this code has developed organically while debugging artdaq issues and could use a cleanup. The main issues are:
- the code uses a mix of c++ and c-style casts, with types and const not being respected. A consistent use of c++ casts would be better. Even better would be if the fragment-related objects provided casting functions themselves, though that is outside the scope of this repo.
- the module uses data members to cache data descriptions (nSamples_ , np_per_hit_, etc) which are set on reading the first valid packet, but aren't reset between runs, where they can change
- the decoding is spaghetti code. A more encapsulated set of functions would be easier to maintain and debug.
- the bounds checks aren't complete, and are interspersed with the decoding. Factorizing out the bounds checks (per fragment, per DTC, etc) would make the code easier to understand.
this code has developed organically while debugging artdaq issues and could use a cleanup. The main issues are: