bug#64784: 30.0.50; Eglot: Lisp error: (wrong-type-argument number-or-marker-p return) in eglot--post-self-insert-hook

[Eli Zaretskii]

> From: Tassilo Horn <thorn@fastmail.fm>
> Date: Sat, 22 Jul 2023 07:06:58 +0200
> 
> 
> I'm in a rust-ts-mode buffer and ran eglot.  When I hit RET (bound to
> newline), I get the following error:
> 
> Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p return)
>   eglot--post-self-insert-hook()
>   self-insert-command(1)
>   newline(nil 1)
>   funcall-interactively(newline nil 1)
>   call-interactively(newline nil nil)
>   command-execute(newline)

This is because eglot--post-self-insert-hook assumes last-input-event
is a character:

 (defun eglot--post-self-insert-hook ()
   "Set `eglot--last-inserted-char', maybe call on-type-formatting."
   (setq eglot--last-inserted-char last-input-event)
   (let ((ot-provider (eglot--server-capable :documentOnTypeFormattingProvider))
         ;; transform carriage return into line-feed
         (adjusted-ie (if (= last-input-event 13) 10 last-input-event)))
                          ^^^^^^^^^^^^^^^^^^^^^^^

But in this case we get the symbol 'newline' as the value.

The easy way out is to use characterp before comparing with =, but I
actually think there's a deeper problem here, because the code which
uses this post-self-insert-hook _wants_ to process inserted newlines.
So I think a better fix here would be to use last-command-event, not
last-input-event.  All the other post-self-insert-hook implementations
use that, AFAICT.

Stefan, am I right?  Can last-command-event, when accessed from
post-self-insert-hook, ever be something other than a character?