On Tue, Sep 24, 2002 at 01:43:09AM +0200, Erich Schubert wrote: > > > That Ã* looks very much like the unicode representation of Ü. > > When you write unicode, do you mean an UTF-8 encoded version of ISO > > 10646? It is not the same as unicode. > Well, i did. > But it isn't. > It's broken... I can see it here, too. > "rm /" gives me (as user) > rm: Verzeichnis (schreibgeschà 1/4 tzt) »/« entfernen? > Where "à 1/4 " is some "A<space>1<slash>4<space>"... > UTF-8 representations of latin-1 non-ascii characters are usually just > two characters; so there was an additional conversion at work; > converting the "1/4" single-character to a string... > A short look at the /usr/share/locale/de/LC_MESSAGES/coreutils.mo > file gives me the impression that the german locales files are broken > for coreutils - some characters are UTF-8, some are latin-1. > The file contains "schreibgeschÃŒtz" and "durchzuführen" > where "ü" and "ÃŒ" are the latin-1 and UTF-8 representations of the same > character. Probably an artifact of the merging done upstream to create the coreutils po files -- given that the first merge attempt resulted in fuzzy translations where they should not have been, it doesn't surprise me that no encoding verification was done, either. Steve Langasek postmodern programmer
Attachment:
pgptdXQOVDFwN.pgp
Description: PGP signature