On Thu, Aug 29, 2002 at 05:57:05AM +0100, Peter Karlsson wrote: > Tomohiro KUBOTA: > > > If copypage.pl is used when conversion from *.wml to *.html , it is > > just opposite to my intension. > > copypage.pl is used to initiate the translation of a page. When it is > run, the English page is copied over to your translation, adding a > translation-check header, creating any required directories and placing > Makefile. And now it will also recode any non-ASCII characters into > entities if the target encoding is not ISO 8859-1. > > The problem is of course proper handling of diff files, but if they can > be loaded using ISO 8859-1 encoding, all should be well.
I did not think about this before, it is certainly very painful; the hack with copypage.pl only helps with initial translation. For this reason, I now fully support Kubota-san initiative, English page should only contain ASCII characters. If we really want iso-8859-1 output, we could recode it when compiling. > This problem would of course go away if we would switch to using UTF-8 > everywhere, but I guess that will never happen... But Kubota-san did commit an utf-8 file, and you broke it ;) Denis
