Fix parsing invalid HTML messages with BOM after <!DOCTYPE> (#1490291)

11 years ago · 759566fe99
2 changed files with 12 additions and 4 deletions
--- a/1
+++ b/1
@ -19,6 +19,7 @@ CHANGELOG Roundcube Webmail
 - Fix so "over quota" errors are displayed also in message compose page
 - Fix duplicate entries supression in autocomplete result (#1490290)
 - Fix "Non-static method PEAR::isError() should not be called statically" errors (#1490281)
+- Fix parsing invalid HTML messages with BOM after <!DOCTYPE> (#1490291)

 RELEASE 1.1.0
 -------------
--- a/program/lib/Roundcube/rcube_washtml.php
+++ b/program/lib/Roundcube/rcube_washtml.php
@ -403,16 +403,23 @@ class rcube_washtml
    {
        // 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'.' &nbsp; '.'\\3',
            '',
            '',
+            '',
            '<html>',
        );
        $html = preg_replace($html_search, $html_replace, trim($html));