Re: Getting rid of prog-indentation-context

[Stefan Monnier]

>> >> Could we just mark them as obsolete in Emacs-26?
>> > Doesn't sound like a good idea to me: something that was introduced
>> > recently and barely had any chance to be used doesn't fall under the
>> > "obsolete" label for me.
>> Eli, I fail to understand what you expect prog-indentation-context will
>> be used for.
> Stefan, you admitted this stuff into Emacs 2 years ago, so why are you
> asking me these questions?

Because you clearly think it's important to have this new variable in
Emacs-26.  The important word in my question is "will".
I don't expect an answer like "it will be used for mmm-support", but
rather, which specific packages do you expect will make use of it?

> I'm sure you asked that yourself back then, and I'm sure you had
> good answers.

I thought it would be used by updating "all" major modes to obey it,
because I thought that relying on narrowing was unworkable (because of
the usual ambiguities linked to narrowing's intentions).

I now understand that narrowing's ambiguities are resolved by locally
overriding the narrowing so it has a precise meaning.  and I also know
from experience that changing "all" major modes to obey
prog-indentation-context is not going to happen.

So, yes, we can keep prog-indentation-context in Emacs-26, but I still
expect that antlr-mode (the only current package which sets
prog-indentation-context) will sooner or later be changed to use
narrowing instead, so prog-indentation-context will be just a weird wart
whose only effect will be to confuse some people.

Note also antlr-mode will work just as well without Emacs defining
prog-indentation-context (it currently tests (boundp
'prog-indentation-context), but it only ever let-binds it, so the
boundp test is not needed).

>> Here's what I'd like to see happen for Emacs-26:
> How is this different from what began this thread quite some time ago?
> Do we want to start all over again?

I'm not sure how the past discussions have changed people's positions on
those different points.  So I don't want to "start over", but rather
I want to see what do we stand now.

Also past discussion have changed my position to some extent which is
reflected in my above presentation of the issues (e.g. mentioning that
we could keep the first element of prog-indentation-context).


        Stefan