Topic: Bitcoin Core / Bitcoin XT (Read 1153 times)

Thanks for the response, and thanks for the work you do. I have sent a donation to help fund Armory development.

AFAIK, XT hasn't changed the names of any of the binaries. You may need to use command line args like "satoshi-datadir" to point to particular directories, and maybe turn off auto-management and run bitcoind manually. (I don't remember offhand if you can tell Armory where to find bitcoind. I think you can.) That should be it.

Pardon my ignorance on the matter, but how does one get Armory to use XT as the engine, rather than Core?

Armory constructs merkle roots, computes block hashes and figures out the longest chain. It will also detect double spends on ZC transaction. Armory does not have an implementation of consensus rules, and thus cannot verify that transactions nor blocks respect them.

Armory relies entirely on the full node that it is connected to. It treats any information it receives as accurate.

This is why armory makes it very difficult to connect to a random node on the network, as it would be very easy to be tricked into thinking that you received a payment when you did not and you could be prevented from sending a transaction that is valid, and it would be difficult to tell that anything is wrong.

Okay. Here's my understanding. Core already does a lot of heavy lifting. It validates blocks, does a lot of work to validate the transactions, etc. It wouldn't make sense for us to turn around and say, "Well, we need to do our own block verification scheme." Instead, we trust whatever comes out of Core, although we obviously need to have some processing logic in order to read the blocks and transactions, along with writing transactions.

In addition, when creating a Tx, we don't actually fire it onto the Bitcoin network ourselves. We form something that should be valid and feed it to Core, letting Core once again do the heavy lifting. If the Tx goes through, cool. If not, we try to give info to the user explaining what might have gone wrong.

Like I said, I could be wrong about some of the details. I'm pretty sure this is correct, though. It just doesn't make sense to, say, take several hours to download the blockchain (including time to verify the blocks) and then take several more re-validating everything. We trust that Core will provide us with good data. This has led to certain issues, like when the blockchain isn't complete. (Core apparently validates a block and then attempts to write it to a file. If the write fails, Core won't care. Obviously, this leads to weirdness with Armory.)

are you saying Armory does not actually validate blocks? or can you elaborate this more pls? (i think you're saying we can switch bitcoin core with XT no problem as Armory just rely on them....)

But i still hope Armory team make a thread to clarify all this.

Thanks

I haven't tested it personally. I think a user tried it with a pre-0.11A version and said it worked fine.

Has Armory been tested with recent versions of BitcoinXT? I take it there are no issues, as XT is pretty much the same as Core, with added capabilities, but I was curious if this setup was already tested by developers and if there's any issue with it.

Thanks, that's good to know.

As I understand things, Armory shouldn't have any problems if it encounters bigger blocks. If there are bugs, fixing them should be pretty trivial. Core/XT does the actual block validation. Armory then runs with the ball and is more concerned with the block's transactions, which aren't affected by the blocksize debate.

(If I'm wrong about any of this, I'm sure other developers will correct me. )

I know Armory runs the Bitcoin client in the background, but what will happen with regards to bigger block sizes being implemented with Bitcoin XT? Will Armory users be able to choose which client they decide to run in the background?

I personally am in favour of larger block sizes.

Thanks.