I’ve got an issue that occurs when using Comp Geom and the Parasite Drag tool together.

Simple Cases
If I run Comp Geom multiple times with Subsurfs disabled on a geometry set (deleting the meshes in between runs), then run the Parasite Drag tool on the same geometry set, it will output all SWet values (and therefore all C_D values) as zero.

I've attached a zip file with a simple vsp file and python script to reproduce the issue:
parasite_drag_apitest.zip

The script does the following:

1. load in the simple vsp geometry
2. run Comp Geom n times in a row with Subsurfs disabled
3. repeatedly run the Parasite Drag tool, printing SWet values for the 3 components in between each run

Using this, it looks like SWet values will be zero for the next n-1 Parasite Drag runs following the n Comp Geom runs. For the screenshot below, I ran CompGeom 5 times in a row, then printed the computed wetted areas of 3 components for a series of 10 Parasite Drag tool runs. I have run this script in versions 3.31.1 and 3.36.0 with identical behavior. The GUI appears to behave in the same way.

Notes:
• I’ve tried this on multiple aircraft models, including a vsp default pod and wing combo, with consistent results
• This problem occurs regardless of whether the geometry components in the set have subsurfaces or not.
• For these simplified cases, all S_wet values are zero, or none of them are; I haven’t seen any partial failures

Complex Use Case
I found this issue while trying to diagnose a similar behavior in a more complex case. In the complex case, CompGeom is first run on multiple geometry sets. Following that, repeated Parasite Drag analysis runs produce varying values of SWet. The image here shows the SWet values for a set of four components over a series of ten parasite drag runs. On the first run, SWet values for all components are zero, but they filter in and out over the subsequent runs.

Notes:
• The SWet values of each component change, implying that only a subset of the components are getting passed to the CompGeom execution within the Parasite Drag analysis.
• Unlike the simple cases, there seem to be partial failures in which some components have zero SWet while others do not.