Re: Rounded Polyline, and other questions?

[anthonym overture com]

On Tue, Jun 29, 2004 at 10:29:03PM +0200, Lars Clausen wrote:
> On Thu, 2004-06-24 at 08:13, anthonym overture com wrote:
> > Not a problem, it's a good thing because I realize I missed something
> > with the previous patch, I didn't adjust the radius if if was bigger than
> > the polyline could handle.  This only manifested itself with acute angles
> > when rounding polylines.
> > 
> > Anyway, this is fixed in the attached patch (which also includes the previous
> > changes, so you can disregard that patch).
> 
> I've applied this patch, with some changes (moved purely geometric
> functions to geometry.c, moved unrounded rendering back to polyline).

Darn, the changes haven't been mirrored to anoncvs yet.  Guess I'll wait
till tomorrow to sync my tree.

> I see one weird behaviour right now:  You calculate the minimum radius
> for the entire line.  This means that if you have a well-rounded
> polyline with >>3 points and then bend one end back, all corners
> suddenly become less rounded, not just the one that's bent.  Guess the
> radius calculation should be local to each corner.

I calculated the minimum radius for the whole line for 2 reasons.
  1) the properties dialog suggests a single corner radius
  2) the box corner rounding had similiar behavior

It should be a straightforward change to do the calculation for each
corner, however given the box behavior and the dialog, should the
calculation be local to to each corner?  And if so should the
properties dialog be somehow different?

-Anthony

-- 
------------------------------------------------------------------------
Anthony Molinaro                                 <anthonym overture com>