You wouldn't be the first.  Right now, this is the cost of the features that come with Armory.  However, I should be able to pull down the system req'ts after I'm done with the wallets (at the expense of using more HDD space).  And eventually I want to have an actual-lite version (also known as "Standard usermode" -- it would be a purely pruned version, and only track your balance, not the transaction history)

I would think this points towards supporting a lighter (but not entirely light) mode in Armory: https://en.bitcoin.it/wiki/BIP_0037

That would of course require that Armory maintains the blockchain itself instead of relying on the reference client.  But as the ref. client is open source, it should be possible to merge the relevant parts into Armory.  That would drop the requirement of running the reference client in the background, and could form the basis for a development where only a partial block chain is downloaded and stored.  I am sure it will take ages before BIP 37 is implemented, but etotheipi has probably lots of other more urgent things to keep him busy anyway :-)

I just want to say that I've been using the latest updates from the git threading branch without issue.

I've done online and offline transactions (creating and broadcasting only, haven't updated the offline signing install yet - building for the rasbpi is slow), creating new keys and receiving coins for both online and offline wallets.

Armory does not support coin-control, but I think it will not be terribly difficult to support.  I just have some other priorities.  I *do* plan to add it when I'm feeling stuck with some major other feature upgrade

Ok, I found something that's broken. When I try to sign a message with a key that's in a watching only wallet, it crashes hard with this in the debug log:

Well then, don't do that!   

To be fair, it sounds like I need to put some more protections on some of the dialogs when the user is using a watching-only wallet.  I thought I had caught a lot... but I guess not yet.   Getting people to test has been hard, so I keep finding stupid stuff like this that has been in there months.  Perhaps I need to start offering 0.1 BTC bounties for finding bugs in the software -- I bet I'd get a lot of people digging in hard!  Or I might go broke...




He actually already contacted me and I suggested that there will be room in the new wallet format for including "extra attributes" with the addresses.  I don't plan to mess with colored coins myself, but it might make it easily extensible by others.



You're not the first person, by far, who has asked for this.  I suppose I could bump it on the priority list for after beta -- because not only is it in high demand, I don't think it will be that hard.  I think I already know how to do it, I just need to devote a day or two to implementing and testing.  Luckily, the underlying libraries already have all the hooks I need (I just create a new wallet in RAM with only the specified addresses, and retrieving its balance and creating transactions with it is trivial).  And I think I already know how to create the window to let you select the set of addresses from which you want to spend coins...

Maybe I'll do it this weekend!  I am already kinda stuck with the new wallet format, while I wait for feedback about my design ideas...

Armory's not very good at "drinking from the firehose", as I've heard it described.

Btw, the latest version should have a warning window pop up now, telling you it detected this condition, and please stop Armory until Bitcoin-Qt is sync'd.  It's not the best solution.  But it makes the user aware that the bizarre behavior is not a "feature."

What is a DVC tip?  I always accept tips

Given the magnitude of implementation complexity, I guess it makes sense to jump into coin-control now... compared to how long it will take to do the new wallets, it looks like noise...

Amazing!  Now that I have all this experience with PyQt and a super-versatile library under-the-hood, making the interface was not hard at all!  How's it look? 



That's purely the interface, showing you all non-zero-balance addresses, and updating the send-BTC dialog to show you the result of your selection.  It doesn't actually work yet (it still selects from all addr).  Getting that part working is my project for after dinner...