[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Testing 1.3.x on really old box

From: Daniel Rall <dlr_at_collab.net>
Date: 2006-03-17 23:14:51 CET

On Fri, 17 Mar 2006, Ben Collins-Sussman wrote:

> On 3/17/06, John Peacock <jpeacock@rowman.com> wrote:
> > Jani Averbach wrote:
> > > If your system has a hardware based random source, use that
> > > (CONFIG_HW_RANDOM) and install rngd which feeds randomness from
> > > hw_random to random.
> >
> > That's a server that doesn't get any hand's on use. When I wiggled the
> > mouse and typed something, the entropy went up to 2 (wow!). I found an
> > old sound card and threw that in and suddenly I have randomness to spare
> > (entropy == 3000)!
> >
> > Is this a significant enough limitation that I should write up a FAQ entry?
>
> I had this very same problem back when running the svn tests on FreeBSD in 2001.
>
> The real problem, I think, is that APR needs to grow a second API.
> Subversion should be able to ask for 'not super high quality' random
> numbers when generating uuids, so that the OS doesn't end up blocking
> for entropy.

Until then, a developer FAQ entry certainly provide some reasonable
interim help.

-- 
Daniel Rall

  • application/pgp-signature attachment: stored
Received on Fri Mar 17 23:06:07 2006

This is an archived mail posted to the Subversion Dev mailing list.