you'd be interested in making a pull request, would you?
Not really, because now
I can mine whole coins in signet, and it already has all things, which testnet5 would have.
But good luck, maybe someone else will do that.
with POW "we" (people that agree on a rule change) don't actually have to have Core merge anything, do we?
Only if it will be compatible with what we have now. Which means, that you can for example reject all blocks with timestamps bigger than 20 minutes from the last block, except difficulty adjustments. Then, blocks could have just "max +20 minutes", and the last block, which adjusts the difficulty, can use the real time, because it will have ASIC difficulty anyway.
In general, it looks like that:
Hard-fork: You need 99.99% network support, quite difficult to get it right. This is what your code does now. The chain is splitted into testnet4 and testnet5.
Soft-fork: Something like 80% or 90% network support, old nodes will follow your chain. Everyone uses the same chain, called testnet4, and testnet5 is not created, as long as miners are on your side.
No-fork: Just 51% hashrate on your side, and you are good to go. No user will need to update any software, only major miners have to do that. All CPU blocks will finally disappear, even if they could be visible for a while by some old nodes.
We just need 80% of 2016 blocks to pass a new rule... as I understand(?)
The more support you have, the easier it will be. But for the safest no-fork option, you need just 51%. And then, nobody else needs to change anything, and users could use the old version, it will work for making transactions. Also, old users could still produce blocks with 20 minutes and one second, but they will disappear, when the next ASIC block will come.
I attempted to open a PR, but forking appears to be disabled on the repository, so I can’t submit it directly.
You can open a Pull Request to your own repository, not to Bitcoin Core. It is all about having a space to put comments.
For example:
https://github.com/stwenhao/bitcoin/pull/1See? I don't have any special rights, to touch Core repository. But, as you can see, I can create Pull Requests inside my own repository. And you can do the same thing:
Just pick "l0rdicon/bitcoin" in your source and destination, and pick your own branch.
No need to make a pull request until there's some discussion in the mailing list, in my opinion.
If you pick a no-fork route, then you won't have to ask Bitcoin Core about anything.
Changing Bitcoin Core is a very slow process, and you don't want to rush and end up being ignored.
You won't be ignored, if your ASIC block will keep reorging all recent CPU blocks, and if it will lead to lowering the difficulty back to the real levels. And if other ASIC miners will do the same, when you will have 51% hashrate on your side, then all old nodes will follow your chain.
No need to make a pull request until there's some discussion in the mailing list, in my opinion.
There was a discussion:
https://groups.google.com/g/bitcoindev/c/iVLHJ1HWhoUThere is no need to create yet another topic for the same thing:
https://groups.google.com/g/bitcoindev/c/Jsv1VYpewuUThe main reason, why nothing happened, is that nobody cares. If someone wants stable mining, then
signet Proof of Work faucets can be used. And if someone wants unstable network, that cannot be centrally controlled, then buggy testnet3 and buggy testnet4 are there, so they can be used, as they are.
So, why anyone should care about fixing testnet4, if signet already can do the same things, without any fixes? Also, for many people, blockstorm is not a bug, but a feature, because they can get things confirmed faster, or they can make more coins than usually, or even they can put more data into the blocks, than normally (up to 24 MB per second).
The good thing with this proposal is that we're "removing features" and making it less complicated, so it should require less effort than otherwise.
If the new chain would be started from scratch, then yes, it would be "less complicated". But if the code is "use old rules for 150k blocks, and then use new rules", then it is more complex, than before.
But, I think this is not the right approach, as it'd render the non-updated Core clients vulnerable to reorg attacks.
It is testnet. And existing Core clients are vulnerable anyway, because a single ASIC block can reorg a lot of CPU blocks. The only reason, why it is not happening here and now, is because ASIC owners don't have a working binary, which would do that by default. And if they use the Bitcoin Core implementation, then they are accepting CPU blocks, and setting them in stone, which makes the attack worse and worse.
If someone mines a block with CPU difficulty, then it is million times easier to produce, and it is unsafe to bet on it anyway, until the next ASIC block will set it in stone. So, it won't make the situation worse, than it is, because the current CPU miners are vulnerable anyway.
Not to mention, that you can still fork some nodes out of the network, if they don't enforce BIP94 rules properly.
Imagine that you run a client that constantly receives difficulty=1 blocks and validates them, and a couple of hours after, your chain is reorged due to the 51% hashrate ignoring your blocks.
This is exactly what you need, to make it a no-fork, and don't care, what Bitcoin Core will do. Also, there are benefits in that approach, because then, old versions could be still used to test, if your block is valid or not. And if you can spend 2^32 hashes, to fully validate any block, then many CPU users will be willing to pay that price.
It'd be a lot less painful to just modify Core.
No, because there is no consensus, how testnet5 should look like. The latest proposal with premine was rejected:
https://bitcointalk.org/index.php?topic=5543921See? In May 2025, people thought, that it is "coming soon". And now we have January 2026, and nothing happened.
Also, think about it: signet already is beyond the first halving. Which means, that at least 10.5 million coins, if not more, are already taken by signet developers. So, why should they create a new testnet5, which will be as centralized, as the current signet?
Not to mention, that you can mine whole coins from signet on a CPU, just like I do. So, what else do you need?
I think it remains a problem that should be fixed in a test network.
There is no reason to make a new testnet5, which will be similar to signet.
But good luck. I wanted to fix it many months ago, and I figured out, why fixing it is not needed, or why it can bring more problems, instead of solving them. So I decided to focus on other things. But good luck, you can try to improve things, if you know how.
