>>110
I was able to trivially remove the minibuffer--require-match hack, the only reason I didn't see it earlier is because I've still yet to commit my self to fully understand the program, or even have a loose understanding for that matter. All that was necessary was declaring a local variable vertico--require-match and having this set based off the advice around completing-read-default and completing-read-multiple. I was hoping this would solve an issue I've noticed with tramp, but unfortunately it does not seem to.
+ (defvar-local vertico--require-match nil
+ "REQUIRE-MATCH parameter of `completing-read'.")
(defun vertico--setup (require-match)
"Setup completion UI."
(setq vertico--input t
vertico--candidates-ov (make-overlay (point-max) (point-max) nil t t)
vertico--count-ov (and vertico-count-format
(make-overlay (point-min) (point-min) nil t t)))
(setq-local resize-mini-windows 'grow-only)
(setq-local max-mini-window-height 1.0)
(setq-local truncate-lines t)
(setq-local completion-auto-help nil)
(setq-local completion-show-inline-help nil)
+ (setq-local vertico--require-match require-match)
(use-local-map vertico-map)
;; Use -90 to ensure that the exhibit hook runs early such that the
;; candidates are available for Consult preview. It works, but besides
;; that I'dont have a specific reason for this particular value.
(add-hook 'post-command-hook #'vertico--exhibit -90 'local))
(defun vertico--advice (orig &rest args)
"Advice for ORIG completion function, receiving ARGS."
- (minibuffer-with-setup-hook #'vertico-setup (apply orig args))
+ (minibuffer-with-setup-hook
+ #'(lambda () (vertico--setup (nth 3 args)))
+ (apply orig args)))
>>113
TRAMP is actually kind of terrible, has it ever worked well? There are some alternatives on EmacsWiki like Sudo Save, or using SSHFS, or Hobo Mode, etc. I may give these another look at a later date. Anyway on that particular bug (which can be worked around by the user) my best guess is that this actually has something to do with my GNU Emacs install. I might try nuking it and see if things improve at all.
I was hoping this would solve an issue I've noticed with TRAMP, but unfortunately it does not seem to.
I might try nuking it and see if things improve at all.
Completion with TRAMP works with icomplete-mode in my version of GNU Emacs, so I assume that it is not the case that my GNU Emacs install is erroneous. I'll attempt to locate the difference in implementation between these two implementations.
>>113-115
The problem with TRAMP is that it thinks its self special, and it will eagerly clobber everything you're doing if it feels like it should. I resolved the issue I was having by removing any special treatment for it. If users want TRAMP to be convenient to use they should write a function using completing-read to establish connections before using find-file. After the connection is established completion in find-file works as expected. Anyway I've been talking to myself for long enough I think. I rebased to master, and here is my port, likely as it will be for the foreseeable future: https://paste.textboard.org/38fd7edf