Re: [Nmh-workers] Fixing decoding of quoted messages in replies

[Valdis . Kletnieks]

On Sat, 06 Nov 2010 11:08:52 +0700, Robert Elz said:
>     Date:        Fri, 05 Nov 2010 21:27:33 +0000
>     From:        Peter Maydell <address@hidden>
>     Message-ID:  <address@hidden>
> 
>   | [Aside: 'columns' is
>   | usually the same as 'characters' except that some Chinese/Japanese
>   | characters are two columns wide on screen.
> 
> Other languages have characters that use no horizontal space on the
> screen at all (they go above, or below, the previous (or potentially,
> I suppose, next, though I have no personal experience with that)
> character).
> 
> Of course, once you start moving the code in this direction, you should
> also start to think about r to l scripts instead of assuming l to r is
> correct for everyone.

Another thing to consider is the increasing difficulty in finding really
good monospaced fonts, especially non-ascii ones, and xterm-ish things
that handle them well (documenting this stuff would probably be useful,
if anybody is in the mood).

For the record, exmh hasn't fully addressed these issues either, and that's
*with* the added advantage of having Tk doing all the canvas management -
it sort of blindly assumes that Tk/Tcl supports it and it can just pretend the
world is left-to-right, top-to-bottom.  There's really no sane way to deal with
bidirectional and vertical scripts unless you throw away the xterm paradigm
and go to a fully functional 2-d canvas.

pgpMySjOBf_h0.pgp
Description: PGP signature