unofficial mirror of bug-gnu-emacs@gnu.org 
 help / color / mirror / code / Atom feed
* bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp
@ 2022-12-27 12:09 Wilhelm H Kirschbaum
  2022-12-28  1:47 ` Yuan Fu
  0 siblings, 1 reply; 4+ messages in thread
From: Wilhelm H Kirschbaum @ 2022-12-27 12:09 UTC (permalink / raw)
  To: 60355

When `treesit-defun-type-regexp` is set with a cons cell (regexp . 
some-pred)
the following case does not get handled as expected when calling 
(end-of-defun):

```elixir
defmodule Example do
  def foo() do ; regexp match, pred match
  end

  ;; <point here>
  @impl true    ; regexp match, pred does not match
  def bar() do  ; regexp match, pred match
  end
  ;; <should jump to here>

  def baz() do
  end
end
;; <jumps to point here>
```

The function `treesit--things-around` only looks at the PRED after 
it searches forward and
would then not be aware of the non matching node.  The following 
change works for my case
where search-forward takes in the PRED, not just the regexp:

diff --git a/lisp/treesit.el b/lisp/treesit.el
index 2130cd0061..7997509f50 100644
--- a/lisp/treesit.el
+++ b/lisp/treesit.el
@@ -1776,7 +1776,13 @@ treesit--things-around
                             node "" nil nil t))
                        node))
          (result (list nil nil nil))
-         (pred (or pred (lambda (_) t))))
+         (pred (if (null pred)
+                (lambda (node)
+                  (string-match-p regexp (treesit-node-type 
node)))
+                (lambda (node)
+                  (and (string-match-p regexp (treesit-node-type 
node))
+                       (funcall pred node))))))
+
     ;; 1. Find previous and next sibling defuns.
     (cl-loop
      for idx from 0 to 1
@@ -1798,12 +1804,10 @@ treesit--things-around
      do (cl-loop for cursor = (when node
                                 (save-excursion
                                   (treesit-search-forward
-                                   node regexp backward 
                                    backward)))
+                                   node pred backward backward)))
                  then (treesit-node-parent cursor)
                  while cursor
-                 if (and (string-match-p
-                          regexp (treesit-node-type cursor))
-                         (funcall pred cursor)
+                 if (and (funcall pred cursor)
                          (funcall pos-pred cursor))
                  do (setf (nth idx result) cursor)))
     ;; 2. Find the parent defun.


Hope this makes sense,
Wilhelm

ps. Sorry about the email swapping, think it should be fixed now. 





^ permalink raw reply related	[flat|nested] 4+ messages in thread

* bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp
  2022-12-27 12:09 bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp Wilhelm H Kirschbaum
@ 2022-12-28  1:47 ` Yuan Fu
  2022-12-28  6:31   ` Wilhelm H Kirschbaum
  0 siblings, 1 reply; 4+ messages in thread
From: Yuan Fu @ 2022-12-28  1:47 UTC (permalink / raw)
  To: Wilhelm; +Cc: 60355


Wilhelm H Kirschbaum <wkirschbaum@gmail.com> writes:

> When `treesit-defun-type-regexp` is set with a cons cell (regexp .
> some-pred)
> the following case does not get handled as expected when calling
> (end-of-defun):
>
> ```elixir
> defmodule Example do
>  def foo() do ; regexp match, pred match
>  end
>
>  ;; <point here>
>  @impl true    ; regexp match, pred does not match
>  def bar() do  ; regexp match, pred match
>  end
>  ;; <should jump to here>
>
>  def baz() do
>  end
> end
> ;; <jumps to point here>
> ```

Thank you very much! The root cause is actually another problem, but
anyway it is fixed now.

May I also suggest you to use this function for the predicate:

(defun elixir-ts-mode--capture-defun (node)
  ;; NODE should be a ‘call’ node.
  (member (treesit-node-text
           (treesit-node-child-by-field-name node "target"))
          elixir-ts-mode--definition-keywords))

This should be more accurate than the current one, and is IMO simpler
too.

Yuan





^ permalink raw reply	[flat|nested] 4+ messages in thread

* bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp
  2022-12-28  1:47 ` Yuan Fu
@ 2022-12-28  6:31   ` Wilhelm H Kirschbaum
  2022-12-28  8:34     ` Yuan Fu
  0 siblings, 1 reply; 4+ messages in thread
From: Wilhelm H Kirschbaum @ 2022-12-28  6:31 UTC (permalink / raw)
  To: Yuan Fu; +Cc: 60355


Thanks Yuan, just checked and both the fix and your suggestion 
works.

I see there is a typo in your last change:

@@ -310,7 +310,7 @@ java-ts-mode
   ;; Imenu.
   (setq-local treesit-simple-imenu-settings
               '(("Class" "\\`class_declaration\\'" nil nil)
-                ("Interface "\\`interface_declaration\\'" nil 
                 nil)
+                ("Interface" "\\`interface_declaration\\'" nil 
nil)
                 ("Enum" "\\`record_declaration\\'" nil nil)
                 ("Method" "\\`method_declaration\\'" nil nil)))
   (treesit-major-mode-setup))


Yuan Fu <casouri@gmail.com> writes:

> Wilhelm H Kirschbaum <wkirschbaum@gmail.com> writes:
>
>> When `treesit-defun-type-regexp` is set with a cons cell 
>> (regexp .
>> some-pred)
>> the following case does not get handled as expected when 
>> calling
>> (end-of-defun):
>>
>> ```elixir
>> defmodule Example do
>>  def foo() do ; regexp match, pred match
>>  end
>>
>>  ;; <point here>
>>  @impl true    ; regexp match, pred does not match
>>  def bar() do  ; regexp match, pred match
>>  end
>>  ;; <should jump to here>
>>
>>  def baz() do
>>  end
>> end
>> ;; <jumps to point here>
>> ```
>
> Thank you very much! The root cause is actually another problem, 
> but
> anyway it is fixed now.
>
> May I also suggest you to use this function for the predicate:
>
> (defun elixir-ts-mode--capture-defun (node)
>   ;; NODE should be a ‘call’ node.
>   (member (treesit-node-text
>            (treesit-node-child-by-field-name node "target"))
>           elixir-ts-mode--definition-keywords))
>
> This should be more accurate than the current one, and is IMO 
> simpler
> too.
>
> Yuan






^ permalink raw reply	[flat|nested] 4+ messages in thread

* bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp
  2022-12-28  6:31   ` Wilhelm H Kirschbaum
@ 2022-12-28  8:34     ` Yuan Fu
  0 siblings, 0 replies; 4+ messages in thread
From: Yuan Fu @ 2022-12-28  8:34 UTC (permalink / raw)
  To: Wilhelm H Kirschbaum; +Cc: 60355



> On Dec 27, 2022, at 10:31 PM, Wilhelm H Kirschbaum <wkirschbaum@gmail.com> wrote:
> 
> 
> Thanks Yuan, just checked and both the fix and your suggestion works.
> 
> I see there is a typo in your last change:
> 
> @@ -310,7 +310,7 @@ java-ts-mode
>  ;; Imenu.
>  (setq-local treesit-simple-imenu-settings
>              '(("Class" "\\`class_declaration\\'" nil nil)
> -                ("Interface "\\`interface_declaration\\'" nil                 nil)
> +                ("Interface" "\\`interface_declaration\\'" nil nil)
>                ("Enum" "\\`record_declaration\\'" nil nil)
>                ("Method" "\\`method_declaration\\'" nil nil)))
>  (treesit-major-mode-setup))

Oops, was in a bit of hurry and forgot to byte-compile. Thanks for the catch!

Yuan




^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2022-12-28  8:34 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-12-27 12:09 bug#60355: 29.0.60; Tree sitter incorrectly handles of PRED for treesit-defun-type-regexp Wilhelm H Kirschbaum
2022-12-28  1:47 ` Yuan Fu
2022-12-28  6:31   ` Wilhelm H Kirschbaum
2022-12-28  8:34     ` Yuan Fu

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).