bug#50256: thing-at-mouse

[Eli Zaretskii <eliz@gnu.org>]

> Cc: juri@linkov.net, 50256@debbugs.gnu.org, larsi@gnus.org
> From: martin rudalics <rudalics@gmx.at>
> Date: Sun, 5 Sep 2021 09:50:14 +0200
> 
>  >>   > That's orthogonal.  (And it seems to issue the warning in an unrelated
>  >>   > case, handled by the 'else' clause?)
>  >>
>  >> Which case?
>  >
>  > This one:
>  >
>  >> +      else
>  >> +	   posint = marker_position (w->pointm);
> 
> You mean when the caller asks for using point in a non-selected window
> that does not show the current buffer.  Why do you think this case is
> unrelated and/or should be handled differently?

I see no reason to display a warning in that case.  It's completely
legitimate; moreover, a non-selected window will almost always display
a buffer that is not the current buffer, and annoying warnings in
those cases will just ... annoy.

>  >> Maybe I have not been making myself clear.  What I would apply is the
>  >> attached (modulo any errors it contains): No change in behavior, just a
>  >> warning in the doc-string and suspicious calls.
>  >
>  > That doesn't fix the problem which started this thread, does it?
> 
> Right.  We have two ways to fix that problem:
> 
> (1) Fix it in a general way.  You don't want to do that because it might
>      harm calls of that function that silently worked until now.
> 
> (2) Fix it by modifying the calls of `pos-visible-in-window-p'.  In that
>      case the warning should help us find the calls that should be fixed.

Once again, the warning is an orthogonal issue.  I'm not opposed to
displaying a warning in cases where the caller might have intended
something else.  But let's please first fix the actual issue that
started this thread, and that wasn't the lack of warning.