Having coinbase TX troubles on RegTest

[McStone]

Hi there,

I'm playing around with some self written mining software (using GPU) on my two node regtest network.


While submitting the first 499 blocks from my miner SW works fine, I've got the 500th block denied:

ERROR: ProcessNewBlock: AcceptBlock FAILED (bad-cb-height, block height mismatch in coinbase)

I've looked into it, and I guess that my height in the sigscript is wrong. Too bad I can't figure out the problem, as it all seems fine for me.

Block 499 coinbase tx serialized data: height is given as f3010000

02000000010000000000000000000000000000000000000000000000000000000000000000fffff fff0a04f30100000401000000feffffff0140be4025000000001976a914cf2a342892313764d1e3 8d790bf37cf3a25f3f0688ac00000000

Block 500 coinbase tx serialized data:height is given as f4010000

02000000010000000000000000000000000000000000000000000000000000000000000000fffff fff0a04f40100000401000000feffffff0140be4025000000001976a914cf2a342892313764d1e3 8d790bf37cf3a25f3f0688ac00000000

any thoughts / hints from your side??

Cheers & thanks

McStone

[1714756492]

1714756492

[1714756492]

1714756492

[1714756492]

1714756492

[Coding Enthusiast]

It is because you have been applying BIP-34 wrongly. The block height has to be pushed with the minimal number of bytes:

Code:

1 -> OP_1
2 -> OP_2
...
499 -> 0x02f301
500 -> 0x02f401

Since BIP-34 activates at block 500 on RegTest, this is ignored on all blocks from height 1 to 499 and from block height 500 it starts being enforced and the invalid push is rejected.

[McStone]

Thanks Coding Enthusiast!

That indeed solved the problem.

Seems there is no difficulty adjustment on a regtest network.
As far as my research goes, there is also no way to trigger difficulty adjustment on regtest?
Or do you know some way to achieve this, @Coding Enthusiast?

Cheers

McStone

[Coding Enthusiast]

Since fPowNoRetargeting is set to true for RegTest the nBits (target) never changes in blocks. I don't think you can change this anywhere apart from editing the code itself.

[McStone]

Thanks again,

this confirms what I thought...