bug#55205: 28.1.50; completion--replace illegally mutates completion candidates

[Daniel Mendler]

> I think it's much better to fix the bug, so I just pushed the patch
> below to `master`.
> 
> I think it's safe enough for `emacs-28`, but I can't claim it's
> "obviously safe", the way I could about that same `copy-sequence` in
> `cl-generic.el`.
>
> diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
> index ef71b4e6be6..fb473cf71b0 100644
> --- a/lisp/minibuffer.el
> +++ b/lisp/minibuffer.el
> @@ -1140,6 +1140,7 @@ completion--replace
>    ;; The properties on `newtext' include things like the
>    ;; `completions-first-difference' face, which we don't want to
>    ;; include upon insertion.
> +  (setq newtext (copy-sequence newtext)) ;Don't modify the arg by 
> side-effect.
>    (if minibuffer-allow-text-properties
>        ;; If we're preserving properties, then just remove the faces
>        ;; and other properties added by the completion machinery.

Thanks, Stefan! This is the proper fix. I also think this is safe enough
for the emacs-28 branch. It will also fix the icomplete bug there, where
minibuffer-force-complete *incorrectly* strips the text properties.

Daniel