usagi32
usagi32
Usually, these kind of problems require a merkle trie. It can be used for maps for O(log n) hashing, while for LogView a simple vector of historical hashes is sufficient...
@rafaelcr
There is a problem, this process is still O(n) and won't make sense if we have, let's say, 10000 blocks in the sender chain even if we are just verifying...
If i am correctly assuming that by the most recent block you meant the previous block that sent the message to the same chain. Then that's exactly how the merkle...
Doesn't the inbox already has the messages, and we just need verify them? Yeah, parent issue was about that.
I know that, but it still leaves these questions. - How will we know the last block that sent the message (i.e. block 1024). - After knowing last block we...
Yes, this backtracing helps I am not denying that. But that's considering the there is a message and current committee at block 2000. - Let's say there is no message...
Yes, we can't just assume convenient and simple facade cases. > Could you add this argument to the discussion in [#3350](https://github.com/linera-io/linera-protocol/pull/3350)? I'll do something about it.
@afck Then it also has to be inevitably included in the `BlockBody`? I was hoping that we remove the execution outcomes altogether from the `Block` as we were doing before....
Also `BlockHeader` would have to be changed with the inclusion of this new field and it's hash.