[Fedora-i18n-bugs] [Bug 668282] PackageKit yum backend uses incorrect encoding for dynamic category names, makes them show up with '?' characters in KPackageKit

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=668282

--- Comment #41 from Nils Philippsen <nphilipp at redhat.com> 2011-10-25 08:54:35 EDT ---
Here's the pimped version of _to_utf8() I plan to commit:

--- 8< ---
def _to_utf8(txt):
    '''convert practically anything to a utf-8-encoded byte string'''

    # convert to unicode object
    if isinstance(txt, str):
        txt = txt.decode('utf-8', errors='replace')
    if not isinstance(txt, basestring):
        # try to convert non-string objects like exceptions
        try:
            # if txt.__unicode__() exists, or txt.__str__() returns ASCII
            txt = unicode(txt)
        except UnicodeDecodeError:
            # if txt.__str__() exists
            txt = str(txt).decode('utf-8', errors='replace')
        except:
            # no __str__(), __unicode__() methods, use representation
            txt = unicode(repr(txt))

    # return encoded as UTF-8
    return txt.encode('utf-8', errors='replace')
--- >8 ---

I've noticed that the packagekit python module also uses unicode objects
instead of UTF-8 encoded strings, but for the moment I'll assume this is in
order and works.

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.