Re: Unicode collation for readline completion in bash

[Linda Walsh]

Rob Hoelz wrote:
> Hello,
>
> I have an idea for a new feature for bash/readline, and I wanted to get
> the community's feedback before I go about implementing it.
---
        Have you looked at the Unicode collation algorithm?
        They have one published with each released.


        You say you want to ignore diacritics...ok... but what order would you 
put
them in?

        How about upper/lower case?  Unicode has a specified ordering that that 
most
OS implementations don't follow.  Bash doesn't follow the Unicode ordering 
saying, well
if POSIX doesn't say I have to, then I won't.  POSIX is a mostly provincial 
standard based
on ascii standard C -- with their official locale being called the C-locale 
until recently,
when they realize how provincial it made their standard look.   A standard that 
only
supports 127 (or less if the anti-invisi-char freaks get their way) chars?

        I would see no problem with implementing a switch that toggles ignoring
diacritics on or off... MS has a similar option in its OS search function -- for
searches, "should" it ignore diacritics or not.


        But I'd also like to see unicode ordering, as it's yet another worldwide
standard that the US isn't participating in (we got POSIX and the antiquated 
English
measurement system...)...

http://unicode.org/reports/tr10/ - Unicode Collation Algorithm, Unicode 
Technical standard #10


http://unicode.org/charts/uca/  - charts to show you the effect of the 
algorithm...



Note A-Z occur before a-z... unlike in most implementations of 
en_US.UTF-8...where they
are intermixed...