AssimpImporter: trim external texture paths

[xqms]

We encountered some further meshes from hell (in .obj format), which contain spaces after the texture path. I think trimming whitespace in the importer is a good solution to make the process more robust.

[xqms]

From what I can tell the CI failures are not my fault ;-)

By the way, a mesh with the mentioned problem can be found here:
http://ycb-benchmarks.s3-website-us-east-1.amazonaws.com/data/google/002_master_chef_can_google_16k.tgz

It has the following .mtl:

newmtl material_0
# shader_type beckmann
map_Kd texture_map.png 

(observe the space after .png)

[mosra]

some further meshes from hell

😂 Great, and thanks for the test file as well 👍 I'll distill that into a minimal thing and add it to the existing tests for assimp featurebugs.

Oh well, yes, I'm having some CI issues right now, caused by cleanups in the main projects and I'm in the middle of writing a press release and didn't have time to update all the repos yet 😅 I'll merge this once I have the CI failures resolved. I hope this doesn't need to be merged today? :)

---

By the way, (I'm not sure if I asked already, in case I did and forgot then I'm sorry), what's the project you're working on? Is it public or meant to be getting published in the future? I'm really interested in what it is.

[xqms]

I hope this doesn't need to be merged today?

No rush, we can easily work with our branch here.

By the way, (I'm not sure if I asked already, in case I did and forgot then I'm sorry), what's the project you're working on? Is it public or meant to be getting published in the future? I'm really interested in what it is.

I'm working on a scene synthesis pipeline for visual perception with deep learning - hopefully there will be a scientific publication and an open-source release quite soon ;-)

[mosra]

Wow! Keep me (and the magnum community) updated please, looking forward :)

[mosra]

Merged in 1902eb0. When adding a minimal test case for this I discovered about five other parsing OBJ inconsistencies / bugs in Assimp, but I'll pretend I didn't see any 😆

Thanks a lot for the contribution!

[xqms]

I discovered about five other parsing OBJ inconsistencies / bugs in Assimp, but I'll pretend I didn't see any

😂 👍