update tasks to respect ECON "shuffle" of ROC links

[tomeichlersmith]

The ECON's channels do not necessarily order the links according to the ROC's ordering since they are just wired as possible. (for example, the ROC0 output elinks 0,1 are mapped to the ECON eRx 3,2 in the HcalBackplane meaning that the final, decoded data has the i_link swapped compared to what we expect from studying the ROC directly with a fiberless connection.)

The ROC->ECON Channel maps are already in pflib, but we need to use them in the tasks so we are doing the appropriate mapping between decoded i_link,channel and roc,channel.

[tomeichlersmith]

Maybe adding a simple callback that can convert from (i_link,channel) space into (roc,channel) space using the information about the ROC elink <-> ECON channel mapping and which ROCs are enabled. Something like

std::pair<int,int> toROCChannel(int i_link, int channel) {
  // look up which ROC the ECON channel i_link corresponds to
  // requires knowledge of which ROCs are enabled since (for example)
  // if only ROC0 is enabled then ECON channel 2  (3) would end up as i_link 0 (1)
}

Then, when needed like in many tasks, folks could insert

auto [i_roc, channel] = tgt->toROCChannel(i_link, channel);

and do checks on i_roc and channel which are the language of the ROC configurations.

[cmantill]

Can try to test this in UVA.

[tomeichlersmith]

This should also handle the "null" shuffle for the fiberless setup where there isn't an ECON-D and link 0 (from the ROC) comes out as link 0 (after decoding). #354

[tomeichlersmith]

Replace the determine_n_links function with a Target function that different targets can implement (so that the Fiberless one does not attempt to access an ECON and the Fiberfull ones can). #360