`Volume EcalEndcapPInsert: Cannot make slices on any axis`

[veprbl]

This message appears every time when geometry is instantiated

Error in <TGeoVoxelFinder::SortAll>: Volume EcalEndcapPInsert: Cannot make slices on any axis

Not clear to me if this is causing any trouble or not.

cc @rymilton @sebouh137

[veprbl]

git bisect says:
5a7dd05 is the first bad commit

[veprbl]

@rymilton @sebouh137 Could you, please, change the geometry to place detector within a well defined envelope (box with cutout)? Currently, layers are being dumped into an assembly, which is not much different from having them in the world volume.

[rymilton]

Thanks for pointing the EcalEndcapPInsert error out. We actually saw during the merge request and thought it was an issue in EICrecon that was causing the issue (see exchange here) and thought it was fixed in an EICrecon PR.

As far as we know this error doesn't impact the simulation performance. A revamp of the forward ECal geometry is actually a pending task that will be addressed soon. We can resolve this error in that effort. If this error is urgent, we can try to fix it earlier but we weren't viewing it as high priority since the geometry is going to change anyways. Let us know your thoughts on this.

[rymilton]

Regarding the assembly vs. envelope, what's the reasoning for putting it into a well defined envelope? When I originally made the insert geometry, I frequently went to @wdconinc's software office hours and he recommended I use an assembly.

[wdconinc]

It's faster for the navigator to determine volume boundary intersections when everything is in simple envelopes. For the insert, due to the slanted and expanding proton beam pipe, that's not easy and would require some subtraction solid. At that point the gain is probably not much and not worth the pain.