DelphesRootReader_ee_91gev test failing

[andresailer]

2/7 Test #5: DelphesRootReader_ee_91gev .....................***Failed 8.78 sec

See log files.

[tmadlener]

The test seems to be pretty flaky. I get a roughly 30% failure rate locally. The main issue is that it's not entirely easy to reproduce, because even with a fixed random seed for delphes I sometimes get it to pass sometimes and to fail the other times.

[andresailer]

We can add a --repeat until-pass:5 to ctest then

[tmadlener]

I didn't even know that is a thing... Let me check quickly whether I find an obvious issue with the seeding just to rule out any stupid issues.

[andresailer]

Although of course it depends on why that test actually fails? Just flaky comparison, or something else.

[tmadlener]

OK. RandomSeed setting doesn't seem to be the problem. The issue with --repeat until-pass:N is that the key4hep CI workflow doesn't expose that possibility to us.

[andresailer]

That missing ctest hook is probably something @jmcarcell can fix.

[tmadlener]

It seems to be very consistent in CI, even --repeat until-pass:5 doesn't do much there: https://github.com/key4hep/k4SimDelphes/actions/runs/11147612771/job/30982426010?pr=130

c.f. #130