Exactly. The whole system falls apart.
Well, we'll see what happens once people read this thread and/or spot the recent block-chain reorganization I guess... 
Topic: [ANNOUNCEMENT] Ixcoin 0.3.24.2 SECOND mandatory update released - page 2. (Read 9386 times)
September 02, 2011, 11:44:29 AM
September 02, 2011, 11:10:00 AM
(Of course, why upgrade if your transactions and mined blocks are then at risk of getting invalidated unexpectedly?)
Exactly. The whole system falls apart.
September 02, 2011, 11:03:26 AM
I think in the upgrade from the old client to the new client that the current blockchain was preserved and coins were preserved. SO... what happens if:
1) The old blockchain is 51%+ of the combined Ixcoin network.
2) The old blockchain is mined quickly while the new blockchain lags.
3) All of the old blockchain users upgrade their client.
Presumably since they now have the new client, the most networking power, and the longest chain, then all of the others who were mining on the new client find their blockchain invalidated?
This is pretty much what should happen, yes. In practice upgrading your client doesn't seem to send the blockchain to the network properly and I ended up having to give it a helping hand. It would probably have happened naturally in the end if enough people had upgraded though. (Of course, why upgrade if your transactions and mined blocks are then at risk of getting invalidated unexpectedly?) 1) The old blockchain is 51%+ of the combined Ixcoin network.
2) The old blockchain is mined quickly while the new blockchain lags.
3) All of the old blockchain users upgrade their client.
Presumably since they now have the new client, the most networking power, and the longest chain, then all of the others who were mining on the new client find their blockchain invalidated?
September 02, 2011, 09:51:29 AM
Still dont get it ?
Get what? Please, clue me in if it's so obvious...Edit: If I'm reading the debug logs right, every IXcoin 0.3.24.1 and 0.3.24.2 nodes I'm connecting to now has 20032 blocks. It's interesting that http://bitcoinx.com/ixcoin/ still says 20029 though.
Edit 2: Oh, and if this is accurate all blocks mined and transactions confirmed on the 0.3.24.1+ branch of the fork since it split off should now be invalid and at 0 confirmations respectively. Too bad for the users on that side, eh?
September 02, 2011, 09:48:13 AM
"...I just can't get it to propagate across the network for some reason..."
And you still don't get it?
I definitely don't. I'm running 0.3.24.2 clients, they're connected to other nodes with the changes, when I pointed a 0.3.24.2 client at another node which had the 20032-block long chain it quite happily downloaded it and accepted it as the valid longest chain, and yet it still doesn't seem to have propagated to the rest of the network. There are even other nodes out there with it now; I can start a fresh 0.3.24.2 node with a clean datadir and pick up 20032 blocks.And you still don't get it?
Still dont get it ?
September 02, 2011, 09:43:57 AM
"...I just can't get it to propagate across the network for some reason..."
And you still don't get it?
I definitely don't. I'm running 0.3.24.2 clients, they're connected to other nodes with the changes, when I pointed a 0.3.24.2 client at another node which had the 20032-block long chain it quite happily downloaded it and accepted it as the valid longest chain, and yet it still doesn't seem to have propagated to the rest of the network. There are even other nodes out there with it now; I can start a fresh 0.3.24.2 node with a clean datadir and pick up 20032 blocks. And you still don't get it?
September 02, 2011, 09:31:52 AM
On Sept 1rst 2011 at 00:12:51 UTC (in under a week), the new client will switch over to a new Ixcoin-specific pchMessageStart peering marker from the original Bitcoin marker. This change was recommended by a few although we haven't had reports of any issues with this yet. Better safe than sorry. The updated Ixcoin nodes will not be able to communicate with non-updated Ixcoin clients after that date. This will require a restart of the Ixcoin client on or after Sept 1rst. (commit)"
That only stops the two halves of the network from communicating directly; it doesn't prevent blocks generated on one half being valid on the other if you can transfer them over in some other way. In fact I've tested this with a pair of nodes locally and 0.3.24.2 does appear to consider the longer 20032-block chain perfectly valid, I just can't get it to propagate across the network for some reason. Something's really quite broken..."...I just can't get it to propagate across the network for some reason..."
And you still don't get it?
September 02, 2011, 09:20:43 AM
On Sept 1rst 2011 at 00:12:51 UTC (in under a week), the new client will switch over to a new Ixcoin-specific pchMessageStart peering marker from the original Bitcoin marker. This change was recommended by a few although we haven't had reports of any issues with this yet. Better safe than sorry. The updated Ixcoin nodes will not be able to communicate with non-updated Ixcoin clients after that date. This will require a restart of the Ixcoin client on or after Sept 1rst. (commit)"
That only stops the two halves of the network from communicating directly; it doesn't prevent blocks generated on one half being valid on the other if you can transfer them over in some other way. In fact I've tested this with a pair of nodes locally and 0.3.24.2 does appear to consider the longer 20032-block chain perfectly valid, I just can't get it to propagate across the network for some reason. Something's really quite broken... September 02, 2011, 09:17:09 AM
I think in the upgrade from the old client to the new client that the current blockchain was preserved and coins were preserved. SO... what happens if:
1) The old blockchain is 51%+ of the combined Ixcoin network.
2) The old blockchain is mined quickly while the new blockchain lags.
3) All of the old blockchain users upgrade their client.
Presumably since they now have the new client, the most networking power, and the longest chain, then all of the others who were mining on the new client find their blockchain invalidated?
1) The old blockchain is 51%+ of the combined Ixcoin network.
2) The old blockchain is mined quickly while the new blockchain lags.
3) All of the old blockchain users upgrade their client.
Presumably since they now have the new client, the most networking power, and the longest chain, then all of the others who were mining on the new client find their blockchain invalidated?
September 02, 2011, 09:14:01 AM
OK, can you point me to the change that makes the two chains incompatible? I can't find it...
"New Peering Marker
On Sept 1rst 2011 at 00:12:51 UTC (in under a week), the new client will switch over to a new Ixcoin-specific pchMessageStart peering marker from the original Bitcoin marker. This change was recommended by a few although we haven't had reports of any issues with this yet. Better safe than sorry. The updated Ixcoin nodes will not be able to communicate with non-updated Ixcoin clients after that date. This will require a restart of the Ixcoin client on or after Sept 1rst. (commit)"
September 02, 2011, 08:52:16 AM
Yeah. One for every brain cell used to copy SolidCoin's code to use as a solution for ixcoin's failure.
LOL this thread is friggen hilarious!
No. Thomas strikes me as the kind of guy who is on top of it all, knows precisely what he's doing, and is in total control. There is abundant evidence.
But yeah, I've not laughed so hard for quite a while.
September 02, 2011, 08:37:04 AM
The 2 chains are incompatible... Some people think, if the old chain grows faster, they can overwrite or invalidate the new chain. That is not true, they are just mining on a dead chain ^^
OK, can you point me to the change that makes the two chains incompatible? I can't find it... September 02, 2011, 08:35:40 AM
Are the block chains actually incompatible yet, or is it just that the two halves of the network can't communicate with each other? If they're not actually incompatible - and I can't see any reason why they should be yet because we haven't reached the 20055 threshhold - then any exchange accepting transactions on the half with the smaller hash rate (and it looks like they are on that half) is risking becoming the victim of a double-spend attack by someone that manages to transfer blocks across.
The 2 chains are incompatible... Some people think, if the old chain grows faster, they can overwrite or invalidate the new chain. That is not true, they are just mining on a dead chain ^^
September 02, 2011, 08:34:41 AM
The new chain may have the value, but the old chain has the hashing power (including my own 6970^^) ;-)
Block count: 20032 vs. 20028
Block count: 20032 vs. 20028
Some people seem not to understand the difference between a network based chain split and a forced software chain split ;-)
i am well aware that "old" ixcoin is even more worthless than "new" ixcoin since all of the exchanges have switched and the sole purpose of mining it is to spite Thomas and the new block chain.
I figure mining the old blockchain is a fairly effective way to prove that i would do something utterly useless rather than supporting Thomas' Ixcoin.
And we're still in the lead:
"Old" Ixcoin block count:
http://ixcoin.info/
"New" Ixcoin block count:
http://bitcoinx.com/ixcoin/
September 02, 2011, 08:31:04 AM
The new chain may have the value, but the old chain has the hashing power (including my own 6970^^) ;-)
Block count: 20032 vs. 20028
Are the block chains actually incompatible yet, or is it just that the two halves of the network can't communicate with each other? If they're not actually incompatible - and I can't see any reason why they should be yet because we haven't reached the 20055 threshhold - then any exchange accepting transactions on the half with the smaller hash rate (and it looks like they are on that half) is risking becoming the victim of a double-spend attack by someone that manages to transfer blocks across. Block count: 20032 vs. 20028
September 02, 2011, 08:11:52 AM
The new chain may have the value, but the old chain has the hashing power (including my own 6970^^) ;-)
Block count: 20032 vs. 20028
Block count: 20032 vs. 20028
Some people seem not to understand the difference between a network based chain split and a forced software chain split ;-)
September 02, 2011, 07:07:30 AM
Nobody But I just pointed out that an exchange accepting coins from the old blockchain would never work out
But I just pointed out that an exchange accepting coins from the old blockchain would never work out Yep, therefore the exchanges have already switched, so at the moment only the new chain coins have some value ;-)
The new chain may have the value, but the old chain has the hashing power (including my own 6970^^) ;-)
Block count: 20032 vs. 20028
September 02, 2011, 05:34:58 AM
Nobody But I just pointed out that an exchange accepting coins from the old blockchain would never work out
But I just pointed out that an exchange accepting coins from the old blockchain would never work out Yep, therefore the exchanges have already switched, so at the moment only the new chain coins have some value ;-)
September 02, 2011, 05:33:27 AM
If I use one copy of my wallet.dat for the old client and another one for the new client, I guess you can spend the Ixcoins in both chains?
Yeah, but who would send you something of value using coins on the old blockchain?
Nobody
But I just pointed out that an exchange accepting coins from the old blockchain would never work out September 02, 2011, 05:25:31 AM
If I use one copy of my wallet.dat for the old client and another one for the new client, I guess you can spend the Ixcoins in both chains?
Yeah, but who would send you something of value using coins on the old blockchain?
Jump to: