Re: "Final" version of tty child frames

[Eli Zaretskii <eliz@gnu.org>]

> Date: Tue, 17 Dec 2024 21:35:35 -0800
> From: Jared Finder <jared@finder.org>
> Cc: Eli Zaretskii <eliz@gnu.org>, emacs-devel@gnu.org, rudalics@gmx.at
> 
> Things otherwise seem fine with tty child frames. There's certainly 
> oddness with mouse interaction, but it's not fundamentally broken in any 
> way, just more things that don't work. In particular:
> 
> With xterm-mouse, as I highlighted earlier, I can select the child frame 
> even if it is set as not selectable. Once a window in a child frame is 
> selected, I can type there normally.
> 
> With gpm mouse, I have the opposite problem. I can never select the 
> child frame and in fact the mouse behaves as if the child frame isn't 
> there. Clicking and tooltip text both pay no attention to the child 
> frame and just act on whatever is behind the child frame.
> 
> For both of these, I couldn't get mouse-face or clicking to work on 
> child frames. I was doing the following:
> 
> (setq button (buttonize "[Click me]" (lambda (&rest _) (message 
> "Clicked!"))))
> (posframe-show " *buffer*" :string (concat "A\n" button "\nB"))
> 
> The posframe would show, but the mouse can't interact with the 
> buttonized text. This may be a limitation of posframe though, it also 
> didn't work in graphical mode.
> 
> That's really it. I don't see any major issues with child frames. As 
> long as we're ok with saying that mouse support is not mature, it seems 
> fine to me.

If we don't see immediate ways of fixing some of these issues, I think
it should be okay to land this on master, if Stefan and Andrea agree.

Thanks.