From 834c4ad0bf82f28949b108eb6c957fde3c18e1ce Mon Sep 17 00:00:00 2001 From: Cash Costello Date: Sat, 20 Apr 2013 11:07:44 -0400 Subject: Fixes #5369 allows ! in urls and adds unit tests --- engine/lib/output.php | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-) (limited to 'engine/lib/output.php') diff --git a/engine/lib/output.php b/engine/lib/output.php index c5a04989b..fe5bbcaaf 100644 --- a/engine/lib/output.php +++ b/engine/lib/output.php @@ -13,28 +13,33 @@ * @param string $text The input string * * @return string The output string with formatted links - **/ + */ function parse_urls($text) { + + // URI specification: http://www.ietf.org/rfc/rfc3986.txt + // This varies from the specification in the following ways: + // * Supports non-ascii characters + // * Does not allow parentheses and single quotes + // * Cuts off commas, exclamation points, and periods off as last character + // @todo this causes problems with // must be in format (no space). // By default htmlawed rewrites tags to this format. // if PHP supported conditional negative lookbehinds we could use this: // $r = preg_replace_callback('/(?"\'\!\(\),]+)/i', - // - // we can put , in the list of excluded char but need to keep . because of domain names. - // it is removed in the callback. - $r = preg_replace_callback('/(?"\'\!\(\),]+)/i', + $r = preg_replace_callback('/(?"\'\(\)]+)/i', create_function( '$matches', ' $url = $matches[1]; - $period = \'\'; - if (substr($url, -1, 1) == \'.\') { - $period = \'.\'; - $url = trim($url, \'.\'); + $punc = \'\'; + $last = substr($url, -1, 1); + if (in_array($last, array(".", "!", ","))) { + $punc = $last; + $url = rtrim($url, ".!,"); } $urltext = str_replace("/", "/", $url); - return "$urltext$period"; + return "$urltext$punc"; ' ), $text); -- cgit v1.2.3