bug#27559: 26.0.50; [PATCH] Add tests for cl-macs.el

[Alex <agrambot@gmail.com>]

Tino Calancha <tino.calancha@gmail.com> writes:

> Alex <agrambot@gmail.com> writes:
>
>> That's more or less what I did at first, but I figured since cl-loop is
>> a large macro, that it's best to quote the macro so that an accidental
>> error doesn't halt testing.
> I prefer just use (eval '...) in those cases where the expansion fails.
> That way the file is more readable, and a reader focus her eye in those
> failing expansions.

I agree that it's more readable, and if I was doing this for my own
project I'd just leave the (eval '...) out where possible.

> Maybe we could add a comment in the file saying that those
> `should-error' are failing expansions to be fixed.

Part of the problem is that the "failing expansions that should be fixed" aren't
wrapped in a should-error, but are wrapped in a should. The test itself
is marked as ":expected-result :failed".

>> It would be nice if the ert tests in general could just continue on with
>> testing even if there are macro-expansion errors.
> Sure, as long as the errors are expected.

I think any macro-expansion errors inside of a
should-error/should/should-not should be similarly to run-time errors.

I also have a patch that appears to do this, but it breaks a few tests
due to Bug#24402[1]. Perhaps a fix to that bug will also make it so that
macro-expansion errors will be caught be should*.

I hope that Bug#24402 can be fixed before committing these tests, so we
can just remove the (eval '...) noise altogether.

Footnotes: 
[1]  https://debbugs.gnu.org/cgi/bugreport.cgi?bug=24402