[PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[gerard.vermeulen]

[-- Attachment #1: Type: text/plain, Size: 1092 bytes --]

Hello,

The Emacs integrated development environments Slime and Sly
communicate by means of similar RPC protocols with different server
programs: Swank in case of Slime and Slynk in case of Sly.  However,
the code in org-babel-execute:lisp always expects to talk with Swank
because of the hard-coded call to swank:eval-and-grab-output.

After starting a Sly REPL, calls to swank:eval-and-grab-output do not
talk with the Slynk server started by the Sly REPL.

The attached patch "soft-codes" eval-and-grab-output to become
swank:eval-and-grab-output when using Slime and
slynk:eval-and-grab-output when using Sly.

For more info: steps to test the patch are:
1. Open a Sly REPL with M-x sly
2. Look in the *sly-events for <LISP>* (<LISP> is sbcl in my case) for
    slynk: prefixes
3. Type C-c C-c on the test block below
4. Look for an addition in *sly-events for <LISP>* looking like
    (slynk:eval-and-grab-output "(let ((...)) (+ 1 1)\n)")

#+name: test-ob-lisp-with-sly
#+begin_src lisp -n :output results
(+ 1 1)
#+end_src

#+RESULTS: test-ob-lisp-with-sly
: 2

Best regards -- Gerard

[-- Attachment #2: 0001-lisp-ob-lisp-fix-org-babel-execute-lisp-for-use-with.patch --]
[-- Type: application/octet-stream, Size: 4335 bytes --]

From 999deeebf473cc39b90ccbaac04d7fb2ea6ad6c2 Mon Sep 17 00:00:00 2001
From: Gerard Vermeulen <gerard.vermeulen@posteo.net>
Date: Wed, 29 Mar 2023 09:21:08 +0200
Subject: [PATCH] lisp/ob-lisp: fix `org-babel-execute:lisp' for use with Sly

* lisp/ob-lisp (org-babel-execute:lisp): by default, Sly and Slime
talk with different RPC server programs with different names.
Specialize the prefix of eval-and-grab-output to Slime (prefix swank:)
or to Sly (prefix slynk:).
---
 lisp/ob-lisp.el | 61 ++++++++++++++++++++++++++-----------------------
 1 file changed, 32 insertions(+), 29 deletions(-)

diff --git a/lisp/ob-lisp.el b/lisp/ob-lisp.el
index 03f23c82d..c45bec20e 100644
--- a/lisp/ob-lisp.el
+++ b/lisp/ob-lisp.el
@@ -90,35 +90,38 @@ current directory string."
   "Execute a block of Common Lisp code with Babel.
 BODY is the contents of the block, as a string.  PARAMS is
 a property list containing the parameters of the block."
-  (pcase org-babel-lisp-eval-fn
-    (`slime-eval (org-require-package 'slime "SLIME"))
-    (`sly-eval (org-require-package 'sly "SLY")))
-  (org-babel-reassemble-table
-   (let ((result
-          (funcall (if (member "output" (cdr (assq :result-params params)))
-                       #'car #'cadr)
-                   (with-temp-buffer
-                     (insert (org-babel-expand-body:lisp body params))
-                     (funcall org-babel-lisp-eval-fn
-                              `(swank:eval-and-grab-output
-                                ,(let ((dir (if (assq :dir params)
-                                                (cdr (assq :dir params))
-                                              default-directory)))
-                                   (format
-                                    (if dir (format org-babel-lisp-dir-fmt dir)
-                                      "(progn %s\n)")
-                                    (buffer-substring-no-properties
-                                     (point-min) (point-max)))))
-                              (cdr (assq :package params)))))))
-     (org-babel-result-cond (cdr (assq :result-params params))
-       (org-strip-quotes result)
-       (condition-case nil
-           (read (org-babel-lisp-vector-to-list result))
-         (error result))))
-   (org-babel-pick-name (cdr (assq :colname-names params))
-			(cdr (assq :colnames params)))
-   (org-babel-pick-name (cdr (assq :rowname-names params))
-			(cdr (assq :rownames params)))))
+  (let (eval-and-grab-output)
+    (pcase org-babel-lisp-eval-fn
+      (`slime-eval (org-require-package 'slime "SLIME")
+                   (setq eval-and-grab-output 'swank:eval-and-grab-output))
+      (`sly-eval (org-require-package 'sly "SLY")
+                 (setq eval-and-grab-output 'slynk:eval-and-grab-output)))
+    (org-babel-reassemble-table
+     (let ((result
+            (funcall (if (member "output" (cdr (assq :result-params params)))
+                         #'car #'cadr)
+                     (with-temp-buffer
+                       (insert (org-babel-expand-body:lisp body params))
+                       (funcall org-babel-lisp-eval-fn
+                                `(,eval-and-grab-output
+                                  ,(let ((dir (if (assq :dir params)
+                                                  (cdr (assq :dir params))
+                                                default-directory)))
+                                     (format
+                                      (if dir (format org-babel-lisp-dir-fmt dir)
+                                        "(progn %s\n)")
+                                      (buffer-substring-no-properties
+                                       (point-min) (point-max)))))
+                                (cdr (assq :package params)))))))
+       (org-babel-result-cond (cdr (assq :result-params params))
+         (org-strip-quotes result)
+         (condition-case nil
+             (read (org-babel-lisp-vector-to-list result))
+           (error result))))
+     (org-babel-pick-name (cdr (assq :colname-names params))
+			  (cdr (assq :colnames params)))
+     (org-babel-pick-name (cdr (assq :rowname-names params))
+			  (cdr (assq :rownames params))))))
 
 (defun org-babel-lisp-vector-to-list (results)
   ;; TODO: better would be to replace #(...) with [...]
-- 
2.40.0

Re: [PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[Ihor Radchenko]

gerard.vermeulen@posteo.net writes:

> The Emacs integrated development environments Slime and Sly
> communicate by means of similar RPC protocols with different server
> programs: Swank in case of Slime and Slynk in case of Sly.  However,
> the code in org-babel-execute:lisp always expects to talk with Swank
> because of the hard-coded call to swank:eval-and-grab-output.
>
> After starting a Sly REPL, calls to swank:eval-and-grab-output do not
> talk with the Slynk server started by the Sly REPL.
>
> The attached patch "soft-codes" eval-and-grab-output to become
> swank:eval-and-grab-output when using Slime and
> slynk:eval-and-grab-output when using Sly.

I do not use ob-lisp, but this patch looks reasonable.

I'd appreciate if other ob-lisp users could test the patch with both SLY
and SLIME.

I have no comments on the Elisp part of the patch.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: [PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[gerard.vermeulen]

On 29.03.2023 11:42, Ihor Radchenko wrote:
> gerard.vermeulen@posteo.net writes:
> 
>> The Emacs integrated development environments Slime and Sly
>> communicate by means of similar RPC protocols with different server
>> programs: Swank in case of Slime and Slynk in case of Sly.  However,
>> the code in org-babel-execute:lisp always expects to talk with Swank
>> because of the hard-coded call to swank:eval-and-grab-output.
>> 
>> After starting a Sly REPL, calls to swank:eval-and-grab-output do not
>> talk with the Slynk server started by the Sly REPL.
>> 
>> The attached patch "soft-codes" eval-and-grab-output to become
>> swank:eval-and-grab-output when using Slime and
>> slynk:eval-and-grab-output when using Sly.
> 
> I do not use ob-lisp, but this patch looks reasonable.
> 
> I'd appreciate if other ob-lisp users could test the patch with both 
> SLY
> and SLIME.
> 
> I have no comments on the Elisp part of the patch.

I have a few additional remarks:

The patch is against main only because it overlaps a region where bugfix
and main diverged.  I am willing to provide a similar patch against 
main.

I have installed Slime and Sly with the same user-emacs-directory using
package-install.  Normally, both autoload when doing M-x sly, but Sly
prompts you whether you want to disable Slime.  I choose "yes" to test
Sly.  I package-delete Sly when I want to test Slime.

See https://github.com/joaotavora/sly/blob/master/CONTRIBUTING.md
describes in detail how Sly uses the RPC protocol.

I hope that my mail reaches David (CC-ed dto@gnu.org  taken from the
file header), but my previous message bounced (non-existent address).

Best regards -- Gerard

Re: [PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[Ihor Radchenko]

gerard.vermeulen@posteo.net writes:

>> I'd appreciate if other ob-lisp users could test the patch with both 
>> SLY
>> and SLIME.
>> 
>> I have no comments on the Elisp part of the patch.
>
> I have a few additional remarks:
>
> The patch is against main only because it overlaps a region where bugfix
> and main diverged.  I am willing to provide a similar patch against 
> main.

I do not think that we in a hurry to put this onto bugfix.
Bugfix is just for trivial and critical fixes.
https://orgmode.org/worg/org-maintenance.html#release-types

Writing the whole extra alternative patch is probably too much to bother.

> I hope that my mail reaches David (CC-ed dto@gnu.org  taken from the
> file header), but my previous message bounced (non-existent address).

I am CCing his alternative (AFAIK) email.
David, if you are still interested to deal with ob-lisp, we can update
the email.

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>

Re: [PATCH] In case of Sly, let org-babel-execute:lisp use Slynk instead of Swank

[Ihor Radchenko]

Ihor Radchenko <yantar92@posteo.net> writes:

>> The patch is against main only because it overlaps a region where bugfix
>> and main diverged.  I am willing to provide a similar patch against 
>> main.
>
> I do not think that we in a hurry to put this onto bugfix.
> Bugfix is just for trivial and critical fixes.
> https://orgmode.org/worg/org-maintenance.html#release-types
>
> Writing the whole extra alternative patch is probably too much to bother.

Since no concerns have been raised, I just applied the patch onto main.
Applied.
https://git.savannah.gnu.org/cgit/emacs/org-mode.git/commit/?id=e58bbded5

-- 
Ihor Radchenko // yantar92,
Org mode contributor,
Learn more about Org mode at <https://orgmode.org/>.
Support Org development at <https://liberapay.com/org-mode>,
or support my work at <https://liberapay.com/yantar92>