3 ms·
It seems like something more modern, like b3sum would be better... no? What about b2sum?
by encryptluks2 4y ago
It seems like something more modern, like b3sum would be better... no? What about b2sum?
- kortex 4y agoI love the performance of blake3 but my understanding is it's still a bit of the new kid on the block. Blake2 is a SHA-3 finalist so should be perfectly sufficient, plus it has variable digest sizes, reasonably fast, and other nice features. Either way, anything relying on hashes for data integrity should at least be flexible to the option of multiple hash algos. But with git, it's going to be hard enough as is to change to SHA-256, and I don't know how parametric it'll be.
- notfed 4y agoActually, BLAKE2 was not a SHA-3 finalist, BLAKE was. Also, BLAKE3 is so significantly faster than it's probably worth waiting for (waiting for it to get vetted cryptanalysis-wise).
- MrStonedOne 4y ago
- rockdoe 4y agoI mean, SHA-3?
- encryptluks2 4y agoBlake3 is faster and more secure. I also really don't trust NIST to be honest when they have a history of promoting solutions with backdoors.
- oconnor663 4y agoI'm not aware of any reason SHA-3 would be less secure than BLAKE3. Quite the opposite: SHA-3 has a very large security margin, so large that the authors cut the number of rounds in half for KangarooTwelve.