From mboxrd@z Thu Jan 1 00:00:00 1970 MIME-Version: 1.0 In-Reply-To: <2777d5bff012c5d1e3cc60d2591d65fe@brasstown.quanstro.net> References: <2777d5bff012c5d1e3cc60d2591d65fe@brasstown.quanstro.net> Date: Wed, 25 Nov 2009 14:32:25 -0800 Message-ID: <13426df10911251432w76ab8bb1m56884356c659ecd0@mail.gmail.com> From: ron minnich To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: [9fans] after a recent pull ... Topicbox-Message-UUID: a0b8e934-ead5-11e9-9d60-3106f5b1d025 On Wed, Nov 25, 2009 at 2:07 PM, erik quanstrom wro= te: > that performance is only for one case. =A0what about the > case where i'd like to know if a local file differs from sources? Then use replica for that. But replica is just too slow to be a useful for package management. It's not great that it takes longer to get the openssl package than it takes to do a full ubuntu install. > (i suggest contrib/diff as an addition to contrib.) =A0that's > really slow and annoying with a tar file. I don't see that. I would be willing to bet (I'll try it at some point) that it is far faster to pull a tar down, mount it via tarfs, and run replica/pull agains that than what we do now. Were we to do this we'd have two ways to use replica. That said, I'm sufficiently impressed with the tinycore tar-based package management that I don't see the need for replica any more. > also, there's no reason replica can't work in parallel. As always, there's no reason lots of things can't be done. They just aren't= :-) ron