Re: completing-read does not display options in new frames

[Platon Pronko]

On 2023-04-30 17:44, MaGelan wrote:
> Hi,
>
> This is the first time I post on newsgroups so please forgive me if I
> am overlooking something and let me know if I break rules.
>
> I have been using Emacs for a year now and I am starting to integrate it
> more and more in my workflow. I am still learning elisp so please bear
> with me because I have a programming issue.
>
> I do not understand how to make the completing-read function work
> properly in a new frame. Whenever I start completing-read in a new frame
> what happens is that I do not see the options right away. I have to
> either press SPC or a input a character and delete it in order to force
> the refresh of the minibuffer and display the options. Here is an
> example that should work on any configuration:
>
>
> (defun test-miniframe-completing-read ()
>      "Test miniframe completing read"
>      (interactive)
>      (with-selected-frame (make-frame '((name . "emacs-miniframe")
>      (minibuffer . only)
>      (width . 120)
>      (height . 33)))
>      (unwind-protect
>          (let ((choices '("first" "second" "third")))
>              (completing-read "Select: " choices))
>          (delete-frame))))
>
> (test-miniframe-completing-read)
>
>
> I am using consult/vertico/orderless framework but my understanding is
> that the completing-read is an independent function. I did notice though
> that the consult-buffer when started with:
>
> emacsclient -c -e '(consult-buffer)'
>
> has the same problem. The buffer selection is not shown right away.
>
> I tried to use C-h f and inspect the function but, apparently, it is a C
> function. So, I do not know how to proceed further here. How do I make
> this code work properly both in new frames and in the existing one?
>
> Thank you very much for any help!

Hi!

Your test function works for me, choices are shown immediately. Can you provide 
more details? (your Emacs version, OS used)

Try starting your emacs in "clean" mode:

emacs -Q -l test.el

(test.el contains the minimized setups code, i.e. relevant parts of init.el and 
test-miniframe-completing-read defun)

This way it will be possible to get a reproducible test case.


I suspect the issue is there's some error or warning, in which case there's a 
message shown in one of minibuffers for couple seconds, and that delays 
completing-read.

--
Best regards,
Platon Pronko
PGP 2A62D77A7A2CB94E