"Liliana Marie Prikler" <liliana.prikler@gmail.com> writes:

> We still only have 97 tests squashed to 5 cases IIUC.

Yes.

> There's some 1000 files in the native-lisp directory.  What I was
> actually hoping for is more or less one test per feature.

Yes, I agree that that would be useful, but doing so was (is) more than
what I was (am) able to do in the time I was (am) able to commit (at
present).

I believe it would be valuable to have such an exhaustive test included
as part of the patch submission to prevent future regressions.

> That is not a good way of checking whether it's Guix' emacs or not.

What would be a better way?  Matching against the --prefix value in the
output of emacs-build-description function?

> I propose doing a per-file deftest instead.

I don't understand the connection between a per-file deftest and the
manner in which guix-emacs-or-not is tested, but otherwise agree on the
utility of per-feature deftests.

> Could you do a MIME attachment next time?

Sure.  For now, for what it's worth and in case it helps, please see
attached a copy of the same file as before.


-- 
Suhail