@yilin.wei10 I don’t understand what you mean by “breaks the diagnostic information”? (equal 'foo #'foo) is t. wrt add-hook, is here an example from racket-xp-mode. It does: (add-hook 'completion-at-point-functions #'racket-xp-complete-at-point t t) In a buffer with racket-xp-mode enabled, C-h v and enter completion-at-point-functions it shows: completion-at-point-functions is a variable defined in minibuffer.el.gz. Value in #<buffer foo.rkt> (racket-xp-complete-at-point) where racket-xp-complete-at-point is clickable and leads to a help page for that variable. (I checked this both byte-compiling racket-xp.el and not.)

Yes @greg, you’re absolutely right wrt to add-hook. I’m not sure what I was doing earlier because I was messing in the REPL for an associated task and managed to get this…

(#f(compiled-function () #<bytecode 0xc9d02d>))

Inside of my hooks.

@greg btw, just to let you know I haven’t forgotten about the open PR - I’ll probably tidy it up in the next few days, because I’ve been dog-fooding it for long enough.

Interesting. idk. Maybe this can happen when a lambda gets byte-compiled?

@yilin.wei10 Sounds good. I felt bad that I hadn’t looked at in awhile. I got busy around the US holidays, then focused on other things, since. But if you update the PR and nudge me again I’ll take another look!