Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 31 Oct 2012 02:15:37 +0400
From:      Peter Vereshagin <peter@vereshagin.org>
To:        freebsd-questions@freebsd.org
Subject:   Re: Famp Server
Message-ID:  <20121030221537.GE7264@external.screwed.box>
In-Reply-To: <k6paae$vne$1@ger.gmane.org>
References:  <1351488821.57168.YahooMailNeo@web162303.mail.bf1.yahoo.com> <k6mos3$36g$1@ger.gmane.org> <20121030071059.GA10291@external.screwed.box> <k6paae$vne$1@ger.gmane.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Hello.

2012/10/30 15:37:55 -0400 Michael Powell <nightrecon@hotmail.com> => To freebsd-questions@freebsd.org :
MP> I do not know. Never ran Gentoo. Never had the time to try every 'distro'. 
MP> Some imprecise and poor wording on my part when I said something about 
MP> 'all', when I should have said the 3 that I did try.  Mistake on my part.

Totally 'portages'-based 'gentoo' with its 'USE flags' feature is a kind of a
freebsd in a linuxes world.  Opps, and an offtopic here, too. But... there
seems to be somebody porting the portages to freebsd, no? Something like
'creating a freebsd kernel portage' as I see this task. If so then 'debian'
guys are not alone foloowing that way with their 'kFreebsd' for whiles
already.

MP> [snip]
MP> > 
MP> > What if being a proper sysadm means ability to deploy a package on
MP> > thousands (tens of thaousands, etc.) of machines? In formal terms this
MP> > means create a package and install on the every machine. This cannot be a
MP> > several packages because this makes the whole task less quick and the
MP> > solution less reliable. And the creation of such a package can be a more
MP> > trivial task for sysadm than the such of a package installation or upgrade
MP> > by itself.
MP> > 
MP> > Such a metaport can be a person-scale/company-wide solution, not a public
MP> > one so no harm for the general freebsd usage approaches/pphilosophy which
MP> > is a kind of a public stuff.
MP>  
MP> I completely agree with this. However, noting the most recent email it looks 
MP> as if he still hasn't quite got the hang of installing software on FreeBSD 
MP> yet. I believe there is a chapter in the Handbook devoted to it. 

I shoudn't ever judge about what the other side did or didn't (read or didn't
read, know or doesn't know). But it's me.

MP> One would still need to learn how to install software on FreeBSD in order to 
MP> make use of a meta port; after all - it is still the same process. I do not 
MP> think a meta port is a solution for not learning how to install software. 

But metaport build/install process can give a clue about the what it is. Ain't
it a 'learning how to install software'?

MP> My suggestion is centered around the idea that learning to install software 
MP> is a prerequisite to using a meta port. I think we should be guiding him 
MP> towards acquiring that understanding, then if such a meta port comes into 
MP> being he will be able to make use of it as well. 
MP> 
MP> I do not disagree with the potential utility of a meta port. This is a 
MP> 'horse before the cart' situation where one cannot replace the other. We 
MP> should be helping him learn how to install software. A meta port should be a 
MP> separate issue of its own, and not be attempting to replace not knowing how 
MP> to install software.

I hate methodologies and teaching. But here are my cents:

  - Such a metaport creation task can motivate him on learning about the
    'porting and installing software for freebsd' topic by himself.  Ain't it
    great for him and easy for us than learning here about his typically
    known beforehand small steps on that way?

  - Horse before the cart is a problem when it's a vice-versa only.

--
Peter Vereshagin <peter@vereshagin.org> (http://vereshagin.org) pgp: A0E26627 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20121030221537.GE7264>