bug#4839: 23.1.50; *Buffer List* -- incorrect handling of multibyte characters?

[Štěpán Němec]

On Mon, Nov 02, 2009 at 04:40:19PM +0100, Andreas Schwab wrote:
> Štěpán Němec <stepnem@gmail.com> writes:
> 
> > On Mon, Nov 02, 2009 at 04:08:16AM -0500, Glenn Morris wrote:
> >> stepnem@gmail.com wrote:
> >> 
> >> > Here is a part of my *Buffer List*:
> >> >
> >> >     desktop-menu.el      20228  Emacs-Lisp        
> >> > ~/.emacs.d/load/desktop-menu.el
> >> [...]
> >> >     面無人色.org         11681  Org               
> >> > ~/Documents/ling/cn/面無人色.org
> >> >     xhc.org               1833  Org               
> >> > ~/Documents/ling/cn/xhc.org
> >> >     词典.org              2940  Org               
> >> > ~/Documents/ling/cn/词典.org
> >> [...]
> >> > All the `Org' lines, with exception of `xhc.org', are misaligned.
> >> 
> >> When I copy and paste your buffer names, the issue seems to be that
> >> `string-width' returns an incorrect value (eg 13 when the buffer name
> >> actually seems to occupy 12). Although some of these characters seem
> >> to have a non-integer width on my system, so I don't see that it is
> >> possible to get this right. But it still looks more wrong than it
> >> should be (off by > 1 character).
> >
> > Yes -- when I view your quotation from my mail above in Mutt, it's
> > aligned correctly, but as soon as I cut & paste it into an Emacs buffer,
> > it's misaligned again, the exact same way as displayed in the *Buffer 
> > List*...
> 
> If the double width characters are not exactly twice as wide as the
> single width characters there is not much that can be done about that.

Indeed, that's what I just realized, too :-)

Probably not much, apart from making Emacs smarter in choosing fonts that do
align properly (i.e. the widths are multiples of some common base).
I'm not sure how other GUI apps go about that... but I'm guessing there
must be ways to align columns properly even in cases like this one -- after
all, there are other than mono-spaced fonts, so the common base might in
fact not be a necessary condition...?