I think this was overengineered and a bad decision. RASTER is a
technical detail, not a semantic label (in contrast with whether it's a
title or a drawing).
Previously code wasn't shared and duplication was reimplemented in an
incomplete way. Now it's shared which means that it's also more
sophisticated (i.e. does all the linked aggregate magic).
Also the duplicate code seems super complicated especially with arrays
and linked aggregates. Moving it into a tool makes it more reusable and
we can test it more granularly in the future.
Historically we overrode the operator, so we had to reinvent native
functionality. Now we only override the hotkey, so we can reuse native
functionality. This also means that things we failed to reinvent like
duplicating unlinked animation data now works properly. This bit me when
I was doing animation work and Bonsai's duplicate messed things up.
E.g. if you would call this operator for 1 specific cost schedule, Blender would remember it's value and would keep on using it even if operator is going to be called outside this schedule
Mentioned in #6545
Override filename to be always empty, so user won't have an expectation that csvs will be saved as a single file (currently they're always saved to a directory).
Because apparently it's not invalid IFC.
IfcStructuralPointConnection documentation says that they should have a topology representation with a IfcVertexPoint but it doesn't restict this representation from having Undefined type if it does have a IfcVertexPoint as it's item.
I can't figure out a more elegant way of doing this. I.e. I don't know
how to detect whether or not the user has explicitly set their own snap
default.
There's now operator to check that all profiles, styles, materials in the current project have a unique name. Which is important since it's used when we append assets to ensure there are no duplicated assets.
Operator location - https://i.imgur.com/bvXjZ8Y.png
Example output:
```
IfcMaterial name 'KnownUnknown' is used by multiple assets:
- # 260=IfcMaterial('KnownUnknown',$,$)
- # 1220=IfcMaterial('KnownUnknown',$,$)
Found issues validating IfcProfileDef assets.
IfcProfileDef name 'DEMO-C' is used by multiple assets:
- # 266=IfcCShapeProfileDef(.AREA.,'DEMO-C',$,0.2,0.1,0.0015,0.03,0.005)
- # 261=IfcIShapeProfileDef(.AREA.,'DEMO-C',$,0.1,0.2,0.005,0.01,0.005,$,$)
Found issues validating IfcPresentationStyle assets.
IfcPresentationStyle name 'Frame' is used by multiple assets:
- # 363=IfcSurfaceStyle('Frame',.BOTH.,(# 362))
- # 1221=IfcSurfaceStyle('Frame',.BOTH.,(# 1222))
```
I'd much rather not expose this option to ifcpatch but I guess models are different and there should be a way to extract elements even if assets name uniqueness is not ensured (since it's not IFC requirement anyway). But it may cause duplicated assets.
Set style name explicitly, this way what user does in UI is what they get in IFC.
Previously confusingly if you set style name to "Glass" and "Glass" already existed, you would get "Glass.001" style internally but "Glass" name in UI.
Just to make sure user will set what was actually saved to IFC right away. To prevent confusion when they see it only after model was reloaded completely.
The problem was that during class assignment, it was setting the object name and `is_renaming` flag but only then linking to IFC. Since object wasn't linked to IFC, name callback wasn't triggered and `is_renaming` wasn't reset and first rename wouldn't work.
Now it's just not setting the flag if object is not linked to IFC, as there's no callback then and don't need to protect from it.
Use temp_override or modify selection not to modify some object by accident, because this operator can also be called not from UI but by providing `object` argument.
Massimo Fabbro
