aboutsummaryrefslogtreecommitdiffstats
path: root/include/Scrape.php
diff options
context:
space:
mode:
Diffstat (limited to 'include/Scrape.php')
-rw-r--r--include/Scrape.php58
1 files changed, 58 insertions, 0 deletions
diff --git a/include/Scrape.php b/include/Scrape.php
index e4f7a0878..bb42c3bdd 100644
--- a/include/Scrape.php
+++ b/include/Scrape.php
@@ -5,12 +5,25 @@ require_once('library/HTML5/Parser.php');
if(! function_exists('scrape_dfrn')) {
function scrape_dfrn($url) {
+ $a = get_app();
+
$ret = array();
$s = fetch_url($url);
if(! $s)
return $ret;
+ $headers = $a->get_curl_headers();
+ $lines = explode("\n",$headers);
+ if(count($lines)) {
+ foreach($lines as $line) {
+ // don't try and run feeds through the html5 parser
+ if(stristr($line,'content-type:') && ((stristr($line,'application/atom+xml')) || (stristr($line,'application/rss+xml'))))
+ return ret;
+ }
+ }
+
+
$dom = HTML5_Parser::parse($s);
if(! $dom)
@@ -77,12 +90,26 @@ function validate_dfrn($a) {
if(! function_exists('scrape_meta')) {
function scrape_meta($url) {
+ $a = get_app();
+
$ret = array();
$s = fetch_url($url);
if(! $s)
return $ret;
+ $headers = $a->get_curl_headers();
+ $lines = explode("\n",$headers);
+ if(count($lines)) {
+ foreach($lines as $line) {
+ // don't try and run feeds through the html5 parser
+ if(stristr($line,'content-type:') && ((stristr($line,'application/atom+xml')) || (stristr($line,'application/rss+xml'))))
+ return ret;
+ }
+ }
+
+
+
$dom = HTML5_Parser::parse($s);
if(! $dom)
@@ -105,12 +132,24 @@ function scrape_meta($url) {
if(! function_exists('scrape_vcard')) {
function scrape_vcard($url) {
+ $a = get_app();
+
$ret = array();
$s = fetch_url($url);
if(! $s)
return $ret;
+ $headers = $a->get_curl_headers();
+ $lines = explode("\n",$headers);
+ if(count($lines)) {
+ foreach($lines as $line) {
+ // don't try and run feeds through the html5 parser
+ if(stristr($line,'content-type:') && ((stristr($line,'application/atom+xml')) || (stristr($line,'application/rss+xml'))))
+ return ret;
+ }
+ }
+
$dom = HTML5_Parser::parse($s);
if(! $dom)
@@ -142,12 +181,31 @@ function scrape_vcard($url) {
if(! function_exists('scrape_feed')) {
function scrape_feed($url) {
+ $a = get_app();
+
$ret = array();
$s = fetch_url($url);
if(! $s)
return $ret;
+ $headers = $a->get_curl_headers();
+ $lines = explode("\n",$headers);
+ if(count($lines)) {
+ foreach($lines as $line) {
+ if(stristr($line,'content-type:')) {
+ if(stristr($line,'application/atom+xml')) {
+ $ret['feed_atom'] = $url;
+ return $ret;
+ }
+ if(stristr($line,'application/rss+xml')) {
+ $ret['feed_rss'] = $url;
+ return ret;
+ }
+ }
+ }
+ }
+
$dom = HTML5_Parser::parse($s);
if(! $dom)