With bulging in 3.0.4 option Config.PROFILE_CONTOUR_REFINEMENT does no longer work as expected

[GRPlan]

In core 3.0.4 a better approximation for bulging of free surfaces was introduced.

This new function has an effect on the Config.PROFILE_CONTOUR_REFINEMENT that was introduced in issue #177.

In some cases the higher more uniform distribution of sampling points on the contur is not applied.

[ChRen95]

The problem resides in the usage of the segmentize function of shapely which by design only works on linear components.
Can you provide an MWE so I can reconstruct the error?

[GRPlan]

MWE_304.ipynb

[ChRen95]

So i looked further into the problem.
Again what seemed easy is quit a hustle in the end.
The problem is that we use refine_cross_section on every Profile due to the structure of PyRolL, the used hook system and so on, the function is sometimes not called. One could force usage by refining every profile, meaning every profile in every iteration loop. This is very resource hungry when applied to such a great scale.
Maybe I can find a easier solution but it's not something that can be expected to be solved soon.