Please sign in to comment.
Read marks as unicode
This is the meat of Christian's all-unicode patch. It comes after the templates, but it must be pulled forward, because it is impossible to tinker with the templates using the available mark corpii, which contain international tags and titles. Clearly it's a bug: we write marks using UTF-8 writer, so we must read them using UTF-8 reader. The current code miraclously works by using safestr(), but the dictionaries of templates mix everything up, thus making the problem apparent.
- Loading branch information...
Showing with 13 additions and 5 deletions.