You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
OCP 8.0.1: BRepCheck_Result.Status() and Bnd_Box.Get() return unbound types #232
Two OCCT return types have no binding in OCP 8.0.1, so the methods that return them raise instead of returning. Both worked on 7.9.3.1, where they came back as BRepCheck_ListOfStatus and a 6-tuple.
On cadquery-ocp-novtk==8.0.1.1.0 (macOS arm64, Python 3.13):
TypeError: Unable to convert function return value to a Python type! The signature was
(self: OCP.OCP.BRepCheck.BRepCheck_Result) -> NCollection_List<BRepCheck_Status>
TypeError: Unable to convert function return value to a Python type! The signature was
(self: OCP.OCP.Bnd.Bnd_Box) -> Bnd_Box::Limits
OCP.collections has no list of BRepCheck_Status (the only *Status instantiation is DataMap_AIS_InteractiveObject_AIS_GlobalStatus), and OCP.Bnd has no Limits. These look like the same gap as #217 (an instantiation the generator didn't pick up) and a nested struct that isn't exposed.
The first one costs the most: Status() is the only way to read whyBRepCheck_Analyzer rejects a sub-shape, so on 8.0 a checker can say "invalid" but not "self-intersecting wire" or "unclosed shell". The workaround we use in text-to-cad re-runs the per-kind checks (BRepCheck_Shell.Closed()/.Orientation(), BRepCheck_Face.IntersectWires()/.ClassifyWires(), BRepCheck_Wire.Closed()), which return a single BRepCheck_Status and still bind, but it misses whatever those don't cover. For Get(), CornerMin()/CornerMax() give the same numbers.
Thanks @jdegenstein, the GetXMin… route works for us too.
One more of the same kind, which we hit porting a downstream app to 8.0.1: the EmissiveFactor field of XCAFDoc_VisMaterialPBR (an NCollection_Vec3<float>) isn't bound anymore. The field is still in OCCT 8, since DumpJson prints it, but there's no attribute to read or set, so the emissive colour of an XCAF material can't round-trip through Python:
importiofromOCP.XCAFDocimportXCAFDoc_VisMaterialPBRpbr=XCAFDoc_VisMaterialPBR()
dump=io.BytesIO()
pbr.DumpJson(dump)
print('"EmissiveFactor"'indump.getvalue().decode()) # True: the field is thereprint(hasattr(pbr, "EmissiveFactor")) # 7.9: True, 8.0.1: False
On 7.9.3.1 both lines print True. On 8.0.1.1 the second prints False. Presumably it's the same cause as the others: the member's type is now an NCollection template instantiation.
Activity
jdegenstein commentedon Oct 21, 2026
We ran into this as well, but used e.g.
GetXMin ... etcsee here https://github.com/gumyr/build123d/pull/1442/changes#diff-899489695e9e952801f560696ff1e4a1c13e7375ebbb91d1a841f7f0deec8623R1124Sean-yh commentedon Oct 25, 2026
Thanks @jdegenstein, the
GetXMin… route works for us too.One more of the same kind, which we hit porting a downstream app to 8.0.1: the
EmissiveFactorfield ofXCAFDoc_VisMaterialPBR(anNCollection_Vec3<float>) isn't bound anymore. The field is still in OCCT 8, sinceDumpJsonprints it, but there's no attribute to read or set, so the emissive colour of an XCAF material can't round-trip through Python:On 7.9.3.1 both lines print
True. On 8.0.1.1 the second printsFalse. Presumably it's the same cause as the others: the member's type is now anNCollectiontemplate instantiation.