Thomas Roessler <[EMAIL PROTECTED]>:
> > The first <@> is in iso-8859-1, and the second <@> is in
> > iso-8859-15. I observed that the @ in the second <@> was not
> > converted ...
>
> What's your local character set, what do you mean by "not
> converted", and does your system have a character set definition
> file for iso-8859-15?
My local character set is UTF-8, of course, so the unconverted
character is replaced by the replacement character, which is a bullet
with a small reverse-field question mark in it with my font.
It looks as though there's something odd with my charmaps:
$ grep 40 /usr/share/i18n/charmaps/ISO-8859-15
<At> /x40 <U0040> COMMERCIAL AT
<commercial-at> /x40 <U0040> COMMERCIAL AT
<@> /x40 <U0040> COMMERCIAL AT
<Oa> /x40 <U0040> COMMERCIAL AT
$ grep 40 /usr/share/i18n/charmaps/ISO-8859-1
<At> /x40 <U0040> COMMERCIAL AT
<commercial-at> /x40 <U0040> COMMERCIAL AT
I wanted to reproduce the problem with an unmodified mutt, just in
case I've broken it in the course of my hackery, but "cvs co mutt" is
doing about 1 bps at the moment ...
For the UTF-8 mutt I tore the charset stuff into two parts with an API
between them that is identical to the standard iconv. I didn't do
anything with the code that wasn't close to the tear, so I probably
haven't broken it, but I'll have another look at home and issue a full
confession if it does turn out to be my fault after all.
Edmund
