Topic: History of Hardforks and Rollbacks in Bitcoin (Read 14154 times)

Can you guys provide sources for your varying definitions of "soft fork" and "hard fork?"

No.

32 MB was the upper limit for network messages, 32 MB blocks were never valid.



Wrong also.

Soft fork=adding new rules that do not conflict with existing rules, hard fork = changing or removing any rules.

In the early days of the currency, blocks could carry up to 36MB of transaction data apiece. However, in 2010, this was reduced to 1MB.

Soft fork=more restrictions. Hard fork = fewer restrictions.

In the early days of the currency, blocks could carry up to 36MB of transaction data apiece. However, in 2010, this was reduced to 1MB.

That is, mining using a node before that change would today produce blocks rejected by recent later versions. This is by definition a hard fork, isn't it?

Well there's a crucial distinction to be made between
a) Sudden, completely unexpected hard forks due to bugs/unexpected behaviour in the code (bitcoin itself, or ancillary stuff such as database management), and
b) Hard forks predicted and planned several months in advance, due to perceived need to improve weaknesses in the protocol or code or both, and implemented by a consensus-of-miners-switchover

The bazillion-btc and broken-db hardforks that happened in 2010 and 2013 are examples of the former. Increasing the blocksize to 20MB, 8MB or whatever is an example of the latter. In the former everyone could see that the 'new' fork had highly undesirable behaviour and so it was straightforward for the devs to tell everyone to roll back - no one thought the 'new' chain was somehow The One True Bitcoin. In the latter, there is a difference of opinion as to the 'brokenness' of each subsequent chain.

(I know the OP probably wasn't 'going there', but with all the angst at the moment, I think it *needs* to be said.)
(In any case I was mostly responding to pereira4's last sentence.)

Interesting post. Its wild how the overflow gave someone the ability to allow a value out that exceed the blockreward at the time.

I think that the only two hardforks were:
- The change in the version message which took effect on February 20, 2012 after two years of advance notice.

- BIP 50

8th August 2010 - 92 billion BTC into existence

On 8th August 2010 bitcoin developer Jeff Garzik wrote what could be mildly described as the biggest understatement since Apollo 13 told Houston: “We’ve had a problem here.”. “The ‘value out’ in this block is quite strange,” he wrote on bitcointalk.org, referring to a block that had somehow contained 92 billion BTC, which is precisely 91,979,000,000 more bitcoin than is ever supposed to exist. CVE-2010-5139 (CVE meaning ‘common vulnerability and exposures’) was frighteningly simple and exploited to the point of farce by an unknown attacker. In technical language, the bug is known as a number overflow error.So instead of the system counting up 98, 99, 100, 101, for example, it broke at 99 and went to zero (or -100) instead of 100. In layman’s terms, someone found a way to flood the code and create a ridiculously large amount of bitcoin in the process.

The fix was the bitcoin equivalent of dying in a video game and restarting from the last save point. The community simply hit ‘undo’, jumping back to the point in the blockchain before the hack occurred and starting anew from there; all of the transactions made after the bug was exploited – but before the fix was implemented – were effectively cancelled.

How serious was it? Bitcoin’s lead developer Wladimir Van Der Laan is pretty blunt about it, telling me: “It was the worst problem ever.”

Source1: http://www.coindesk.com/9-biggest-screwups-bitcoin-history/
Source2: https://bitcointalksearch.org/topic/strange-block-74638-822

11/12 March 2013 - Chain Fork Information
What happened: A bitcoin miner running version 0.8.0 created a large block (at height 225,430) that is incompatible with earlier versions of Bitcoin. The result was a block chain fork, with miners, merchants and users running the new version of bitcoin accepting, and building on, that block, and miners, merchants and users running older versions of bitcoin rejecting it and creating their own block chain.

What is being done:Large mining pools running version 0.8.0 were asked to switch back to version 0.7, to create a single block chain compatible with all bitcoin software.

Questions & Answers

I'm not a miner or a merchant, what should I do?
Nothing. Your bitcoin software will switch to the correct chain automatically, no matter which version you are running.

Are my bitcoins safe?
Yes.

What will be done
The core developers have investigated what caused the old versions to reject the new blocks, and have released a 0.8.1 version that avoids creating blocks that are incompatible with older versions. A full post-mortem document has been published.

Source1: https://bitcoin.org/en/alert/2013-03-11-chain-fork
Source2: http://bitcoinmagazine.com/3668/bitcoin-network-shaken-by-blockchain-fork/