Changing merge-mining format for BIP9

[domob]

Bitcoin merged BIP9 last week - https://github.com/bitcoin/bitcoin/comm ... fbec6bc5ed. Miners will start to set the corresponding high bit in nVersion as soon as they upgrade (no activation time), but that's not an immediate problem for us as far as I can tell.

Nevertheless, I think we should take this as a motivation to speed up the fork process. The code is already mostly ready in https://github.com/domob1812/namecore/tree/hardfork, just missing regression tests and adding other things to the hardfork (removing the BDB lock limit, getting rid of Namecoin nVersion for transactions is my plan).

Any thoughts? Testing and review of the code is already welcome, of course (as always).

[biolizard89]

Since it looks like we need to deal with the hardfork fairly soon, I've set up a GitHub milestone for the nc0.12.0 release. These are all the issues I'm aware of that I think we should consider fixing for the nc0.12.0 release. I'd lean toward releasing the hardfork sometime after nc0.12.0 (let's say nc0.12.1), so that if we discover an issue in nc0.12.0 that prevents people from abandoning 0.3.80, we can get it fixed before the 0.3.80 users get kicked off the network.

https://github.com/namecoin/namecoin-co ... s/nc0.12.0

I'm aware that a significant number of these issues involve me. I will do my best to deal with them ASAP. If anyone disagrees with my evaluation of an issue for this milestone (in either direction), please tag me in the corresponding issue (@JeremyRand).

[domob]

Proposed PR: https://github.com/namecoin/namecoin-core/pull/75

[biolizard89]

Quote from IRC earlier today, which may be of interest.

<qpm> freenode:<btcdrak> qpm: plddr: the first BIP9 deployments are scheduled pretty soon https://lists.linuxfoundation.org/piper ... 12567.html - while the starttime is set to May 1st, code will be released before. And from the CLTV softfork, 20% of miners upgraded to master before any official release.

[domob]

Quote from IRC earlier today, which may be of interest.

<qpm> freenode:<btcdrak> qpm: plddr: the first BIP9 deployments are scheduled pretty soon https://lists.linuxfoundation.org/piper ... 12567.html - while the starttime is set to May 1st, code will be released before. And from the CLTV softfork, 20% of miners upgraded to master before any official release.

Yes - the code is actually already submitted (at least the one setting the "flagging" high bits for general BIP9-support, I'm not sure about others). As far as I can tell, this code is not yet an immediate problem for Namecoin - disallowed parent blocks are only those which either set bit 0x100 or have the higher 16 bits exactly equal to 1. Nevertheless, I think that we should try to proceed as fast as possible (without hastening too much and screwing it up). I have disabled the creation of BIP9-signalling blocks from Namecoin itself for the moment, as those would not be valid until the fork is in.