Merge bitcoin/bitcoin#24976: netgroup: Follow-up for #22910

e5d183151709ab59d2fa6fe9e0243000e8d6abbe [netgroup] Use nStartByte as offset for the last byte of the group (dergoegge)

Pull request description:

  This addresses my review [comments](https://github.com/bitcoin/bitcoin/pull/22910#discussion_r856095896) I left on #22910.

  This has no effect on the current logic as `nStartByte` is only used for internal addresses which only ever add 10 whole bytes to the returned group. However to avoid future bugs, I think we should use `nStartByte` as offset for the last byte as well, in case we ever add a new address type that makes makes use of `nStartByte` and adds fractional bytes to the group.

ACKs for top commit:
  jnewbery:
    Code review ACK e5d183151709ab59d2fa6fe9e0243000e8d6abbe
  theStack:
    Concept and code-review ACK e5d183151709ab59d2fa6fe9e0243000e8d6abbe

Tree-SHA512: 4c08c7d6cb38b553e998798b3e3b790177aaa2141a48e277dfd538e01a7fccadf644329e93c5b0fb5e7e4037494c8dfe061b94eb52c6b31dc21bdf99eb0e311a
This commit is contained in:
fanquake 2022-05-04 18:57:49 +01:00 committed by pasta
parent 656f525855
commit 1288494d4a
No known key found for this signature in database
GPG Key ID: E2F3D7916E722D38

View File

@ -71,7 +71,7 @@ std::vector<unsigned char> NetGroupManager::GetGroup(const CNetAddr& address) co
// ...for the last byte, push nBits and for the rest of the byte push 1's
if (nBits > 0) {
assert(num_bytes < addr_bytes.size());
vchRet.push_back(addr_bytes[num_bytes] | ((1 << (8 - nBits)) - 1));
vchRet.push_back(addr_bytes[num_bytes + nStartByte] | ((1 << (8 - nBits)) - 1));
}
return vchRet;