bug#61312: Patch for eglot: scala LSP binary name

bug#61312: Patch for eglot: scala LSP binary name

[skykanin via Bug reports for GNU Emacs, the Swiss army knife of text editors]

[-- Attachment #1: Type: text/plain, Size: 447 bytes --]

Hello,

It seems that eglot expects the scala metals LSP server binary to be named `metals-emacs` instead of `metals`. The included patch fixes this behaviour. At least in the nix package manager the `metals` binary is simply call 'metals'. However if other package managers distribute an alias for the binary as 'metals-emacs' as well we could instead do something like:
```
(scala-mode . ,(eglot-alternatives '("metals" "metals-emacs"))
```

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-Patch-eglot-scala-LSP-server-binary-name.patch --]
[-- Type: text/x-patch; name=0001-Patch-eglot-scala-LSP-server-binary-name.patch, Size: 1225 bytes --]

From 215d54ff5ee3352bcda01d88d7a167925cd8dbeb Mon Sep 17 00:00:00 2001
From: skykanin <3789764+skykanin@users.noreply.github.com>
Date: Sun, 5 Feb 2023 21:41:19 +0100
Subject: [PATCH] Patch eglot scala LSP server binary name

Ensure that eglot picks up the correct
binary name for the scala metals LSP server.
---
 lisp/progmodes/eglot.el | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lisp/progmodes/eglot.el b/lisp/progmodes/eglot.el
index 4fa886bc78..ca025cf7d8 100644
--- a/lisp/progmodes/eglot.el
+++ b/lisp/progmodes/eglot.el
@@ -221,7 +221,7 @@ chosen (interactively or automatically)."
                                               "--client-id" "emacs.eglot-dart"))
                                 (elixir-mode . ("language_server.sh"))
                                 (ada-mode . ("ada_language_server"))
-                                (scala-mode . ("metals-emacs"))
+                                (scala-mode . ("metals"))
                                 (racket-mode . ("racket" "-l" "racket-langserver"))
                                 ((tex-mode context-mode texinfo-mode bibtex-mode)
                                  . ,(eglot-alternatives '("digestif" "texlab")))
-- 
2.39.0

bug#61312: Patch for eglot: scala LSP binary name

[Eli Zaretskii]

> Date: Sun, 05 Feb 2023 21:15:48 +0000
> From:  skykanin via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> 
> It seems that eglot expects the scala metals LSP server binary to be named `metals-emacs` instead of `metals`. The included patch fixes this behaviour. At least in the nix package manager the `metals` binary is simply call 'metals'. However if other package managers distribute an alias for the binary as 'metals-emacs' as well we could instead do something like:
> ```
> (scala-mode . ,(eglot-alternatives '("metals" "metals-emacs"))
> ```

Looking at this site:

  https://github.com/rossabaker/lsp-scala

I see this:

  (defcustom lsp-scala-server-command "metals-emacs"
    "The command to launch the Scala language server."
    :group 'lsp-scala
  :type 'file)

So maybe we do need to support both names.

João, WDYT?

bug#61312: Patch for eglot: scala LSP binary name

[João Távora]

[-- Attachment #1: Type: text/plain, Size: 724 bytes --]

Hi, Eli,

As usual, I defer this decision to you. I think it's reasonable to
support both names, and I also think it's reasonable to stick to
just the one we think is most used or representative of the
program.

In this case, I think "metals-emacs" is a contradiction of LSP's
stated goal, which is to have editor-agnostic servers. But I
don't know what the reasons were for doing this, I haven't
investigated.

It would be even more reasonable, I think, if distributions
settled -- or mostly settled - on names for their binaries they
distribute, much like *nix toolchains do. Of course we do not
control that process, but maybe we could influence it instead
of being constantly influenced by it.

João

[-- Attachment #2: Type: text/html, Size: 1011 bytes --]

bug#61312: Patch for eglot: scala LSP binary name

[Eli Zaretskii]

> From: João Távora <joaotavora@gmail.com>
> Date: Mon, 6 Feb 2023 18:45:39 +0000
> Cc: skykanin <skykanin@proton.me>, 61312@debbugs.gnu.org
> 
> As usual, I defer this decision to you. I think it's reasonable to
> support both names, and I also think it's reasonable to stick to 
> just the one we think is most used or representative of the 
> program.
> 
> In this case, I think "metals-emacs" is a contradiction of LSP's
> stated goal, which is to have editor-agnostic servers. But I 
> don't know what the reasons were for doing this, I haven't
> investigated.
> 
> It would be even more reasonable, I think, if distributions 
> settled -- or mostly settled - on names for their binaries they 
> distribute, much like *nix toolchains do. Of course we do not 
> control that process, but maybe we could influence it instead
> of being constantly influenced by it.

Thanks, I went with supporting both names in Emacs 29.  I cannot see
any harm in supporting both, once "metals" is the first name to check.

With that, I'm closing this bug.