> 2.6.2 is a realistic target.

I tried to your patch 03-properties.diff, then I found two problems.

1. browsers autodetection make a mistake.

<dtml-var "u' '"> makes some web browsers confuse, it ignores meta tag
charset and autodetect UTF8.So I will remove <dtml-var "u' '">, when
REQUEST has another charset value.

2. "management_page_charset_tag" really needs?

Because "management_page_charset" is approach of not using unicode.
But management_page_charset_tag exists, REQUEST will convert strings to
unicode strings.field2string has not convert_unicode method, then raise
unicodeerror. It is contradiction.

HTTPRequest.py line:499
if flags&CONVERTED:
        if character_encoding:
            # We have a string with a specified character encoding.
            # This gets passed to the converter either as unicode, if it can
            # handle it, or crunched back down to latin-1 if it can not.
            item = unicode(item,character_encoding)
            if hasattr(converter,'convert_unicode'):
                item = converter.convert_unicode(item)
                item = converter(item.encode('latin-1'))

I made a patch for fix those problems. It looks good work on my zope.


Yusei Tahara                    "So it goes"

Attachment: properties.diff
Description: Binary data

Reply via email to