GNU bug report logs - #62086
29.0.60; ruby-ts-mode regressions

Previous Next

Full log

Message #20 received at 62086 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Juri Linkov <juri <at> linkov.net>, Yuan Fu <casouri <at> gmail.com>
Cc: 62086 <at> debbugs.gnu.org
Subject: Re: bug#62086: 29.0.60; ruby-ts-mode regressions
Date: Fri, 10 Mar 2023 18:37:08 +0200

On 10/03/2023 09:35, Juri Linkov wrote:
>>> I don't know if the second bug is related to this, but while
>>> in the same file, also type 'C-M-l' ('reposition-window').
>>> It raises the error:
>>>     Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p
>>> nil)
>>>       treesit-end-of-defun()
>>>       end-of-defun(-1)
>>>       reposition-window(nil nil)
>>>       reposition-window(nil 89)
>>>       funcall-interactively(reposition-window nil 89)
>>>       command-execute(reposition-window)
>>> This regression is also recent.
>> I've managed to reproduce this, but only once. Do you see this every time?
> I see it only in some files in test/lisp/progmodes/ruby-mode-resources/
> e.g. ruby-parenless-call-arguments-indent.rb, ruby-method-call-indent.rb,
> ruby-block-indent.rb.  But not in e.g. ruby-after-operator-indent.rb.
> Also everywhere in test/lisp/progmodes/js-resources/js-indent-init-dynamic.js,
> js-indent-init-t.js.  But not in e.g. js-chain.js.

Thanks, I can repro. I might have been trying the wrong binding at the 
end last night (C-l instead of C-M-l).

The fix seems to be easy:

diff --git a/lisp/treesit.el b/lisp/treesit.el
index c118f5d52a4..b271a1f0c4b 100644
--- a/lisp/treesit.el
+++ b/lisp/treesit.el
@@ -1882,6 +1882,7 @@ treesit-end-of-defun
 `treesit-defun-skipper'."
   (interactive "^p\nd")
   (let ((orig-point (point)))
+    (if (or (null arg) (= arg 0)) (setq arg 1))
     (catch 'done
       (dotimes (_ 2) ; Not making progress is better than infloop.


But I'm not quite sure if that is what we want to do.

More naturally, I think, would be to remove the argument from 
treesit-end-of-defun altogether (and adjust the code accordingly), 
because end-of-defun-function is documented to take no arguments.

The only other place where treesit-end-of-defun seems to be used is the 
<remap> <end-of-defun> binding set up by treesit-major-mode-setup.

Why not keep the default bindings for these? When 
beginning-of-defun-function and end-of-defun-function are set 
appropriately, they should work fine. Don't they?

Cc'ing Yuan on that subject.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.