I second this!  Both the +1 and the suggestion.

In fact, I think the trading history could be published in a way that would make it easier to refute at least some future attacks on your site's honesty. 

There could be a link generating a list of all transactions in the last week, with timestamps, amount and prices, and all clearings done.  The list should be autogenerated on the fly so it contains even the last transaction, and it should be cryptographically signed.  Then after trading, users could in principle download the list, verify that their trades are present and as expect - and later verify against a newer list that the logs have not been tampered with in between.  The cryptographic signature prevents a malicious user from altering the list and claiming tampering.

No, it is all save dialogs on the online computer.  Both when saving the unsigned TX, and when saving a CSV file with all the transactions in a wallet.

I am on Mac OS X 10.8.2.  The offline computer is running Ubuntu, and there I have the last official release of Armory.

Maybe I should try reinstalling Qt.

EDIT: Recompiling qt and pyqt seems to have helped, at least partly.  It is too late to test more now, I will test more in the weekend (no time tomorrow, unfortunately).

You didn't - you said it almost looked malicious.


I am sure you are - and don't apologize, we really appreciate all your hard work!

In principle, the problem with the save dialog could be a Qt problem, although I do not think I have updated Qt.

I updated, and tried again.  I wanted to make a test transaction from the offline wallet.  Now I can specify the payment, and get all the way to the window where I should save the transaction.  Unfortunately, that window is totally unresponsive, and I can neither save the transaction nor get rid of the window (short of killing the Python process with Ctrl-C).

The second time I tried, the file name did not appear in the Save window.  I could type in a name myself, but again I could not save the file or close the window, and had to kill Python.


Your words "it almost looks malicious" makes me slightly nervous - and I actually refrained from testing the online wallets! 

What are the odds that a part of a file drops out in such a way that it is still syntactically correct??  On the other hand, if it was malicious, why remove the part of the code leading the user to enter the password - why not just steal the password....

In any case, I suppose git can show you exactly what was changed, and when.

Because those of us not doing anything illegal and not being overly paranoid(*) think that using TOR is inconvenient and slow. 

(*) And that may be a significant fraction of all bitcointers. 

I pulled the latest version (git pull) and compiled it.  Unfortunately, it does not work very well :-(  I did a "make clean" followed by a "make" just in case something was not recompiled, but that did not help.


BUG 1:

I tried to generate a transaction from an offline wallet.  I clicked "Send Bitcoins" and selected the offline wallet.  Then I pressed OK and got this error:

(yes, the error appears twice).

I get the same error when sending from a normal wallet, by the way.  Just tried it.

BUG 2:

Previously to this, I tried to export all transaction from an online wallet.  I chose File / Export Transactions, chose .csv format and selected a wallet (an unencrypted online wallet).  When I click export, a normal Save window appears, but it is dead.  It does not react to pressing any body, and I cannot get rid of it or quit Armory.  In the end I had to press Ctrl-C in the Terminal window where I started Armory.



I am running Mac OS X version 10.8.2 (Mountain Lion).  It is the first time I have any trouble with Armory's basic functionality (although I have never tried the export transaction menu point before).

EDIT:  I use the threading branch.

Yes, it is not obvious at all that it is a tag associated with the address, it looks like it is just an ad being run by your site.  Perhaps you should add a label saying something like "owner-supplied address tag".  I am sure you can find a better wording

I missed the "fun" tonight by around 15 min.

Has the evil manipulator returned, or is everything calm and serene?

So you are a sensible person!

Around 2%  

Bitcoin may crash tomorrow.  Or it may make us rich (no, I don't really believe it will, but it could happen).

Until that happens, it is a fun and elegant experiment.

I think Fireballs explanation makes sense.


Nobody can place money at any exchange and claim that they "couldn't have known" if the site runs off with the money.  Unfortunately.  That severely limits how much money I place anywhere. 

With ICBIT, the risk that they run off with my money is real.  So is the risk that they are closed by the authorities, or are hacked.  I still think the largest risk to my money is the trades I make.   

I don't complain!!

I just thought you stopped doing it because any kind of manual intervention opens you to being attacked for being biased.  But I agree it was a good thing when it happened.

You say you are not TORwallet.  Perhaps you are not, perhaps you are and are just back for more.  You offer the same service, both are anonymous, both hold all the customer funds.

An exchange or a site like SR can get a large revenue from fees compared to the amount they can gain by running away with the customers money, since most people don't store money there for long.  An online wallet is in the opposite situation.  And an anonymous online wallet - quite honestly, forget it!  Only a fool will place money with you guys - even if you are legit, you have set yourself up to look like scammers, including the sign saying "SUCKER-BAIT" in five-feet tall, burning letters.

Bitstamp, Bitcoin-24, Intersango.  But then I am in Europe, transferring to these is cheap.  In the US I would consider using BitFloor and CampBX.  BitFloor's deposit and withdrawal support is often down, though.  And CampBX reserve the right to do all the nasty things MtGox are doing, but I have no experience with them, and don't know if they actually do it.

This makes me wonder if I am doing something wrong.  I have compiled the git source on Mac OS X, and start the program with

Am I missing an executable I should run instead?  If I am starting the Python executable that will load libpython - if Armory then has the same library built-in, that should be asking for trouble.

No.  You saw behavior that you find consistent with scamming.  Others saw otherwise - at least your proof is far from convincing.  It may be convincing what you saw, but certainly not what has been posted here or on the blog.


I am an actual customer.  I have been able to place orders during these manipulation incidents.  And yes, I have also been in situations where I have not been able to place orders for a minute or so while other people obviously were.  But since absolutely nothing interesting happened during those times, I assume it was some kind of network problem.

If Fireball was behind these manipulations, why would he have undone them by an extra clearing one of the first times it happened?   I lost a significant amount due to a manipulation, but instead of liquidating my position (although I was below the limit for a margin call), a second clearing occurred "due to high volatility" half an hour or an hour later.  You can probably still find the announcement in the Twitter feed.  Fireball quickly stopped doing this, probably he realised that this kind of ad-hoc manual interventions are not a good idea. 


That is indeed one interpretation.  Another one is that more and more people (including me) were trying to get a profit from the predictable manipulations, and the same day the spot market price moved against the manipulator.  In combination, this made it too expensive to continue manipulating the market - the manipulator lost a lot that evening!  Of course it did not help him that the clearing was delayed due to maintenance - whether that maintenance was real or just Fireball's excuse for thwarting him.


Don't bother, the moderators will not do anything.  MPOE is entitled to her opinions, and is entitled to post them here.  We are entitled to disagree (and the shrill tone of her posts certainly don't help, even if her points were valid - which they might be!)

It would be very bad indeed if the moderators started censoring people calling a scam - even if they appear to be wrong!!!

If you use some kind of Linux, you probably need to install the python-dev package (the name may vary, it contains stuff to compile Python extensions).

They certainly reserve their right to freeze your account without warning until you provide ID documents.

So before you deposit anything, consider what you will do in that situation.  Even if you don't mind them knowing who you are (I am OK with that!) you may not want to email PDF's of personal documents to a foreign company.  If these documents are sufficiently to prove your ID, they seem to be sufficient to steal your ID too.  And I do mind setting myself up for identity theft.  Hence I have emptied my MtGox account and use other exchanges.

[edit for spelling]

Erm... no...

Although I guess you call it off-site mirroring or something like that if it happens real-time.  Which IMHO is a minimal requirement for a payment infrastructure.  Not that I expect it to be implemented in even a minority of bitcoin-related sites.  But it should be.

Not if the existing platform is copyrighted.

Bitcoinica imploded before I had even heard about it, but wasn't the code released as open source?  If so, it should be OK - at least if the guys releasing it as open source had the copyright when they did it (and I seem to remember that there is some controversy hidden here too).  If not, reusing code that someone else has copyrighted it ... well ... asking for trouble.

Ouch!  That should never be necessary.  Remember the three ways to avoid loosing data: Backups, backups and backups.