Interesting. It seems right to implement this on a per-dab basis.

But it looks a bit strange to add a "mirror axis" to stroke_to(), a method
that only takes tablet input as arguments so far.  Maybe mirror paint should
be a property of the surface (tiledsurface.hpp), instead of the brush class?

Brushlib calls draw_dab() and get_color() on the surface class (which is
kind of its "public" API, see brushlib wiki page).  The surface class could
decide to render two dabs instead of one, for every call of draw_dab().  And
for every call of get_color() it would have to average the color over two
areas, instead of one.  This way brushlib doesn't need to know about the
mirroring GUI used in MyPaint.

The "dab_angle" parameter should also be mirrored. And this is a good time
to think about a fix for https://gna.org/bugs/?17400

On Sat, Jan 07, 2012 at 03:00:28PM -0600, Martin Franco wrote:
> +    bool ret1(false), ret2(false);

Please don't use C++ syntax when avoidable. I would like to convert brushlib
to pure C (C99) eventually.

-- 
Martin Renold

_______________________________________________
Mypaint-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/mypaint-discuss

Reply via email to