Re: $RANDOM not Cryptographically secure pseudorandom number generator

From	Ole Tange <tange@gnu.org>
Newsgroups	gnu.bash.bug
Subject	Re: $RANDOM not Cryptographically secure pseudorandom number generator
Date	2019-01-02 02:29 +0100
Message-ID	<mailman.6674.1546392582.1284.bug-bash@gnu.org> (permalink)
References	(5 earlier) <4bc5800d-0dfb-17a5-0b20-9f4bef5a60b6@case.edu> <CA+4vN7yTJRqc=8eCJWQMXu7nZu7ZreLTEp56SC-LTavSVW-d1A@mail.gmail.com> <CAOSMAus5OvX91r-y8K3Bw7e-T1aj9U72+uE64iM346AV+uSwAA@mail.gmail.com> <CA+4vN7wS_PPH60j_gw2t_2kokds0eZxCzgqAydqiNz5=JLTMxw@mail.gmail.com> <8d46616b-7618-be59-3651-1a01bc3bf7e6@case.edu>

Show all headers | View raw

On Mon, Dec 31, 2018 at 8:12 PM Chet Ramey <chet.ramey@case.edu> wrote:
:
> Thanks for the patch. I'll take a look after I release bash-5.0. One
> question: can you reproduce the same random sequence by using the same
> seed? That's for backwards compatibility, even if the sequences themselves
> differ.

Yes. Seeding with a value will give the same sequence:

$ RANDOM=4; echo $RANDOM $RANDOM
21584 22135

$ RANDOM=4; echo $RANDOM $RANDOM
21584 22135

For backwards compatibility integers are supported, but they are
really parsed as strings. Strings make it easier to seed with more
than 64-bits:

RANDOM=`cat GPLv3.txt`

So these give the same value on 4.4.23, but differs with the patch applied:

RANDOM=$(echo 2^64 | bc );echo $RANDOM
RANDOM=$(echo 2^65 | bc );echo $RANDOM

RANDOM="foo"; echo $RANDOM
RANDOM="bar"; echo $RANDOM

RANDOM=2; echo $RANDOM
RANDOM=" 2.0"; echo $RANDOM
RANDOM=" 2.0noise"; echo $RANDOM


/Ole

Back to gnu.bash.bug | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

Re: $RANDOM not Cryptographically secure pseudorandom number generator Ole Tange <tange@gnu.org> - 2019-01-02 02:29 +0100
  Re: $RANDOM not Cryptographically secure pseudorandom number generator Luuk <luuk@invalid.lan> - 2019-01-02 12:01 +0100

csiph-web