From a9000b03443d64ac5d7137c868862a5e1496e871 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Wed, 19 Jun 2013 19:40:36 +0400
Subject: feedparser: check if initial xpath query for root element returns
 anything

---
 classes/feedparser.php | 34 +++++++++++++++++++---------------
 1 file changed, 19 insertions(+), 15 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index d93c575b2..eb8606de9 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -51,24 +51,28 @@ class FeedParser {
 
 		$this->xpath = $xpath;
 
-		$root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)")->item(0);
+		$root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)");
 
 		if ($root) {
-			switch (mb_strtolower($root->tagName)) {
-			case "rdf:rdf":
-				$this->type = $this::FEED_RDF;
-				break;
-			case "channel":
-				$this->type = $this::FEED_RSS;
-				break;
-			case "feed":
-				$this->type = $this::FEED_ATOM;
-				break;
-			default:
-				if( !isset($this->error) ){
-					$this->error = "Unknown/unsupported feed type";
+			$root = $root->item(0);
+
+			if ($root) {
+				switch (mb_strtolower($root->tagName)) {
+				case "rdf:rdf":
+					$this->type = $this::FEED_RDF;
+					break;
+				case "channel":
+					$this->type = $this::FEED_RSS;
+					break;
+				case "feed":
+					$this->type = $this::FEED_ATOM;
+					break;
+				default:
+					if( !isset($this->error) ){
+						$this->error = "Unknown/unsupported feed type";
+					}
+					return;
 				}
-				return;
 			}
 
 			switch ($this->type) {
-- 
cgit v1.2.3-54-g00ecf


From 4f00f55ca2ecd2e5a75c2c4ef37ca0e1143a7ac7 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Thu, 11 Jul 2013 15:40:09 +0400
Subject: parser: add charset recoding hack for systems where libxml is build
 without support for iconv (handles libxml error 32)

---
 classes/feedparser.php | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index eb8606de9..53f6c52a9 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -20,6 +20,24 @@ class FeedParser {
 
 		$error = libxml_get_last_error();
 
+		// libxml compiled without iconv?
+		if ($error && $error->code == 32) {
+			if (preg_match('/^(<\\?xml .*?)encoding="(.+?)"(.*?\\?>)/', $data, $matches) === 1) {
+				libxml_clear_errors();
+
+				$enc = $matches[2];
+
+	         $data = iconv($enc, 'UTF-8//IGNORE', $data);
+	         $data = preg_replace('/^<\\?xml .*?\\?>/', $matches[1] . $matches[3] , $data);
+
+				$this->doc = new DOMDocument();
+				$this->doc->loadXML($data);
+
+				$error = libxml_get_last_error();
+		   }
+		}
+
+		// some terrible invalid unicode entity?
 		if ($error && $error->code == 9) {
 			libxml_clear_errors();
 
-- 
cgit v1.2.3-54-g00ecf


From f8160106af762da1a2f4b4a2579b8165cb2005e2 Mon Sep 17 00:00:00 2001
From: wltb <github.11.el_robb_duderino@spamgourmet.com>
Date: Mon, 15 Jul 2013 00:25:45 +0200
Subject: Feedparser/encoding change: More general regular expression, set
 encoding to UTF-8 explicitly

---
 classes/feedparser.php | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 53f6c52a9..651ee010f 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -22,13 +22,13 @@ class FeedParser {
 
 		// libxml compiled without iconv?
 		if ($error && $error->code == 32) {
-			if (preg_match('/^(<\\?xml .*?)encoding="(.+?)"(.*?\\?>)/', $data, $matches) === 1) {
+			if (preg_match('/^(<\?xml[\t\n\r ].*?encoding=["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
 				libxml_clear_errors();
 
 				$enc = $matches[2];
 
-	         $data = iconv($enc, 'UTF-8//IGNORE', $data);
-	         $data = preg_replace('/^<\\?xml .*?\\?>/', $matches[1] . $matches[3] , $data);
+				$data = iconv($enc, 'UTF-8//IGNORE', $data);
+				$data = preg_replace('/^<\?xml[\t\n\r ].*?\?>/s', $matches[1] . "UTF-8" . $matches[3] , $data);
 
 				$this->doc = new DOMDocument();
 				$this->doc->loadXML($data);
-- 
cgit v1.2.3-54-g00ecf


From f612dbe8a097544cdb06e08f528b43195e911670 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Sat, 3 Aug 2013 14:45:27 +0400
Subject: improve support for feeds with invalid unicode entities

---
 classes/feedparser.php | 36 +++++++++++++++++++++++-------------
 1 file changed, 23 insertions(+), 13 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 651ee010f..4a2c6c2da 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -18,36 +18,46 @@ class FeedParser {
 		$this->doc = new DOMDocument();
 		$this->doc->loadXML($data);
 
+		mb_substitute_character("none");
+
 		$error = libxml_get_last_error();
 
 		// libxml compiled without iconv?
-		if ($error && $error->code == 32) {
+		if ($error && ($error->code == 32 || $error->code == 9)) {
 			if (preg_match('/^(<\?xml[\t\n\r ].*?encoding=["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
-				libxml_clear_errors();
-
 				$enc = $matches[2];
 
-				$data = iconv($enc, 'UTF-8//IGNORE', $data);
+				$data = mb_convert_encoding($data, 'UTF-8', $enc);
+
 				$data = preg_replace('/^<\?xml[\t\n\r ].*?\?>/s', $matches[1] . "UTF-8" . $matches[3] , $data);
 
-				$this->doc = new DOMDocument();
-				$this->doc->loadXML($data);
 
-				$error = libxml_get_last_error();
+				// apparently not all UTF-8 characters are valid for XML
+				$data = preg_replace('/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', ' ', $data);
+
+				if ($data) {
+					libxml_clear_errors();
+
+					$this->doc = new DOMDocument();
+					$this->doc->loadXML($data);
+
+					$error = libxml_get_last_error();
+				}
 		   }
 		}
 
 		// some terrible invalid unicode entity?
 		if ($error && $error->code == 9) {
-			libxml_clear_errors();
+			$data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
 
-			// we might want to try guessing input encoding here too
-			$data = iconv("UTF-8", "UTF-8//IGNORE", $data);
+			if ($data) {
+				libxml_clear_errors();
 
-			$this->doc = new DOMDocument();
-			$this->doc->loadXML($data);
+				$this->doc = new DOMDocument();
+				$this->doc->loadXML($data);
 
-			$error = libxml_get_last_error();
+				$error = libxml_get_last_error();
+			}
 		}
 
 		$this->error = $this->format_error($error);
-- 
cgit v1.2.3-54-g00ecf


From 5fa36a640168d3eaf27f3d6630ffd71527bcd724 Mon Sep 17 00:00:00 2001
From: wltb <wltb@localhost.com>
Date: Mon, 9 Sep 2013 00:44:55 +0200
Subject: Feedparser: Change handling of libxml error 9 (cycle all errors)

---
 classes/feedparser.php | 37 ++++++++++++++++++++-----------------
 1 file changed, 20 insertions(+), 17 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 4a2c6c2da..22052bdb6 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -23,18 +23,12 @@ class FeedParser {
 		$error = libxml_get_last_error();
 
 		// libxml compiled without iconv?
-		if ($error && ($error->code == 32 || $error->code == 9)) {
-			if (preg_match('/^(<\?xml[\t\n\r ].*?encoding=["\'])(.+?)(["\'].*?\?>)/s', $data, $matches) === 1) {
-				$enc = $matches[2];
-
-				$data = mb_convert_encoding($data, 'UTF-8', $enc);
+		if ($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);
 
-
-				// apparently not all UTF-8 characters are valid for XML
-				$data = preg_replace('/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', ' ', $data);
-
 				if ($data) {
 					libxml_clear_errors();
 
@@ -43,20 +37,29 @@ class FeedParser {
 
 					$error = libxml_get_last_error();
 				}
-		   }
+			}
 		}
 
 		// some terrible invalid unicode entity?
-		if ($error && $error->code == 9) {
-			$data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
+		if ($error) {
+			foreach(libxml_get_errors() as $err) {
+				if ($err->code == 9) {
+					// remove dangling bytes
+					$data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
+					
+					// apparently not all UTF-8 characters are valid for XML
+					$data = preg_replace('/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', ' ', $data);
 
-			if ($data) {
-				libxml_clear_errors();
+					if ($data) {
+						libxml_clear_errors();
 
-				$this->doc = new DOMDocument();
-				$this->doc->loadXML($data);
+						$this->doc = new DOMDocument();
+						$this->doc->loadXML($data);
 
-				$error = libxml_get_last_error();
+						$error = libxml_get_last_error();
+					}
+					break;
+				}
 			}
 		}
 
-- 
cgit v1.2.3-54-g00ecf


From d3305ff8de6ca9881021646221827d27240dba38 Mon Sep 17 00:00:00 2001
From: wltb <wltb@localhost.com>
Date: Tue, 24 Sep 2013 16:43:47 +0200
Subject: Feedparser: Only format fatal errors

---
 classes/feedparser.php | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 22052bdb6..07de73a9d 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -62,8 +62,16 @@ class FeedParser {
 				}
 			}
 		}
-
-		$this->error = $this->format_error($error);
+		
+		$this->error = "";
+		if($error) {
+			foreach(libxml_get_errors() as $error) {
+				if($error->level == LIBXML_ERR_FATAL) {
+					$this->error = $this->format_error($error);
+					break; //break here because currently we only show one error
+				}
+			}	
+		}
 		libxml_clear_errors();
 
 		$this->items = array();
-- 
cgit v1.2.3-54-g00ecf


From 4d49863f65e14190c8e460700f2bd197e724666c Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Wed, 25 Sep 2013 14:26:45 +0400
Subject: fix style in feedparser

---
 classes/feedparser.php | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 07de73a9d..2942fb27d 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -42,11 +42,11 @@ class FeedParser {
 
 		// some terrible invalid unicode entity?
 		if ($error) {
-			foreach(libxml_get_errors() as $err) {
+			foreach (libxml_get_errors() as $err) {
 				if ($err->code == 9) {
 					// remove dangling bytes
 					$data = mb_convert_encoding($data, 'UTF-8', 'UTF-8');
-					
+
 					// apparently not all UTF-8 characters are valid for XML
 					$data = preg_replace('/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', ' ', $data);
 
@@ -62,15 +62,15 @@ class FeedParser {
 				}
 			}
 		}
-		
+
 		$this->error = "";
-		if($error) {
-			foreach(libxml_get_errors() as $error) {
-				if($error->level == LIBXML_ERR_FATAL) {
+		if ($error) {
+			foreach (libxml_get_errors() as $error) {
+				if ($error->level == LIBXML_ERR_FATAL) {
 					$this->error = $this->format_error($error);
 					break; //break here because currently we only show one error
 				}
-			}	
+			}
 		}
 		libxml_clear_errors();
 
-- 
cgit v1.2.3-54-g00ecf


From 54f23d38a6d9e994575d2159b2277674149ae5d8 Mon Sep 17 00:00:00 2001
From: wltb <wltb@localhost.com>
Date: Thu, 26 Sep 2013 19:56:45 +0200
Subject: Feedparser: Store libXML fatal error messages in an array, repair
 error reporting

---
 classes/feedparser.php | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 2942fb27d..4eaeb2429 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -2,6 +2,7 @@
 class FeedParser {
 	private $doc;
 	private $error;
+	private $libxml_errors = array();
 	private $items;
 	private $link;
 	private $title;
@@ -63,12 +64,12 @@ class FeedParser {
 			}
 		}
 
-		$this->error = "";
 		if ($error) {
 			foreach (libxml_get_errors() as $error) {
 				if ($error->level == LIBXML_ERR_FATAL) {
-					$this->error = $this->format_error($error);
-					break; //break here because currently we only show one error
+					if(!isset($this->error)) //currently only the first error is reported
+						$this->error = $this->format_error($error);
+					$this->libxml_errors [] = $this->format_error($error);
 				}
 			}
 		}
@@ -216,6 +217,10 @@ class FeedParser {
 		return $this->error;
 	}
 
+	function errors() {
+		return $this->libxml_errors;
+	}
+
 	function get_link() {
 		return $this->link;
 	}
-- 
cgit v1.2.3-54-g00ecf


From 2c6f3c2eb304c227bbb1d4bbbf1df8a9879715b1 Mon Sep 17 00:00:00 2001
From: wltb <wltb@localhost.com>
Date: Fri, 27 Sep 2013 00:13:55 +0200
Subject: Feedparser: Add important check for unknown feed type detection

---
 classes/feedparser.php | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 4eaeb2429..1c97e496b 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -93,7 +93,7 @@ class FeedParser {
 
 		$root = $xpath->query("(//atom03:feed|//atom:feed|//channel|//rdf:rdf|//rdf:RDF)");
 
-		if ($root) {
+		if ($root && $root->length > 0) {
 			$root = $root->item(0);
 
 			if ($root) {
-- 
cgit v1.2.3-54-g00ecf


From 4ad04ee227dd7d704f417aaf9d6762f5cfdf4c1f Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Tue, 29 Oct 2013 12:15:26 +0400
Subject: report all libxml errors in updater debug output force utf8 encoding
 if devforceupdate is on parser: try to convert non-unicode feeds with
 specified encoding to utf8 before trying to remove dangling utf8 characters
 in case of utf8-related libxml errors because doing so produces garbage
 content

---
 classes/feedparser.php | 29 +++++++++++++++++++----------
 classes/feeds.php      |  2 +-
 include/rssfuncs.php   |  8 +++++++-
 3 files changed, 27 insertions(+), 12 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 1c97e496b..de6c56542 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -13,6 +13,16 @@ 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();
@@ -25,19 +35,15 @@ 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();
+				$this->doc->loadXML($data);
 
-					$error = libxml_get_last_error();
-				}
+				$error = libxml_get_last_error();
 			}
 		}
 
@@ -45,6 +51,9 @@ 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');
 
diff --git a/classes/feeds.php b/classes/feeds.php
index 7f5fd10af..2c17a2257 100644
--- a/classes/feeds.php
+++ b/classes/feeds.php
@@ -148,7 +148,7 @@ class Feeds extends Handler_Protected {
 					$override_order = false, $include_children = false) {
 
 		if (isset($_REQUEST["DevForceUpdate"]))
-			header("Content-Type: text/plain");
+			header("Content-Type: text/plain; charset=utf-8");
 
 		$disable_cache = false;
 
diff --git a/include/rssfuncs.php b/include/rssfuncs.php
index bfbec0919..bc6048217 100644
--- a/include/rssfuncs.php
+++ b/include/rssfuncs.php
@@ -1110,7 +1110,13 @@
 
 			$error_msg = db_escape_string(mb_substr($rss->error(), 0, 245));
 
-			_debug("error fetching feed: $error_msg", $debug_enabled);
+			_debug("fetch error: $error_msg", $debug_enabled);
+
+			if (count($rss->errors()) > 1) {
+				foreach ($rss->errors() as $error) {
+					_debug("+ $error");
+				}
+			}
 
 			db_query(
 				"UPDATE ttrss_feeds SET last_error = '$error_msg',
-- 
cgit v1.2.3-54-g00ecf


From d71ac5d3dabde7dda4aef6efa2af3a68ee2095c6 Mon Sep 17 00:00:00 2001
From: wltb <wltb@localhost.com>
Date: Wed, 18 Dec 2013 17:05:43 +0100
Subject: implemented get_comments_url() and Atom Threading Extension

---
 classes/feeditem/common.php | 17 +++++++++++++++--
 classes/feedparser.php      |  1 +
 2 files changed, 16 insertions(+), 2 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feeditem/common.php b/classes/feeditem/common.php
index 58065b198..80bebf8fb 100644
--- a/classes/feeditem/common.php
+++ b/classes/feeditem/common.php
@@ -44,13 +44,26 @@ abstract class FeedItem_Common extends FeedItem {
 		}
 	}
 
-	// todo
 	function get_comments_url() {
+		//RSS only. Use a query here to avoid namespace clashes (e.g. with slash).
+		//might give a wrong result if a default namespace was declared (possible with XPath 2.0)
+		$com_url = $this->xpath->query("comments", $this->elem)->item(0);
 
+		if($com_url)
+			return $com_url->nodeValue;
+
+		//Atom Threading Extension (RFC 4685) stuff. Could be used in RSS feeds, so it's in common.
+		//'text/html' for type is too restrictive?
+		$com_url = $this->xpath->query("atom:link[@rel='replies' and contains(@type,'text/html')]/@href", $this->elem)->item(0);
+
+		if($com_url)
+			return $com_url->nodeValue;
 	}
 
 	function get_comments_count() {
-		$comments = $this->xpath->query("slash:comments", $this->elem)->item(0);
+		//also query for ATE stuff here
+		$query = "slash:comments|thread:total|atom:link[@rel='replies']/@thread:count";
+		$comments = $this->xpath->query($query, $this->elem)->item(0);
 
 		if ($comments) {
 			return $comments->nodeValue;
diff --git a/classes/feedparser.php b/classes/feedparser.php
index de6c56542..187875b5f 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -97,6 +97,7 @@ 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;
 
-- 
cgit v1.2.3-54-g00ecf


From b69d94eeeb6ed539267a4328265354a45d23ba86 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <noreply@fakecake.org>
Date: Mon, 3 Mar 2014 21:04:28 +0400
Subject: parser: experimentally set preservewhitespace to false

---
 classes/feedparser.php | 3 +++
 1 file changed, 3 insertions(+)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index 187875b5f..efef3bca4 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -27,6 +27,7 @@ class FeedParser {
 		libxml_use_internal_errors(true);
 		libxml_clear_errors();
 		$this->doc = new DOMDocument();
+		$this->doc->preserveWhiteSpace = false;
 		$this->doc->loadXML($data);
 
 		mb_substitute_character("none");
@@ -41,6 +42,7 @@ class FeedParser {
 				libxml_clear_errors();
 
 				$this->doc = new DOMDocument();
+				$this->doc->preserveWhiteSpace = false;
 				$this->doc->loadXML($data);
 
 				$error = libxml_get_last_error();
@@ -64,6 +66,7 @@ class FeedParser {
 						libxml_clear_errors();
 
 						$this->doc = new DOMDocument();
+						$this->doc->preserveWhiteSpace = false;
 						$this->doc->loadXML($data);
 
 						$error = libxml_get_last_error();
-- 
cgit v1.2.3-54-g00ecf


From 4ce778677ef16353d5b0b8342a29bd65575310f5 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <noreply@fakecake.org>
Date: Tue, 4 Mar 2014 16:30:46 +0400
Subject: Revert "parser: experimentally set preservewhitespace to false"

This reverts commit b69d94eeeb6ed539267a4328265354a45d23ba86.
---
 classes/feedparser.php | 3 ---
 1 file changed, 3 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feedparser.php b/classes/feedparser.php
index efef3bca4..187875b5f 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -27,7 +27,6 @@ class FeedParser {
 		libxml_use_internal_errors(true);
 		libxml_clear_errors();
 		$this->doc = new DOMDocument();
-		$this->doc->preserveWhiteSpace = false;
 		$this->doc->loadXML($data);
 
 		mb_substitute_character("none");
@@ -42,7 +41,6 @@ class FeedParser {
 				libxml_clear_errors();
 
 				$this->doc = new DOMDocument();
-				$this->doc->preserveWhiteSpace = false;
 				$this->doc->loadXML($data);
 
 				$error = libxml_get_last_error();
@@ -66,7 +64,6 @@ class FeedParser {
 						libxml_clear_errors();
 
 						$this->doc = new DOMDocument();
-						$this->doc->preserveWhiteSpace = false;
 						$this->doc->loadXML($data);
 
 						$error = libxml_get_last_error();
-- 
cgit v1.2.3-54-g00ecf


From 31bd6f7643bf139802a224f4584caca3cbbcc9b8 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <noreply@fakecake.org>
Date: Tue, 4 Mar 2014 16:38:04 +0400
Subject: parser: trim some some feed-extracted data link titles and links

---
 classes/feeditem/atom.php | 10 +++++-----
 classes/feeditem/rss.php  | 12 ++++++------
 classes/feedparser.php    |  8 ++++++--
 3 files changed, 17 insertions(+), 13 deletions(-)

(limited to 'classes/feedparser.php')

diff --git a/classes/feeditem/atom.php b/classes/feeditem/atom.php
index 74be03d80..48e3aa567 100644
--- a/classes/feeditem/atom.php
+++ b/classes/feeditem/atom.php
@@ -43,9 +43,9 @@ class FeedItem_Atom extends FeedItem_Common {
 				$base = $this->xpath->evaluate("string(ancestor-or-self::*[@xml:base][1]/@xml:base)", $link);
 
 				if ($base)
-					return rewrite_relative_url($base, $link->getAttribute("href"));
+					return rewrite_relative_url($base, trim($link->getAttribute("href")));
 				else
-					return $link->getAttribute("href");
+					return trim($link->getAttribute("href"));
 
 			}
 		}
@@ -55,7 +55,7 @@ class FeedItem_Atom extends FeedItem_Common {
 		$title = $this->elem->getElementsByTagName("title")->item(0);
 
 		if ($title) {
-			return $title->nodeValue;
+			return trim($title->nodeValue);
 		}
 	}
 
@@ -106,13 +106,13 @@ class FeedItem_Atom extends FeedItem_Common {
 
 		foreach ($categories as $cat) {
 			if ($cat->hasAttribute("term"))
-				array_push($cats, $cat->getAttribute("term"));
+				array_push($cats, trim($cat->getAttribute("term")));
 		}
 
 		$categories = $this->xpath->query("dc:subject", $this->elem);
 
 		foreach ($categories as $cat) {
-			array_push($cats, $cat->nodeValue);
+			array_push($cats, trim($cat->nodeValue));
 		}
 
 		return $cats;
diff --git a/classes/feeditem/rss.php b/classes/feeditem/rss.php
index 28f4a388d..bf08a1dfe 100644
--- a/classes/feeditem/rss.php
+++ b/classes/feeditem/rss.php
@@ -33,20 +33,20 @@ class FeedItem_RSS extends FeedItem_Common {
 					|| $link->getAttribute("rel") == "alternate"
 					|| $link->getAttribute("rel") == "standout")) {
 
-				return $link->getAttribute("href");
+				return trim($link->getAttribute("href"));
 			}
 		}
 
 		$link = $this->elem->getElementsByTagName("guid")->item(0);
 
 		if ($link && $link->hasAttributes() && $link->getAttribute("isPermaLink") == "true") {
-			return $link->nodeValue;
+			return trim($link->nodeValue);
 		}
 
 		$link = $this->elem->getElementsByTagName("link")->item(0);
 
 		if ($link) {
-			return $link->nodeValue;
+			return trim($link->nodeValue);
 		}
 	}
 
@@ -54,7 +54,7 @@ class FeedItem_RSS extends FeedItem_Common {
 		$title = $this->elem->getElementsByTagName("title")->item(0);
 
 		if ($title) {
-			return $title->nodeValue;
+			return trim($title->nodeValue);
 		}
 	}
 
@@ -90,13 +90,13 @@ class FeedItem_RSS extends FeedItem_Common {
 		$cats = array();
 
 		foreach ($categories as $cat) {
-			array_push($cats, $cat->nodeValue);
+			array_push($cats, trim($cat->nodeValue));
 		}
 
 		$categories = $this->xpath->query("dc:subject", $this->elem);
 
 		foreach ($categories as $cat) {
-			array_push($cats, $cat->nodeValue);
+			array_push($cats, trim($cat->nodeValue));
 		}
 
 		return $cats;
diff --git a/classes/feedparser.php b/classes/feedparser.php
index 187875b5f..239fdb7a6 100644
--- a/classes/feedparser.php
+++ b/classes/feedparser.php
@@ -205,6 +205,10 @@ class FeedParser {
 				break;
 
 			}
+
+			if ($this->title) $this->title = trim($this->title);
+			if ($this->link) $this->link = trim($this->link);
+
 		} else {
 			if( !isset($this->error) ){
 				$this->error = "Unknown/unsupported feed type";
@@ -252,7 +256,7 @@ class FeedParser {
 
 			foreach ($links as $link) {
 				if (!$rel || $link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
-					array_push($rv, $link->getAttribute('href'));
+					array_push($rv, trim($link->getAttribute('href')));
 				}
 			}
 			break;
@@ -261,7 +265,7 @@ class FeedParser {
 
 			foreach ($links as $link) {
 				if (!$rel || $link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
-					array_push($rv, $link->getAttribute('href'));
+					array_push($rv, trim($link->getAttribute('href')));
 				}
 			}
 			break;
-- 
cgit v1.2.3-54-g00ecf