Re: Logic Gates in SVG format for Dia

[Cyrille Chepelov <cyrille chepelov org>]

> I see the fault in the fact that we don't have rotation and flipping as a
> general thing for all shapes.  It's not something the shape itself really
> should have to worry about, it's something that should be at toolbox
> level.  I'm thinking by now that the best way is to have a
> rotate/flip/scale mode that is separate from the 'normal' modify mode.  I
> don't think some in-diagram handles for rotation would work well, the
> objects would just be overloaded with handles.

Agreed. And most of it ultimately stems from the Core X font handling.
Keith P. and friends are now addressing that. I'd wager that in 3-6
months, a Cairo renderer should be something realistic (either that, or
GTK will simply wrap Cairo for us, which would actually be better: on
Win32, it should be a no brainer given that GDI has had this capability
for, what, 10 years now? (I'm talking about NT 3.1's GDI, 16-bit windows
GDI was a little bit too short). 
Remember my old Display Postscript renderer? I had a quick look at
Cairo's interface. This will be much, much easier (and our code is
cleaner nowadays that it was when I did my DPS experiment).

By the way another nice thing everybody but X has, and Cairo will give
us, is an alpha channel. Yes, Venn diagrams.

> I have a fairly strong theory that that's not possible without either
> crippling the objects fatally or making the SVG markup generic beyond
> usability.  But we've discussed that.

Right. Let proponents show code.

	-- Cyrille

--