From 759566fe99fe1fcf2857afc51d0b343b67aa4d44 Mon Sep 17 00:00:00 2001
From: Aleksander Machniak <alec@alec.pl>
Date: Fri, 27 Feb 2015 03:13:39 -0500
Subject: [PATCH] Fix parsing invalid HTML messages with BOM after <!DOCTYPE> (#1490291)
---
program/lib/Roundcube/rcube_washtml.php | 15 +++++++++++----
1 files changed, 11 insertions(+), 4 deletions(-)
diff --git a/program/lib/Roundcube/rcube_washtml.php b/program/lib/Roundcube/rcube_washtml.php
index e0cce68..b042f5f 100644
--- a/program/lib/Roundcube/rcube_washtml.php
+++ b/program/lib/Roundcube/rcube_washtml.php
@@ -403,16 +403,23 @@
{
// special replacements (not properly handled by washtml class)
$html_search = array(
- '/(<\/nobr>)(\s+)(<nobr>)/i', // space(s) between <NOBR>
- '/<title[^>]*>[^<]*<\/title>/i', // PHP bug #32547 workaround: remove title tag
- '/^(\0\0\xFE\xFF|\xFF\xFE\0\0|\xFE\xFF|\xFF\xFE|\xEF\xBB\xBF)/', // byte-order mark (only outlook?)
- '/<html\s[^>]+>/i', // washtml/DOMDocument cannot handle xml namespaces
+ // space(s) between <NOBR>
+ '/(<\/nobr>)(\s+)(<nobr>)/i',
+ // PHP bug #32547 workaround: remove title tag
+ '/<title[^>]*>[^<]*<\/title>/i',
+ // remove <!doctype> before BOM (#1490291)
+ '/<\!doctype[^>]+>[^<]*/im',
+ // byte-order mark (only outlook?)
+ '/^(\0\0\xFE\xFF|\xFF\xFE\0\0|\xFE\xFF|\xFF\xFE|\xEF\xBB\xBF)/',
+ // washtml/DOMDocument cannot handle xml namespaces
+ '/<html\s[^>]+>/i',
);
$html_replace = array(
'\\1'.' '.'\\3',
'',
'',
+ '',
'<html>',
);
$html = preg_replace($html_search, $html_replace, trim($html));
--
Gitblit v1.9.1