The result of innerMaterial and outerMaterial depends on if the other function is called before

[jmcarcell]

Check duplicate issues.

• Checked for duplicates

Goal

When calling outerMaterial() after innerMaterial() has been called, the result is different for some surfaces than when outerMaterial() is called first. This is happening because the TGeoManager keeps some internal state that changes after innerMaterial() is called, leading to it not being the same when calling outerMaterial(). For CLD_o2_v07, it seems that about 20 surfaces are affected, so it seems it's not a very large effect.

Operating System and Version

AlmaLinux 9

compiler

GCC 14

ROOT Version

6.32.08 and 6.34.04

DD4hep Version

master

Reproducer

source /cvmfs/sw-nightlies.hsf.org/key4hep/setup.sh
python issue.py
python issue.py --call-inner-material

If we now diff the two output files

diff surfaces_good.txt surfaces_bad.txt

We will see that in one file some surfaces are Silicon and in the other Air. For example:

< 1 Silicon
---
> 1 Air

The file issue.py:

import os
import argparse
import dd4hep
import DDRec

argparser = argparse.ArgumentParser()
argparser.add_argument('--call-inner-material', action='store_true', default=False)

det = dd4hep.Detector.getInstance()
det.fromCompact(f'{os.getenv("K4GEO")}/FCCee/CLD/compact/CLD_o2_v07/CLD_o2_v07.xml')

args = argparser.parse_args()

sm = det.extension[DDRec.SurfaceManager]()
multimap = sm.map('world')
d = {}
for k, v in multimap:
    if args.call_inner_material:
        innermat = v.innerMaterial()
    outermat = v.outerMaterial()
    d[k] = outermat.name()

with open(f'surfaces_{"bad" if args.call_inner_material else "good"}.txt', 'w') as f:
    for k, v in d.items():
        f.write(f'{k} {v}\n')

Additional context

No response