Thanks to hard work over the past few months, the guys over at Paper.js have built out a brand new path boolean system that should work in ways the old one did not, possibly allowing for ditching the current fill method, and all of it's issues (very slow, inaccurate, #35, #33, #37 to name a few).
Right now there's no proof it solves all the issues with direct boolean intersection required to make this work, but it's quite promising, and they have it on their roadmap to work as expected which is exactly what we'd need to make this work.
Investigative work would require testing all self-intersection and other forms of open path intersection in the creation of closed inner paths via iterative path build boolean. Once the work is proven it can likely be implemented directly over the existing work and will likely result in speed increases beyond 100x, and near 100% accuracy to the centerline position even for relatively small areas.
Slimmy82
commented
8 years ago
Thanks to hard work over the past few months, the guys over at Paper.js have built out a brand new path boolean system that should work in ways the old one did not, possibly allowing for ditching the current fill method, and all of it's issues (very slow, inaccurate, #35, #33, #37 to name a few).
Right now there's no proof it solves all the issues with direct boolean intersection required to make this work, but it's quite promising, and they have it on their roadmap to work as expected which is exactly what we'd need to make this work.
Investigative work would require testing all self-intersection and other forms of open path intersection in the creation of closed inner paths via iterative path build boolean. Once the work is proven it can likely be implemented directly over the existing work and will likely result in speed increases beyond 100x, and near 100% accuracy to the centerline position even for relatively small areas.