There's no way we can change upstream AMP to 32-bit prefixes; it's an incompatible change. We could propose amp32 as an additional protocol, or we could write negotiation code that starts on 16-bit and switches to 32-bit if supported.
The former would be fine: it's something we can try though it may be rejected, but it doesn't block this branch.
The latter is, at a guess, several days work instead of ~6 hours that went into this. I'm okay to do it, but I guess we've got more important things to do.
There's no way we can change upstream AMP to 32-bit prefixes; it's an incompatible change. We could propose amp32 as an additional protocol, or we could write negotiation code that starts on 16-bit and switches to 32-bit if supported.
The former would be fine: it's something we can try though it may be rejected, but it doesn't block this branch.
The latter is, at a guess, several days work instead of ~6 hours that went into this. I'm okay to do it, but I guess we've got more important things to do.