0.96.4 RC1

[Searinox]

Another hour spent fruitlessly trying to scan the transaction history for a wallet, at 99% and 28 seconds left admorydb.exe quits and armoryQt says nothing about even a connection loss, it just sits there. I only found in armorycpplog.txt "ERROR - 1486672480: (SwigClient.cpp:61) POLLERR error in readAndWrite", that is just the Qt losing connection to ADB but again - Qt said nothing about it. It will display tray messages about connection to bitcoind being lost but not ADB, which coupled with the ADB's lack of relevant reports, conspires to make these crashes as silent and obscure as possible. Even moreso, when I relaunch everything the history is rescanned from scratch.

dbLog.txt last states "-INFO  - 20:33:49.875: (..\BlockchainScanner.cpp:852) scanned from block #498700 to #499207", whatever it was doing at the time before it dropped and how far it got, and after that nothing. I don't think I have anything more to give.

[goatpig]

Delete your databases folder and try from scratch.

[Searinox]

I already did that thrice. What I've done now is deleted the last few block files from bitcoin core corresponding to about where that recurring 99%-ish point would be(going by the logs), and restarted it with -reindex. Once it's back up to sync I will -AGAIN- rebuild the armory DB.

[Mr.Vice]

I already did that thrice. What I've done now is deleted the last few block files from bitcoin core corresponding to about where that recurring 99%-ish point would be(going by the logs), and restarted it with -reindex. Once it's back up to sync I will -AGAIN- rebuild the armory DB.

Your issue reminds me on mine serveral months ago. Are you using any Anti-Virus software, NAS or encrypted HDD?

[Mr.Vice]

I've found an error regarding tx preview. I have sent some tx multiple times by using RBF and CPFP via Armory and cleared all unconfirmed tx also multiple times, which could have messed up Armory's indices.

armorylog.txt

Code:

2017-12-17 17:30:29 (WARNING) -- Transaction.pyc:1263 - Unexpected nested type for TxIn 16
2017-12-17 17:30:29 (WARNING) -- Transaction.pyc:1263 - Unexpected nested type for TxIn 16
2017-12-17 17:30:29 (ERROR) -- Traceback (most recent call last):
  File "ui\TxFrames.pyc", line 897, in createTxAndBroadcast
  File "ui\TxFrames.pyc", line 862, in validateInputsGetUSTX
  File "qtdialogs.pyc", line 4546, in __init__
  File "armoryengine\Transaction.pyc", line 2727, in verifySigsAllInputs
  File "armoryengine\Transaction.pyc", line 1579, in verifyAllSignatures
  File "armoryengine\Transaction.pyc", line 1863, in evaluateSigningStatus
IndexError: list index out of range

Could this be a damaged wallet file or is it "just" a tx index that's not working. I'm currently doing a rebuild and rescan.

Kind regards :-)

[goatpig]

It looks like stuff that's already fixed in the testing branch. I'll put out another release soon.

[Searinox]

I already did that thrice. What I've done now is deleted the last few block files from bitcoin core corresponding to about where that recurring 99%-ish point would be(going by the logs), and restarted it with -reindex. Once it's back up to sync I will -AGAIN- rebuild the armory DB.

Your issue reminds me on mine serveral months ago. Are you using any Anti-Virus software, NAS or encrypted HDD?

No encrypted HDD, no NAS, and antivirus already excludes the Armory folder(which cotains both ADB and Bitcoin blockchain) due to some well-known trolling events where ppl deliberately included transaction bytes to resemble malware and flag false-positives.

[Mr.Vice]

No encrypted HDD, no NAS, and antivirus already excludes the Armory folder(which cotains both ADB and Bitcoin blockchain) due to some well-known trolling events where ppl deliberately included transaction bytes to resemble malware and flag false-positives.

I've used McAfee AV some time ago. Previously everything worked pretty well, but all in a sudden I've gotten corrupted files and crashes of ArmoryDB. I also tried to set folder exemptions, but it didn't really prevent these errors from happening. So I've tried to deinstall my AV and gotten rid of all of my AV services (except Win Defender) and then it worked like it should.
But I have to admit it was hard to drop something that's already paid in order to get Armory running. So I can't tell if that might solve your issue, but again it looks pretty similar to mine.

[Searinox]

Nay, I ain't using McAffee either.

EDIT: I found something that may be useful. I had generated 22 P2SH-P2WPKH addresses by clicking on the "receive bitcoins" button and another 2000 using compute. However, on the next restart of Qt, it moved ALL those addresses that I hadn't spent from into the P2PKH area, where they looked completely different. The only way I was able to tell was because they carried the labels I had given them. And the unused P2SH-P2WPKH field became empty. They were effectively all converted. I now had 2022 P2PKH addresses and none others.

[goatpig]

I'm pretty that's fixed in testing as well.

[Searinox]

So if there's another testing build is it gonna be an RC or a CR? (christmas release)

[goatpig]

Yeah, sometimes next week probably.

[alomar]

having never had to do this before with a non-RBF enabled tx, what's the fastest, safest, most reliable way to unstick a too low fee stuck tx from the mempool using this version?  specific details would be appreciated.

[goatpig]

CPFP. Spend the change of the stuck ZC with a big fee back to yourself.

[alomar]

CPFP. Spend the change of the stuck ZC with a big fee back to yourself.

crap, now Coin Control button freezes again!

[goatpig]

Remove the dlg size entry from Settings.txt

[alomar]

Remove the dlg size entry from Settings.txt

ok, i totally screwed that up somehow the last time you told me to do that and i had to recompile.  this is what i see in the Settings.tx file.  what exactly do i delete?:

ccDlgGeometry                        | 01d9d0cb000100000000000000000000000004ff000003050000000000000000fffffffefffffff e000000000200

everything after the "|" ?  or do i make it a true or false?  or delete the entire line?

[goatpig]

The entire line.

[alomar]

The entire line.

ok, i'm off.  will let you know how CPFP goes...

[alomar]

everything worked like a charm.  thx goatpig!

how do i make the ccDlgGeometry line deletion persistent?