Huge speed-up for loading meshes with loadObj function.

[ferreram]

Added hash function for Vertex / TexturableVertex. Fix errors in TexturableVertex struct (uv init with glm::vec3 & tangent data not checked for equality operator). Simplify default constructor of Vertex / TexturableVertex & equality operator of those & Face struct.

[patrykcieslak]

Hello @ferreram ! I appreciate your contribution but sadly something is not working well :(
When I tried to merge it to the master and run some simulations, the meshes are messed up completely (see screenshot).
I export meshes using Blender as triangulated and with normals and UV coordinates.

[ferreram]

Hi @patrykcieslak. Sorry for the inconvenience, I forgot a minus one to access vector elements from the IDs of vertices / normals / uvs in the obj files. It should properly work now.

Please, note that the speed-up is mostly significant on big models.

My results from running a few tests:

• model 1: 46895 vertices and 46895 normals -> before: 0.056 secs / after: 0.069 secs
• model 2: 150561 vertices and 150561 normals -> before: 3.1 secs / after: 0.51 secs
• model 3: 451371 vertices and 601556 normals -> before: 4536.23 secs / after: 1.5 secs