eglot with pyright reports errors on random non-existing variables in buffer

[x0xntor33alx0x]

when running eglot on a python in a python project, sometimes the server complains about random non-existing variables not being defined, here is a showcase of the problem:

steps to reproduce:

1. start emacs
2. open a python file
3. run eglot
4. after a while working in the buffer typing random text then deleting the server starts reporting a non-existing variables not being defined as above

I cannot reproduce the issue using emacs -Q.

transcript: https://gist.github.com/x0xntor33alx0x/e5f6ec874c3a36448f6fa74445a7e8ff

software versions
emacs: GNU Emacs 29.0.92 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.24.20, cairo version 1.16.0) of 2023-06-25 also tested on 28.0.50
eglot: 1.12.29 (builtin)
pyright: 1.1.315. also tested with version 1.1.166
OS: ubuntu 20.04

PS: I've been using pyright with lsp-mode but never had this problem

[joaotavora]

I cannot reproduce the issue using emacs -Q.

So this means that your step "1. start emacs" is actually "1. start emacs and load all my configuration", right?

[x0xntor33alx0x]

yes, it works normally without any configurations.

[joaotavora]

yes, it works normally without any configurations, so I guess the problem is in eglot itself.

How do you reach that conclusion? If emacs -Q + Eglot "works normally without any configurations", how can the problem be in Eglot?

[x0xntor33alx0x]

sorry I meant in the configuration, I just removed that.

[x0xntor33alx0x]

I was wondering if there is a fix for that, I tried disabling some of the modes in the buffer but the problem persists.

[joaotavora]

I think the next step is for you to follow the 7 step quige guide @ https://joaotavora.github.io/eglot/#Troubleshooting-Eglot closely.

You already have some of the elements, it seems. Pay special attention to Emacs so now you just have to package it altogether. Pay special attention to step number 7, i.e. following your own recipe from scratch.