what do I need to update?

ok think i got it

.merge+ 8805

Added

sech1: I assume your PR is now merge ready from your side?

yes

I reduced it to just 2 lines of code :D

not counting comments and formatting

and jeffro's separate PR isn't important for this release?

I will be gone for a few days starting tomorrow so hopefully we can get this wrapped up tonight

<sech1> "I reduced it to just 2 lines..." <- Correct me if wrong, but sech's PR in its current state will still drop nodes who notify with incoming blocks that contain transactions with tx_extra that is too big

Not downloading, which should be covered, but incoming block notifications

no

kept_by_block bool prevents it

I'm running this PR on my nodes, and I have stable number of outgoing connections

do you have testnet nodes running this

no, only mainnet

later im going to have 1 testnet node runnin release + your patch. and another 18.2.0 . i will broadcast nfts from the 18.2.0 node and see i get banned. (i will add-priority-node=release+patch to the nft node) does this sound like a sane test? and check if i get banned?

you will not get banned even without the patch

it just drops connection without it

ah ok so the ban check is redundant, so the release+patch node 'will not have those errors we seen with release' of tx failing verification

minted 27 ordinals. sitting in mempool. made 1 normal tx. the node with release+patch see's only the 1 real tx (nobody is mining nfts on testnet atm)

connections stable on mainnet release+patch. i will leave it an hour or 2 then mine my own morbinals because all the testnet miners must be using the patch

my normal txs appear in other nodes mempools / mined blocks

hmm, my home node was failing to start. It turned out I had node.majesticbank.su in my long list of priority nodes in bitmonero.conf - this failed to resolve, and monerod was failing to start, and the error message couldn't be seen until I increased log level. Not the best UX

I've also found one more critical bug that makes it nearly impossible to sync from 0 on Windows under certain conditions (but they're rare, so we're lucky). I'll make a PR with the fix tomorrow.

I can confirm, I can sent 1kb jpegs but cent send over that

I only found this sync bug because my node wasn't starting and I tried to delete data.mdb and sync again

sech1: my apologies 🙏

the above image is 985bytes, it was broadcast/minted by a node running the tx limit patch (i wrongly assumed it would limit sending an image of 32 bytes or 255 if the image was split over all outputs). i will now mine the ordinals in my mempool - mainnet connections stable, looks good sech1