Failure to build swept surface along a closed spline

[marcocecchiscmgroup]

The enclosed model tube.f3d was created with Fusion 360, simply by using the 'tube' command, which is nothing but a swept of a circle along a given curve. The curve was a closed spline.
When converting, an exception 'Geom_BSplineCurve: Weights values too small' is raised. Problem is that the weights are all 1.0, so they are ok, in number as well. The real deal is that, in the SAT representation, the spline curve is given as open. In fact, with this change in Acis.py/createBSplinesCurve(nubs, sense) the model is imported correctly:

			if (nubs.rational):
				bsc.buildFromPolesMultsKnots(		\
					poles		  = nubs.poles,		\
					mults		  = nubs.uMults,	\
					knots		  = nubs.uKnots,	\
					periodic	  = True, # False, nubs.uPeriodic,	\
					degree		  = nubs.uDegree,	\

The original value is False. On top of my head, I'd say it should have been nubs.uPeriodic, but there must surely be a good reason for that. If that reason is that you noticed that the 'open'/'periodic' tags in the acis format are not realiable, well, then I think this is yet another example. Of course the actual periodicity cannot be known before building the spline, but from the cardinality of the parameters it should probably be feasibile, i.e. n+1 control points and n+p+2 knots as described here below?
Closed B-spline

There is another problem, however, which worries me more. When I dump the sat representation of the f3d model and try to convert it directly, a step file missing any kind of geometries is produced. Does reading from sat trigger a different round of ready_to_build, None checks and the like?

tube.zip

[marcocecchiscmgroup]

My bad, I was mislead by the hard-coded False and by the fact that I knew the spline was closed: changing to True is not the real reason for the model being imported correctly, it just triggers a caught exception so that createBSplinesCurve(nubs, sense) returns None and the conversion can continue!
Instead, False doesn't immediately trigger an exception, it does it in another place which must prevent resuming, at this point.

The question about the value True / False / nubs.uPeriodic, remains, however.

[marcocecchiscmgroup]

I am afraid this is not the most effective 'fix' to do:

			elif (self.subtype == 'sweep_spl_sur'):
				profile = self.profile.build(None, None)
				if (profile):
					path = self.path.build(None, None)
					if (path):
						try:
							self.shape = Part.makeSweepSurface(path, profile)
						except:
							pass

waiting for feedback.

[marcocecchiscmgroup]

Dear Jens,

it's a pity that you are not working on this project anymore.

What are you plans?

[jmplonka]

„What are your plans?“ - From time to time, make some improovements ;)