completion list duplication

Hello I'm getting a strange completion list duplication issue:

rec21-10-24_11-30

This is my config on Emacs 29

(use-package mct :straight (:host gitlab :repo "protesilaos/mct" :files ("*.el"))
  :config
  (setq mct-hide-completion-mode-line t)
  (setq mct-minimum-input 2)
  (setq enable-recursive-minibuffers t)
  (setq mct-apply-completion-stripes t)

  (mct-mode 1)

Thank you @dinkonin for reporting this! It must be a regression from what we pushed yesterday with the help of @phikal. I will investigate it right away.

added bug label

This seems to fix the immediate issue:

 mct.el | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mct.el b/mct.el
index fe7f25d..b1617e2 100644
--- a/mct.el
+++ b/mct.el
@@ -256,7 +256,6 @@ (defun mct--clean-completions ()
 (defun mct--fit-completions-window ()
   "Fit Completions' buffer to its window."
   (setq-local window-resize-pixelwise t)
-  (select-window (mct--get-completion-window))
   (fit-window-to-buffer (mct--get-completion-window)
                         (floor (frame-height) 2) 1))
 
@@ -376,11 +375,12 @@ (defun mct--get-completion-window ()
 
 (defun mct--show-completions ()
   "Show the completions' buffer."
+  (minibuffer-hide-completions)
   (let ((display-buffer-alist
          (cons (cons mct-completion-windows-regexp mct-display-buffer-action)
                display-buffer-alist)))
     (save-excursion (minibuffer-completion-help)))
-  (fit-window-to-buffer (mct--get-completion-window)))
+  (mct--fit-completions-window))
 
 ;;;###autoload
 (defun mct-focus-mini-or-completions ()

But maybe there are side-effects. One I noticed is that depending on how fast I type, I get this error:

Error running timer ‘mct--live-completions’: (error "#<window 347> is not a live window")

This happens when I type quickly C-x C-f and then, say, ~/D for "Documents", "Downloads", "Desktop". I type that slowly, so that the Completions' buffer shows up, then there is no error.

"Protesilaos Stavrou (@protesilaos)" gitlab@mg.gitlab.com writes:

But maybe there are side-effects. One I noticed is that depending on how fast I type, I get this error:

Error running timer ‘mct--live-completions’: (error "#<window 347> is not a live window")

This happens when I type quickly C-x C-f and then, say, ~/D for "Documents", "Downloads", "Desktop". I type that slowly, so that the Completions' buffer shows up, then there is no error.

I cannot replicate this issue either.

...

-- Philip Kaludercic

mentioned in commit 51c1e17e

Philip just sent a patch which should fix this issue:

 mct.el | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mct.el b/mct.el
index fe7f25d..e0651bf 100644
--- a/mct.el
+++ b/mct.el
@@ -121,7 +121,7 @@ (defcustom mct-completion-passlist nil
   :group 'mct)
 
 (defcustom mct-display-buffer-action
-  '((display-buffer-at-bottom))
+  '((display-buffer-reuse-window display-buffer-at-bottom))
   "The action used to display the Completions' buffer.
 
 The value has the form (FUNCTION . ALIST), where FUNCTIONS is

[ I even have this in my dotemacs---should have known better. ]

At any rate, the error with the timer still exists. We shall consider that separately and I'm closing this now.

Thanks again @dinkonin for reporting this and sorry for the bad first experience :)

closed

I can confirm the patch works, thank you both. Looking forward to the evolution of mct.

At any rate, the error with the timer still exists. We shall consider that separately and I'm closing this now.

Strangely I cannot reproduce the timer error on my side.

I can confirm the patch works

Very well!

Looking forward to the evolution of mct.

Feel welcome to suggest any ideas or send patches (if I remember correctly, you have signed the FSF papers). These are early days so now is the right time to make any big changes in the overall direction.

Strangely I cannot reproduce the timer error on my side.

Maybe I misconfigured something else on my end. Will continue testing.

Just to confirm that the error with the timer no longer exists. I must have accidentally evaluated something... Restarting Emacs fixed things.