XMIR format for bytes with data should be fixed

[maxonfjvipon]

Check this for more details. In general: current representation of org.eolang.bytes with data in XMIR is wrong, we miss one anonymous abstract object so it matches to the next PHI expression:

{[x -> Ф.org.eolang.string(
  a0 -> Ф.org.eolang.bytes(
    a0 -> [\Delta -> 00-00-00-00] # anonymous abstract object as argument
  )
)]}