Closed neilpa closed 1 month ago
It actually appears slice
doesn't work at all, get the same error with cube(1).slice(0.5)
. I suspect something is broken with the WASM bindings.
I debugged this a bit, the failure is here in polygons2vec
.
I've got limited WASM experience, but it looks like a WASM-bridged C++ object is getting passed to the function rather than an expected JS array?
If I'm reading correctly, it's coming directly from Manifold._Slice
.
Thanks! Looks like we must be missing a WASM test example that uses slice. @pca006132 wasn't there a recent refactor of this? Maybe something got missed?
Yeah, probably something simple, will check later when I have time.
I noticed project
as the same problem. Seems like this is the culprit change - 85d71f5cef43d9b026bc67f9ddf9d64e7872a8e8
There must be something I don't understand here. I just tried
const box = Manifold.cube(1)
const result = Manifold.extrude(box.slice(0), 0.1);
in ManifoldCAD.org and it worked just fine - made a thin box as expected. What is the bug?
Weird. This works for me now too on manifoldcad.org. However, I double checked on my local build and I can still repro the issue w/out my changes in #930.
This screenshot is with the most recent changes on master (ced53e2)
I don't know if this is intended, but the reason manifold.org is working is cause it seems to have to fixed code from the pr? at least I'm able to see the new example on there
I don't know if this is intended, but the reason manifold.org is working is cause it seems to have to fixed code from the pr?
Nope and that would do it. I’m guessing one of the CI actions on PRs is inadvertently deploying/updating manifoldcad.org
Ha! This is a security issue.
I think we should only upload the artifact when it is triggered from the master branch.
Ha! Wow, that's not how I thought THAT worked. Besides, I didn't think uploading an artifact did anything - I thought only our deploy action that downloads said artifact and packs it into our Github pages deployment would actually put it on a server. I'm so confused...
Well, uploading an artifact probably does nothing, but when you have a race condition and the deployment job is running on the master branch...
Oh, that's interesting. I feel like there's supposed to be a way to ensure you only get an artifact from your same job - am I crazy or did we I mess that up?
https://github.com/orgs/community/discussions/106300
Seems a bit complicated.
It appears the
slice
function doesn't work as advertised for extracting the bottom faces of a manifold. I'm hitting an undefined error that I can reproduce with a trivial example on manifoldcad.orgProduces
Related, is there a way to use
slice
to extract the top faces?Assuming this bug gets resolved, I can work around it by mirroring and slicing the bottom, but that feels like a kludge.