Re: fix for bug 10994 breaks ido customizations in major way

[Vitalie Spinu]

 >> Le Wang <address@hidden>
 >> on Tue, 7 May 2013 22:47:58 +0800 wrote:

 > Hi Drew,
 > On Tue, May 7, 2013 at 10:44 PM, Drew Adams <address@hidden> wrote:
 >> I don't care much about Ido, so I don't have an opinion about the question 
 >> at
 >> hand.
 >> 
 >> However, I will say that it is NOT at all the case that there is nothing out
 >> there, in the wild, that takes advantage of completion candidates that have 
 >> the
 >> same text (same string chars)

Indeed, it is not difficult to imagine an application when same names
have different faces.

 > We can say with certainty that there is "nothing in the wild" because
 > there is a bug in the latest released Emacs and ido that does not work
 > with duplicate strings.

This is your subjective opinion. The current state is a very reasonable
default, it does *work* (aka cycles) with duplicated strings, and call
this behavior a bug is at least inappropriate.

To have useless properties that are not needed by your application and
rely on IDO to delete those for you is a bad and lazy style. So the
applications that you mention are buggy, not IDO.

Don't want duplicated names, pass through delete-dups before calling
ido-completing-read, or even better, don't collect those properties in
the first place. Not a big deal.

What you propose will force *every* implementation out there to
explicitly deal with duplicated strings. As said before, if they have
duplicated strings they have them for a reason (even if it is simple
laziness). Don't you see that you propose to cut some of the
programmers' and users' flexibility?

    Vitalie