GNU bug report logs - #21644
24.4; completing-read acts differently on functional collection

Previous Next

Full log

View this message in rfc822 format

From: Ryan <rct <at> thompsonclan.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>, Oleh Krehel <ohwoeowho <at> gmail.com>
Cc: 21644 <at> debbugs.gnu.org
Subject: bug#21644: 24.4; completing-read acts differently on functional collection
Date: Mon, 12 Oct 2015 18:34:40 -0700

On 10/12/15 6:23 PM, Dmitry Gutov wrote:
> On 10/08/2015 01:13 PM, Oleh Krehel wrote:
>
>> I can reproduce this as well, although "TAB RET" works where "RET"
>> doesn't. I attach a patch with a fix. If someone can check it I can push
>> it. The patch doesn't solve all the intricacies of
>> `completion-ignore-case' that were mentioned in that function's
>> comments, just this specific use case.
>
> Is there really a bug here?
>
> The example code didn't handle `test-completion' (the case when ALL 
> would be `lambda').
>
> With this definition, the behavior is the same in either case:
>
> (defun collection-as-function (collection)
>   "Return a function equivalent to COLLECTION.
>
> The returned function will work equivalently to COLLECTION when
> passed to `all-completions' and `try-completion'."
>   (if (functionp collection)
>       collection
>     ;; Capture collection in a closure
>     (lambda (string pred all)
>       (funcall
>        (cond
>         ((eq all 't)
>          #'all-completions)
>         ((eq all 'lambda)
>          #'test-completion)
>         (t
>          #'try-completion))
>        string collection pred))))
>
>
I wasn't aware of `test-completion'. I was going based on the docstring 
for `completing-read', which only mentions `try-completion' and 
`all-completions'. Is there any other documentation I should have consulted?

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.