>>>>> Paul Eggert <eggert@cs.ucla.edu> writes:

> So although it'd be fine to have async-signal-safe autosave as an option, as
> a user I'd like to also retain the option for Emacs to keep doing the same
> old unsafe stuff that it's been doing for decades, at least until after we
> get in-the-field experience with the new approach.

Agreed. There should absolutely be an option to retain the unsafe behavior,
since at the very least it can be useful to have Emacs stick around while
exploring the behavior of a bug in situ. So we won't delete the longjmp code,
we just might change the default behavior to something more consistent with
safer practices.

Also completely agree on the in-the-field experience part. I'm not even sure
we'll change the default behavior right away; we might give it a few releases
to see how it goes, and to hear back from people who've decided to turn it on.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2