Set correctly geometry and conditions versions in tests

LHCb!3946 (merged) and lhcb-datapkg/PRConfig!278 (merged) make the selection of DD4hep geometry and conditions version more explicit and precise, so downstream tests have to be adapted.

I also fixed the way test_davinci_tupling_from_raw_data is enabled in DD4hep builds, similar to what was done with Moore!2040 (merged).

Requires and is required by LHCb!3946 (merged).

marked this merge request as draft

mentioned in merge request LHCb!3946 (merged)

• [2023-02-03 15:45] Validation started with lhcb-master-mr#6930
• [2023-02-03 16:49] Validation started with lhcb-master-mr#6934
• [2023-02-06 15:06] Validation started with lhcb-master-mr#6963
• [2023-02-07 11:15] Validation started with lhcb-master-mr#6968

marked this merge request as ready

added Configuration label

requested review from @erodrigu

Morning @clemenci, I'm going to approve this as I see the changes upstream have at least been discussed with @cburr and probably others for Lbexec and the way configuration is done. But I would welcome some info email around so that people know why we now need to have these 2 new properties as well as the usual 2 DB tags. I wonder myself why. Maybe show a slide at the next joint RTA-DPA meeting (cc @nskidmor) or indeed send a mail around? Advance thanks.

approved this merge request

added 14 commits

• 0405419b...9d75e6cc - 12 commits from branch master
• 5acfe1dd - fix exclusion for dd4hep specific test
• 50e95746 - Set correctly geometry and conditions versions in tests

Compare with previous version

The reason for the duplication is that the old and new detector description and conditions database tags are not (and cannot be) identical. One may think about using the same fields and change the values according to the UseDD4hep variable, but this cannot work in PRConfig because the same TestFileDB is used in master and run2-patches, so it is much easier to have both the run2 and run3 versions of the tags. The added advantage is that you can easily identify test files that can only be used in either DetDesc or DD4hep mode (for the time being if the DD4hep fields are not specified we use the DetDesc ones, but the plan is to drop the backward compatibility so that if you do not specify the new fields the file cannot be used with DD4hep).

For the lbexec case the problem is that you cannot put conditionals in a YAML file, so either you use two YAML files per test (one with the DetDesc tags and one with the DD4hep ones) or you have a single file with both set of tags (which does not prevent you to use two files if you prefer).

Thank you Marco for the comprehensive and clear reply.

mentioned in issue Moore#518 (closed)

With this MR I said I enable test_davinci_tupling_from_raw_data, but the test fails with

IODataManager                         ERROR Error: connectDataIO> Cannot connect to database: PFN=mdf:root://eoslhcb.cern.ch//eos/lhcb/point8/lhcb/data/2022/RAW/TURBO/LHCb/COLLISION22/255620/255620_00150000_0000.raw FID=mdf:root://eoslhcb.cern.ch//eos/lhcb/point8/lhcb/data/2022/RAW/TURBO/LHCb/COLLISION22/255620/255620_00150000_0000.raw
EventSelector.DataStreamTool_1        ERROR Failed to connect to:mdf:root://eoslhcb.cern.ch//eos/lhcb/point8/lhcb/data/2022/RAW/TURBO/LHCb/COLLISION22/255620/255620_00150000_0000.raw

which cannot be related to my changes, so I'm inclined to have this merged at the same time as LHCb!3946 (merged) and let some expert fix the test.

resolved all threads

mentioned in issue #98 (closed)

merged

mentioned in commit 5725e113

mentioned in issue Moore#527 (closed)