DPL: write support for TTree using arrow::Dataset API

[ktf]

No description provided.

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1

[ktf]

@sawenzel@jackal1-66 do you understand the errors about

545/549 Test #506: o2sim_checksimkinematics_G3 ..........................................................Subprocess aborted***Exception: 1.48 sec

which are starting to appear since around yesterday?

Do you also understand the other one about the missing Pythia8.h?

[ktf]

Does not touch anything in production just a test (which works correctly). Merging.

[jackal1-66]

I'm running a test on o2sim_checksimkinematics_G3. I see that it fails on the TPCtracks when checking for the trackref size, but I want to see if this fails also locally (and understand why the mcreader it's getting a 0). Regarding the Pythia8 issue instead I have no clue right now. The only recent modification in GeneratorPythia8.h is unrelated to this. Were there some updates on the building machine side instead? Locally on my slc9 pc I don't get this issue.

[jackal1-66]

After talking to @sawenzel I opened #13721 to fix the issue with the missing Pythia8 header. Still investigating on the o2sim_checksimkinematics_G3 problem instead.