August 31^st, 2006

I had to revise my post titled “UTF8-encoded Unicode support“, because I found out that django’s unicode support has it’s own problems. Some are, of course, connected to the character-set handling of their database code, but generally they currently handle all strings as binary, assuming that everything works within the DEFAULT_ENCODING setting.

That way, foreign character sets can break functions (#1355) in django and even worse, you can’t hook up a legacy database that uses a different character-set without patching the database driver. At least, current work on SQLAlchemy integration might make that less of a concern. I’m just hoping that django 1.0 will include full unicode support. For more information read the update on the old post.

Update (02/15/2008)

Django has come a long way since this post. An update can be found here.

or, more appropiately: “The Administration of Crooks and Liars“, compiled by Mother Jones Magazine. A must read!