[ANN] Armory Multi-Sig with Simulfunding [BOUNTY 0.03 per bug]

[helgabutters]

1) In armoryd if you don't give backupwallet a valid file path it throws: "global name 'IOERROR' is not defined" 

Code:

armory@armory-vm:~/armoryD$ python /usr/lib/armory/armoryd.py --testnet backupwallet testes
{
    "Error": "An error occurred in backupwallet", 
    "Error Type": "NameError", 
    "Error Value": "global name 'IOERROR' is not defined"
}

1.1) This is also thrown when you try to save a file to a folder where you do not have proper permissions.

Code:

armory@armory-vm:/usr/lib/armory$ python armoryd.py --testnet backupwallet /root/test.za
{
    "Error": "An error occurred in backupwallet", 
    "Error Type": "NameError", 
    "Error Value": "global name 'IOERROR' is not defined"
}

2) In armoryd you can create a 1 of 1 lockbox. You can't do a 1 of 1 in the armoryqt so you can't edit the lockbox you created.

https://i.imgur.com/1m24oqY.png

3) In Linux: When you select "Request Payment" in the lockbox manager nothing happens and it throws this error:

Code:

2014-07-16 17:18 (ERROR) -- Traceback (most recent call last):
  File "/usr/lib/armory/ui/MultiSigDialogs.py", line 1158, in funcReqPayment
    DlgRequestPayment(self, self.main, p2shAddr).exec_()
  File "/usr/lib/armory/qtdialogs.py", line 9896, in __init__
    linkText = hex_to_binary(self.main.getSettingOrSetDefault('DefaultLinkText', defaultText))
  File "/usr/lib/armory/armoryengine/ArmoryUtils.py", line 1846, in hex_to_binary
    bout = h.replace(' ','')  # copies data, no references
AttributeError: 'long' object has no attribute 'replace'

4) In Linux: When you select "Create Clickabe Link" in Recieve Bitcoins > New Recieving Address nothing happens and it throws this error:

Code:

2014-07-16 17:20 (ERROR) -- Traceback (most recent call last):
  File "/usr/lib/armory/qtdialogs.py", line 2772, in openPaymentRequest
    dlg = DlgRequestPayment(self, self.main, addrStr, msg=msgTxt)
  File "/usr/lib/armory/qtdialogs.py", line 9896, in __init__
    linkText = hex_to_binary(self.main.getSettingOrSetDefault('DefaultLinkText', defaultText))
  File "/usr/lib/armory/armoryengine/ArmoryUtils.py", line 1846, in hex_to_binary
    bout = h.replace(' ','')  # copies data, no references
AttributeError: 'long' object has no attribute 'replace'

5) In armoryd you when encrypt and unlock a wallet, instead of a displaying the wallet ID the output is "<armoryengine.PyBtcWallet.PyBtcWallet object at ... >"

Code:

armory@armory-vm:/usr/lib/armory$ python armoryd.py --testnet unlockwallet password1
"Wallet <armoryengine.PyBtcWallet.PyBtcWallet object at 0x9add54c> has been unlocked."
armory@armory-vm:/usr/lib/armory$ python armoryd.py --testnet encryptwallet password
"Wallet <armoryengine.PyBtcWallet.PyBtcWallet object at 0x9add54c> has been encrypted."

[etotheipi]

@helgabutters

#3 and #4 are actually the same bug.  And I recognize the issue... it's a rare condition that's leading to the data being saved in the settings file being read later, incorrectly.  Can you tell me what data is in the "DefaultLinkText" line of the ArmorySettings.txt file?  Either way, you can close Armory, remove that line, and then restart Armory and it will be fixed.  I'll figure out how to avoid that condition.

[doug_armory]

Thanks to all our testers, especially helgabutters!   It took us a while to get through all his bug reports, but I think we knocked them off all of the non-OSX bug reports.  Although we may not have fixed all the OSX bugs, doug_armory was able to integrate some Qt4 patches into the OSX build system that should resolve a ton of stability issues.  No promises, but he insists that he experienced a notable improvement!  Please let us know.

Hello. Don't worry, everybody, I'll be cranking out OS X fixes as quickly as possible! I probably would've gotten them done by now had I not spent a few days researching the stability patch (I tried an earlier version months ago, and it crashed and burned), not to mention cleaning up some other OS X build stuff that's invisible to 99.9% of you but critical for anybody building Armory, ourselves included.

First armoryd bug!

When running "help" they all return "Error": "The function description is malformed."
http://pastie.org/pastes/9395585/text?key=fbvn2gyp5xzxicil4reag

Since this is the last release before 0.92 is there any specific date you are aiming to have it out by?

How did I know you'd be on the case? Odd that it's no longer working. Worked like a charm last I checked, although the code is a bit fragile. (Nature of the beast for now, unfortunately.) I'll look into this and the other bugs ASAP.

[helgabutters]

@helgabutters

#3 and #4 are actually the same bug.  And I recognize the issue... it's a rare condition that's leading to the data being saved in the settings file being read later, incorrectly.  Can you tell me what data is in the "DefaultLinkText" line of the ArmorySettings.txt file?  Either way, you can close Armory, remove that line, and then restart Armory and it will be fixed.  I'll figure out how to avoid that condition.

Ah, that makes sense.

The value for DefaultLinkText is: 66657766616566

[PRab]

Testnet bug.

Open in testnet mode and send a transaction. It asks if you want to donate, say yes. It will fill in the mainnet address.

[PRab]

Enhancement request.

If you send from 1 wallet to another (both controlled by the same armory instance) the outgoing transaction doesn't always show up before the incoming transaction.

 

[SimonBelmond]

I wouldn't expect this qualifies for the 0.03 bounty, but I noticed a little something. At the end of the first paragraph is the text "Untitled".

...

A brief overview of how I got here, in case it matters:
I created two promissory notes (0.1 BTC each) to pay an address in my own wallet, created the TXSIGCOLLECT from that, and then imported that into the Review & Sign window (these are still unsigned transactions).

Whoops, that's a pretty obvious error.  You definitely deserve a bounty for that ... or at least we deserve to lose a bounty for not noticing it ourselves

This was already pointed out here: https://bitcointalk.org/index.php?topic=607046.msg7606266#msg7606266

[etotheipi]

I wouldn't expect this qualifies for the 0.03 bounty, but I noticed a little something. At the end of the first paragraph is the text "Untitled".

...

A brief overview of how I got here, in case it matters:
I created two promissory notes (0.1 BTC each) to pay an address in my own wallet, created the TXSIGCOLLECT from that, and then imported that into the Review & Sign window (these are still unsigned transactions).

Whoops, that's a pretty obvious error.  You definitely deserve a bounty for that ... or at least we deserve to lose a bounty for not noticing it ourselves

This was already pointed out here: https://bitcointalk.org/index.php?topic=607046.msg7606266#msg7606266

...and he still gets the bounty because it should've been fixed but still exists in the new testing version.   I think we got so distracted by helgabutters' bugs, we forgot to go back through some of the other reports.  Whoops.  [EDIT: scratch that... it was fixed, but TimS's report was made against an earlier testing version]

Enhancement request.

If you send from 1 wallet to another (both controlled by the same armory instance) the outgoing transaction doesn't always show up before the incoming transaction.

...

You know, that has always bugged me, but just not enough to be motivated to go find and update the sorting code.  Thanks for encouraging me to do it, finally!

Perhaps .9 won't be the last testing release, but I also won't make a big deal about doing a couple tiny intermediate releases.  My release scripts are mature enough that it's not a huge burden anymore.  And as simple as these changes are, it always makes me uncomfortable right before a release (but none of them are the kind that would lead to money loss if we botched something).  

[CircusPeanut]

I wouldn't expect this qualifies for the 0.03 bounty, but I noticed a little something. At the end of the first paragraph is the text "Untitled".

Thought I fixed it here:

https://github.com/etotheipi/BitcoinArmory/commit/29944272499dd167f3a26af6d164750fd3266af7

[TimS]

I wouldn't expect this qualifies for the 0.03 bounty, but I noticed a little something. At the end of the first paragraph is the text "Untitled".

Thought I fixed it here:

https://github.com/etotheipi/BitcoinArmory/commit/29944272499dd167f3a26af6d164750fd3266af7

Yes, as has been mentioned in a couple of edits, I was testing against an older version. Oops

[Searinox]

0.91.99.9: The warning popup is now gone but the "minimize to tray on open" now no longer works and Armory pops up to full window on launch instead of discretely activating in tray. The option appears to be broken. The whole idea for me waiting for the update was something that could launch on startup or sleep quietly(yes I actually made a script to restart it on resume sleep), and now I've gone from annoying warning to program UI popping in my face.

Also an older issue: on both graceful and ungraceful shutdown, Armory leaves a ghost icon of itself in the tray which vanishes on mouseover.

OS: Windows 7 x64 SP1 + All Updates

[TimS]

I'm using .9 now.

"oragnizer" instead of "organizer":


From https://github.com/etotheipi/BitcoinArmory/blob/doug-devel/ui/MultiSigDialogs.py#L1727 (note the branch...the main branch doesn't have this file at all yet, and I'm unfamiliar with the branch structures you use) has an unnecessary comma (I haven't found yet where, or if, this is shown in the app, so no screenshot).

Once it has enough signatures, any device, can broadcast the transaction to the network.

Should be

Once it has enough signatures, any device can broadcast the transaction to the network.

This is in both lines 1727 and 1791.

[TimS]

When spending from a 2 out of 3 lockbox, Armory allows you to provide all 3 signatures. When it tries to broadcast this, it is rejected by the network (presumably because it sent all 3 though it had promised just 2). It should either stop you at 2 signatures, or allow all 3 and then only broadcast any 2 of them.




Also, when exporting signed/unsigned sigcollect.tx files, the suffix can appear twice (I'm using Win 7 SP1 x64). When I actually click Save, it saves with just one suffix. If the file (with the single suffix) already exists, this will mean that you can overwrite the file without a confirmation telling you the file already exists.

[doug_armory]

From https://github.com/etotheipi/BitcoinArmory/blob/doug-devel/ui/MultiSigDialogs.py#L1727 (note the branch...the main branch doesn't have this file at all yet, and I'm unfamiliar with the branch structures you use) has an unnecessary comma (I haven't found yet where, or if, this is shown in the app, so no screenshot).

That's an old branch I haven't synced in a month and really haven't developed on since May. Any issues there that aren't in the "devel" branch should be ignored. (In fact, the entire branch should be ignored!)

[etotheipi]

From https://github.com/etotheipi/BitcoinArmory/blob/doug-devel/ui/MultiSigDialogs.py#L1727 (note the branch...the main branch doesn't have this file at all yet, and I'm unfamiliar with the branch structures you use) has an unnecessary comma (I haven't found yet where, or if, this is shown in the app, so no screenshot).

That's an old branch I haven't synced in a month and really haven't developed on since May. Any issues there that aren't in the "devel" branch should be ignored. (In fact, the entire branch should be ignored!)

On that note, after this release we're going to go through and clean up all the random side branches.  I'd like to reduce to just "master", "testing", "dev", and a couple others that are works in progress.  There's a lot of abandoned branches that were really for one bug or feature that took way longer than it should have.   A little belated spring cleaning. 

[Searinox]

Was something changed in the way Armory handles the minimize on startup and I'm doing it wrong? Because I continue to fail to get it to start in tray, even with a vbscript sending it alt+f4. When the program launches, it pops up the UI and completely disregards shortcuts that have "start minimized" in the properties, armory's options, VBscripts that try to change the window state, or anything of the sort. AFTER the UI has been brought up it responds to closing, it closes nicely in tray, but ONLY after it's done slapping itself on the screen and its resilience is annoying me.

[helgabutters]

Sorry for the delay, works been rather busy the past few days so I haven't been able to put in as much time as I'd like. Also, could you give me some insight on which ones I've submitted that are definitely not bugs and a few that you aren't sure of? I'm just trying to gauge what exactly you want and what I can safely not worry about finding/reporting.

1) You can't "importprivkey" as it will always error. I tried both hex and base58

http://pastie.org/pastes/9403588/text?key=1qunxbmpxlqa8arkoitgxq

2) There is some calculation error in listaddrunspent. You can see once you hit 3 characters in the string, it will still say 3 once you've put in a fourth and continue counting up from there.

http://pastie.org/pastes/9403585/text?key=roc4hfgkcsc6k9igcmmgq

3) You get different errors depending on the number of characters in "getblock". It seems dependent on if it's even or odd.

http://pastie.org/pastes/9403583/text?key=wucl7ytepivyq4uywdq5w

4) If you have a very long wallet description, the "Delete Wallet Options" window gets overly large and it won't show the whole thing. It also can't be resized down at all (This happens on Win and Linux, haven't tested OSX)

https://www.youtube.com/watch?v=xxl08nkXNPk

5) You can't "verifysignature" as it always interprets the "-----" as an argument for armoryd (and the only compatible signature types have "-----" in them).
I've tried in just the command line and in a bash script to no avail. If this is just an issue with my understanding of the command or how to properly give the command the information, please let me know.

Address: mkT61iEMnSSFgoCJkXFqbV9SiMHKnHDktN
Message: This is to test armoryd verifysignature for mkT61
Base64 Sign

http://pastie.org/pastes/9403589/text?key=dwsxu9jjzfyrgwlooqija

Same as above but Clearsign

http://pastie.org/pastes/9403590/text?key=zcw015sdcylrwu7k1hkniq

6) For a verified signature there is an extra ' in front of "The owner of the following ..."

https://i.imgur.com/zcbMDn8.png

7) You get an "Unexpected BLKSTRING" error when trying to signasciitranscaction. I tried both right in the command line and from a file output

http://pastie.org/pastes/9403594/text?key=zu0cydlii6zwkfyqozpq

8 ) This one I'm not sure of the cause. I thought it was originally because the wallet was locked but that isn't the case. One of my wallets if I try to "getledger" or "getledgersimple" I will get an "AttributeError":

http://pastie.org/pastes/9403608/text?key=qtztvorf7wni6c8tj9jtbg

I can provide you with this wallet if you'd like, just let me know.

[etotheipi]

Final Testing Version before 0.92 (0.91.99.11-beta)
This will be the released version of 0.92 unless important/dangerous bugs are found (gotta stop polishing and just release it at some point).  Plan to rename this to 0.92 on Tues or Weds.

Installers for version 0.92 (pre-release 0.91.99.11-beta):
  Armory 0.91.99.11-beta for Windows XP, Vista, 7, 8+ 32- and 64-bit
  Armory 0.91.99.11-beta for MacOSX 10.7+ 64bit
  Armory 0.91.99.11-beta for Ubuntu 12.04+ 32bit
  Armory 0.91.99.11-beta for Ubuntu 12.04+ 64bit
  Armory 0.91.99.11-beta for RaspberryPi (armhf)


Offline Bundles:
  Armory 0.91.99.11-beta Offline Bundle for Ubuntu 12.04 32bit
  Armory 0.91.99.11-beta Offline Bundle for Ubuntu 12.04 64bit
  Armory 0.91.99.11-beta Offline Bundle for RaspbianPi (armhf)

Signed Hashes:
  Armory 0.91.99.11-beta: Signed hashes of all installers


Honestly, not a lot has changed from the .9 testing version except:

• Proper sorting of simultaneous ledger entries now (should've done that like 2 years ago)
• Offline message format upgrade warning when you attempt to create and unsigned transaction from the "Send Bitcoins" dialog (matching the one when you click "Offline Transactions" from the main window)
• Strips extraneous signatures from transactions just before broadcast -- it will carry the extra sigs around with it so you can see who has signed, but the extras are removed when you hit "Broadcast".

[doug_armory]

Sorry for the delay, works been rather busy the past few days so I haven't been able to put in as much time as I'd like. Also, could you give me some insight on which ones I've submitted that are definitely not bugs and a few that you aren't sure of? I'm just trying to gauge what exactly you want and what I can safely not worry about finding/reporting.

Thanks for posting some more bugs to fix. I've started going through them. At least one (#2) isn't a bug because you're passing in a Base58 value, while the error message is counting the number of bytes in the decoded value. Subtle, I know, but it doesn't really count as a bug.

In the meantime, I'll see what I can get into 0.92. Gonna go pester Alan.

EDIT: Also, as I'm going through these, I'm noticing that some aren't actual bugs. For example, #3 isn't a bug. You need to supply a block hash. So, the following would be valid. (I just tested it.)

Code:

python armoryd.py --testnet getblock 00000000000072da6566808cce7e50c0b4f26fd56993a2a22d8d26e7cc23bad8

Use the following to see what the parameters should be. You can also poke around the source to see some examples.

Code:

python armoryd.py --testnet help

EDIT 2: Regarding #5, this is a bug, but not in the sense that you may think. One thing I forgot to point out in the help text is that the message needs to be passed in with quotation marks so as to avoid tripping up the command line argument detector. I've included an example of what's valid (other than the obvious nonsense).

Code:

python armoryd.py verifysignature \"-----BEGIN BITCOIN SIGNED MESSAGE-----$'\n'Comment: Hello.$'\n'-----BEGIN BITCOIN SIGNATURE-----$'\n'$'\n'junkjunkjunk$'\n'-----END BITCOIN SIGNATURE-----\"

That being said, it turns out that the unit test we wrote didn't take into account the fact that, on the command line, the spaces would cause the sig to be split into multiple pieces. I fixed that.

EDIT 3: #7 is also not a bug. The formatting isn't quite right. The following shows a format that will work. It's basically the same as the one from #5 with the exception that quotation marks aren't required (and will, in fact, mess things up, although I might slip in some code to handle that particular case.)

Code:

python armoryd.py --testnet signasciitransaction =====TXSIGCOLLECT-GynG9f37======================================$'\n'AQAAAAsRCQcAAAAAAf18AQEAAAALEQkHchUHvHxM29fPeY02InKy5ZQeYZ8vMA9G$'\n'rJVpM8tCGBEAAAAA/QEBAQAAAAEhOzQQbizqkYkkOvs8sxROO1j198R/ls8zebRC$'\n'TgGNagMAAACMSTBGAiEA5nWwoKcAU7g3dhs/5dkq6eaw+yY7T+uJiqKbnmjwkjAC$'\n'IQCARNrs4jaNx5iNFPyObj05olSK710suLl+YbfKe8ihUgFBBM3zuTqCoy/uKAh0$'\n'Kgn5LeWNYpMv/WJvqUsJYT2+/QtuG73EQhswCxN7aqyOaOF8EHN7mLUT93p9G0st$'\n'eWtEUof/////AlA/R9wVAAAAGXapFCOOhZJjttO/Z1siFJighsAM3huMiKwAypo7$'\n'AAAAABepFEsgPEt7f4VFiKFCrGqixxSWvpZVhwAAAAAAAAD/////AUEEYjJpOVJc$'\n'aXeB3CgM2Apdk/GpLtSA4wrfvuvhsbweFhs/BqIi3MpiugRcdTbTox3iVUhCpYon$'\n'BrmWu9/jvMRqjwAAAjQBAAAACxEJBxl2qRQApYwGgX3k4qNc5Y6XjJCDxNbZd4is$'\n'UF5R1hUAAAAAAAROT05FAAAANAEAAAALEQkHGXapFENn/3Vq4y59O99BpXN6HZTH$'\n'Zbh8iKwA4fUFAAAAAAAABE5PTkUAAAA=$'\n'================================================================

EDIT 4: (Final edit, I hope!) Hmmm. I'm not able to reproduce #8. Works for me. Sanity check: Are you using a testnet wallet in that particular example? If you are, PM me and we'll discuss what's going on.

With that said, everything other than #4 and #8 has been handled one way or another. I'll get the rest tomorrow.

[PRab]

Looks like one of my testnet wallets is producing invalid transactions with the latest version (0.91.99.11-beta). Every time I try to send bitcoind disconnects then reconnects and gives the log message "2014-07-20 00:36:46 ERROR: AcceptToMemoryPool : inputs already spent". I was doing some rapid spending/receiving before I did the upgrade, but everything was already confirmed before I shutdown armory to do the upgrade. I'm guessing that doing a "Rescan Database" will clear it up, but I'll hold off doing that in case you guys need some other log/trace data.