add Capella support to Forked*

[tersec]

• a large no-op, essentially, the goal of which is from this point to allow safe, incremental Capella development/fleshing-out within the unstable branch
• general pattern was to flesh out spec/forks to its almost Capella-final form, but the rest to stub where feasible, to preserve the no-op aspect in a PR which already in an essentially minimal form adds 500 lines and changes 42 files
• capellaImplementationMissing serves both as a grep (rg, ag, etc) anchor, but also as a compile-time anchor. Once the capellaImplementationMissings in an area of the codebase is complete, and it compiles without it, it implies that the Capella aspects should be done there
• where the data is under internal control, it does raiseAssert $capellaImplementationMissing to be as visible as possible, because after all, checkForkConsistency() ensures that nimbus_beacon_node can't run yet with a Capella network:
  

  nimbus-eth2/beacon_chain/spec/datatypes/base.nim

  Lines 955 to 958 in 3ef09ff

  	func checkForkConsistency*(cfg: RuntimeConfig) =
  	doAssert cfg.CAPELLA_FORK_EPOCH == FAR_FUTURE_EPOCH
  	doAssert cfg.SHARDING_FORK_EPOCH == FAR_FUTURE_EPOCH

However, beacon_chain/spec/eth2_apis/eth2_rest_serialization has places where otherwise, this would allow random unauthenticated external input to crash Nimbus at will, obviously suboptimal -- therefore it returns as a more controlled parsing error, but still both greppable and dependent on this symbol that allows compile-time checking. When writing, it stays with the raiseAssert, because that is under internal control and if that happens, that's a logic error.

• There's a change between Nim 1.2 and 1.6 whereby

proc f() =
  raiseAssert "foo"
  echo "hello world"

changes from an error around the unreachable echo statement in Nim 1.2 to a warning. Therefore, in several places, for compatibility between Nim 1.2 and 1.6, this PR uses:

proc f() =
  if true: raiseAssert "foo"
  echo "hello world"

[github-actions]

Unit Test Results

         9 files  ±    0       849 suites  ±0   23m 31s ⏱️ - 1m 7s
  2 481 tests +  96    2 236 ✔️ ±0  245 💤 +  96  0 ❌ ±0 
10 970 runs  +144  10 653 ✔️ ±0  317 💤 +144  0 ❌ ±0 

Results for commit fa830d1. ± Comparison against base commit 3ef09ff.

[arnetheduck]

time for OnForkyBlockAdded I think

[tersec]

This isn't something I do routinely now, since nim-lang/Nim#18095 still isn't fixed and it's a gamble each time whether attempting to do so will have been a waste of time, but in this case, bf785d6 seems to work on both Nim 1.2 and Nim 1.6.

[arnetheduck]

cruft

[tersec]

b91875d