Geant4 version compatible with macOS 26

[soleti]

Hello,

Just putting this here for people who might encounter the same problem. The recommended Geant4 version, geant4-v11.0.2, does not compile on newer version of macOS (I tried with macOS 26). However, geant4-v11.3.2 does work, so that's a possible solution. We might want to update the wiki, with the caveat that the results of the simulation will most probably change.

[kvjmistry]

Thanks Roberto, do you know if there is anything of importance in the release change? If no breaking changes then I think it's worthwhile to update nexus the new G4 version.

[soleti]

Well, the changelog between the two versions is huge, but ChatGPT made a good summary...

Yes—there are a few changes between Geant4 11.0.p02 and 11.3.p02 that can matter for NEXT-style HPXe/optical simulations. Here are the ones worth caring about, plus what to do:

1. Thin-film optics at boundaries
   Geant4 11.1 adds explicit support for thin-film coatings in the optical boundary process and the material/surface property tables (parameters + MPT ordering fixes). If you model TPB/WLS coatings, ITO, dielectric-metal stacks, etc., you can now implement them natively (and you’ll see new example macros like coated.mac). Action: revisit your G4MaterialPropertiesTable/G4OpticalSurface to use the thin-film parameters; make sure your property vectors are sorted by increasing energy (Geant4 now enforces this). 
2. EMLOW dataset upgrades (affects low-energy gammas)
   Required EM data jumped from G4EMLOW-8.2 (v11.1) to G4EMLOW-8.5 (v11.2). This touches low-energy photoelectric/Livermore models—relevant for Kr-83m lines (~41 keV), Ba-133 lines, X-rays, etc., used in calibration/validation. Action: install the new datasets and rerun low-energy validation plots; expect small but real cross-section changes. 
3. Multiple scattering/transport coupling (steping/track shapes)
   11.1 introduces G4TransportationWithMsc (enabled by default in G4EmStandardPhysics_option1) and other MSC tweaks. That can subtly change electron step limitation and track blobs/topology, which you care about in NEXT (track+blob separation, diffusion systematics). Action: fix your physics-list choice and re-tune step limits/production cuts if you rely on option1; document any AUC/ROC changes in topology classifiers. 
4. New X-ray surface reflection process
   11.2 adds G4XrayReflection using EMLOW-8.5 data. Mostly for grazing-incidence X-ray optics; probably not needed for HPXe VUV optics, but be aware if you import examples or see unexpected surface scattering at higher photon energies. 
5. Radioactive-decay time threshold change
   Default “ignore very long decays” threshold shifted to 1 year (was “~age of the universe”). If you simulate activation chains or very long-lived backgrounds in materials, you may need to raise it back via UI/C++ (docs provide exact command). Action: set /process/had/rdm/thresholdForVeryLongDecayTime 1e60 year (post-init) for studies needing essentially no cut. 
6. Minor but nice optics/infra fixes
   Warnings when a step is just larger than geometry tolerance at optical boundaries (helps diagnose boundary artifacts), and reordering/locking of optical property vectors. Action: watch logs during long optical runs; fix any boundary step issues. 
7. Hadronic model tunings (Bertini/FTF)
   11.3 retunes Bertini and adds controls to revert to 11.2-like behavior if needed. Likely negligible for pure EM/optics work, but can matter if you simulate cosmics/neutrons and their secondaries for shielding studies. Action: stick with defaults unless you validated to 11.2 and need reproducibility—then use the provided flags to match older behavior. 
8. Practical: supported toolchain bumps
   Newer CLHEP/PTL/VecGeom requirements across 11.1–11.3; rebuild your app and (if you use VecGeom) re-check navigation/overlap tests.

[kvjmistry]

Oh boy, ok I think we will discuss this at the next SW meeting

[gonzaponte]

Indeed. We need to produce data and verify that the current implementations still work as intended. This would be a good time to develop a release verification protocol/recipe.