On Dec 21, 2004, at 10:06, Igor Lobanov wrote:
Hmm, problem not with utf8 output. I think it works. Problem that source data are in native charset (it could be not utf8) and must be converted to UTF8 for only task. In other places they must be in native charset...
I would tend to convert all strings to native perl encodings the first time I see them, and convert them back to bytes as late as possible. But then I suffer from pain caused by retrospectively having to bolt proper charset support onto a huge app that previously assumed everything was just latin1. You really don't want to have to do that.
_______________________________________________ templates mailing list [EMAIL PROTECTED] http://lists.template-toolkit.org/mailman/listinfo/templates
