Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!aramis.rutgers.edu!paul.rutgers.edu!sparky.rutgers.edu!gaynor
From: gaynor@sparky.rutgers.edu (Silver)
Newsgroups: comp.emacs
Subject: Re: Why does completion require an alist?
Message-ID: <Sep.1.03.08.23.1990.10801@sparky.rutgers.edu>
Date: 1 Sep 90 07:08:24 GMT
References: <9008311659.AA14901@STEVE-DALLAS.MIT.EDU>
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 16

bjaspan@ATHENA.MIT.EDU ("Barr3y Jaspan") writes:
> Why does read-completion require an alist instead of an ordinary list of
> strings?  Considering that only the car of each element of the alist is used
> or returned, it doesn't seem to be a sensible limitation.

Just the way things are.  Probably turned out that the predominating case was
one of using completing-read to grab an association from an alist, eg,

  (assoc (completing-read "Well, which one? " some-alist) some-alist)

I probably don't need to say this, but an easy way to use completing read with
a list of strings is to simply mapcar list on the list of strings, as in:

  (completing-read "Well, which one? " (mapcar 'list list-of-strings))

Regards, [Ag]