parser: experimentally set preservewhitespace to false

[tt-rss.git] / classes / feedparser.php

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 1c97e496b25b09bd4fd69b572f5c47530128bd8c..efef3bca41a0914c01b5f1c85e3588840ee95d2b 100644 (file)
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -13,10 +13,21 @@ class FeedParser {
        const FEED_RSS = 1;
        const FEED_ATOM = 2;
 
+       function normalize_encoding($data) {
+               if (preg_match('/^(<\?xml[\t\n\r ].*?encoding[\t\n\r ]*=[\t\n\r ]*["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
+                       $data = mb_convert_encoding($data, 'UTF-8', $matches[2]);
+
+                       $data = preg_replace('/^<\?xml[\t\n\r ].*?\?>/s', $matches[1] . "UTF-8" . $matches[3] , $data);
+               }
+
+               return $data;
+       }
+
        function __construct($data) {
                libxml_use_internal_errors(true);
                libxml_clear_errors();
                $this->doc = new DOMDocument();
+               $this->doc->preserveWhiteSpace = false;
                $this->doc->loadXML($data);
 
                mb_substitute_character("none");
@@ -25,19 +36,16 @@ class FeedParser {
 
                // libxml compiled without iconv?
                if ($error && $error->code == 32) {
-                       if (preg_match('/^(<\?xml[\t\n\r ].*?encoding[\t\n\r ]*=[\t\n\r ]*["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
-                               $data = mb_convert_encoding($data, 'UTF-8', $matches[2]);
-
-                               $data = preg_replace('/^<\?xml[\t\n\r ].*?\?>/s', $matches[1] . "UTF-8" . $matches[3] , $data);
+                       $data = $this->normalize_encoding($data);
 
-                               if ($data) {
-                                       libxml_clear_errors();
+                       if ($data) {
+                               libxml_clear_errors();
 
-                                       $this->doc = new DOMDocument();
-                                       $this->doc->loadXML($data);
+                               $this->doc = new DOMDocument();
+                               $this->doc->preserveWhiteSpace = false;
+                               $this->doc->loadXML($data);
 
-                                       $error = libxml_get_last_error();
-                               }
+                               $error = libxml_get_last_error();
                        }
                }
 
@@ -45,6 +53,9 @@ class FeedParser {
                if ($error) {
                        foreach (libxml_get_errors() as $err) {
                                if ($err->code == 9) {
+                                       // if the source feed is not in utf8, next conversion will fail
+                                       $data = $this->normalize_encoding($data);
+
                                        // remove dangling bytes
                                        $data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
 
@@ -55,6 +66,7 @@ class FeedParser {
                                                libxml_clear_errors();
 
                                                $this->doc = new DOMDocument();
+                                               $this->doc->preserveWhiteSpace = false;
                                                $this->doc->loadXML($data);
 
                                                $error = libxml_get_last_error();
@@ -88,6 +100,7 @@ class FeedParser {
                $xpath->registerNamespace('slash', 'http://purl.org/rss/1.0/modules/slash/');
                $xpath->registerNamespace('dc', 'http://purl.org/dc/elements/1.1/');
                $xpath->registerNamespace('content', 'http://purl.org/rss/1.0/modules/content/');
+               $xpath->registerNamespace('thread', 'http://purl.org/syndication/thread/1.0');
 
                $this->xpath = $xpath;