Kyle

@rhash I'll provide some sample outputs from the cfv util in the coming days. It might be possible to reuse existing localisation if the options were mutually exclusive, as the...

👍 +1 for making the xpi available in releases - at least then users can manually install an update? I feel its a good workaround until something else more robust...

Precursor Q: Are you going with disk encryption?

All good, interesting project to input on :) Its the kinda thing I'd do if I had the time/hardware/etc. So its nice to give input on something I can't do...

> very nice, however i think its a bit tricky to make the default algorithm dynamic: if one node has xxhash and the other node hasn't , it will look...

Additional thoughts. Perhaps the stdout should contain a predictable header (defining the specification of the operation - how `zfs-check` was invoked) which contains the spec of the invocation, the options/arguments...

> That sounds good, first line as a comment with the "hashing" config? (Things like the hash and chunksize and skip option?) Yes, along those lines. I was thinking that...

After studying the `zfs_autobackup/BlockHasher.py` code in a bit more detail I couldn't help but thinking that it "looks OK". In theory sparse file optimisation should be performed at a lower...

Hey. Acknowledged. It might be useful to offer the user control of the hash class via an option. And perhaps default to a (currently) secure but faster hash like sha512....

> i think you're confusing password hash security with block-hash security. > > even with a chunksize of only 1MB, it would be impossible to guess the original file from...