bug#62694: 30.0.50; eglot-tests fails with recent pylsp

[Michael Albinus <michael.albinus@gmx.de>]

João Távora <joaotavora@gmail.com> writes:

Hi João,

>> Don't know, I have no knowledge about language servers. But I would
>> expect, that you first check whether pylsp is installed (you do it
>> already), and then you could connect to pylsp, send a specific request,
>> and decide, based on the ansewer, whether it is a proper language
>> server.
>
> Connect to a language server just to discover if I can run a test?

Sure. See tramp-tests.el, which run at the very beginning

--8<---------------cut here---------------start------------->8---
(and
 (file-remote-p ert-remote-temporary-file-directory)
 (file-directory-p ert-remote-temporary-file-directory)
 (file-writable-p ert-remote-temporary-file-directory))
--8<---------------cut here---------------end--------------->8---

A sanity check, in order to know that further tests make sense.

> All to get around (presumably) faulty Debian packaging of
> an obscure language server that almost noone uses when a perfectly
> good alternative to install said obscure already exists?

Forget Debian. Any language server could behave wrong, remember the
(whereever it came from) clangd on my laptop. I believe, testing just
that an executable with a given name exists is not sufficient in the
world of language servers, which is still rapidly evolving.

But I don't fight with you. Do whatever you believe is proper, and live
with the fact that your tests fail on Emba. We have more discussed
about, as it would require to write a small check.

I have more serious problems to be solved.

Best regards, Michael.