TALInterpreter uses str() when it shouldn't, leads to unicode error

[Bug Importer]

When using an i18n:name directive on a tag that gets unicode content (with non-ascii chars) using tal:replace, the TAL machinery barfs because it tries to call str() on the content. Example:

> <span tal:replace="here/something_that_returns_nonascii_unicode" i18n:name="somename" />

The offending str() call seems to be in TAL/TALInterpreter.py, line 520:

> if not structure:
> value = cgi.escape(str(value))

Not entirely sure about my fix, but if I change this to the following it does work:

> if not structure:
> if type(value) not in [str, unicode]:
> value = str(value)
> value = cgi.escape(value)

Note that I don't have any PlacelessTranslationService or anything installed.

[Andres Herrera (andresh)]

Hi, continues even with the problems in the newer versions?

[Hanno Schlichting (hannosch)]

Zope2 uses the zope.tal implementation since 2.10 and uses unicode for page templates ever since.