[Octave-bug-tracker] [bug #58800] BIST for rng sometimes fails

[Markus Mützel]

Follow-up Comment #23, bug #58800 (project octave):

If we always re-initialized with the same state, the sequence of numbers would
always be the same. As we have established, this is not what we want. (Always
re-initializing the random generator with the same state basically corresponds
to zero entropy.)

Later on we raised the requirements and wanted to have a "non-predictable"
sequence of random numbers. Even if that might not be easy, the PID and clock
at the time of initialization of the initial state can be guesstimated.
That is why we add some additional "entropy" from random_device.
While it might be that we add a little bit less entropy on Linux after the
change, 1024 bit should be more than enough if I correctly got Michael
Leitner's point.
The new approach is definitely better on Windows where we didn't add any
additional "entropy" before.

    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?58800>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/