From d7057441b823a42b1378eadb0b4488d6fb7069f5 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Jo=C3=A3o=20T=C3=A1vora?= Date: Sun, 5 Sep 2021 20:05:45 +0100 Subject: [PATCH] Fixup last commit Per https://github.com/joaotavora/eglot/issues/726. I'm still not entirely convinced using all-completion here is a good idea. As usual the completion list we get from the server is pre-filtered to whatever the server wishes. Letting the completion style do its own filtering (most completion styles use completion-regexp-list and all-completions themselves) is completely useless here. Let's hope it's not harmful. * eglot.el (eglot-completion-at-point): Fix all-completions call --- lisp/progmodes/eglot.el | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/lisp/progmodes/eglot.el b/lisp/progmodes/eglot.el index e8b7ffbbb13..7ebc4224519 100644 --- a/lisp/progmodes/eglot.el +++ b/lisp/progmodes/eglot.el @@ -2371,8 +2371,7 @@ is not active." (filterText (plist-get item :filterText))) (and (or (null pred) (funcall pred proxy)) (string-prefix-p - probe (or filterText proxy) completion-ignore-case)))) - (funcall proxies))))) + probe (or filterText proxy) completion-ignore-case)))))))) :annotation-function (lambda (proxy) (eglot--dbind ((CompletionItem) detail kind) -- 2.39.2