search

bSI-InfraRoom / IFC-Specification

IfcDoc baseline to produce the documentation
24 stars 25 forks source link

IFC4X4 issues #807

Open iegorychev opened 4 months ago

iegorychev commented 4 months ago

Hello, found following:

  1. Double semicolon appears in current IFC4X4 schema sometimes in such places: Values : ARRAY [1:SELF\IfcVoxelData.GridSize] OF IfcInteger := IfcListToExpandedArray(ValueData,1,SELF\IfcVoxelData.GridSize,SELF\IfcProduct.Representation.Representations[1].Items[1]\IfcVoxelGrid.Voxels);; This looks incorrect from the point of view of grammar.

  2. Voxels : ARRAY [1:NumberOfVoxelsXNumberOfVoxelsYNumberOfVoxelsZ] OF IfcBoolean; Again, NumberOfVoxelsX (or NumberOfVoxelsY and NumberOfVoxelsZ) are unset, so ARRAY can not be instantiated as upper bound is indeterminate. Looks like LIST or combination of LIST+ARRAY.

  3. CorrectGridSize : SIZEOF(SELF\Voxels) = NumberOfVoxelsX NVL(NumberOfVoxelsY, NumberOfVoxelsX) NVL(NumberOfVoxelsZ, NumberOfVoxelsX);

SIZEOF(SELF\Voxels) - doesn't look as qualified_attribute (SELF\entity_ref.attribute_ref), seems should be just SIZEOF(Voxels) as this isn't redeclared (as in other places within schemas).

SergejMuhic commented 4 months ago

To 1, that semicolon slipped through because I tested the expression only and IfcDoc when generating the schema automatically ads it. Fixes here. To 2, this is allowed according to grammar. Checked and validated with an author of EXPRESS: image there will be an update to this part though, if you mean the optional version of "NumberOfVoxelsY and NumberOfVoxelsZ". NumberOfVoxelsX is set since it is not optional. Could go either way here, calling NVL or making them mandatory. I lean towards the latter i. e. making them mandatory. To 3, Indeed, have we not had this issue already? Apparently with the merges it got reverted.

Thanks. Will correct and commit. But please wait for the merges as two PRs are still open. After I merge, I will also run the intepreter.