b1pi test fails, likely caused by new TOF timewalk correction

[aaust]

Since mid-December, the b1pi test fails using the nightly-build software stack. The danarest file still looks ok, but the last step which creates the histograms throws an exception.
It can be easily reproduced:

> gxenv /volatile/halld/gluex/nightly/2025-01-03/Linux_Alma9-x86_64-gcc11.4.1/version.xml
> export JANA_CALIB_CONTEXT="variation=mc"
> hd_root -PPLUGINS=monitoring_hists /volatile/halld/gluex/b1pi/2024-12-30/Linux_Alma9-x86_64-gcc11.4.1/30480/dana_rest.hddm

JANA >>In DTOFHit_factory, loading constants...Hz  (avg.: 0.0Hz)     
JANA >>TOF: USE WALK CORRECTION TYPE 3
JANA >> Factory Call Stack
JANA >>============================
JANA >> JEventLoop:OneEvent  (evnt)  --   line:695  src/JANA/JEventLoop.cc
JANA >>----------------------------
src/JANA/JEventLoop.cc:698  EXCEPTION : map::at
src/JANA/JApplication.cc:1386  EXCEPTION caught for thread 140102153754176 : std::exception
... 
JANA ERROR>> The program will quit now.

This is the main reason why we are holding off on the latest release version_5.22.0.xml. The error does not appear with version_5.21.0.xml, which is still the default:

> gxenv $HALLD_VERSIONS/version.xml
> export JANA_CALIB_CONTEXT="variation=mc"
> hd_root -PPLUGINS=monitoring_hists /volatile/halld/gluex/b1pi/2024-12-30/Linux_Alma9-x86_64-gcc11.4.1/30480/dana_rest.hddm