Re: highlighting non-ASCII characters

[Ted Zlatanov]

On Wed, 24 Mar 2010 23:06:48 +0800 Jason Rumney <address@hidden> wrote: 

JR> Yes, within that [social engineering] context suspicious has a clear
JR> enough meaning. But the cases others were discussing seemed more
JR> like cases of mismatched encodings that they wanted to be flagged
JR> more obviously.

All of those use cases are valid, but the social engineering is IMO the
most important reason why we should consider it for first-class
inclusion in Emacs.

On Wed, 24 Mar 2010 17:21:40 +0100 Lennart Borgman <address@hidden> wrote: 

LB> Maybe "Recommended Identifier Profiles for IDN" should be implemented
LB> in Emacs? (See http://www.unicode.org/reports/tr39/data/idnchars.txt)

LB> How about a bool vector (see make-bool-vector) for this?

I don't know if we should be planning the storage mechanism yet, but you
can certainly put that together.  Don't optimize this prematurely, just
use a list at first.

It's important to decide what we're doing with this feature; as you see
there's several use cases and I want to be sure we can cover them all
(hence my proposal with several classes of characters and matching
rules) before jumping into the implementation.  I'm pretty sure we'll at
least need the "confusables" or homoglyphs so maybe that's a good start.

Ted