From: Antoine Beaupré <anarcat@koumbit.org>
Date: Wed, 10 Sep 2014 03:11:51 +0000 (-0400)
Subject: do not double-decode unicode in CGI forms
X-Git-Tag: 3.20140916~52^2
X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/commitdiff_plain/feb21ebfacb341fc34244e1c9b8557fd81d1dfc1?hp=6057107d71e9944bd6fd7093060e4297e617733e

do not double-decode unicode in CGI forms

this works around a behavior change introduced in Encode.pm 2.53
shipped with the Perl 5.20 release described here:

http://ikiwiki.info/bugs/garbled_non-ascii_characters_in_body_in_web_interface/
---

diff --git a/IkiWiki/CGI.pm b/IkiWiki/CGI.pm
index c0d8f598b..cb83319e6 100644
--- a/IkiWiki/CGI.pm
+++ b/IkiWiki/CGI.pm
@@ -110,11 +110,23 @@ sub decode_cgi_utf8 ($) {
 	}
 }
 
+sub safe_decode_utf8 ($) {
+    my $octets = shift;
+    # call decode_utf8 on >= 5.20 only if it's not already decoded,
+    # otherwise it balks, on < 5.20, always call it
+    if ($] < 5.02 || !Encode::is_utf8($octets)) {
+        return decode_utf8($octets);
+    }
+    else {
+        return $octets;
+    }
+}
+
 sub decode_form_utf8 ($) {
 	if ($] >= 5.01) {
 		my $form = shift;
 		foreach my $f ($form->field) {
-			my @value=map { decode_utf8($_) } $form->field($f);
+			my @value=map { safe_decode_utf8($_) } $form->field($f);
 			$form->field(name  => $f,
 			             value => \@value,
 		                     force => 1,