[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ITP: 1-mb-random-data -- one megabyte of pseudo-random data

On Wed, Nov 12, 2003 at 11:17:57AM +0100, Henning Makholm wrote:

| Please, please, no! /dev/urandom does not reliably deliver
| pseudo-random data.  There is a chance that fresh entropy will arrive
| in the middle of the computation and mess up with the pseudoness. 

No, I already covered that in another message.  See:

| This may happen at a buildd even if the package you build yourself
| does seem to be pseudo. How you you imagine fixing that? Or even
| discover that something has gone wrong?

But there are porting teams to handle that.  The kernel /dev/urandom
device already has to handle pseudoness on each architecture anyway, so
if it doesn't work we're really just exposing a bug in the buildd, which
is a Good Thing.

My package has been made and I'm testing it right now.  It works fine
for me.  I know how pseudo randomness works, so RTFMing about entropy
can be done later.  It's not a fundamental flaw in my package, it's just
a bug which I'll soon fix.  You're welcome to file a bug against my
package if it's still there once it enters the archive.

Besides, you're just hassling me about all these problems that the
current randomness in Debian already has!  I'm not making a truly new
randomness package, the maintainer behind the scenes is God who's
already written entropy for the upstream laws of Physics works now.  I
just make the package that is based on his patches.

| The mere fact that you make this suggestion without sufficiently
| researching the subject matter clearly shows that you're inherently
| unable to create Debian packages reliably.

That's objective!

My contribution to the Debian project is being discouraged and ignored
by trolls who clearly don't understand the informal standards of how
Debian works.  There's no point in continuing this thread if all you're
going to do is provide slanderous arguments.

Reply to: