Re: [elpa] externals/pabbrev 6e3c38e 1/2: Safe access of subsequence.

[Phillip Lord]

Stefan Monnier <address@hidden> writes:

>>     Emacs 25 crashes when the indexes for cl-subseq are out-of-bounds,
>                ^^^^^^^
> Just checking: you mean "signals an error" or "burps", right?  If it
> really crashes, we have a problem.

Sorry, very poor wording indeed. Signals an error, yes.


>>     while previous Emacs silently ignored. Better to handle this here.
>
> This change in cl-subseq was controversial, so if you feel that it
> should be reverted, you might like to file a bug report.

I just asked about this (see "behaviour change in cl-subseq"), and I am
sorry that I missed the controversy previously.

I am inclined to think that the new behaviour is better than the
existing behaviour, although suspect that it may cause a fair few
breakages when Emacs 25 is released (fortunately, Eric Fraga is a
pabbrev user running on Emacs 25 and complained).

As I said to Leo in the other thread, I think the docstrings need
patching. Exceptional behaviour was undocumented before and is
undocumented now. Happy to do this, but will do via a merge-request if
that is okay by you. Then I can get feedback on my patch before it goes
onto trunk.

Phil