Re-implement the l1topodebug chain in Run-3 Trigger software

This is a full re-implementation of the HLT_l1topodebug chain from Run 2 in the Run-3 Trigger software paradigm.

The chain provides online monitoring for L1Topo hardware by comparing RoIB path, DAQ path, and simulated results. It can also accept events where mismatches are detected (into the L1TopoMismatches stream).

The main methods of the monitoring algorithm follow exactly the logic and structure of the Run-2 implementation:
https://gitlab.cern.ch/atlas/athena/-/blob/archive/21.1/Trigger/TrigAlgorithms/TrigGenericAlgs/src/AcceptL1TopoMonitor.cxx
but significant changes are made to the chain configuration and the histogramming code. Namely:

• All histogramming uses the Run-3 online monitoring framework (the Monitored namespace).
• The chain is split into standard Run-3 chain building blocks: input maker, a "reconstruction" (monitoring) algorithm, and a hypothesis algorithm with a hypothesis tool.
• The monitoring and the chain ported here from r21 are relevant for the Run-2 L1Topo hardware, still in use at the beginning of Run 3 while new hardware is under commissioning. Therefore, the chain is called HLT_l1topodebug_legacy_L1All and the algorithm L1TopoLegacyOnlineMonitor. Equivalent solution for Run-3 hardware without the "legacy" in the names is expected in the future.
• L1Topo (legacy) simulation is executed as part of the chain, unless it's been globally enabled to run before HLT with the doL1Sim=True option. The complete L1 simulation is never run in P1 data taking, so online the L1Topo simulation will be always part of the chain.

Closes ATR-23432