Re: Problem with unicode and bash line re-edit.

[Pierre Gaston]

On Mon, Sep 7, 2009 at 2:27 PM, Linda Walsh <bash@tlinx.org> wrote:

> I was wanting to enter a colon in a file name that had to be readable on
> both Unix and Dos.
>
> Unicode has a 'display' colon: At U+FF1A, which works find -- I can enter
> it
> in bash in linux -- it displays as a colon on linux and on windows as part
> of the filename.
>
> I can either cut/paste from the Win Charmap, OR hold down ALT and type
> 65306
> on the number pad.  There may be a way to enter the hex directly, but I
> forget the method off hand.
>
> Anyway, the problem comes on line-re-editing.  Even though the displayed
> character takes up 1 character just as a colon would, re-editing the line
> to
> change it is really a mess.  It would be a WAG, but maybe bash is counting
> bytes and using those as 'logical character widths), when in this case, it
> needs to do UTF-8 decoding when the terminal (and languages) are set to
> UTF-8, and only move the cursor by the appropriate width.
>
> Any plans to move toward UTF-8 support, since that seems the way the web is
> going?
>
> Thanks,
> -linda
>
> I don't remember having problems with utf-8 chars and readline,
I don't notice something strange trying echo and your char, using
gnome-terminal, a utf-8 locale
and bash 3.2.25 with a centos 5 distribution.

Consider using bash-bug that reports useful things like  your operating
system,
the bash version used etc...

Your locale settings and terminal used can be useful too.
Also a precise description of the steps to reproduce the problem, as well as
a description
of the problem can help.