Conversation
|
Was this causing a failed test? Looks good either way, after the merge conflict. |
I'm not sure if these are super important, but it's probably good to use these instead of using `cos()` and `sin()`.
This was the logic in the original `impl.cpp` file, but was missed during the port. This dramatically improves the performance of things with high vert counts. Most noticeable in the `Properties.ToleranceSphere` Manifold test
JaminKoke
force-pushed
the
mini_changes
branch
from
293efe4 to
c209571
Compare
|
I should have made note of this in the PR comment, but there are 2 main changes here. The first, porting Second, other changes such as with the |
|
Got it thanks. I must've falsely assumed that |
| } | ||
|
|
||
| impl<'a> Boolean3<'a> { | ||
| pub fn result(self, op: OpType) -> MeshBoolImpl { |
There was a problem hiding this comment.
I just noticed this change here while fixing some merge conflicts for my project. Consuming self was intentional here: Manifold establishes a pattern in this function of freeing memory as early as possible to avoid running out of RAM during very large mesh boolean operations. Was this a tidying task or was something else not working?
2 participants
A collection of small changes