Stefan Monnier <monnier@iro.umontreal.ca> writes:

> IMO, we should have a proper separate pdump test, instead.

I tend to agree with Stefan about the relatively low importance of
supporting overlays owned by buffers in the dump.  I have deferred that
work, though the tests I have added will be useful if someone wants to
take it up in the future.

See attached patch, which:

a) Adds a test that exercises portable dumping from a batch mode
subprocess, using the Emacs executable hosting the test as a basis.

b) Performs the dump while a buffer with overlays is live in the
subprocess.

c) Verifies the dump by running a second subprocess and checking how
many overlays are in the buffer that had overlays.

The test took most of the time.  I'm quite open to suggestions there; my
elisp knowledge is novice level.