 12:46:10 +02:00 committed by Roland Haeder
parent f2ce7f51cd
commit abf6fb10a5
No known key found for this signature in database
GPG key ID: B72F8185C6C7BD78
1 changed files with 43 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

49
include/Scrape.php
View file

@ -356,7 +356,7 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
$result = array();
if(! $url)
if (!$url)
return $result;
$result = Cache::get("probe_url:".$mode.":".$url);
@ -365,6 +365,7 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
return $result;
}
$original_url = $url;
$network = null;
$diaspora = false;
$diaspora_base = '';
@ -393,7 +394,12 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
else
$links = lrdd($url);
if(count($links)) {
if ((count($links) == 0) AND strstr($url, "/index.php")) {
$url = str_replace("/index.php", "", $url);
$links = lrdd($url);
}
if (count($links)) {
$has_lrdd = true;
logger('probe_url: found lrdd links: ' . print_r($links,true), LOGGER_DATA);
@ -440,12 +446,21 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
// aliases, let's hope we're lucky and get one that matches the feed author-uri because
// otherwise we're screwed.
$backup_alias = "";
foreach($links as $link) {
if($link['@attributes']['rel'] === 'alias') {
if(strpos($link['@attributes']['href'],'@') === false) {
if(isset($profile)) {
if(($link['@attributes']['href'] !== $profile) AND ($alias == ""))
$alias = unamp($link['@attributes']['href']);
$alias_url = $link['@attributes']['href'];
if(($alias_url !== $profile) AND ($backup_alias == "") AND
($alias_url !== str_replace("/index.php", "", $profile)))
$backup_alias = $alias_url;
if(($alias_url !== $profile) AND !strstr($alias_url, "index.php") AND
($alias_url !== str_replace("/index.php", "", $profile)))
$alias = $alias_url;
}
else
$profile = unamp($link['@attributes']['href']);
@ -453,6 +468,9 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
}
}
if ($alias == "")
$alias = $backup_alias;
// If the profile is different from the url then the url is abviously an alias
if (($alias == "") AND ($profile != "") AND !$at_addr AND (normalise_link($profile) != normalise_link($url)))
$alias = $url;
@ -769,6 +787,9 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
if (($baseurl == "") AND ($poll != ""))
$baseurl = matching_url(normalise_link($profile), normalise_link($poll));
if (substr($baseurl, -10) == "/index.php")
$baseurl = str_replace("/index.php", "", $baseurl);
$baseurl = rtrim($baseurl, "/");
if(strpos($url,'@') AND ($addr == "") AND ($network == NETWORK_DFRN))
@ -816,8 +837,24 @@ function probe_url($url, $mode = PROBE_NORMAL, $level = 1) {
}
// Only store into the cache if the value seems to be valid
if ($result['network'] != NETWORK_PHANTOM)
Cache::set("probe_url:".$mode.":".$url,serialize($result), CACHE_DAY);
if ($result['network'] != NETWORK_PHANTOM) {
Cache::set("probe_url:".$mode.":".$original_url,serialize($result), CACHE_DAY);
/// @todo temporary fix - we need a real contact update function that updates only changing fields
/// The biggest problem is the avatar picture that could have a reduced image size.
/// It should only be updated if the existing picture isn't existing anymore.
if (($result['network'] != NETWORK_FEED) AND $result["addr"] AND $result["name"] AND $result["nick"])
q("UPDATE `contact` SET `addr` = '%s', `alias` = '%s', `name` = '%s', `nick` = '%s',
`name-date` = '%s', `uri-date` = '%s' WHERE `nurl` = '%s' AND NOT `self`",
dbesc($result["addr"]),
dbesc($result["alias"]),
dbesc($result["name"]),
dbesc($result["nick"]),
dbesc(datetime_convert()),
dbesc(datetime_convert()),
dbesc(normalise_link($result['url']))
);
}
return $result;
}