The BOLT #9 specification shows a 52-bit field that represents various feature flags. That should require no more than 7 bytes, or less.
When I run the
listnodes command, some of the nodes include features fields that are in the range of expected lengths, such as
808252a1 which is less than 7 bytes, but others are unexpectedly long, always beginning with the byte
80 followed by a long sequence of
00 bytes and then what appear to be the actual feature flags.
One of them looks like this:
Is this a bug of some sort, or is there some other information to be interpreted from these long sequences of feature flags?