bug#61726: [PATCH] Eglot: Support positionEncoding capability

[Augusto Stoffel]

On Thu, 23 Feb 2023 at 11:32, João Távora wrote:

> So it's nice that this new capability popped up. But, as far as I
> understand, the only benefit of leveraging it is for better
> efficiency.  Right? Or are we risking incompatibility with some
> servers until we implement support for it? Please confirm this,
> Augusto.

There is no real risk in not implementing this.

I don't know how many servers out there are nonconforming, because, in
practice, the problem is very rare and basically will only appear if the
user is operating on a line containing Emoji or uncommon math
characters.  So there may well be a lot of nonconforming server out
there but we don't see the consequence of that very often.

Digestif (which is not very important) is nonconforming because I didn't
want to implement such a dumb spec.

IIUC correctly, clangd was the group that pushed for this new LSP spec,
but being a such a big project they surely support the official way of
counting.

Anyway, more than efficiency, this to me is an aesthetic question.  The
UTF-16 way of counting of the original LSP spec is a totally misguided
idea and should be avoided.