> FWIW, I don't see any major benefit to having "-test" in the test name,
> as the fact that it's a test is immediately obvious in all contexts
> where you see them. So it mostly just makes the test names longer.
For me, who doesn't know the eglot codebase, it is helpful to see,
whether a Lisp object belongs to eglot or eglot-tests.
Prior my changes, there were even the defun `eglot-lsp-abiding-column'
and the ert-deftest `eglot-lsp-abiding-column'. Terrible!
FWIW this doesn't seem "terrible!" to me at all. The way I (and those
who teached me) see Lisp, symbols can have multiple bindings in
different domains. So the same symbol may have a function binding,
a value binding and why not a test binding? A unit-test for testing a
given function may well be named after the function it is testing.
Namespacing is a good alternative to separate domains, sure, but
namespacing in Emacs is... err... not without its challenges.
João