The issue happens when a wall has a "ATPATH" connection and has a
negative direction sense. It mixes the profile points in an incorrect
order. This commit seems to fix it, but maybe there is a better solution.
In some cases remove_representation got 2x times faster (when there were a lot representation items that are not IfcTessellatedFaceSets).
getattr with attribute that might not be present on the element could be much slower then checking it's class explicitly.
If user has fake-bpy-module for type hints, it will fail with `ImportError: cannot import name 'Vector' from 'mathutils' (unknown location)` instead of `ModuleNotFoundError`.
To support obtuse x_angle a refactoring had to be made, which helped
improve the general code for slab addition.
This is challenging because there are a few features that interact
with each other to create slabs, like `depth`, `direction_sense`, `offset`
and `x_angle`. In addition, these interactions can happen in different parts
of the code. This refactor improves the coherence between those different parts.
Files changed:
- `api/geometry/add_slab_representation.py`
- `model/slab.py` - inside the function `change_thickness()`
File to be changed in a following commit:
- `model/wall.py` - inside the operator `ChangeExtrusionXAngle` - To-do
1) `do_not_delete` performs best when it's set
2) also_consider when `element` related elements go first, so there will be no need to traverse all other elements to see if they cover `element`'s inverses.
check schema version instead of accessing attribute directly (hasattr under the hood is just doing getattr and checking whether it returns AttributeError), so one less IFC access
Similar thing with element.Item to access IFC just once.
Andrej730