I pinged Christoph and he pinpointed the issue to be likely in the
seam-line vectorization or de-vectorization code. He has added a
debugging switch to the Enblend code [1]
Christoph Spiel wrote:
> Your patch has proven so valuable that I
> integrated it into the mainline.
I can't believe it! I wish I could pin-point more precisely the code
where the bug lies. To my understanding: somewhere before line 880 in
mask.h?
> the problem sits in the seam-line vectorization
> or de-vectorization code. Any other opinions
> out there? Differential diagnosis!
I think there is more brain to be picked on Hugin-PTX than in a private
conversation with half a dozen developers.
For users: a possible temporary workaround is on my blog.
Yuv
[0]
http://panospace.wordpress.com/2009/11/10/the-most-welcoming-bed-and-breakfast-in-london-ontario/
[1] http://enblend.hg.sourceforge.net/hgweb/enblend/enblend/rev/c2b918407a36