bug#73279: [PATCH] Eglot: use immediate-mode eglot--request for :completionItem/resolve

["Brennan Vincent" <brennan@umanwizard.com>]

Hi João,

Just in case you're not familiar with Rust, here are the detailed repro
instructions:

1. Install rust, cargo, and rust-analyzer: https://rustup.rs
2. Create a new project: cargo new my-project
3. Paste the repro recipe from my patch commit message into
my-project/src/main.rs

As for the regression: yes, it's a regression from 29.4, and in fact I
have bisected to the commit that regressed it:
a8c1559a663d9bc21776e33303251e244f86f0d7

So this is broken for me in both 30 and 31.

João Távora <joaotavora@gmail.com> writes:

> On Mon, Sep 16, 2024 at 12:28 PM Eli Zaretskii <eliz@gnu.org> wrote:
>>
>> > From: Brennan Vincent <brennan@umanwizard.com>
>> > Date: Sun, 15 Sep 2024 15:22:05 -0400
>
>> João, any comments?
>
> Yes, I have yet to try the reproduction recipe of this alleged problem
> but one thing stands out.  Brennan, you wrote "as  it did in Emacs 29.4".
> So this is a recent regression of sorts, is that what you're saying?  I don't
> think that code has changed much over the past years.
>
> I also don't recognize the timing you're describing, but I'll comment
> on that after I've done my investigation.  Usually there's a low
> chance a patch will easily go through in this very sensitive area
> of the code.   But a low change is nonetheless non-0 :-)  Maybe
> Dmitry can also have a look, as he's been tinkering with Eglot
> completion lately.
>
> João Távora