Resolve "Validate tt+jets for Sherpa 3.0.0"

Description

• Revert changes to RSTrigger, see #490 (closed), because it broke the total cross section for RS 2->4 @PeterMeinzinger @shoeche If you need this commit to be re-reverted, we'd need to understand first why it breaks the TTBar benchmark This is now properly fixed, retaining the restructuring of the RSTrigger, but fixing a bug in the treatment of the real term.
• Fix numerical comparison-to-zero issues in AMEGIC's external state construction, which led to NaNs and crashes in the BVI part
• This includes a fix-up by @max.knobbe for the recent modifications to the massive splitting kernels

Test Plan

• Run HL_TTLHC benchmark, don't get the many issues found and discussed in #490 (closed)

Example Result

https://www.theorie.physik.uni-goettingen.de/~bothmann/sherpa-benchmarks/rel-3.0.0+ddb05cfc/HL_TTLHC/TTBAR_MCatNLO_Ahadic_Rivet/

Additional Checklist Items

• I have updated documentation as needed
• I considered if this should be cherry-picked/ported somewhere else (rel-2-x or master), and if yes, opened a corresponding issue or MR

Closes #490 (closed)