Added Project() #638
Added Project() #638
Conversation
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## master #638 +/- ##
==========================================
- Coverage 91.62% 91.58% -0.05%
==========================================
Files 36 36
Lines 4682 4693 +11
==========================================
+ Hits 4290 4298 +8
- Misses 392 395 +3 ☔ View full report in Codecov by Sentry. |
src/manifold/src/face_op.cpp
Outdated
| @@ -194,17 +196,12 @@ void Manifold::Impl::Face2Tri(const Vec<int>& faceEdge, | |||
| * For the input face index, return a set of 2D polygons formed by the input | |||
There was a problem hiding this comment.
need to update documentation
|
Yeah I envisioned this epsilon-valid issue when converting our 3D model into 2D, I think similar things will happen for slice as well. Not sure how clipper handles this. |
| } | ||
| glm::vec2 size = box.Size(); | ||
| if (glm::abs(area) > glm::max(size.x, size.y) * epsilon) { | ||
| filtered.push_back(poly); |
There was a problem hiding this comment.
This is how I took care of the tiny sliver polygons we were getting in the bracelet.
There was a problem hiding this comment.
I also went ahead and updated Clipper2 to the latest release - I didn't actually notice any difference, but they said they fixed some bugs and I don't think we were on a release at all before.
* added Project * updated Clipper2 to 1.3.0 * fix simplify to remove more cruft * update docs
Addresses one part of #426
Pretty simple algorithm. However, I'm getting lots of micro holes in the bracelet projection, probably from rounding errors, which are messing up the triangulation. I haven't determined if the CrossSection is actually epsilon-valid, and either way, I'm surprised that Clipper's positive fill rule isn't getting rid of them.