Policy inputs checks

[shargon]

Close #1937

[erikzhang]

UT failed.

[roman-khimov]

Looks good, but how about 128K limit for Block.MaxTransactionsPerBlock?

[shargon]

Looks good, but how about 128K limit for Block.MaxTransactionsPerBlock?

I think that this need more tests, it can be related with the time per block reduction issue. #375 @roman-khimov could you create a different issue for this?

[shargon]

UT Fixed

[roman-khimov]

I can, but to me it doesn't seem to be related to #375, it's just a matter of size match between PayloadMaxSize and MaxTransactionsPerBlock. With current MaxTransactionsPerBlock one block can't use all of PayloadMaxSize capacity irrespective of MillisecondsPerBlock.

[shargon]

For me it's related because both pretend to increase TPS changing the tx per block or the seconds per block, but it's more or less the same. Maybe I am wrong

[roman-khimov]

both pretend to increase TPS

I don't. I'm specifically not mentioning TPS here, it's just a matter of match between different components of the system. There is a P2P payload limit, there has to be some, the maximum possible block size should probably match it and the maximum possible block size should be practically reachable with MaxTransactionsPerBlock, that's it.

it's more or less the same

And #375 is more about latency to me. Of course, there is a natural MemoryPoolMaxTransactions/MillisecondsPerBlock ratio, but for mainnet/testnet this ratio doesn't seem to be a problem to me from the TPS perspective.

[shargon]

What do you think @erikzhang ?

public const int MaxContentsPerBlock = 134_000; // (PayloadMaxSize=0x02000000)/(AverageTxSize=250)

Or maybe we can remove this variable and compute it according to PayloadMaxSize

[erikzhang]

PayloadMaxSize is not for blocks only.

[erikzhang]

And I want to limit the average block size to 256KB.