Pomax
I've pushed a v2.4.4 which makes the previous jsbin render correctly: the arc now has start and end angles that line up with where they should be, even when the...
good find, that error would come from https://github.com/Pomax/bezierjs/blob/gh-pages/lib/utils.js#L540 which would suggest that https://github.com/Pomax/bezierjs/blob/gh-pages/lib/utils.js#L519 is yielding a `false`... which implies a line/line intersection test is failing pretty hard over in https://github.com/Pomax/bezierjs/blob/gh-pages/lib/utils.js#L203-L209...
yeah there is no linearity detection at the moment, so I'm okay with those yielding bad arcs. The others are more interesting, I'll make some coffee and investigate what is...
Running your test cases through the jsbin, with `threshold = 1.2143080752705958`, I get the following results: `172, 158, 108, 124, 29, 82` appears to be linear but does not show...
Just in case this is caused by maths engines: which OS/Browser are you seeing things go wrong on, and do you have code that you can drop in a JSbin...
Ah! good point, using 0.5 seems to be a magical value for the threshold.
I'll happily take a PR, but be aware that there is no symbolic solution to arc length parameterization of cubic (or higher) curves due to needing to solve a polynomial...
This looks like a very curious bug. If I edit the coordinates in http://jsbin.com/vadayaneli/edit?html,console,output just a little, the offsetting works fine, so there is bound to be a rounding issue...
that's certainly an opinion, but unless you want to plead a case to have it changed, I have no reason to change it because it works just fine the way...
on the "it's wrong": yeah that definitely looks wrong...marking as bug. As for getting a non-intersecting outline, offsetting cannot generate a G1 or higher continuous offset, but you can perform...