Re: New jrpc.el JSONRPC library

["Clément Pit-Claudel" <cpitclaudel@gmail.com>]

On 2018-05-20 19:11, João Távora wrote:
>> Doesn't LSP-mode let you use language-specific additions, for modes
>> that require more complex interaction than the default style?
> 
> If your question is about lsp-mode.el, I'm not the best person to ask.

Sorry, I meant LSP. I'm not sure why I added -mode after this.

> That said, as Josh points out in the other thread, eglot.el's plan is
> not to shut the door on these experiments in any way. That hypothetical
> code should go into the major-mode's .el file, hopefully briefly, until
> eventually it ends up in eglot.el as a part of the official spec.

Very nice.

The specific scenario I'm curious about is proof assistants like Coq or F*. In these, the buffer is divided into a number of regions (definitions), each of which has a state: unsent, sent, processed.  The user tells emacs which fragments to send and when, and the server responds as it processes the fragments.

I'd love to be able to move to LSP, instead of having to recode clients every time.

Clément.