Better UTF-8 escaped entity handling. Fixed bad chars in all RSS feeds.

darcs-hash:20081104084027-462f3-ea7ab93938358bf90a1c1851d6f665973beae767.gz

3 files changed, 17 insertions, 10 deletions

diff --git a/lib/rssaction.php b/lib/rssaction.php
index 76859a876..0152f25bc 100644
--- a/lib/rssaction.php
+++ b/lib/rssaction.php
@@ -123,7 +123,7 @@ class Rss10Action extends Action {
 		$nurl = common_local_url('shownotice', array('notice' => $notice->id));
 		$creator_uri = common_profile_uri($profile);
 		common_element_start('item', array('rdf:about' => $notice->uri));
-		$title = $profile->nickname . ': ' . $notice->content;
+		$title = $profile->nickname . ': ' . common_xml_safe_str($notice->content);
 		common_element('title', NULL, $title);
 		common_element('link', NULL, $nurl);
 		common_element('description', NULL, $profile->nickname."'s status on ".common_exact_date($notice->created));
diff --git a/lib/twitterapi.php b/lib/twitterapi.php
index d4b6fff4f..378716eaa 100644
--- a/lib/twitterapi.php
+++ b/lib/twitterapi.php
@@ -60,7 +60,7 @@ class TwitterapiAction extends Action {
 		$profile = $notice->getProfile();
 
 		$twitter_status = array();
-		$twitter_status['text'] = $notice->content;
+		$twitter_status['text'] = common_xml_safe_str($notice->content);
 		$twitter_status['truncated'] = 'false'; # Not possible on Laconica
 		$twitter_status['created_at'] = $this->date_twitter($notice->created);
 		$twitter_status['in_reply_to_status_id'] = ($notice->reply_to) ? intval($notice->reply_to) : NULL;
@@ -91,8 +91,8 @@ class TwitterapiAction extends Action {
 
 		$server = common_config('site', 'server');
 		$entry = array();
-
-		$entry['content'] = $profile->nickname . ': ' . $notice->content;
+		 
+		$entry['content'] = $profile->nickname . ': ' . common_xml_safe_str($notice->content);
 		$entry['title'] = $entry['content'];
 		$entry['link'] = common_local_url('shownotice', array('notice' => $notice->id));
 		$entry['published'] = common_date_iso8601($notice->created);
@@ -115,14 +115,14 @@ class TwitterapiAction extends Action {
 		$entry['title'] = sprintf('Message from %s to %s',
 			$message->getFrom()->nickname, $message->getTo()->nickname);
 
-		$entry['content'] = $message->content;
+		$entry['content'] = common_xml_safe_str($message->content);
 		$entry['link'] = $message->uri;
 		$entry['published'] = common_date_iso8601($message->created);
 		$entry['id'] = "tag:$server,2008:$entry[link]";
 		$entry['updated'] = $entry['published'];
 
 		# RSS Item specific
-		$entry['description'] = $message->content;
+		$entry['description'] = $entry['content'];
 		$entry['pubDate'] = common_date_rfc2822($message->created);
 		$entry['guid'] = $entry['link'];
 
@@ -137,8 +137,8 @@ class TwitterapiAction extends Action {
 		$to_profile = $message->getTo();
 
 		$twitter_dm['id'] = $message->id;
-		$twitter_dm['sender_id'] = $message->from_profile;
-		$twitter_dm['text'] = $message->content;
+		$twitter_dm['sender_id'] = $message->from_profile;		
+		$twitter_dm['text'] = common_xml_safe_str($message->content);
 		$twitter_dm['recipient_id'] = $message->to_profile;
 		$twitter_dm['created_at'] = $this->date_twitter($message->created);
 		$twitter_dm['sender_screen_name'] = $from_profile->nickname;
@@ -569,5 +569,5 @@ class TwitterapiAction extends Action {
 		}
 		return $source_name;
 	}
-
+	
 }
\ No newline at end of file
diff --git a/lib/util.php b/lib/util.php
index 88841d3f6..23abb1426 100644
--- a/lib/util.php
+++ b/lib/util.php
@@ -708,7 +708,7 @@ function common_render_content($text, $notice) {
 }
 
 function common_render_text($text) {
-	$r = htmlentities($text, ENT_NOQUOTES, 'UTF-8');
+	$r = htmlspecialchars($text);
 
 	$r = preg_replace('/[\x{0}-\x{8}\x{b}-\x{c}\x{e}-\x{19}]/', '', $r);
 	$r = preg_replace_callback('@https?://[^\]>\s]+@', 'common_render_uri_thingy', $r);
@@ -745,6 +745,13 @@ function common_render_uri_thingy($matches) {
 	return '<a href="' . $uri . '" class="extlink">' . $uri . '</a>' . $trailer;
 }
 
+function common_xml_safe_str($str) {
+	$xmlStr = htmlentities(iconv('UTF-8', 'UTF-8//IGNORE', $str), ENT_NOQUOTES, 'UTF-8');
+	
+	// Replace control, formatting, and surrogate characters with '*', ala Twitter
+	return preg_replace('/[\p{Cc}\p{Cf}\p{Cs}]/u', '*', $str);
+}
+
 function common_tag_link($tag) {
 	$canonical = common_canonical_tag($tag);
 	$url = common_local_url('tag', array('tag' => $canonical));