Testnet Merged Mining Block Number

[zamgo]

Many people are interested in testing out the merged mining system before it goes live.

Namecoin testnet chain is an ideal place for, well, testing. To speed things up in regards to merged mining, may I suggest those of us active on testnet set a date and/or block number to upgrade to merged mining clients.

This should be moderately easy to accomplish, as there's only handful of people using testnet now. When merged mining testing gets into full swing, I'd hope the use increases.

comments?

[nodemaster]

Count me in. I'm already on Testnet with RC3 From my point of view we can set merged mining block number to the next block.

[smoothie]

Count me in. I'm already on Testnet with RC3 From my point of view we can set merged mining block number to the next block.

Do you think we have done enough testing to ensure no exploits exist?

What errors or anomalies have you seen if any?

[jackjack]

Count me in. I'm already on Testnet with RC3 From my point of view we can set merged mining block number to the next block.

Do you think we have done enough testing to ensure no exploits exist?

What errors or anomalies have you seen if any?

Exploits?
Isn't that the purpose of Testnet? Testing?

[nodemaster]

Count me in. I'm already on Testnet with RC3 From my point of view we can set merged mining block number to the next block.

Do you think we have done enough testing to ensure no exploits exist?

What errors or anomalies have you seen if any?

None, but it might be this is due to the fact that we can't generate any NMC on Testnet because all POW are rejected due to the small block number. Thus without upgrading testnet we can't say for sure.

[zamgo]

I'm not sure on this... but are the testnet problems because:

main.cpp: CheckProofOfWork() switches on GetAuxPowStartBlock() ... which ends up calling GetAuxPowStartBlock() in namecoin.cpp as:

Code: Select all

virtual int GetAuxPowStartBlock()
{
    return 24000;
}

But I don't see any checks on main vs testnet along the way... do we need:

Code: Select all

virtual int GetAuxPowStartBlock()
{
    if( fTestNet ) 
    {
         return 0;
    }
    return 24000;
}

?

[nodemaster]

I'm not sure on this... but are the testnet problems because:
But I don't see any checks on main vs testnet along the way... do we need:

Code: Select all

virtual int GetAuxPowStartBlock()
{
    if( fTestNet ) 
    {
         return 0;
    }
    return 24000;
}

?

Yes, IMHO this fixes it. I patched RC3 and am giving it a try on http://alpha.masterpool.eu/

[nodemaster]

@vinced: Could you please add this to RC4?

Code: Select all

virtual int GetAuxPowStartBlock()
{
    if( fTestNet ) 
    {
         return 13722;
    }
    return 24000;
}

I effectively splitted the testnet blockchain at this block. Everybody who wants to join in is more than welcome. You can add the following to your namecoins bitcoin.conf of the modified RC3 client:

Code: Select all

testnet=1
connect=78.47.40.55:18334

This will bootstrap your client from alpha.masterpool.eu. Furthermore please set up some miners on this blockchain. If you don't want to do it solo just register at https://alpha.masterpool.eu.

[nodemaster]

alpha.masterpool.eu is still running on the new splitted blockchain. No problems so far. I cloned vinced repo and added the merged mining block number on testnet. Credits go to zamgo. If you like to test and join in please use this repo:

https://github.com/nodemaster/namecoin

At the moment we're seven blocks ahead the original NMC Testnet blockchain. If things go utterly wrong, we can revert to the old testnet blockchain. If everything is going fine we should agree on a very soon block number for testnet and use a fast but smooth migration path.

[zamgo]

Looks like things are going well with the testing so far. Merged-mining-testnet is up about 40 blocks more then "old" testnet

One bootstrapping suggestion: don't use connect, use addnode:

Code: Select all

testnet=1
addnode=78.47.40.55:18334