Bitcoin Forum

Bitcoin => Bitcoin Technical Support => Topic started by: rafik35 on February 17, 2021, 03:54:25 PM



Title: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 03:54:25 PM
Hello guys,

I'm new to wallet business. I Already send multiple transactions in bitcoin network, but never had an issue like i'm having now.
First my tx is 87f45f4054eb12c8dcbc7e14ba71cca166ddc270640fdb8a11b5893a0f2ec634
There are 2 outputs, lower one is probably my change, because wallet showed me excalty that balance.
Most of the time I sent my btc was between exchanges and some android wallets. Never had issues because most of the time my fees were 0.0002 up to 0.001 btc. This time is different.
So i went to bitcoin atm to buy some bitcoin. It got stuck with my last bill. I could repeat the process but fee of the atm would consume 90% of the bill value so i didn't use it. That said that bill was for my fee to send bitcoin after i receive it.
So i did get 0.04209 btc. I thought it's not bad, not my usual fee, i can wait a little bit. So I'm using mycelium android wallet for this transfer, set up my fee to 34sats/byte which is low, but it should be confirmed in few hours. After i hit send button i saw there are some leftovers in my account and that tx fee was set for 4sats/byte. Don't know how this happened but here i am.
More than 2 days later i'm waiting for confirmation and not sure what to do.
Of course i read https://bitcointalk.org/index.php?topic=1802212.0
Besides that i'm reading all over the internet and not sure if there is any help in there for me.
Tried to import my wallet into electrum, which i did successfully but couldn't do rbf on my transaction.
Mycelium gives me option to bump fee(cfbf) - but as i read cfbf is only helpful if i'm receiver, which i'm not. There's support not responding, there is no faq/help about wallet and it's options.
Many of you suggest to wait it out. Or is there any other possibility? Installing bitcoin-core and zaping my transaction would help? Or when i used wallet without rbf it's not an option for me and shouldn't waste time?
I read that there might be also double spend transaction not being processed by miners problem. Because i'm new and not to technical not sure if that would help me or hurt me.
Also, am I right thinking that logging into wallet is rebrodcasting my tx? Can't find straight answer for this one all over the internet.
If it does, how waiting tx out to be forgotten is any good for me if my wallet will brodcast it again anyway?
Or do i don't understand something here.
What can i do at this point? Someone is waiting for their payment and i don't know what to do.
Any advice would be much appreciated.
Using viabtc with paid service is last resort for me, but don't want to go there because right now i can't afford it.

cheers,
Rafal

p.s. I know i've chosen poorly in wallet segment area so please do not instruct me to choose other one in future because i know that already :)



Title: Re: stuck tx, looking for viable solution
Post by: ranochigo on February 17, 2021, 03:59:44 PM
CPFP works best if you are the receiver as it doesn't require another replacement transaction to be made or if you didn't select RBF(Replace By Fee) during the creation of the transaction. You can absolutely do a CPFP using the wallet to give miners an incentive to confirm both of your transactions. I'm not sure if Mycelium rebroadcasts your transaction, but it would likely take up to two weeks for majority of the network to drop your transaction and I would think that is way too long and possibly get you a confirmation sooner.

Installing Bitcoin Core and zapping your transaction is absolutely not a solution for your current situation. Rebroadcasting the transaction to keep it in the mempool is not helpful if your intention is to make another replacement transaction but it helps to keep it in the mempool and getting it confirmed once the fee gets lower.

Either:
1. Wait for the fees to drop which should hopefully be within the coming weekend.
2. Do a CPFP using the change that you currently have access to at the caveat of paying more fees.


There's no double spend with your transaction, not relevant to your issue.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 04:18:01 PM
Thank you for quick reply.

By doing CPFP did you meant that i should do it in my mycelium wallet or can i go elswhere with this? Is there any tutorial regarding this or this is strictly bound with wallet software?
Im asking because:
1. When i imported my wallet into electrum it shows 0 balance, but i see my tx's. Tried that with samourai, but balance was 0 with no history.
2. Mycelium says my balance is to low, which is strange if they allowed my to send tx with 44 cents and they say my leftover 4$ is too low. Support not responding isn't helpful in it that regard.
3. Because of this brodcasting problem, i deleted (locally) the tx for it not to appear in my wallet, and balance went from 4 $ to 0. I read that it might appear again. Not sure if that's an issue for me regarding doing CPFP.

From theoretical stand point, if I would ask receiver to do CPFP is it possible to do it with any wallet? And if they do, miners are obligated to confirm my tx to because they both connected?
How can one spend btc they don't have yet? Or if the network see unconfirmed transaction it takes it into account?


Title: Re: stuck tx, looking for viable solution
Post by: hosseinimr93 on February 17, 2021, 04:22:43 PM
As stated by ranochigo in the post above you can implement CPFP method by spending the change.

Using CPFP, you can make your transaction more prioritized, but you can't make that confirmed very fast.

Your change address contains only 8089 satoshi.
Even if you send the lowest possible amount (547 satoshi) and pay the remaining balance (7542 satoshi) as transaction fee, the total fee would be 8453 satoshi.
The total size of transactions (the one already made and the one going to be made for CPFP) would be 133 + 222 = 355 vbyte.
So, the total fee would be about 23 satoshi per vbyte.

This amount of fee will put your transaction about 16 MB from the tip. It's not good for a fast confirmation but it will at least increase the chances of confirmation. Your transaction is now about 80 MB from the tip.


Title: Re: stuck tx, looking for viable solution
Post by: Charles-Tim on February 17, 2021, 04:25:23 PM
I thought it's not bad, not my usual fee, i can wait a little bit. So I'm using mycelium android wallet for this transfer, set up my fee to 34sats/byte which is low, but it should be confirmed in few hours. After i hit send button i saw there are some leftovers in my account and that tx fee was set for 4sats/byte.
The leftover you mentioned, are they unconfirmed transactions?
How many input are they in total?
How many output are they also in total?
When did you make the transaction? Hopefully not more than two weeks. You will need to know that these can be the reason why your transaction not confirmed in time in addition to the congested mempool and low fee paid in this present transaction.

It will be good to wait till weekend, Sunday precisely. During the weekend, mempool are less congested, and many low transaction fee that are not confirmed in the weekdays have high chance to be confirmed.

Also, know that if you want to use CPFP, another transaction fee will be required entirely in addition to the fee that will get the previous unconfirmed transaction confirmed. I mentioned this in order to let you know how useful replace-by-fee can be. Using RBF will only just make you to pump or cancel the transaction using higher fee (but lower if compared to CPFP if you are sending to another change address in your wallet). It will be better to get yourself a wallet that enable RBF to be used, one the best wallet that support RBF is electrum so that you will be able to just pump the fee next time for ease.


Title: Re: stuck tx, looking for viable solution
Post by: ranochigo on February 17, 2021, 04:36:36 PM
By doing CPFP did you meant that i should do it in my mycelium wallet or can i go elswhere with this? Is there any tutorial regarding this or this is strictly bound with wallet software?
It's not. The gist of CPFP is to make another transaction with a higher than proportionate fee to subsidize the "parent" transaction so that the total fee rates for both transaction would be reasonable. It is definitely not bounded by wallet software but doing it in the wallet which you sent from could be simpler.

1. When i imported my wallet into electrum it shows 0 balance, but i see my tx's. Tried that with samourai, but balance was 0 with no history.
When importing BIP39 seed phrases, there are specific configuration for the addresses to be generated and can differ greatly between the wallets. IIRC, MyCelium uses BIP44[1] as the standard to generate the addresses. This is transparent to you and should no matter as long as you can do a CPFP within the wallet.
2. Mycelium says my balance is to low, which is strange if they allowed my to send tx with 44 cents and they say my leftover 4$ is too low. Support not responding isn't helpful in it that regard.
CPFP requires a high fee rate to compensate for the low fees of the parent transaction so that might be where your issue is. In this case, CPFP is not viable either unless you are able to send more funds into the wallet and spend both the change and the Bitcoins added. I don't think this is particularly feasible either.

3. Because of this brodcasting problem, i deleted (locally) the tx for it not to appear in my wallet, and balance went from 4 $ to 0. I read that it might appear again. Not sure if that's an issue for me regarding doing CPFP.
You can't really tell if it rebroadcasts or someone else rebroadcasts it for you until it reaches the mempool expiry of the nodes. I would advice you to continually check the Blockexplorers and once you get an error from searching the transaction ID, make a new transaction.

From theoretical stand point, if I would ask receiver to do CPFP is it possible to do it with any wallet? And if they do, miners are obligated to confirm my tx to because they both connected?
How can one spend btc they don't have yet? Or if the network see unconfirmed transaction it takes it into account?
Yes. CPFP works as the first transaction has to be confirmed for the miner to collect the fees from both of the transaction.

Unconfirmed transactions are still spendable, it is just that it cannot be confirmed until the transaction from which it is spending from gets a confirmation.

[1] https://github.com/bitcoin/bips/blob/master/bip-0044.mediawiki


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 04:41:52 PM
As stated by ranochigo in the post above you can implement CPFP method by spending the change.

Using CPFP, you can make your transaction more prioritized, but you can't make that confirmed very soon.

Your change address contains only 8089 satoshi.
Even if you send the lowest possible amount (547 satoshi) and pay the remaining balance (7542 satoshi) as transaction fee, the total fee would be 8453 satoshi.
The total size of transactions (the one already made and the one going to be made for CPFP) would be 133 + 222 = 355 vbyte.
So, the total fee would be about 23 satoshi per vbyte.

This amount of fee will put your transaction about 16 MB from the tip. It's not good for a fast confirmation but it will at least increase the chances of confirmation. Your transaction is now about 80 MB from the tip.

So i have to do another transaction that is somehow connected to original tx? But how to do that?
Don't get me wrong, because i know i'm the one that got myself into that mess, but all i read all over the internet is that i can do CPFP (as you and ranochigo stated) but no one really says how to do it :)
I already lost more then 20 hours trying to solved it past 48 hours so either i'm stupid or instructions aren't realy user/newbie friendly.

I thought it's not bad, not my usual fee, i can wait a little bit. So I'm using mycelium android wallet for this transfer, set up my fee to 34sats/byte which is low, but it should be confirmed in few hours. After i hit send button i saw there are some leftovers in my account and that tx fee was set for 4sats/byte.
The leftover you mentioned, are they unconfirmed transactions?
How many input are they in total?
How many output are they also in total?
When did you make the transaction? Hopefully not more than two weeks. You will need to know that these can be the reason why your transaction not confirmed in time in addition to the congested mempool and low fee paid in this present transaction.

It will be good to wait till weekend, Sunday precisely. During the weekend, mempool are less congested, and many low transaction fee that are not confirmed in the weekdays have high chance to be confirmed.

Also, know that if you want to use CPFP, another transaction fee will be required entirely in addition to the fee that will get the previous unconfirmed transaction confirmed. I mentioned this in order to let you know how useful replace-by-fee can be. Using RBF will only just make you to pump or cancel the transaction using higher fee (but lower if compared to CPFP if you are sending to another change address in your wallet). It will be better to get yourself a wallet that enable RBF to be used, one the best wallet that support RBF is electrum so that you will be able to just pump the fee next time for ease.

As i said i'm not too technical about all this stuff but i'll try to do my best to answer.
I'm not 100% sure about the leftover. I mean how it can be or not be unconfirmed? I see it in my mycelium wallet and also it appears in the tx. I don't get how that is possible.
https://btc.com/87f45f4054eb12c8dcbc7e14ba71cca166ddc270640fdb8a11b5893a0f2ec634
I see 2 outputs here, one is the receiver, second one should (that i'm not sure) be the change. I'm not sure about it for 2 reasons:
The change appear only in my mycelium wallet, electrum dosen't seem to see it - balance is 0.
I don't understand how my original tx took 0.00009 from my wallet, put 0.00000911 into a fee and rest into change. Why not just do 1 output and leave my change with my wallet?
Tx was sent 2 days and 6 hours ago.
There is one input and 2 outputs if I understand correctly the link i posted.
I know mempool get stuck, the low fee use wasn't on purpose.


Title: Re: stuck tx, looking for viable solution
Post by: o_e_l_e_o on February 17, 2021, 04:42:27 PM
1. When i imported my wallet into electrum it shows 0 balance, but i see my tx's.
Although your bitcoin came from a legacy address, it looks like Mycelium sent your change to a nested segwit address. When entering your seed phrase in to Electrum, check the box under "Options" entitled "BIP39 seed", and on the next screen click on the middle option entitled "p2sh-segwit (p2wpkh-p2sh)". This should change the bottom box to m/49'/0'/0'. Click next, and hopefully you should see your change address show up in that wallet.

As hosseinimr93 says though, a CPFP will at most boost your fee to 23 sats/vbyte, which would currently still put you around 19 MB from the tip which is unlikely to confirm until the mempool empties a little at the weekend.

Don't get me wrong, because i know i'm the one that got myself into that mess, but all i read all over the internet is that i can do CPFP (as you ranochigo stated) but no one really says how to do it :)
You simply locate the unconfirmed bitcoin in your change address, and send that bitcoin with a higher fee to another address that you own.

Why not just do 1 output and leave my change with my wallet?
Because outputs can't be split like that to just take what is needed and leave the rest behind. Each output must be spent in its entirety, so if you don't want to send it all to the receiving address, then the rest must be sent back as change to another address that you own.


Title: Re: stuck tx, looking for viable solution
Post by: hosseinimr93 on February 17, 2021, 04:44:00 PM
1. When i imported my wallet into electrum it shows 0 balance, but i see my tx's. Tried that with samourai, but balance was 0 with no history.
When importing BIP39 seed phrases, there are specific configuration for the addresses to be generated and can differ greatly between the wallets. IIRC, MyCelium uses BIP44[1] as the standard to generate the addresses. This is transparent to you and should no matter as long as you can do a CPFP within the wallet.
I just checked this in my android phone.

I see three addresses in my wallet.
1. A legacy address located in m/44'/0'/0'/0/0   
2. A nested segwit address located in m/49'/0'/0'/0/0   
3. A native segwit address located in m/84'/0'/0'/0/0

In the transaction posted by OP, the input is a legacy address and the change is a P2SH address.
So, seems that mycelium automatically sends the change to P2SH address of the wallet.


To OP:

The change is now in your P2SH address.
So, for importing your wallet into electrum (as stated by o_e_l_e_o in the post above) you need to select p2sh-segwit when importing your seed phrase.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 04:56:30 PM
I've imported into electrum as instructed and I see my change. Thanks for that.
Tommorow i'll have few more bucks in my account but it would appear in my legacy address, but maybe i can reroute that.

Bear with me for last question. Given that i have this change, how to create another tx that will bump my previous tx? Don't mind the amounts im spending, that it'll be too low. I want to know how to do it.


Title: Re: stuck tx, looking for viable solution
Post by: Charles-Tim on February 17, 2021, 05:01:27 PM
Bear with me for last question. Given that i have this change, how to create another tx that will bump my previous tx? Don't mind the amounts im spending, that it'll be too low. I want to know how to do it.
If truly that address is the change address, then just send the transaction to new address on your wallet with high fee. The reason CPFP is possible is because you have a change (UTXO) from the parent transaction (the first transaction) back into an address on your wallet.


Title: Re: stuck tx, looking for viable solution
Post by: o_e_l_e_o on February 17, 2021, 05:07:47 PM
Bear with me for last question. Given that i have this change, how to create another tx that will bump my previous tx? Don't mind the amounts im spending, that it'll be too low. I want to know how to do it.
In your Electrum wallet, right click on the "Coins" tab (you may have to click on View -> Show Coins first), right click on your change output, and click spend. Then click on the "Send" tab, enter another address you own in the "Pay to" tab, click on "Max", click on "Pay", and select the highest fee you can.

Tommorow i'll have few more bucks in my account but it would appear in my legacy address, but maybe i can reroute that.
The best thing to do would be to export the private key of your change address and the private key of the legacy address which is receiving the new coins tomorrow, import both in to a new wallet together, and then spend them both in the same transaction. This will mean you don't have to make yet another transaction and spend yet more on fees sending your legacy address coins to your nested segwit wallet.

To do this, in each of your two Electrum wallets (the legacy one and the nested segwit one), right click on the necessary address in the "Addresses" tab, and click on "Private key". Copy the private key, including the prefix before the colon. Then create a brand new Electrum wallet, selecting "Import Bitcoin address or private keys, and paste the two private keys in. Once you've done that, you can spend the coin in both address (your change and your new coin) together in the same transaction with a high fee to bump your first transaction.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 05:18:45 PM
That is brilliant. Thank you guys so much.
Thank you o_e_l_e_o for that detailed response.
I see the CPFP option now.
I might be able to redirect the transfer to my segwit address so probably no need to import my wallets into one. But i'll do it anyway just to try that option.
You saved my day and i learned some new things today. Thanks a lot everybody! :)


Title: Re: stuck tx, looking for viable solution
Post by: hosseinimr93 on February 17, 2021, 05:25:05 PM
I see the CPFP option now.
If you click on "Child pays for parent" in right-click menu, electrum will spend the unspent output (the 8089 satoshi) to your own address.
As you need to combine another UTXO with that 8089 satoshi, you shouldn't use that option.
Instead of that, make a transaction to another address manually. In this way, you can spend the 8089 satoshi and the amount you are going to receive in a single transaction.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 17, 2021, 06:16:27 PM
will do,thank you for that clarification


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 18, 2021, 11:35:33 AM
Hello guys,
I'm still waiting for additional money but i wonder if i can do transaction now with change i've got, and then (if it dosen't get confirmed) do a rbf with that new transactions when satoshis arrives. I see i can check that rbf box for new tx, but not sure if my thinking is right while doing cpfp. Or if these two operations work separatly and cpfp wont affect rbf in the future?
Other option is to do cpfp and leave 1 satoshi from that change, and do another cpfp when i get additional resources.
I don't like second option because it would add additional bytes to my tx, so i think if i would choose 2nd option i would rather wait for my satoshis.
But never done rbf, its just changing the fee or is there any catch to it? Does it change weight of the tx or am i thinking right to try confirm with light tx and then add via rbf if it doesn't confirm?

edit:
also should i take into account that the cpfp would be done from segwit address? Would it affect rbf in any way? I mean if it would be impossible for that type of btc address.


Title: Re: stuck tx, looking for viable solution
Post by: o_e_l_e_o on February 18, 2021, 01:51:34 PM
I'm still waiting for additional money but i wonder if i can do transaction now with change i've got, and then (if it dosen't get confirmed) do a rbf with that new transactions when satoshis arrives.
Yes, this is possible.

You can make a CPFP transaction now, and then later use RBF to bump the fee of the CPFP transaction if it does not confirm. The only limitation here is that your RBF transaction can only include an unconfirmed input if the transaction it is replacing also contained that same unconfirmed input. What this means is that any additional funds you want to add in to the RBF transaction (such as those you are waiting on arriving) must have confirmed before you can include them.

Other option is to do cpfp and leave 1 satoshi from that change, and do another cpfp when i get additional resources.
The minimum amount you can leave as change is 546 satoshi in a legacy address or 294 satoshi in a segwit address.

But never done rbf, its just changing the fee or is there any catch to it? Does it change weight of the tx or am i thinking right to try confirm with light tx and then add via rbf if it doesn't confirm?
The weight of the transaction would change if your replacement transaction included more inputs or outputs than the one you are replacing, such as if you were including an extra input of the coins you are waiting to receive.

Having said all that, I don't think there is much point. We know that at most you will be able to boost the fee to around 23 sats/vbyte, which will still put you around 8 MB from the tip. Either wait until 23 sats/vbyte may get you confirmed, or wait until you have more funds available to bump the fee higher.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 18, 2021, 03:23:44 PM
Great, thanks.
I got another idea because i may come across another problem.
My new funds will be sent with 0.00022 (or just 2, cant get a hold of the sender) fee. So probably it will be confirmed in early morning i hope(in europe, when fees are low) or on sunday.
Would double cpfp be possible in that scenario?
I mean i would create tx with one output(myself) that would include both unconfirmed change from my original transaction and those unconfirmed funds from the transfer i'm waiting for.
As far as i understand all this i think it would, but please correct me if i'm wrong.
cheers


Title: Re: stuck tx, looking for viable solution
Post by: o_e_l_e_o on February 18, 2021, 03:26:47 PM
Would double cpfp be possible in that scenario?
I mean i would create tx with one output(myself) that would include both unconfirmed change from my original transaction and those unconfirmed funds from the transfer i'm waiting for.
Yes, that is also possible, but the fee for your CPFP transaction would need to be even higher since it is bumping up the fee for two unconfirmed transactions rather than just one.


Title: Re: stuck tx, looking for viable solution
Post by: NotATether on February 18, 2021, 03:31:52 PM
~
Would double cpfp be possible in that scenario?
I mean i would create tx with one output(myself) that would include both unconfirmed change from my original transaction and those unconfirmed funds from the transfer i'm waiting for.

Yes you can. A CPFP involves spending an unconfirmed output, so if you spend unconfirmed outputs from multiple transactions at the same time, all of them should be included at the same time in a block.

Keep in mind that when you choose the fee for your third transaction, you want to make the average of those three fees something reasonable for miners. No point doing a CPFP 60 sats/vbyte if your transaction fees for your other two transactions are 2 and 10 sats/vbytes for example  ::)


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 18, 2021, 03:58:37 PM
So in other words i would have to sum fee that would be used for 3 transactions and divide that sum by total size of 3 tx's in vbyte and conclude if that would push my tx?

I have now 134, 222 and probably 323 vbyte transaction that gives me total sum of 679.
I would use 911, 22000 and around 27000 sats that is 49 911.
That would give me 49911/679 = 73,5 sats/vbyte. Not bad considering were i started.
Am i calculating this right?
The 323 number is taken from first tx that was generated from bitcoin atm, the sender is the same. This time it would be done by human so it might be lower than this.

Just called the guy from atm company and he said he'll do the transfer tomorrow, i'm so pissed.


Title: Re: stuck tx, looking for viable solution
Post by: rafik35 on February 18, 2021, 09:54:23 PM
So i'm not sure, but i think i managed to screw up my situation again.
I saw on btc.com that my tx is no longer in mempool. So i "ran" to my wallet (before thinking to check other explorers), deleted local tx and broadcasted new one, with 47sats/byte with rbf active in case i need it tomorrow.
Some of the explorers see new transaction (218eaf4dbce7d50e20355c5ae29d5f4cc2a9f4abcb0158044926bf344ceb3878), some of them don't - still see previous one. One of them marked it even as double spend. https://www.smartbit.com.au/tx/87f45f4054eb12c8dcbc7e14ba71cca166ddc270640fdb8a11b5893a0f2ec634
What will happen if electrum catch that first transaction from network again?
Should i worry about it? Or i just have to wait for rest of them to drop old tx and accept new one?


Title: Re: stuck tx, looking for viable solution
Post by: NotATether on February 20, 2021, 05:19:07 AM
I saw on btc.com that my tx is no longer in mempool. So i "ran" to my wallet (before thinking to check other explorers), deleted local tx and broadcasted new one, with 47sats/byte with rbf active in case i need it tomorrow.
...
What will happen if electrum catch that first transaction from network again?

You deleted the first transaction from Electrum and that is what matters. It should no longer be related by full nodes or in their mempools. The behavior you are seeing is just the block explorer databases keeping the old transactions cached in their systems, and the ones that are showing the second transaction are giving the correct result.

(By the way, this transaction has since been confirmed so you don't have a problem anymore.)


Title: Re: stuck tx, looking for viable solution
Post by: o_e_l_e_o on February 20, 2021, 07:24:01 AM
So in other words i would have to sum fee that would be used for 3 transactions and divide that sum by total size of 3 tx's in vbyte and conclude if that would push my tx?
That is correct. You need to consider your final transaction and every unconfirmed parent (including if those unconfirmed parents have unconfirmed parents, and so on) for your final calculation.

Am i calculating this right?
Provided the sizes and fees you are listed are accurate, then yes, your calculations are correct.

You deleted the first transaction from Electrum and that is what matters. It should no longer be related by full nodes or in their mempools. The behavior you are seeing is just the block explorer databases keeping the old transactions cached in their systems, and the ones that are showing the second transaction are giving the correct result.
This is not necessarily true. Although some nodes dropped his first transaction, and would therefore accept his second one as it wasn't flagged as a double spend, some nodes did keep his first transaction in their mempools, and they are not "incorrect" for doing so.