automintspark & gasfee settings leads to unmined spark mints

[mobias17]

automintspark & gasfee settings leads to unmined spark mints

[mobias17]

By coincidence the topic continues...

above was performed on a Windows env. Today I wanted to crosscheck a few things on an Ubuntu 18.04. env and restored my test wallet with the seed phrase without a wallet back up file. So the fresh qt-wallet synched with the chain.

I thought the anonymization coins were stuck. Well, bear with me, they are and they aren't. (also confusing to me at this point)

(Top Windows were I intentionally stuck as described above, bottom fresh synch on Ubuntu)

Okay, the timestamps, for whatever reason, are slightly out of synch for the transactions, however I assure you it is the same wallet I synched.

Next, some might claim, in the top the pending private balance + available transparent balance != available transparent balance in the lower picture. Correct so, but the delta is the gas fee that in the end has never been paid to a miner. So i'd be willing to claim that the figures indeed add up.

To continue let's have a look at the 4 test cases from initial post. Again, as the wallet file was not imported I lost the labels and between the wallets there is time delta of about 2 min for each transaction. This also raises a flag with me, but could also be my wrongdoing....
Luckily we are on a blockchain an we trust in immutable hashes, so lets compare the transaction ID, which adds up.

Note how the presumed lost/broken/blocked transactions are not there anymore in the history of the fresh key phrase-restored wallet. In then end it makes sense, because I paid so little gas fees that they never got mined onto a block and are not on chain.
Still, a) IMO both wallets should show the same and b) latest when I query directly via the console I should get the same results in both (but the same) wallets... I don't.

Summary

Finally, it seems like that the mempool is indeed giving back the unmined transaction or they are still in there and are not synched in new wallet until competition. The latter could open a new dimension of challenges, but I was unlucky finding out more on this end. In a few days I might reindex the wallet and see what this does.

There seem to be a (uneasy) way to rectify it when the mint did not go through. Still, also that the transaction history is not the same can be argued whether it is right or wrong. To be positive, what happed on chain is there.

In the end and as for myself, I am not getting any closer to the solution at this point. Happy for some comments on this.

just to be sure the versions for comparison:

[justanwar]

Please try with the latest spark commit. For Qt, please try latest commit TheArcadiaGroup@92d4c4e

[mobias17]

Retest with commit suggested.

Reproducable.

Okay... What else has been learned during reproduction.

Breaking transaction are resulting from Transaction Fee --> Custom with active checkbox.

What puzzles me, although I uncheck the box and then use the 0.00001000 FIRO/kB, this goes through.

When it is checked it doesn't

The other observations (deviation in time stamp, deferent balances) from before remain.

In the end sure, when checking the box I should read the tooltip. Let's call me a naive critic for now, however when unexperienced users are clicking around the wallet (like i did during GUI shakedown test) and then later go to the magic button "Anonymise all", not sure they'd be happy if the transaction gets stuck.

[justanwar]

Please reply to this post with a mainnet Firo address for the bug bounty payout. Use a new address that has never been used before. Make sure to anonymize after receiving!

[mobias17]

@justanwar sorry for the delay in responding and thank you! aF2FvfKrHZfpf1if4m1RFiAogxz4UFHz74

[justanwar]

TXID: 557a01c5a26fcd406d578bcd12a21f42095cec3c996a44e7f41298a8e87162db

Bounty paid:
Type: UI
Amount: 40