bug#19774: 25.0.50; Emacs crashes when C-u <N> on a Gnus folder (reproducible)

[Sebastien Vauban]

Óscar Fuentes wrote:
> I can see a problem on `master' with a setup that does not involve Gnus.
>
> Get key-chord.el:
>
> https://github.com/emacsmirror/key-chord/raw/master/key-chord.el
>
> Get guide-key.el:
>
> https://github.com/kai2nenobu/guide-key/raw/master/guide-key.el
>
> Get popwin.el (required by guide-key):
>
> https://github.com/m2ym/popwin-el/raw/master/popwin.el
>
> Put everything on some place reachable from your load-path.
>
> emacs -Q
>
> Eval this code on *scratch*:
>
> (require 'key-chord)
> (key-chord-mode 1)
> (require 'guide-key)
> (key-chord-define-global "x1" 'delete-other-windows)
> (guide-key-mode 1)
>
> Now press `C-u 1'. A message appears:
>
> Error running timer `guide-key/polling-function': (wrong-type-argument
> sequencep #<EMACS BUG: INVALID DATATYPE (PVEC 0xffffffffffffffff) Save
> your buffers immediately and please report this bug>)

Thanks for giving a non-Gnus recipe, which shows that the problem is in
fact more generic...

> After (key-chord-define-global "x1" ...), key-chord will watch for any
> input that consists on "x" or "1" and handle it (it uses
> `input-method-function' for that). When the user inputs `C-u 1', the `1'
> is what triggers the problem.
>
> My guess is that those two packages are messing with the guts of Emacs'
> input processing and when you combine them on the same setup things
> break.

What's "funny" is that I used the same setup for months, and it's since
the last update of key-chord that Gnus began to crash.

Best regards,
  Seb

-- 
Sebastien Vauban