From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Thompson, David" Subject: Re: [PATCH] gnu: Add libsodium Date: Thu, 26 Jun 2014 09:03:18 -0400 Message-ID: References: <87zjh0tsfu.fsf@izanagi.i-did-not-set--mail-host-address--so-tickle-me> <87a98zu99f.fsf@gnu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:55451) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1X09Kr-000552-6s for guix-devel@gnu.org; Thu, 26 Jun 2014 09:03:35 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1X09Ki-0003f7-I7 for guix-devel@gnu.org; Thu, 26 Jun 2014 09:03:29 -0400 Received: from na3sys009aog124.obsmtp.com ([74.125.149.151]:44164) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1X09Ki-0003eZ-B1 for guix-devel@gnu.org; Thu, 26 Jun 2014 09:03:20 -0400 Received: by mail-pb0-f42.google.com with SMTP id ma3so3133629pbc.15 for ; Thu, 26 Jun 2014 06:03:18 -0700 (PDT) In-Reply-To: <87a98zu99f.fsf@gnu.org> List-Id: "Development of GNU Guix and the GNU System distribution." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org Sender: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org To: =?UTF-8?Q?Ludovic_Court=C3=A8s?= Cc: guix-devel On Thu, Jun 26, 2014 at 8:48 AM, Ludovic Court=C3=A8s wrote: > However, the web page reads: > > In order to pick the fastest working implementation of each primitive, > NaCl performs tests and benchmarks at compile-time. Unfortunately, > the resulting library is not guaranteed to work on different hardware. > > Which means that Hydra would end up building a version that uses the > specific ISA extensions that happened to be available on the build > machine, which in turn might be unavailable on the user=E2=80=99s machine= . > > Is there a way to disable the compile-time magic, and instead let > libsodium make the choice at run time? GMP has --enable-fat for that > purpose. Well, Sodium is a fork on NaCL, and underneath that paragraph it reads: Sodium performs tests at run-time, so that the same binary package can still run everywhere. So, I think we are okay! This is a good advantage of Sodium over NaCl. With that concern out of the way, okay to push? > PS: Apparently this NaCl is unrelated to Google=E2=80=99s NaCl sandboxing > thing; terrible! Confused me as well. Thanks. I appreciate your thorough reviews. - Dave