emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Drop the Copyright Assignment requirement for Emacs


From: João Távora
Subject: Re: Drop the Copyright Assignment requirement for Emacs
Date: Sat, 9 May 2020 16:52:05 +0100

On Sat, May 9, 2020 at 4:41 PM Dmitry Gutov <address@hidden> wrote:
>
> On 09.05.2020 18:25, João Távora wrote:
> > It's not for discoverability.  It's, among other reasons, so that major 
> > modes
> > can adding to Eglot's hooks, for example.
>
> I'm not sure for what purposes, but hooks are flexible enough for this
> to be possible today already (without Eglot in the core).

Yes, that works.  But Eglot doesn't have only these kinds of interfaces.  It
also has generic functions.  I guess we could have just a "lsp-interface.el"
in the core that defines the generic functions and no implementations.

But in general, that's a fair stance, to have little interdependence and
suitable indirections betweenmodules.

> More generally, I oppose collecting more and more code inside Emacs.
> Lots of features can live just as well as packages.

That's true.  Some things are desirable in the core though.  In my view
a proper completion tooltip that lives in the core and uses capf
exclusively is a nice thing to have.  I don't want to M-x package-install
completion-thingy.

> > Or that someday font-locking and
> > or indentation can be done via the LSP server.
> That's the goal of TreeSitter, isn't it? Or are there some new additions
> to the LSP protocol I haven't heard about?

I think so yes.

> But even the usual argument to have stuff in the core ("what if I don't
> have Internet?") doesn't work for Eglot, considering it needs to
> download external programs anyway (or have the user download them).

True, to a point.  But the user could have those programs already,
or use Eglot to connect through the network. But that argument sucks,
I agree.  I do think once something is in the core it's more
discoverable/taken more seriously.  At least until we start bundling
packages.

In other words, we could be farther from that ideal of modularity, but
we're just not there yet.

João



reply via email to

[Prev in Thread] Current Thread [Next in Thread]