[geeklog-devel] Feed handling of Danish

[Michael Jervis]

Dirk,

Checking back in on this:
http://www.geeklog.net/forum/viewtopic.php?showtopic=82353

I'm confused!

In 1.611 of lib-common.php you committed support for UTF-8 feeds.
COM_rdfimport checks the charset of a Geeklog instance:

        $charset = COM_getCharset();

        // format articles for display
        $readmax = min($maxheadlines, count($feed->articles));
        for ($i = 0; $i < $readmax; $i++) {
            if (empty($feed->articles[$i]['title'])) {
                $feed->articles[$i]['title'] = $LANG21[61];
            }

            if ($charset == 'utf-8') {
                $title = $feed->articles[$i]['title'];
            } else {
                $title = utf8_decode($feed->articles[$i]['title']);
            }

So, we get the charset of a Geeklog instance,and if it's UTF-8 we do
nothing to the title, if it's not, we utf8_decode the title.

The feed in question is iso-8859-1 and my Geeklog instance is
iso-8859-1. So in this case we're UTF8 decoding something that isn't
UTF-8 and loosing the special characters.

I believe the correct option is to utf8_decode if and only if the feed
is UTF-8 and the site is not. But I'm confused due to the deliberate
implementation of this logic.

Opinions?

-- 
Michael Jervis
mjervis at gmail.com
504B03041400000008008F846431E3543A820800000006000000060000007765
62676F642B4F4D4ACF4F0100504B010214001400000008008F846431E3543A82
0800000006000000060000000000000000002000000000000000776562676F64
504B05060000000001000100340000002C0000000000