Make the EM Q2 threshold (kMinQ2Limit) configurable via CommonParam

[LiangLiu212]

Motivation

The minimum Q2 for electromagnetic scattering events is a hardcoded constant in src/Framework/Utils/KineUtils.h:

static const double kMinQ2Limit = 0.02;  // GeV^2

Any study that needs a different EM low-Q2 boundary (e.g. electron-scattering validation work where the generation threshold must be varied) currently requires patching and rebuilding GENIE.

Change

Replace the constant with a small proxy object that reads EM-MinQ2Limit from CommonParam.xml [Lepton] lazily on first use and caches it for the lifetime of the program:

• KineUtils.h — kMinQ2Limit becomes extern const EMMinQ2LimitProxy, a struct with operator double(), so every existing call site compiles and behaves unchanged.
• KineUtils.cxx — the proxy implementation: one read of CommonParam[Lepton]/EM-MinQ2Limit via AlgConfigPool (lazy, so the pool is guaranteed initialised), cached in a function-local static, logged once at pNOTICE for provenance.
• CommonParam.xml — adds EM-MinQ2Limit to the [Lepton] param set with the historical default 0.02.

Backwards compatibility

Default behaviour is identical: if the parameter (or the registry) is absent the proxy falls back to the historical 0.02 GeV^2, and the shipped CommonParam.xml sets exactly that value. No call-site changes anywhere in the codebase.

Validation

An equivalent patch has been exercised in our local electron-scattering builds: tunes overriding EM-MinQ2Limit between 0.02 and ~1.2 GeV^2 via per-tune CommonParam.xml overlays show the expected generation-level Q2 cut in both gmkspl splines and gevgen event samples (e- on C12, O(10M)-event productions), while builds without an override reproduce the previous hardcoded behaviour.

🤖 Generated with Claude Code

[LiangLiu212]

I currently work on (e, e'p) analysis. I create this PR via claude code just with prompt.

Although I open this PR, I do not intend to include it in the v3.8 release.

[jtenavidal]

This work is already ongoing in #492 and was presented last week at the GENIE core meeting. It had a lot of discussion and feedback. Pablo is already working on it.

We discussed the implementation at the GENIE meeting with specific input needs from e4nu. I would recommend avoiding duplication of efforts.