Re: Brand new clojure support in Emacs ;-)

["João Távora" <joaotavora@gmail.com>]

"Bozhidar Batsov" <bozhidar@batsov.dev> writes:

>  Sure I got that.  And currently CIDER is strongly coupled to
>  implementation details of clojure-mode.
>
> You say it like it was some design failure,

No, I didn't.  You're again imagining things.  I'm just stating
objective facts which are pretty relevant if one wants to add CIDER
interoperation capability to the new major mode that is the subject of
this thread.

> The APIs that are being used are things like "find the current Clojure
> namespace", sexp-related functions, etc.

Yes, that's what I meant by "the existing informal API between the two
realities: major mode and CIDER overlay minor mode".  Those things you
mention are very similar to the coupling between SLIME/SLY and
lisp-mode.  If someone wanted to make a new Common Lisp mode (say, a
TreeSitter one) and they wanted to use SLY with it, the same problem
would arise.  And yes, it is easily solved.

>  I clearly wrote about integrating with CIDER, not rewriting it.
>  So, again, you're really reading things that I didn't write.
>
> Well, you did speak about RCP and there's nothing at the RCP level
> happening between clojure-mode and CIDER. (all the RCP-related code is
> in CIDER) The APIs that are being used are things like "find the
> current Clojure namespace", sexp-related functions, etc.

What I wrote is that there's nothing transcendental that I can detect
about the architecture of an RPC system for Emacs like CIDER.  There's a
major mode, a collection of minor modes and a reflection backend in the
host language.

> experts". I'm not an expert in LSP and I wouldn't dare to lecture you
> how to do things in eglot, but I do know a bit about Clojure, nREPL
> and all the Emacs packages and I happen to know what we've tried and
> didn't work out. You want to brush this aside, for reasons that are
> still unclear to me.

That's the thing: Noone is lecturing you on anything.  Noone's even
asking you for anything.  If anything it's the other way round, you seem
want us to literally sit through a lecture of yours on the subject.

>  If all this is not a paternalistic attempt at dissuasion, it's at
>  least odd from someone who's already said he's not interested
>  in contributing to this endeavor, and who's not really being
>  asked to anymore.
>
> What "endeavour" is that exactly? Bringing
> clojure-mode/clojure-ts-mode to core or replacing them with 2 lines of
> code?

Writing a new major mode, of course, a direct:

* consequence of Richard's request
* your refusal to integrate clojure-mode or clojure-ts-mode
* Philip's interesting suggestion

It's the top message of this discussion, have you read it?  And yes it's
two lines of code, but you have to start somewhere and that's what I
did.  TreeSitter could also be useful later on, for font-locking, etc...

> And who started the conversation in a hostile manner?  If I didn't
> happen to be subscribed to emacs-devel I wouldn't even know what was
> being discussed here, given how events unfolded. Is it really
> surprising I'd be unhappy about the communication on the topic so far?

What hostile manner?  Do we need permission from you to start discussing
Clojure editing facilities here?

João