bug#61479: 28.2; icomplete-in-buffer --- Error in post-command-hook: args-out-of-range

[Fernando de Morais <fernandodemorais.jf@gmail.com>]

Hi Juri,

Juri Linkov <juri@linkov.net> writes:

>>> Does it also select a completion candidate?
>>
>> Well, when I hit C-j, it selects the candidate, but then:
>>
>>   - The buffer is modified in a way that only the selected candidate is
>>     visible;
>>   - The same is true for multiple C-M-i when in the candidate list;
>>   - I need to use `undo' to go back to the previous state of the buffer.
>
> Here is a short patch that completely fixes this problem.
> But probably it's too late for the release branch
> since icomplete never worked correctly in the buffer.
>
> diff --git a/lisp/icomplete.el b/lisp/icomplete.el
> index f7a91599f3b..aadc2e32787 100644
> --- a/lisp/icomplete.el
> +++ b/lisp/icomplete.el
> @@ -215,7 +215,11 @@ icomplete-force-complete-and-exit
>         ;; calculated, This causes the first cached completion to
>         ;; be taken (i.e. the one that the user sees highlighted)
>         completion-all-sorted-completions)
> -      (minibuffer-force-complete-and-exit)
> +      (if (window-minibuffer-p)
> +          (minibuffer-force-complete-and-exit)
> +        (minibuffer-force-complete (icomplete--field-beg)
> +                                   (icomplete--field-end)
> +                                   'dont-cycle))
>      ;; Otherwise take the faster route...
>      (minibuffer-complete-and-exit)))

This patch fixes the problem when we hit C-j to select a candidate, but
the problem persists if we type multiple C-M-i when in the candidate
list, unfortunately.

At least, now, `icomplete' in buffer is very usable.  😀

Thank you!

-- 
Regards,
Fernando de Morais.