Got this one figured out: io-streams crate gives me unbuffered input, so no secrets will be leaked via buffers here. Now to the next secret leak…
Unbelievable but true: I have it all ironed out. All the implicit input/output buffers, all the timing issues, and even most of the OS-specific weirdness when it comes to searching a process’ memory for leftover secrets. 🥳
And apparently, the answer is: I compile with my own allocator. This way I can not only log all allocations, I can also ignore deallocations to make sure no two data structures share a memory location. Rather smelly code but it works.
https://github.com/palant/pfp-cli/commit/23c3cea5716257dfa0548a795ea5457d47b8b0ca
Well, I have a dilemma: reading a password from stdin via usual I/O leaves that password in memory, due to a libc buffer I think. Reading the password properly via rpassword does not but it isn’t compatible with integration tests (the ones searching memory for secrets). Heh…