WASM warp #221
WASM warp #221
Conversation
| @@ -72,6 +73,13 @@ Manifold Transform(Manifold& manifold, std::vector<float>& mat) { | |||
| return manifold.Transform(matrix); | |||
| } | |||
|
|
|||
| Manifold Warp(Manifold& manifold, void* funcPtr) { | |||
| void (*f)(glm::vec3&) = reinterpret_cast<void (*)(glm::vec3&)>(funcPtr); | |||
There was a problem hiding this comment.
This part feels the most wrong to me. I'm slightly hazy on the distinction between a reference and a pointer, but I feel like the JS wants a pointer and the C++ is giving a reference to the glm::vec3 and I'm not really sure how to resolve that. And I'm not even sure that's the error I'm running into.
There was a problem hiding this comment.
Well, it's working now, so apparently that distinction is pretty moot. I'm honestly amazed this works.
There was a problem hiding this comment.
Well, references are basically implemented as pointers under the hood, but with a bit more restrictions (cannot be null, cannot have T&&&&&& while you can have T********, for example).
bindings/wasm/bindings.cpp
Outdated
| Manifold Warp(Manifold& manifold, void* funcPtr) { | ||
| void (*f)(glm::vec3&) = reinterpret_cast<void (*)(glm::vec3&)>(funcPtr); | ||
| Manifold out = manifold.Warp(f); | ||
| EM_ASM({removeFunction($0)}, f); |
There was a problem hiding this comment.
I was following this example: https:/emscripten-core/emscripten/tree/main/test/interop
There was a problem hiding this comment.
I kind of feel like that the removeFunction call should be called in the binding code? So in case we want to reuse some warp with js functions, we can do so with minimal overhead. This is not really C++ code anyway.
There was a problem hiding this comment.
Good call, I was following that example a little too blindly.
Codecov ReportBase: 92.23% // Head: 92.23% // No change to project coverage 👍
Additional details and impacted files@@ Coverage Diff @@
## master #221 +/- ##
=======================================
Coverage 92.23% 92.23%
=======================================
Files 32 32
Lines 3424 3424
=======================================
Hits 3158 3158
Misses 266 266 Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
| @@ -72,6 +73,13 @@ Manifold Transform(Manifold& manifold, std::vector<float>& mat) { | |||
| return manifold.Transform(matrix); | |||
| } | |||
|
|
|||
| Manifold Warp(Manifold& manifold, void* funcPtr) { | |||
| void (*f)(glm::vec3&) = reinterpret_cast<void (*)(glm::vec3&)>(funcPtr); | |||
There was a problem hiding this comment.
Well, references are basically implemented as pointers under the hood, but with a bit more restrictions (cannot be null, cannot have T&&&&&& while you can have T********, for example).
bindings/wasm/bindings.cpp
Outdated
| Manifold Warp(Manifold& manifold, void* funcPtr) { | ||
| void (*f)(glm::vec3&) = reinterpret_cast<void (*)(glm::vec3&)>(funcPtr); | ||
| Manifold out = manifold.Warp(f); | ||
| EM_ASM({removeFunction($0)}, f); |
There was a problem hiding this comment.
I kind of feel like that the removeFunction call should be called in the binding code? So in case we want to reuse some warp with js functions, we can do so with minimal overhead. This is not really C++ code anyway.
|
I tested the example at the top. We don't have a great home for JS examples at the moment, so I'm going to add this along with yours when I do #219, which is coming soon. |
* builds, doesn't run * runs! * addressing feedback
Fixes #197 (well, it's mostly fixed by earlier PRs)
This works! In the editor I tried this, which clearly created a squashed ball: