Re: master e21a1da: Allow completion styles to adjust completion metadata

["João Távora" <joaotavora@gmail.com>]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I assumed it is.  Do you know that it isn't?  nil?
> I think it can be nil, yes.

I just discovered this in the beginning of completion--nth-completion.
Doesn't it quash your concerns?

(defun completion--nth-completion (n string table pred point metadata)
  "Call the Nth method of completion styles."
  (unless metadata
    (setq metadata
          (completion-metadata (substring string 0 point) table pred)))

          
>>> BTW, I just changed the bootstrap so that cl-defgeneric should now be
>>> usable in minibuffer.el.
>> Out of energy right now.  Be my guest, the switch to cl-generic is
>> trivally outlined in the now-revert commit.
> I was about to do that when I saw the issue with `metadata` being
> potentially nil.

See above.  Also, I don't see the change to the bootstrap. Maybe you
forgot to push?

João