Michael Albinus <michael.albinus@gmx.de> writes:

> João Távora <joaotavora@gmail.com> writes:
>
> Hi João,
>
>>> Honestly, I'm lost with all the patches back forth and back. Could you
>>> pls show me the change you propose, based on the current master state? 
>>
>> OK.  Based on the current master state, I'm sending you 4 (four
>> patches).  I known it looks a lot, but is quite simple.
>
> I've applied the 4 patches. As expected, eglot-tests, tramp-tests, and
> filenotify-tests run successfully.
>
>> * bug#61350 should be fixed -- even without the Eglot workaround.  Let
>>   me know if you need help testing this (I have a nifty dockerfile and
>>   some command-line incantations that make testing very easy).
>
> The example given by Thomas blocks, again. So we are back to the
> beginning, aren't we?

Doh! I gave you a wrong patch-0003.  It needs a tweak.

You can either apply this one-line diff or re-aplply attached patchset
of 4 over origin/master again.

Then please try again, and sorry for the extra confusion.

diff --git a/lisp/net/tramp.el b/lisp/net/tramp.el
index 885b29f9471..a7406a9d80e 100644
--- a/lisp/net/tramp.el
+++ b/lisp/net/tramp.el
@@ -5804,7 +5804,7 @@ tramp-accept-process-output
 	     (v (process-get proc 'vector)))
     (dolist (p (delq proc (process-list)))
       (when (tramp-file-name-equal-p v (process-get p 'vector))
-	(while (accept-process-output p 0 nil t)))))
+	(while (accept-process-output p 0)))))
   (with-current-buffer (process-buffer proc)
     (let ((inhibit-read-only t)
 	  last-coding-system-used


I've tested Thomas's example 5 times in a row, moving around energically, in
the Emacs -Q, exercising multiple Eglot features, etc.  0% failures.