Re: Platform independent graphical display for Emacs

[Dmitry Gutov]

On 25.12.2021 15:06, Eli Zaretskii wrote:
> > Cc: stefankangas@gmail.com, luangruo@yahoo.com, drew.adams@oracle.com,
> >   emacs-devel@gnu.org
> > From: Dmitry Gutov <dgutov@yandex.ru>
> > Date: Sat, 25 Dec 2021 13:57:40 +0200
> >
> > > So you'd suggest to the OP to develop the software in the hope that
> > > all of the above will happen?  And if it doesn't, just agree for the
> > > results to be abandoned?  The OP would have to agree to that.
> >
> > I suppose.
> >
> > And I think the BeOS port had been accepted under the same conditions
> > recently.
>
> Because we did have someone who volunteered, and because nothing would
> be lost on platforms of interest to us if the BeOS port bitrots.

Which would be also true when someone initially proposes the new "single 
toolkit" port for inclusion.

> > > And I fail to see how that solves the long-term maintenance problem,
> > > once we do accept the code.  This happened in the past, more than
> > > once.
> >
> > We should be able to drop unmaintained ports.
>
> Once again, a single toolkit on which all the platforms depend cannot
> be dropped.  So it's a huge difference from the BeOS port.

Once again: we would only get to that point after the "single toolkit" 
port is functional, accepted by the majority of us, and probably 
"adopted" by a few regular contributors as their area of interest.

Only then we could realistically vote to drop all other ports.

I'm sure a contributor who would propose such new port would understand 
all of this.

> > Even if we're reluctant, in general, to remove features that someone
> > is using. After all, the history of changes is saved, so as soon as
> > a volunteer arrives to resurrect it, they can start with 'git
> > revert' and continue.
>
> That's generally not a practical possibility, because the code changes
> very rapidly, and what you revert won't even compile in most cases.

Yes and no: I'm sure having the previous code available will be a lot of 
help. As opposed to rewriting the same port from scratch.