mnml_

``` c:\blank>where git && git --version C:\Program Files\Git\cmd\git.exe git version 2.19.1.windows.1 c:\blank>where npm && npm --version C:\Users\d\nodejs\npm C:\Users\d\nodejs\npm.cmd 6.5.0 ```

looks like https://github.com/rustwasm/rust-webpack-template/issues/95 might be related...

it gets better: it seems to work when I bang in the create-rust-webpack.js into a node instance manually: ``` c:\blank>node > > const { spawnSync } = require("child_process"); undefined >...

i'd use `sgn(a-4)`: it's -1 if a4

i cloned master, and i included the source (daef.rs) for my test-binary

Would it be possible to consistently break ties like that (same starting point, same direction) by another measurement (e.g. squared length, x component, ...)?

While using a release-build and ignoring the missed events works out often, I also get results like this: the union of yellow and green:  results in the orange and...

I thought the bigger numerical instabilities arising from smaller floats call for more robust code in the library - therefore it should be beneficial. eg dont calculate intersectionpoints - but...

I've implemented [rank fixing](https://github.com/bluenote10/rust-array-stump/pull/2) for the new data structure and the [respective calls there](https://github.com/bluenote10/rust-geo-booleanop/pull/1) in subdivide_segments after possible mutations of the events. This all happens in a pretty naive(=unconditional) way,...

OT: I should probably include (geo-)clipper in my [comparison](https://daef.github.io/poly-bool-comparison/dissection/). How does it compare performance wise?