This is an announcement for those who use Perlmonks' “Tickers”, pages that export site content in an xml format intended to be parsed by scripts, not shown in a browser. One of these is the chatterbox ticker that most external chatterbox clients use.

You probably know that posts and most other text content in the site are stored in cp-1252 encoding (aka windows-1252). Previously this used to be iso-8859-1 (aka latin-1), but as cp-1252 is mostly a conservative extension of this (not counting some rarely used control characters), many browsers already assumed cp-1252 encoding, and in fact some post titles have used the extra characters before the change. (You can use ampersand escapes in posts to display characters that are not in this character set.) When the change happened, all previous content was retroactively considered cp-1252 instead of iso-8859-1, so as the perlmonks server sees, nothink in the actual contents were changed, only some headers that name the character set needed to be changed. See the Tidings entry at Jul 31, 2008 about this change.

There used to be a bug about this change though, namely the XML declaration in the content served by the XML tickers still claimed iso-8859-1. (The http headers said cp-1252.) This was now fixed, so most of these ticker headers should say cp-1252.

While this change makes the ticker content more correct, and might magically make ticker clients work with cp-1252 characters, it can also break some clients that assume all the content is iso-8859-1. So, if you have any client that uses these tickers, please make sure they can handle such content. For example, if you have a chatterbox client, type things like “–…” in the chatterbox (directly, not using html escapes), and make sure it does not crash your client, and that it displays them correctly. Sorry for the possibly incompatible change and thanks for your time.