Armory - Discussion Thread

[freewil]

I'd like to imitate `listsinceblock <blockhash>` from bitcoind in armoryd which is similar to `listtransactions` except it only shows new transactions since the given blockhash. Any help/advice on how I might be able to get started with this?

[1715300043]

1715300043

[freewil]

Updates to 0.91-dev branch: (just pushed)

• armoryd.py updates -- a bunch of functions that broke in the update to 0.90 have been fixed.  Most importantly, "listtransactions" and "getledger" now work.  CircusPeanut is working on building a better test environment for it to improve the reliability.

the two parameters for listtransactions don't seem to have any effect on the output of the command - it outputs the same thing regardless.

UPDATE: submitted a pull request to fix this: https://github.com/etotheipi/BitcoinArmory/pull/160

[etotheipi]

UPDATE: submitted a pull request to fix this: https://github.com/etotheipi/BitcoinArmory/pull/160

The change was simple enough I just updated it in my working branch.  It will appear in the next push of 0.91. 

[ncs0ne]

Don't know if it has been mentioned before. Sorry if so.

Armory 0.90-beta on linux

Click on 'send Bitcoins'
Enter address, amount and description
Put into the description one or more foreign characters like é, ö, ..

Click on 'Send!' and nothing happens.
No message, hint or error is displayed - you just wonder why it has no effect.

Log shows

Code:

(ERROR) Traceback (most recent call last):
  File "/usr/lib/armory/qtdialogs.py", line 5244, in createTxAndBroadcast
    txdp = self.validateInputsGetTxDP()
  File "/usr/lib/armory/qtdialogs.py", line 5395, in validateInputsGetTxDP
    self.comments.append(str(self.widgetTable[i][COLS.Comm].text()))
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe4' in position 5: ordinal not in range(128)

Seems like the issue is a missing transfromation toUTF-8, only.
When I change line 5395 in file '/usr/lib/armory/qtdialogs.py' from:

Code:

self.comments.append(str(self.widgetTable[i][COLS.Comm].text()))

to

Code:

self.comments.append(str(self.widgetTable[i][COLS.Comm].text().toUtf8()))

I can enter those foreign characters and get my confirmation popup, but no clue whether sending really works, not tested.



Further I saw this line when starting via bash:

Code:

sni-qt/769" WARN  12:26:19.849 void StatusNotifierItemFactory::connectToSnw() Invalid interface to SNW_SERVICE

And when closing:

Code:

-INFO  - 1389440887: (BlockUtils.cpp:4000) Saving wallet history for next load
(ERROR) armoryengine.py:13216 - Resetting BDM and all wallets
(WARNING) armoryengine.py:661 - Killing process pid=1881

On restart some seconds later:

Code:

(ERROR) ArmoryQt.py:4592 - Error in checkSatoshiVersion
Traceback (most recent call last):
  File "/usr/lib/armory/ArmoryQt.py", line 4555, in checkSatoshiVersion
    self.checkForLatestVersion()
  File "/usr/lib/armory/ArmoryQt.py", line 1220, in checkForLatestVersion
    versionLines = urllib2.urlopen(HTTP_VERSION_FILE, timeout=CLI_OPTIONS.nettimeout)
  File "/usr/lib/python2.7/urllib2.py", line 127, in urlopen
    return _opener.open(url, data, timeout)
  File "/usr/lib/python2.7/urllib2.py", line 404, in open
    response = self._open(req, data)
  File "/usr/lib/python2.7/urllib2.py", line 422, in _open
    '_open', req)
  File "/usr/lib/python2.7/urllib2.py", line 382, in _call_chain
    result = func(*args)
  File "/usr/lib/python2.7/urllib2.py", line 1214, in http_open
    return self.do_open(httplib.HTTPConnection, req)
  File "/usr/lib/python2.7/urllib2.py", line 1187, in do_open
    r = h.getresponse(buffering=True)
  File "/usr/lib/python2.7/httplib.py", line 1045, in getresponse
    response.begin()
  File "/usr/lib/python2.7/httplib.py", line 409, in begin
    version, status, reason = self._read_status()
  File "/usr/lib/python2.7/httplib.py", line 365, in _read_status
    line = self.fp.readline(_MAXLINE + 1)
  File "/usr/lib/python2.7/socket.py", line 476, in readline
    data = self._sock.recv(self._rbufsize)
timeout: timed out

Armory was starting up quite fast, seems like the error in 'checkSatoshiVersion' it's not critical at all, nevertheless it would be nice if you could take a look onto it. Maybe simple try/catch ?

Thanks for listening and your awesome work!

[Ente]

[..]
Put into the description one or more foreign characters like é, ö, ..
[..]

Confirmed, I think Alan commented on this a long time ago already..
I can't change the description afterwards to something with foreign characters neither.

Ente

[Carlton Banks]

[..]
Put into the description one or more foreign characters like é, ö, ..
[..]

Confirmed, I think Alan commented on this a long time ago already..
I can't change the description afterwards to something with foreign characters neither.

Ente

Also confirmed to be fixed in the release version of 0.91 (full Unicode character set)

[goatpig]

We dont currently support unicode characters in the wallet files (where the comments on txn are saved) so these are currently turned off. We plan on having full unicode support for the new wallet format.

As for the other errors:

The BDM closing thing is not an error, it simply is broadcasted through LOGERROR() which starts the line with an (ERROR) tag.

The other errors relate to old code that kinda go left alone and now needs some fixing, but is in no way fatal.

[justusranvier]

Is connecting to a non-local bitcoind something that's coming soon?

[bitpop]

Is connecting to a non-local bitcoind something that's coming soon?

No security is too risky and parsing blocks is too hard at this point

[goatpig]

Is connecting to a non-local bitcoind something that's coming soon?

Not any time soon, but possibly

[etotheipi]

Just a general update:

0.91-dev is mostly broken.  But that's because we've got 3 people pushing changes that require a little bit of interoperation to work properly.  We'll have it sorted out soon.  After all, this is the development branch for a reason.  But if you want a 0.91-dev that works, you'll have to go about a week back in the repo...

On the other hand, lots of great stuff has already been merged/updated in 0.91-dev.  Most of this isn't actually functional yet on 0.91-dev, but will be shortly.

• (Goatpig) MSVS project will now compile with WinXP support if MSVS 2010 Express is installed
• (Goatpig) Should have fully operational unicode support throughout the app.  Wallets themselves will not support it until the 0.92-beta, but then we'll finally be able to start full internationalization
• (Goatpig) Got a solid instructions ironed out for cross-compiling an offline Raspberry Pi build.  Yes, I can now compile the RPi binaries on my Ubuntu system and create the offline bundles from there!
• (Goatpig) Progress bars for a variety of long-running operations
• (Goatpig) Integrated wallet-recovery tool from the menu.
• (CircusPeanut) Full code refactoring.  No more 20k line armoryengine.py
• (CircusPeanut) A plethora of armoryd.py upgrades.   Includes the ability to setup email notifications if you have a source email address
• (CircusPeanut) Upgraded send-bitcoins dialog to include wallet selection
• (CircusPeanut) Fixed occasional send-bitcoins failure on wallets with lots of tiny inputs that was driving some users insane
• (etotheipi) Making Armory P2SH-aware.  Will enable sending to P2SH addresses
• (etotheipi) Improved efficiency in wallet scanning and zero-conf tx handling.  Some improvement already, will be doing more.

Very shortly, I will finally get back to the new wallet format, which will do all sorts of cool new things.  Including multi-sig, which has become my number one priority after 0.91 (especially since we now have employees holding down the other forts ). 

[Ente]

Just a general update:

0.91-dev is mostly broken.  But that's because we've got 3 people pushing changes that require a little bit of interoperation to work properly.  We'll have it sorted out soon.  After all, this is the development branch for a reason.  But if you want a 0.91-dev that works, you'll have to go about a week back in the repo...

On the other hand, lots of great stuff has already been merged/updated in 0.91-dev.  Most of this isn't actually functional yet on 0.91-dev, but will be shortly.

• (Goatpig) MSVS project will now compile with WinXP support if MSVS 2010 Express is installed
• (Goatpig) Should have fully operational unicode support throughout the app.  Wallets themselves will not support it until the 0.92-beta, but then we'll finally be able to start full internationalization
• (Goatpig) Got a solid instructions ironed out for cross-compiling an offline Raspberry Pi build.  Yes, I can now compile the RPi binaries on my Ubuntu system and create the offline bundles from there!
• (Goatpig) Progress bars for a variety of long-running operations
• (Goatpig) Integrated wallet-recovery tool from the menu.
• (CircusPeanut) Full code refactoring.  No more 20k line armoryengine.py
• (CircusPeanut) A plethora of armoryd.py upgrades.   Includes the ability to setup email notifications if you have a source email address
• (CircusPeanut) Upgraded send-bitcoins dialog to include wallet selection
• (CircusPeanut) Fixed occasional send-bitcoins failure on wallets with lots of tiny inputs that was driving some users insane
• (etotheipi) Making Armory P2SH-aware.  Will enable sending to P2SH addresses
• (etotheipi) Improved efficiency in wallet scanning and zero-conf tx handling.  Some improvement already, will be doing more.

Very shortly, I will finally get back to the new wallet format, which will do all sorts of cool new things.  Including multi-sig, which has become my number one priority after 0.91 (especially since we now have employees holding down the other forts ). 

This sounds delicious! Awesome!
New wallet format? Yeah!

Does that mean CircusPeanut is officially on board too?

Ente

[bitpop]

Can old wallets be upgraded? If so, New backup? Also new db is manual?

[inbox]

When importing the root key, if i misspell 1-2 characters, it gives me the same id, and same wallet. Should this be of any concern?
v.0.9

[cp1]

When importing the root key, if i misspell 1-2 characters, it gives me the same id, and same wallet. Should this be of any concern?
v.0.9

That's to help you out in case you incorrectly wrote down a couple letters.

[TierNolan]

When importing the root key, if i misspell 1-2 characters, it gives me the same id, and same wallet. Should this be of any concern?
v.0.9

They have a checksum.  If you make a mistake, it doesn't match.  The software goes through each position and tests all possible options for that letter.  If it finds one that the CRC matches, then it assumes that is correct.

It can only correct one letter per line.  Were your 2 errors on different lines (or maybe beside each other)?

[inbox]

When importing the root key, if i misspell 1-2 characters, it gives me the same id, and same wallet. Should this be of any concern?
v.0.9

That's to help you out in case you incorrectly wrote down a couple letters.

Interesting, can someone help me out find more information about this?

It can only correct one letter per line.  Were your 2 errors on different lines (or maybe beside each other)?

I quickly tested it, they were probably next to each other.

[TierNolan]

Interesting, can someone help me out find more information about this?

I did a reimplementation of the system (back when chaincodes existed).

However, you can also read the armory source itself.

It is a special hex encoding.

Instead of "0123456789abcdef" for the 16 values, it uses "asdfghjkwertuion".

These were picked so that there is less chance of a typo.

Each line is 16 data bytes + 2 CRC bytes.  This means that there are 36 characters (2 characters per byte).

The CRC is the first 2 bytes of SHA256(SHA256(data)).

[inbox]

Interesting, can someone help me out find more information about this?

I did a reimplementation of the system (back when chaincodes existed).

However, you can also read the armory source itself.

It is a special hex encoding.

Instead of "0123456789abcdef" for the 16 values, it uses "asdfghjkwertuion".

These were picked so that there is less chance of a typo.

Each line is 16 data bytes + 2 CRC bytes.  This means that there are 36 characters (2 characters per byte).

The CRC is the first 2 bytes of SHA256(SHA256(data)).

I'll look into it, much appreciated.

[maxmint]

I tried to install the 0.90-beta Offline Bundle for Ubuntu/Debian 12.04-64bit on a Ubuntu 13.10 machine. Other than expected I can start the application and it seems to be running fine.
The download page suggests to use the exact Ubuntu version (which would be 12.04) but unfortunately I'm only able to get 13.10 running on my Macbook Air booting from an USB drive.

So my question is: if the Armory app starts fine, can I then use it for production on this 13.10 system? Or could there be any hidden bugs caused by using the wrong Ubuntu version?