> - If we want to support Windows at some point (and I was thinking
> fltk might be a great help here although the recent announcement
> of Digital Domains new management ceasing support and closing
> the source is very bad news) we will need to abstract the
> workarea a little better. Hopefully the wonderful new Painter
> class could simply be reimplemented for each major system (X11
> now, with possible Windows and OS/2 native also).
Yes, the new painter should be portable. That was a mayor point of the
exercise :-)
So now, what needs to be done to make LyX truly portable is this:
1) Design and implement new metric-stuff for insets, such that they use the new
painter.
2) Finish the moving of popups to frontends - in particular an "rgrep -B
FORMS_LOCATION *" in the src directory should come up with very few hits.
3) Fix the cursor in screen.C to use a painter.
4) Implement picture/pixmap support in the painter (it's there, but untested).
5) Implement copying between two painters. This would allow screen.C to be
fully portable, thus abstracting all drawing on the canvas.
6) Implement an abstraction of the canvas.
7) Make the figinset use the new painter.
8) The metric and drawing stuff in LyXFont.h has to go.
Some of these are big: 1), 2) and 6). The rest should be small ones, and
doable in a few hours. Please go ahead and work on it if you have time.
Unfortunately I'm finding it hard to find time to do it myself, but I'll always
find time to answer questions.
--
Since the state of Fltk is in a flux, another serious candidate for a toolkit
is GTK--. The 1.0 release is close (sic), and GTK+ has been ported to MS
Windows.
GTK-- has the best toolkit interface I have ever seen. It's better than Qt,
IMO, and you would never think that it's a wrapper around GTK+ unless you knew
it.
Greets,
Asger
