bug#71170: 30.0.50; hs-hide-all in python mode not works as expected

[kobarity <kobarity@gmail.com>]

lorniu wrote:
> 
> 
> > Is this what you want?
> 
> What I want is, for example, this python code:
> 
>   ```
>   if TYPE_CHECKING:
>       Base64FileInput = Union[IO[bytes], PathLike[str]]
>       FileContent = Union[IO[bytes], bytes, PathLike[str]]
>   else:
>       Base64FileInput = Union[IO[bytes], PathLike]
>       FileContent = Union[IO[bytes], bytes, PathLike]  # PathLike is not  subscriptable in Python 3.8.
> 
> 
> 
> 
>   class aaa:
>       ...
> 
> 
> 
>   class bbb:
>       ...
>   ```
> 
> will fold as:
> 
>   ```
>   if TYPE_CHECKING:...
>   else:...
> 
>   class aaa:...
> 
>   class bbb:...
>   ```
> 
> instead of:
> 
>   ```
>   if TYPE_CHECKING:...
>   else:...
> 
> 
> 
> 
>   class aaa:...
> 
> 
> 
>   class bbb:...
>   ```
> 
> That is too much newlines between hidden blocks.
> 
> Maybe the patch I suggested is not well enough. Some update like this
> may be better?
> 
> diff --git a/lisp/progmodes/python.el b/lisp/progmodes/python.el
> index 79e383a1c1a..009e0c4fb3a 100644
> --- a/lisp/progmodes/python.el
> +++ b/lisp/progmodes/python.el
> @@ -5734,7 +5734,13 @@ python-describe-at-point
>  (defun python-hideshow-forward-sexp-function (_arg)
>    "Python specific `forward-sexp' function for `hs-minor-mode'.
>  Argument ARG is ignored."
> -  (python-nav-end-of-block))
> +  (python-nav-end-of-block)
> +  (end-of-line)
> +  (let ((pt (point)))
> +    (skip-chars-forward " \n\t")
> +    (re-search-backward "\n\n" pt t)
> +    (unless (eolp) (goto-char (1- (line-beginning-position))))
> +    (when (< (point) pt) (goto-char pt))))
> 
>  (defun python-hideshow-find-next-block (regexp maxp comments)
>    "Python specific `hs-find-next-block' function for `hs-minor-mode'.
> 
> 
> 
> -- 
> Mess.

The patch works as expected.
I guess it depends on the person whether this kind of behavior is
desirable or not.  Wouldn't it be better to make it a customized
setting or something?