4 ms·
I recall reading in the Bitcoin docs that the Forth-like scripting language was non-Turing complete. In addition, nothing particularly complex was using the scr
by lordnacho 10y ago
I recall reading in the Bitcoin docs that the Forth-like scripting language was non-Turing complete. In addition, nothing particularly complex was using the scripting at the time.
I guess this sort of thing would be the reason. One thing is finding bugs in ordinary software, where the bugs are accidents. It's hard.
Another thing entirely is where you are looking for adversarial bugs. Just look at security articles that appear on HN now and again. They're incredibly complex, and it's not like you can turn off the firehose. When you fix one gap, someone will find another.
I haven't done a lot of reading on ETH, but I would imagine the smart thing to do would be to have some small number of contract types that a lot of people can stare at and try to break. The more attention is distributed among various bespoke contracts, the harder it gets to secure them.
It's like everyone building their own awesome cars, with special bells and whistles, and then asking these non-security engineers to design a lock. Everyone will end up re-learning some painful lessons.
- wscott 10y agoThey are debugging it right now. The ultimate bug bounty program.