On Tue, 2009-02-17 at 13:29 -0800, Ben Jackson wrote: > On Tue, Feb 17, 2009 at 09:08:33PM +0000, Peter Clifton wrote:
> It usually boils down to numerical instability in the polygon code. > I may have an example. I save all the strange boards people send in with > bug reports to use as regression tests. If that is due to the "snap rounding" problem, then the code _ought_ to avoid it. If it is real instability, then presumably we'd end up having trouble with cases other than coincident lines. (The line caps on the end of two lines have a higher order of symmetry than the line its-self). I always have difficulty distilling simple test-cases on polygon problems. Once some combination of geometry conspires to break one of the invariants PCB's polygon code expects to hold true, all hell breaks loose, and it can be hard to track down which piece of geometry was the first aggressor. I probably have some of my own boards saved which cause various printf / failures, but it is very hard to bisect the board to get a single failing case. Self-intersecting contours probably aren't a well-tested case, I'd guess. -- Peter Clifton Electrical Engineering Division, Engineering Department, University of Cambridge, 9, JJ Thomson Avenue, Cambridge CB3 0FA Tel: +44 (0)7729 980173 - (No signal in the lab!) _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
