Exception using IFC4.3ADD2 Bridge

[ehtick]

need to check it is available in the tree, with ifc 4.3Add2 they can exist only in the root.
if (localPlacement.PlacementRelTo != null && Nodes.ContainsKey(localPlacement.PlacementRelTo.EntityLabel)) //resolve parent

in IFC4.3ADD2 samples this gives exception if no check on dictionary

[Ibrahim5aad]

I would consider this fix just a way around the fundemental issue that causes the exception, which is that the linear placements are not processed into the placements tree.

So if a local placment has a linear placement as PlacementRelTo it will throw exception. But it should be addressed with our IFC4x3 work as we now support processing of linear placements. (prerelease package: https://www.nuget.org/packages/Xbim.Geometry.Engine.Interop/6.1.801-netcore)

[ehtick]

I would consider this fix just a way around the fundemental issue that causes the exception, which is that the linear placements are not processed into the placements tree.

So if a local placment has a linear placement as PlacementRelTo it will throw exception. But it should be addressed with our IFC4x3 work as we now support processing of linear placements. (prerelease package: https://www.nuget.org/packages/Xbim.Geometry.Engine.Interop/6.1.801-netcore)

The code is a simple check on null pointers, often ifc models are not correct and can lead to exceptions.
Better is to have a kind of error log or validation on the model.

Will try the new engine on ifc 4.3 models to see if there are any failures.

[Ibrahim5aad]

I would consider this fix just a way around the fundemental issue that causes the exception, which is that the linear placements are not processed into the placements tree.
So if a local placment has a linear placement as PlacementRelTo it will throw exception. But it should be addressed with our IFC4x3 work as we now support processing of linear placements. (prerelease package: https://www.nuget.org/packages/Xbim.Geometry.Engine.Interop/6.1.801-netcore)

The code is a simple check on null pointers, often ifc models are not correct and can lead to exceptions. Better is to have a kind of error log or validation on the model.

Will try the new engine on ifc 4.3 models to see if there are any failures.

I got your point (There is already a PR that tries to address this with extra logging), I was just making a point in the scope of this PR which is IFC4.3ADD2 placements; in which I would still assume that the fundemental issue was the unprocessed linear placements.

I would also avoid updating packages as it is out of scope of the PR.