I'm to cheap to buy a 1TB disk, I just used 3 500Gb disks and RAIDed them together to get a 1TB RAID 5 disk.

P.S. Great work etotheipi!  I can't wait for the updates your working on.

Let's do this!

https://bt.irlbtc.com/view/299684.0

Please don't use it with large amounts of money yet.  It's probably still chock full of bugs, but I've done quite a bit of internal testing and feel pretty good about the major features.

Besides fixing the rescan-on-every-load issue, the feature set is frozen.  I'd like to fix what's in it, rather than add anything new.  A whole bunch of less-critical things (unicode fixes, message signing, tx size limits), will go into the next version (and that next version will not take 3 months to release ).

Please discuss bugs & general observations in that other thread, to not clutter this one!

Not yet.  That will be 100x easier to implement now, but I wanted to minimize the interface changes on this update, since there was enough turbulence with the new database engine.  So it still uses the local blk*.dat files like the 0.88.1 did (though it doesn't crash on blk*.dat file splits anymore!).  But now that it won't start any operations until Bitcoin-Qt/bitcoind is sync'd, and it's maintaining it's own blockchain database, it should be straightforward to upgrade it.  Especially, because I was careful to make sure that the operations did not depend on the contents/structure of the blk*.dat files in anyway -- it reads them once, and then 100% of ops after that is based on its own DB.  It should even handle switching to a different set of blk*.dat files between loads, as it redetects the blocks every load (part of the headers-first implementation).

Also, I'm slightly concerned about any anti-DoS measures built into Bitcoin-Qt/bitcoind which might cause unexpected issues for the user.  For instance, if I restart a few times, and constantly disconnect, reconnect, and download tons of data, over and over, it might just blacklist the Armory instance.  I don't expect that to happen, but I always expect something silly to crop up and eat another week of my life

bump.

I appreciate the RAM reduction effort happiness (numbers look GREAT!), but the message signing thing is really bugging me ATM. The unencrypt/re-encrypt didn't work with BTC-qt 8.3 or 8.5, so I had to export my addresses to blockchain, which I would really like to avoid. Anything else that can be done?

Thanks.

By the way:  in case you didn't subscribe to the other thread:  I finished the last major feature/bug.  Armory now properly saves all its data between loads.   On my system, managing Bitcoin-Qt/bitcoind myself, I can close and reopen Armory in 19 seconds!  and 260 MB of RAM. 

There still a lot of missing optimizations, but this is no doubt the most usable version yet.  Please try it out!

https://bt.irlbtc.com/view/299684.0

Finished building armory with latest version of LevelDB on msvc11. There are code files and instruction files to pass on, as well as some implementation specifics to discuss, maybe we should proceed in private messages so as to not bloat this thread any further.

1/ I won't mess with

2/ looks like a good temporary measure, I'll have to fiddle with that.

That's an improvement, thanks for sharing!

Can you see the transaction on blochchain.info? If so, the transaction has been transmitted correctly but may depend on unconfirmed inputs, have a huge size and too low fee, or something.

If it is not on blockchain.info, then armory has not sent it (or bitcoind has not relayed it), and it will never confirm.

Close Armory (and Bitcoin-Qt, if running it yourself).   Delete mempool.bin from your Armory home directory,  then restart Armory.   The tx will be forgotten and you can try again. 

It's probably not actually connecting and reconnecting. It doesn't seem to handle notifications correctly in W7 (at least). If you receive a payment, the same can happen, where it "flickers" repeatedly on the screen for 2-10 minutes. I'd guess it lost connection once, then started "flickering." This happens to me more often than not using same version on same OS.