glad to see that at least we would have a mitigation path in case of sigop spam attack. The worried me a bit.

What is the rational on the default of 20 though ?

During current rules, if all transactions max out the MAX_STANDARD_TX_SIGOPS (4000) we can get 5 transactions in the block.

In size, the maximum those 5 transactions would be is 20000 bytes.
This is 5 bytesperop currently. (without this option)

Not sure if it is relevant though. Just interested to know why you chose 20.

Concept ACK

ut ACK

Concept ACK

utACK

Needs trivial rebase.

Nit addressed, and rebased.

Ggggggbgh

On Saturday, November 28, 2015, Luke-Jr [email protected] wrote:

Nit addressed, and rebased.

—
Reply to this email directly or view it on GitHub
#7081 (comment).

Sent from Gmail Iphone

I haven't thought hard enough yet about the value of 20 as a default, but conceptually I think something like this makes sense. Could you add a unit test or RPC test that exercises the new logic?

utACK

ACK, with nit. Do something about that help message.

Done

utACK but agree with previous nit that it would be nice to have a brief explanation of the choice of default.

re-ACK

Changed milestone back to 0.12, added "Needs backport" label as it needs backport to 0.11

utACK 45b8e27 but would be nice to see some RPC tests.

@luke-jr Not sure what you want to do here instead of rebasing, but this somehow needs to be made suitable for merge if you still want it included in 0.12

This has been backported to 0.12 via #7323, removing "needs backport" label

So, are we metacoin guys supposed to go back to p2pkh encoding - or should we just fragment our user's wallets into more inputs, before continuing to use op_multisig encodings?

There were repeated requests in this PR for a justification of the limit chosen which were ignored. It was a mistake to merge the PR without first understanding why 20 was the right number. I apologize for not being more forceful about it at the time.

For what it's worth, I've reverted to pub key hash encoding for the time being: https://github.com/17Q4MX2hmktmpuUKHFuoRmS5MfB5XPbhod/dropzone_ruby/blob/master/lib/dropzone/connection.rb#L167-L173

I'd rather not engage in a cat and mouse game on this, and I would expect that doing so would probably just further waste the resources that you guys are generally try to protect.