bug#66782: 29.1; ERT tests report test redefined depending on loading sequence

[Eli Zaretskii]

> From: Xiyue Deng <manphiz@gmail.com>
> Cc: Mattias Engdegård <mattiase@acm.org>,
>   66782@debbugs.gnu.org
> Date: Sat, 28 Oct 2023 02:39:46 -0700
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > If test2 requires test1, why are both of them explicitly run from the
> > command line?  Isn't that redundant, since running test2 will also run
> > the tests defined by test1?
> 
> IIUC most projects just run all tests without explicit dependency
> detection.  For example, in lsp-mode it uses eask ert-runner to run all
> tests[1].

OK, but why do that in a single Emacs session?  That is unclean for
much more important reasons: some left-overs from previous tests could
affect the following tests.  IMO, each test should be run in a fresh
Emacs session, i.e. no command line should ever load more than a
single test file.