At PR review club I learned about BIP-50 where a change from BerkeleyDB to LevelDB unintentionally caused a chain split because LevelDB could handle blocks w more to inputs.
Incidents like these explain why core development is so conservative but as a developer, I do find it comforting that even the best projects make mistakes. A rite of passage for every young engineer is to accidentally break prod (for me this was at an AdTech company so maybe it was actually a good thing?)
@staciewaleyko I’d rather not have every new developer break production, thanks 🤣
@BitcoinAustrian wait I thought SV’s motto was “build fast and break things”?
@staciewaleyko breaking prod in a serious way will really make you learn the importance of thinking about bigger picture.
@staciewaleyko it wasn't clear at the time BIP50 was written, but the BDB issue had been causing some nodes to fall out of consensus for almost a year before the 0.8 release with the leveldb change. The issue had just never affected a large miner before. If things were a little bit different, it could've blown up before the 0.8 release, or maybe everyone could've updated to leveldb and the bug never discovered until it was irrelevant.
@htimsxela @harding @staciewaleyko It was Part 1 of interview with sipa. Transcript is here: https://diyhpl.us/wiki/transcripts/chaincode-labs/2020-01-28-pieter-wuille/
The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!