Based on D9342, we can potentially register the proof earlier depending on the proofs ValidationState at init.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Mar 23 2021
Mar 22 2021
rebase on master after landing D9342
drop the if
I suspect that the if (!g_avalanche) test will never be true. In init.cpp, g_avalanche is set in step 7 before the node actually starts connecting to peers (step 12).
move & add comments
Mar 19 2021
address most of the review items
rename "Challenge" -> "Sighash" to keep the c++ and python code consistent
fix typo: I accidentaly turned MSG_CMPCT_BLOCK into a tuple by appending a comma to the line
must fix another test broken by change to CInv
Mar 18 2021
add a basic test
I just noticed that buildRemoteChallenge is already implemented in the test framework and is called get_sighash: https://reviews.bitcoinabc.org/source/bitcoin-abc/browse/master/test/functional/test_framework/messages.py$998
add a test (contributed by @Fabien)
rebase on top of D9329 to factor out the remote challenge generation
rebase after merging D9306
Mar 17 2021
implement missing step in the avalanche handshake:
after receiving a avahello message, a node requests the proof corresponding to the proofid in the delegation,
and now the other node responds to this request with its proof.
don't forget to define the AVAPROOF net msg type
nits
Mar 16 2021
fix build failure
if proof validation fails due to missing UTXO, don't give up. Maybe the UTXO will be in a future block.
Mar 15 2021
Mar 11 2021
In the summary, change PR21274 -> core#21274.
rebase after landing D9290
fix the discrepency for max number of stakes between the node and the test, increase the max number of stakes to 1000.
My quick calculation gives me 178 bytes per signed stake. We can ignore the rest of the proof, as it is does not change much depending on the number of stakes.
That is a huge number of stakes in 2 MB: 11781.
Mar 10 2021
reap the benefits of the simpler API, and remove uneccessary const
address comments: add a QUORUM_NODE_COUNT=16 to test, return veector of peers instead of passing a reference
Mar 9 2021
@bot build-tsan
specialize Misbehaving messages: unsolicited-avahello, unsolicited-avapoll and unsolicited-avaresponse
Mar 8 2021
minor optimization: don't do and additional check for each message
Mar 7 2021
Mar 5 2021
fix functional test by sorting the list of node ids before comparing with range(1, 17)
Mar 4 2021
change icon to something a bit more readable
It is not very readable. But I created the diff anyways to start discussing solutions. I know nothing about HTML.
apparently the additional NodeSet index is not necessary. You can do a partial search on only the first key in a composite_key.
Mar 3 2021
try not to loop over all nodes for each peer just to find its node.
address review: fix typo in RPC command name in help message