Writing large integration tests for core Emacs

[Matt Armstrong]

Stefan Kangas <stefan@marxist.se> writes:

> Matt Armstrong <matt@rfc20.org> writes:
>
>> I did a manual test.  I edited a file, then changed the file outside
>> emacs, then ran revert-buffer.  The file's lock file was still removed,
>> even with the patch below applied.
>
> Perhaps we should write some automatic tests before changing anything in
> this area?

This is in-reply-to:
<CADwFkmmKSsXD20USQ=Fyvk8cXrk=+3Z+SvgUt8JEGfk+CbzsOA@mail.gmail.com>
from Steve (i.e. bug#46701) but the topic is broader than the bug.

In the context of that bug I thought about writing tests but saw little
precedent.  Adding unit tests for simple functions is one thing, but in
my experience larger integration tests have a whole different cost
vs. value proposition.

Do we have other examples of tests that launch emacs as a subprocess of
itself, manipulate files outside of Emacs' knowledge, etc.?