Re: Emacs and GTK.

[William M. Perry]

"Jan D." <address@hidden> writes:

> Richard Stallman wrote:
> 
> >     Annoying, but doable and unclean.  I would prefer not to see time spent 
> > on
> >     an implementation like that when a native GTK port would not be much 
> > more
> >     difficult (20%?).
> > What is a "native GTK port", and what it would entail?
> > Which parts of Emacs would have to be changed?
> > 
> 
> Not speaking for wmperry, but for me it means replacing all X calls with
> GTK calls.  Basically this is the functionallity of lwlib and at least
> the functionallity of these files:
>
> widget.c
> xfaces.c
> xfns.c
> xmenu.c
> xrdb.c
> xselect.c
> xterm.c

Exactly!

> I believe it is possible to have a mixed X/GTK world while developing,
> thus replacing piece by piece.  Even so, this is a bit of work.

Well, when I did the XEmacs/GTK port, to get the bare minimum redisplay
working took about 5 days.  This didn't have all the image formats
supported, or scrollbars, menubars, or toolbars.  But it was enough to
bootstrap and the rest of the time it took to do the fun widgets and the
emacs-lisp GTK bindings was done with the current source tree.
Self-hosting as they say in kernel development, right? :)

Note that the 5 days I outlined above was pretty much 8a-5p and 10p-1a
because I took time off from work to do the project.

I seriously think it would take more time to get the mixed X/GTK working
well than just biting the bullet and doing the redisplay.

I am trying to spend all my hacking time working on the URL library and
getting Emacs/W3 into better shape.

I would be willing to take some more time off to help with this, but it
would have to be after christmas I think.

-bp
-- 
Ceterum censeo vi esse delendam