>>>>> "Martin" == Martin Schwenke <[EMAIL PROTECTED]> writes:
Martin> #("Großjohann" 0 10 (vm-string t vm-charset "iso-8859-1" vm-coding
iso-latin-1))
Martin> Bingo! I'm looking at the "fixed" BBDB (none of the above
Martin> crud). When I hit ":" when I'm looking at a message from
Martin> Kai, it asks if I want to change his name from "Kai
Martin> Großjohann" to "Kai Großjohann". Strange! If I say yes,
Martin> then the above gets put in for his family name...
Martin> If it's relevant, I'm running with:
Martin> (standard-display-european 1) ; 8-bit european characters.
Kai commented that Ted might like to try using multibyte mode instead
of unibyte mode. This got me thinking... although I don't really
understand the whole MULE/multibyte character thing very much...
In my BBDB, the 4th character of Kai's family name is character 223.
In a From: header of a message from him I see character 2271 (= 8 *
256 + 223). Now, I thought that (standard-display-european 1) put me
into unibyte mode. Checking the value of enable-multibyte-characters,
I find that it is t in the Presentation buffer for the mail message,
but nil in the BBDB buffer(s) (file and BBDB mode buffer). Is VM
forcing the mail buffers into multibyte mode? I haven't checked...
In bbdb.el/bbdb-annotate-message-sender, I can make the problem go
away by doing:
(name (string-make-unibyte (bbdb-clean-username (car data))))
Now, in general, this is very obviously the *WRONG* way of doing
things. However, if the BBDB buffer is in unibyte mode then it is
probably a very good thing to do...
Perhaps there should be a check in bbdb-string-trim to check for this
and, if "necessary", do a string-make-unibyte.
By the way, I don't understand why bbdb-clean-username only does a
bbdb-string-trim if the name is an (e-mail) address. Shouldn't it be
unconditional?
peace & happiness,
martin
_______________________________________________
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/bbdb-info
BBDB Home Page: http://bbdb.sourceforge.net/
