bug#57907: 29.0.50; Using keywords with cl-loop

[Lars Ingebrigtsen]

Philip Kaludercic <philipk@posteo.net> writes:

> I phrased that clumsily -- while it is a new language, it doesn't always
> have to feel like one.

And I'm saying that that feeling is a misleading feeling -- we're
leading the users down the garden path.

> E.g. the repeat-collect example I gave before.
> If you *can* use keywords, you *can* make some cl-loop invocations
> appear to be a macro or a function call with keyword-arguments.
>
> But of course this wouldn't be the case, as the macro should stay
> backwards compatible.

I understand that.  But it means that the stated rationale for this
change does the users a disservice.

So I'm against extending cl-loop in this way.  Gerd has explained why
it works this way in Common Lisp (because CL doesn't have any other
choice because that's the way the language works), but we should not
shoehorn in this accident into Emacs, too.