Topic: Requesting Testnet4 tBTC - page 4. (Read 2333 times)

Actually, it should be as easy as just requesting some tBTC4 for free.

Eh, theoretically, we could use testnet4 for the same purpose, without "touching" the main net at all. It just becomes too complicated for little reason.

---

Some people, apparently, try to recreate or move Ordinals on testnet4: https://bitcointalksearch.org/topic/utxo-sat-range-on-btc-testnet4-5515748.

Even worse than that: you'll be testing things in your real Bitcoin wallet. You could of course create a new wallet, but I like to keep testing and real Bitcoin strictly separated.

If only you would have read more than just the OP of this 4.5 months old topic, you would have known OP received 1 fresly mined block months ago. Normally, I wouldn't have responded to a shitpost on the Tech boards, and just click Ignore (which I did), but it reminded me of something else: it shouldn't be necessary to justify the use of testnet coins before being able to get some! Testnet should be available for anyone who wants to use it, no matter what they use it for. And that brings me back here:

well you should give a proper insight about the project, before demanding for tBTC!

It does not allow you to test things across long periods of time. It rather sounds like a complicated way to achieve a temporary regtest network.

However, the regtest starts sounding like the most appropriate way to resolve this problem. If we don't want the units of a network to become valuable, we need to simply make these networks abundant, instead of trying to ignore scarcity of tBTC, which is economic reality.

If you have a stale block, then it allows you to test things, while at the same time, it is not part of the heaviest chain, so it quickly disappears (and any assignable value is then lost, after a single confirmation).

Because blocks with minimal difficulty are accepted in such test network. And then, when a new mainnet block kicks in, then they become invalid.

You won't "acquire" anything permanently. For example: you can use my code now, and CPU-mine some blocks on top of testnet4. Then, some time later, all of your blocks disappear (unless you will win network propagation games). Imagine having a similar feature, but on top of mainnet, where all amounts and network conditions will automatically follow the real network, so you will start with 3.125 tBTC, instead of 50 tBTC. And then, you mine some new coins, you test some new things, and then, a new mainnet block arrives, and the whole test chain is automatically destroyed, to prevent any trading.

• How does this disincentivize speculators from giving value to testnet4?
• How does it allow you to mine coins from your CPU? If we all start doing this, wouldn't it also become infeasible to practically acquire some yourself by mining with your CPU because of competition?

The spam stopped again. I thought of a "counter argument": it's going to take weeks to add that much to the blockchain (at 1 block every 10 minutes). The chain will grow, but it still won't be more than ~50 GB per year. That's annoying for a quick test, but it's manageable. Anyone else can still get their transaction in, as long as they raise the fee. When that happens, 50 tBTC won't be able to add 5 GB of spam anymore.

Just make stale blocks, on top of mainnet blocks. Then, they will always disappear, but they can be shared between interested parties.

For example, you can see the latest mainnet block: https://mempool.space/pl/block/00000000000000000002a120cece3ee55bda9db73c67dedff14650da5a581c37

And then, you mine a new block, on top of that, with your CPU:
Then, you can be 100% sure, that the block 0000000071d804b9fdd91ff6475b242dde9725b87b4d06e099922a61df828606 will disappear, because it doesn't meet the real network difficulty. But: you can also left this field unmodified, and still grind it, then you will have a small chance, to get a valid mainnet block, and a quite huge chance, to get just another test block.

And when I told people, that each 50 tBTC will give everyone the ability to put 5 GB of spam, then they ignored it, just started a new test network, without any halvings anyway, and now they are surprised, that people are spamming. They do it, because they can. The initial fees in mainnet were something around 0.01 BTC, and "coinage" allowed a lot of users to process their transactions for free, if their coins were "old" enough. Starting with the current rules of 1 sat/vB, and with 50 coins, will lead just to an altcoin, and will not reproduce mainnet conditions from 2009 (if someone wanted that, by having no halvings).

On real Bitcoin, I always assumed the data spam was meant to sell their made-up tokens to gullible people, at a profit. By my estimate, fees for data spam add up to hundreds of millions of dollars.

Maybe.

My guess is speculation on a future price increase. Compared to the fees spent on on-chain Bitcoin data spam, this is cheap.

I'm starting to think it's futile.

What not? We frequently notice this behavior in main net as well, where large fees correspond to real world, hard money. In here, you can fill blocks, essentially, for free. So perhaps Ordinal users, or whoever thought of a new bloating-chain idea, test their software in the new testnet4.

I still don't understand why people give value to large sums of the testnet4 and what we can do to disincentivize this activity. Previously, you talked about changing to CPU mining algorithm, but besides the fact that it is no longer Bitcoin, we'd still create an environment, where people would need to spend computational power, and therefore, it'd still involve some cost, which would create just another market.

Testnet4 blocks right now:

Is someone trying to just make the blockchain heavy? Blocks are filled with transactions with 997 inputs and 997 outputs. Mempool shows 0.01 tBTC fee per block, which means finding 1 block is enough "ammo" to completely fill 5000 blocks. That's 5 GB of blockchain. What's the point?

Exactly. And it works in all cases in testnet3. But: in testnet4, there is one specific case, when it wouldn't drop: during difficulty adjustment. Then, I turn my miners off, when I see a different value than 0x1d00ffff.

It matters, because if you put it 20 minutes forward, then you can make six blocks in a 2-hour window. But if you put 1 hour forward, then you can fit only two blocks. And then, when everyone is working 2 hours in the future, the whole trick stops working, and only those, who are accepting things more than 2 hours in the future, can actually mine anything, at lower difficulty.

Yes, but you can locally accept a weaker block, which will then always disappear. And by doing so, you can learn for example, if a particular non-standard transaction is valid or not. Or: if a given block exceeded sigops limit or not. Or some other things, which you can only check in a bigger context.

For anchors, introduced in the latest version, you have only this single address. But: there are many possible scripts, where no keys are required, for example: https://mempool.space/testnet4/address/tb1qft5p2uhsdcdc3l2ua4ap5qqfg4pjaqlp250x7us7a8qqhrxrxfsqaqh7jw

Because in this way, you peg coins into other networks, based on Proof of Work. This kind of locking script is just "lock to a sidechain output", which would be moved, when sidechain users would produce small enough signature, to move those coins somewhere else.

Because it is profitable. Normally, everyone would mine just Bitcoin, and nothing else, because there is no reason to mine anything else in the first place. But: because the whole game is just a "winner takes all" game, and because making altcoins, mining them, and then selling, is profitable, then that's what people do. And now, testnets are just altcoins, signed by Bitcoin Core developers, so unless the whole model behind testnet will be changed, they will be pumped and dumped endlessly, because it is profitable.

In the ideal case, you wouldn't have other mining algorithm than double SHA-256 in altcoins. And all of them would use Merged Mining, so reorging any altcoin would be as hard, as doing 51% attack on Bitcoin. And all chains would calculate "the global difficulty", to make sure, that their chain cannot be attacked, if you don't have 51% mining power in the whole crypto space.

But: instead of that ideal model, you have just completely separate chains, and you can get more BTCs by mining and selling altcoins, than by mining BTC directly. Sad, but true. And that's why people are mining testnet3, testnet4, or any other altcoin: to sell it for BTCs. And of course, some miners can have some alts in their wallet, but then, they keep them only, if they can increase their profits (for example by having bigger control over a given altcoin).

1. You wouldn't need any centralized mining pools, if it would be possible to directly mine some weak block, and get a fraction of the coinbase reward, for doing that.
2. You wouldn't need altcoins, created only because of mining. You would just mine Bitcoin, and lock your coins in a different way, to peg them into any subnetwork you want.
3. You wouldn't need any trading, to switch from feature A into feature B. You would just move your coin directly from address A into address B.

1. It could work, but nobody put enough effort, to make it real. There are only some local test networks here and there, but nothing, that you can practically use.
2. There is no need to create any dust. If you would have just off-chain transactions, paying you all, what you mined, then you could accumulate your rewards, and withdraw all of that in a single transaction, when it will be worth it. Nobody needs to withdraw a single satoshi immediately.
3. If you make some LN transaction, then it costs you some small fees. If you are a miner, you can be rewarded just in free (or discounted) transactions, inside LN.

Well, for better or worse, Garlo Nicon's code still works, and is compatible with many networks (yes, it works on testnet3, too). You can build as long local chain, as you want. Of course, all of that will disappear, but you can build such chains on top of any networks, including mainnet (if you enable minimal difficulty rule, which is just a single boolean in the code). By the way: doing it on top of mainnet has a lot of sense, because then, you have a lot of real data, to test transaction batching, checking if some non-standard transaction is valid or not, measuring your real hashrate, and so on.

In general, if you call "getblocktemplate", then you can get everything you need, to make a valid block. If you would pass all of that into some regtest client, which accepts blocks with any difficulty, then you will get those "coins appear out of thin air", you will have the chance to explore, how transactions are picked, how they are ordered, how they are selected, how long it takes, before your newly-started node will fetch enough transactions, to fill a block, and so on.

Well... I was thinking about how this "fix" actually works, and the only reasonable explanation is that it only works because setting the timestamp this far in the future virtually resets the difficulty whilst it is still an acceptable window for the other nodes...

In order to mine a valid block, you need a block header whose sha256d hash is under the current target... Let's dig a bit deeper and simplify a bit

The current blockchain is at height 100 (a fictional nuber).
If you execute this function: sha256(sha256()) you get (fictional) hash 1234567890abcdef1234567890abcdef1234567890abcdef1234567890abcdef (which, for some magic reason is under the current target).

The ONLY way you can ever generate block 101 is if you have the sha256d hash of the header of block 100 because block 101's header includes said hash: https://en.bitcoin.it/wiki/Block_hashing_algorithm
you clearly see that hashPrevBlock is an integral part of the block header of the current block you're trying to solve.

This means that, whilst we are at block height 100, it is simply impossible to pre-mine a valid block for height 102. You can only mine a block valid for height 102 once there's a valid block for height 101 found.

What the fix discussed in this thread does (if i'm not mistaking offcourse) is setting the timestamp far into the future... This means that the generate command will think more than 20 minutes has passed since the last block was found, so it'll reset the difficulty to 1 whilst it's generating... The rest of the ASICs think the difficulty is much higher and they try to find a block that's valid at this heigh difficulty whilst people running this fix are trying to solve a block at exactly the same height as the ASICs, but they're trying to solve it at difficulty 1... This means that, eventough both are working at a block at the same height, the cpu miner is solving a much easyer hash and has much better odds of solving at difficulty 1... Once the 20 minutes since the last block have actually passed, and the ASIC's are also readjusting their difficulty, the cpu miner *might* have already solved a block at diff 1, since he has already been trying for 20 minutes whilst the ASIC was trying to solve at a much higher difficulty.

So, all in all, if my theory is correct, it shouldn't matter if you change the timestamp by increasing it with 21 minutes, 1 hour or 4 hours... As long as the other nodes accept the time difference and it's bigger than 20 minutes, the effect should remain the same.

At least, that's how i think this fix actually works... And it seems like it does work (at least, it does for the testnet... it would never work on the main net since there is no diff adjustment if no block has been found for 20 minutes)

I'll quote part of that post:
I saw this yesterday in another topic:
It's the first time I've seen "Bitcoin anchors". It makes me wonder how much more is out there that I've never seen in GUI wallets.

It looks more like a puzzle than something developers should be doing.

Why would ASIC miners do that? Or better: why are they mining testnet coins in the first place?

How cool would it be if real Bitcoin has this feature? I know, I know, it won't work and will only create dust, but I still regret missing out on those early days. Best I could do was mine a shitcoin years ago, and mine testnet recently. There's some charm to seeing coins appear out of thin air.

Well, there were plans, to pick much higher initial difficulty. But then, I guess testnet4 coins would be much more valuable, and CPU miners wouldn't mine it, so some issues would never be noticed in the first place (and there would be much more trading than today, because coins would be much harder to get).

There are coins lying around, for a lot of confirmations, on keyless addresses: https://mempool.space/testnet4/address/tb1pfees9rn5nz

Also, you can try to grind signatures in addresses like this one: https://mempool.space/testnet4/address/tb1qk3endeq6x0xj4pjt4zwag8wf3a629rzqr8jxd7jnmlac902wa5ysqxm9wt

Then, ASIC miners could just allocate some coins for CPUs, and then, CPU miners can sweep them. Because if you have a hashed message, which is not aligned into 80 bytes, then you have to create a slightly different ASIC, and it would take some time, to re-design the hardware, and mine such things.

Another thing is that sometimes, some people would want to make custom block templates, and in that case, I think developers could simply "buy" a given block template, by moving some of their coins into ASIC miner's address.

There is no need for ASIC resistance. People should still be able to CPU-mine blocks, but they should simply get a smaller block reward, for doing that. For example: I can imagine CPU miners getting single satoshis, while ASIC miners will get 50 tBTC.

Also, it is very unlikely to be successful in solo-mining mainnet blocks on CPU, so different kind of mining is needed, if you want to have any chances, and avoid "mine altcoin and dump them" scheme. For example: LN-based mining could be possible, when you could sell some millisatoshis for a valid share, produced by a second party from the channel. Also, mining shares could be traded as well, and miners could receive discounted or free LN transactions.

Still, this doesn't feel right: testnet coins shouldn't only be mineable by ASIC, making developers ask for handouts. But I can't think of a solution, switching to ASIC resistant mining would make testnet too different from Bitcoin.

Scheduled testnet resets may work against gaining value, say every 6 months, but that would mean testnet will never experience any halvings again.

Yes, see the link above. And also note, that I didn't provide any binaries, only sources, and instructions. Which means, that if you have a working binary, with my changes, then you either built it by yourself, or got it from someone else.

My change is compatible with testnet4. You can use the official version, to have a working node, or you can use my version, if you want to mine blocks on your CPU.

1. If you call "generatetoaddress" or any other "generate" command, it will return you the hash of the block, when it is successful.
2. ASIC nodes produce a new block, with CPU difficulty, when for 20 minutes, there is no new block. This is when you should broadcast your block to the rest of the network.
3. You mine your block upfront, and just release it, at the right time. Nothing else is needed. If your block is first-seen, then you won. If not, then you lost. It is that simple.
4. I didn't yet write the code for network propagation wars, because I mined enough coins without it. But I know the theory behind it. However, I think about other ways of CPU mining, for example using OP_SIZE on signatures, because in case of testnets, developers will soon patch this CPU-mining hole, so it is not worth it, in the long term.

1. You prepare a block in advance, and keep it somewhere.
2. You schedule sending a P2P message to the network, at the right time.
3. You avoid all kind of "already seen block, nothing to be done" cases, where your own node will do nothing, because it already has that particular block.
4. You make sure, that you are connected with the original network, and not only with a bunch of nodes, enforcing 20 hours rule.

So, if I would do that, then I guess I would use at least two connections: one for the 20 hours node, to be a block producer, and one "proxy-only", to just serve the right blocks, at the right time.

But, as I said, I don't think it is worth the effort in the long term, so I didn't touch those parts of the code yet.

1. I am not Satoshi, that should be obvious.
2. As you can see, anyone can make changes. And then, users decide, which chain they want to run.

And, knowing the whole context of what happened in testnets, you can read this post again: https://bitcointalksearch.org/topic/m.64645672

Definitely, I didn't expect that ASIC miners will follow my changes. I didn't expect triggering over 500 blocks chain reorganization on testnet4 either. But well, I learned many new things, so I think it was worth it. Also, this is the purpose of testnet: to learn things. And my code allows that learning, by letting you to construct your local chain, and see, how much mining power you have, is your non-standard transaction valid, how chain reorganization works in practice, and so on.

But obviously, all of those CPU-mined blocks will vanish soon, when ASICs will adopt new rules for picking the right chain, to build new blocks on top of. And that's why I think about different models, because the path for CPU mining will be soon closed again (which is also expected; I didn't plan to be a CPU miner forever, and I knew, that it is just a nice hack, which will expire sooner or later, when ASIC miners will adopt those strategies as well, and bring the situation back to normal).

Pardon me for asking, but is the source code most of us run in this place altered by you? Are you the one who set difficulty=0x1d00ffff (including the 20 hours rule) and distributed the source code? Are you testnet4 Satoshi, or did testnet4 began prior this alteration?

So how do I do that? Do I just let it run, and expect to stand lucky, or is there a constant re-broadcasting every few seconds that takes place, which gives me advantage?

No. It is because of this:

Let's see:

It depends on this particular ASIC. It can keep the whole chain, and reorg only "future blocks" (for example: the last six blocks, from the 2 hours future time window), or it can reorg everything. In general, ASIC difficulty is the true difficulty, and ASICs can rule this network. CPU-mined blocks are just an exception, which is present only in testnets. Even signet doesn't have that rule.