Bug#537952: Use more random random seed than localtime

[Jonathan Yu]

Good news everyone.

I've slightly rewritten Don's patch; I was getting some test failures,
presumably because it didn't use the phrase way of setting it. Anyway,
I wrote a simple function to grab the seed from Perl_seed which is
then passed into salfph

Here is some output:

newbuildd:/home/jon/tmp9# for a in $(seq 1 50); do perl
-MMath::Random=random_uniform -e 'print join(q( ),
random_uniform(4)),qq(\n)'; done;

0.665908828658169 0.454013021472426 0.488777318757992 0.607060520723529
0.554647520717718 0.128943547587936 0.16766312543832 0.496622666815727
0.458289686103642 0.465999027532487 0.049702776700601 0.934324689403921
0.863641360499671 0.400219364100437 0.230988093947055 0.234758506042172
0.340603666357376 0.458890598735633 0.314381090794873 0.678264332773382
0.920569428824075 0.82850050340525 0.758234171406322 0.442743852098113
0.949115248245558 0.0698253535363614 0.525481791079954 0.321469143649972
0.043541077385187 0.148428738869933 0.457866965755211 0.0477105989378881
0.587249929046372 0.37406720304662 0.833306070338477 0.523574002321712
0.130949275163323 0.458964866125963 0.747197927679785 0.583421630594339

I'm a bit concerned about exposing Perl's seed value to Perl programs
though. It's a bit dangerous since knowing this value for a particular
instance could lead to an algorithmic complexity denial of service
attack.

I would still recommend using Math::TrulyRandom to get some random
data. This data definitely looks good enough for use in simulations
though :-)

Cheers,

Jonathan

On Wed, Jul 22, 2009 at 10:14 AM, Jonathan Yu<jonathan.i.yu at gmail.com> wrote:
> On Wed, Jul 22, 2009 at 6:40 AM, Don Armstrong<don at donarmstrong.com> wrote:
>> On Wed, 22 Jul 2009, Don Armstrong wrote:
>>> On Wed, 22 Jul 2009, Damyan Ivanov wrote:
>>> > I believe the patch Don is about to send will require smaller
>>> > changes. Having some conditional code would be nice so that it
>>> > doesn't break 5.8 installs, but that's it. Then we forward it
>>> > upstream and let them decide.
>>>
>>> Attached is the fairly trivial change to do this.
>>
>> Updated now to work on 32 bit architectures properly.
> Sweet!
>
> Thanks very much for the patch. I'll work on integrating it, though I
> want to do what Damyan suggested and make sure Perl > 5.8 before doing
> so (actually, I don't know if this is something ppport.h provides
> backward compatibility for)
>>
>>
>> Don Armstrong
>>
>> --
>> Your village called.
>> They want their idiot back.
>>  -- xkcd http://xkcd.com/c23.html
>>
>> http://www.donarmstrong.com              http://rzlab.ucr.edu
>>
>> _______________________________________________
>> pkg-perl-maintainers mailing list
>> pkg-perl-maintainers at lists.alioth.debian.org
>> http://lists.alioth.debian.org/mailman/listinfo/pkg-perl-maintainers
>>
>