Phillip Lord <phillip.lord@newcastle.ac.uk> writes:

> I've just had a bug report for pabbrev. The root cause is a change in
> the implementation of cl-subseq. 
>
> In Emacs 24:
>
> (cl-subseq '() 10 10) ;; nil
>
> While in Emacs 25
>
> In Emacs 25 `cl-subseq' has been redefined in terms of seq-subseq.
>
> (seq-subseq '() 10 10) ;; errors
> (cl-subseq '() 10 10) ;; errors
>
> Which is a reasonably substantial change in the interface of cl-subseq.
> The actual (exceptional) behaviour of cl-subseq is not documented, so
> it's reasonable, but perhaps not sensible. It's easy to fix in pabbrev
> (and anywhere) in a way which does not require me to probe for Emacs
> versions.
>
> I offer no opinions, just wanted to check whether this was intended.

Hi,

The change was apparently intended, and was done in commit
253d44bd27b7d90b614b6b968a3b125eeb0a48f2

Cheers,
Nico
-- 
Nicolas Petton
http://nicolas-petton.fr