From: Martin Mares <mj@ucw.cz>
Date: Fri, 18 Apr 2014 10:44:53 +0000 (+0200)
Subject: UCW::CGI: Let url_param_escape() encode non-ASCII characters
X-Git-Tag: v6.0~53
X-Git-Url: http://mj.ucw.cz/gitweb/?a=commitdiff_plain;h=77c84c65dc22afe21bc52d95cd3996b8e1c7d065;p=libucw.git

UCW::CGI: Let url_param_escape() encode non-ASCII characters

RFC 3986 (generic URI syntax) explicitly says this must be done
and some URI parsers choke on them.

Anyway, I am keeping bare url_escape() as is for the time being,
because it's used for production of user-friendly URLs, too. Needs
cleanup later.
---

diff --git a/ucw/perl/UCW/CGI.pm b/ucw/perl/UCW/CGI.pm
index f81bd3ea..a18ee8ea 100644
--- a/ucw/perl/UCW/CGI.pm
+++ b/ucw/perl/UCW/CGI.pm
@@ -50,8 +50,10 @@ sub url_deescape($) {
 sub url_param_escape($) {
 	my $x = shift @_;
 	defined $x or return;
-	$x = url_escape($x);
+	utf8::encode($x) if $utf8_mode;
+	$x =~ s/([^-\$_.!*'(),0-9A-Za-z])/"%".unpack('H2',$1)/ge;
 	$x =~ s/%20/+/g;
+	utf8::decode($x) if $utf8_mode;
 	return $x;
 }