I’ve been digging into this and here’s what’s happening. I check the validity of the original mesh minus the extracted faces and the extracted mesh. If either of them is not valid the command silently fails. I realize this if kind of stupid based on the meshes we need to work with (I 3d print RC airplanes as a hobby, see https://3dlabprint.com/, for example). So I have a few suggestions and you can let me know which direction to go.
A) Leave the command as is but give the user feedback as to why it failed. This is my last choice.
B) Check the input mesh, ie. the mesh we’re extracting faces from. If it is valid then do validity checks on the results. Usually if the input mesh is good the output will be too. Give feedback if it fails.
C) Don’t do any validity checks, or at least don’t use them to determine the command flow. I think this goes a little against the “don’t add crap to the document” policy but here I think it’s warranted. I could give feedback if either of the meshes aren’t valid. This would be my first choice.
D) Do something else altogether. Give me your thoughts about the choices above or what you think might be best.
In the meantime, I’m going to hook up option C but I’ll follow this thread. I’ll make a YT issue for the change and post it here soon.
I hooked up option C above. If we have a new public build for 9.x tomorrow it should be in there. Try it and let me know if you want it some other way.
Thank you! I think what you did is a great choice.
The use case is that I have a big mesh with a local problem, but ops on the whole mesh take forever so I want to cut out the bad bit to fix it, and now I can do that.
If it helps, I rationalize that this isn’t propagating crap or creating new crap, it’s allowing me to isolate the crap for repair or replacement.