so-long.el updates (was Re: 27.0.50: How can I test a buffer-local window-configuration-change-hook in batch mode?)

[Phil Sainty <psainty@orcon.net.nz>]

Thanks for all the input regarding the window change functions.

I've settled on keeping things more or less as I had them when I asked
the question, but I've updated the documentation to clarify that it's
possible for long lines in invisible buffers to cause performance
issues.

I've stuck with a default behaviour of "do nothing until the buffer is
displayed" as I believe that is the safest approach for end users in
general.

https://lists.gnu.org/archive/html/emacs-diffs/2019-11/msg00126.html

If there are no objections, I'll rebase & merge this and the rest of
the scratch/so-long-updates branch sometime soon.
https://lists.gnu.org/archive/html/emacs-diffs/2019-11/msg00119.html

The other notable changes in the branch are to make the unload-feature
support more complete, plus some backwards-compatibility improvements
so that users with old versions installed will not encounter problems
with the new version (which gets this closer to an ELPA release, too).

I've also added all of flymake-mode, flyspell-mode, and flycheck-mode
to so-long-minor-modes, so they will be disabled by default in any
buffer where so-long triggers.  I'm pretty sure that modes which are
liable to annotate the buffer in some way are sensible to switch off
by default.

Once these changes are merged, I'll ask people to do some testing in
their own environments, and I expect there would be adjustments to the
default config resulting from that; and then we could likely make an
ELPA release.


-Phil