From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from pigeon.gentoo.org ([69.77.167.62] helo=lists.gentoo.org) by finch.gentoo.org with esmtp (Exim 4.60) (envelope-from ) id 1Jny4M-0005Dh-NK for garchives@archives.gentoo.org; Mon, 21 Apr 2008 15:36:38 +0000 Received: from pigeon.gentoo.org (localhost [127.0.0.1]) by pigeon.gentoo.org (Postfix) with SMTP id 95F7AE052C; Mon, 21 Apr 2008 15:36:37 +0000 (UTC) Received: from [12.146.151.254] (unknown [12.146.151.254]) by pigeon.gentoo.org (Postfix) with SMTP id 3E93BE052D for ; Mon, 21 Apr 2008 15:36:37 +0000 (UTC) Received: from no.name.available by [12.146.151.254] via smtpd (for pigeon.gentoo.org [69.77.167.62]) with SMTP; Mon, 21 Apr 2008 08:32:49 -0700 Received: by inertia.localdomain (Postfix, from userid 1000) id 212B6890198; Mon, 21 Apr 2008 08:32:38 -0700 (PDT) Subject: Re: [gentoo-catalyst] proper way to deviate in system use flags? From: Chris Gianelloni To: gentoo-catalyst@lists.gentoo.org In-Reply-To: <20080419031948.GA22283@bbone> References: <20080418032929.GC30602@dev.local> <1208556728.4501.91.camel@cgianelloni.quova.com> <20080419031948.GA22283@bbone> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-DEDB7JT8nyKRNhUezTwj" Date: Mon, 21 Apr 2008 08:32:37 -0700 Message-Id: <1208791957.31037.5.camel@inertia.twi-31o2.org> Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Id: Gentoo Linux mail X-BeenThere: gentoo-catalyst@lists.gentoo.org Reply-to: gentoo-catalyst@lists.gentoo.org Mime-Version: 1.0 X-Mailer: Evolution 2.22.1 X-Archives-Salt: 93877c3e-d0c9-42db-9b4e-9bc0e5ec5e27 X-Archives-Hash: 2e22f455dd5cd846ea3f5c4a55a67c1f --=-DEDB7JT8nyKRNhUezTwj Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Sat, 2008-04-19 at 11:19 +0800, Max Arnold wrote: > On Fri, Apr 18, 2008 at 03:12:08PM -0700, Chris Gianelloni wrote: > > > My steps are: stage3-i686-2007.0 + recent snapshot -> stage1 -> stage= 2 -> stage3 -> stage4 > >=20 > > OK, start from an x86 stage3, though. >=20 > Is it critical? I'm shipping only stage4 and it will be deployed on fixe= d=20 > hardware configuration. If you're building a stage1, you should always use a generic stage3 as your seed. It likely isn't critical for you, but it's good practice to get used to doing. > > > So it is seems that my stage4/use and portage_confdir does not affect= system packages > > > (I guess that catalyst does only --emptytree when emerging stage4). > >=20 > > Ehh, not quite. The USE you set in your spec only affects the stage > > build. It doesn't affect what happens *after* the stage build. It als= o > > doesn't affect packages built before your current stage build. One > > thing that you did not mention is what version of catalyst that you're > > using. Newer versions of catalyst, like the currently masked > > 2.0.6_pre17, support writing out spec_prefix/use to make.conf so the > > changes "stick" in your stage. >=20 > I'm using catalyst-2.0.5, but at least on stage4 it puts my stage4/use in= to make.conf > and copies content of portage_confdir to /etc/portage. I've just verifie= d it in > resulting stage4 tarball. Sure, but none of your earlier stages did this correctly and even the stage4 target doesn't get it quite right. Upgrade your catalyst version. > > > So, there are my questions: > > >=20 > > > 1. Am I correct in assumption that use flags are stacked during stage= 4 like this:=20 > > > profile -> stage4/use -> package.use (increasing priority from left t= o right)? > >=20 > > That's over-simplified, but yes. >=20 > The reason I'm asking this is a phrase "If you set the portage configurat= ion dir: package.use=20 > doesn't work how you think it should in catalyst. It's a known bug, but o= ne that won't=20 > likely be fixed for some time. Basically, if you use $target/use, then it= will ignore=20 > package.use settings." here: http://gentoo-wiki.com/HOWTO_build_a_LiveCD > Is that already fixed? It should be *only* in newer catalyst, not the current stable. > > > 2. What is the proper (and simple to maintain) way to deviate slightl= y in system use flags? > >=20 > > The way that you're doing it works fine if you're planning on shipping = a > > stage4 all the time. Were I doing this myself, I'd make a new profile > > with the desired changes, but that's just me. > >=20 > > > My guesses: > > > 1. Add 'hostuse' variable to earlier specs (stage3?) > >=20 > > Already done in newer catalyst versions... >=20 > 2.0.5 uses it: generic_stage_target.py adds hostuse to valid_values list,= and > stage4_chroot.sh mentions $clst_HOSTUSE. But is it safe to simply defin= e hostuse > in stage spec? Is it additive (I see x86.py assigns some values to this = variable) or > arch specific value will be simply overridden and it is better to do not = redefine it? You cannot redefine it. It is defined exactly once. HOSTUSE is literally *only* for subarchitecture-specific USE which controls CPU capabilities/features. It is only designed to be used internally, so if you try to do anything with it, I cannot guarantee what will happen. > > > 4. Create my own profile (don't know where to put it and how to maint= ain during tree updates) > >=20 > > This is what you should do. Simply stick it in $portdir/profiles where > > you'd like it before running your snapshot. To keep "emerge --sync" > > from wiping it out, add > > PORTAGE_RSYNC_EXTRA_OPTS=3D"--exclude=3Dprofiles/myprofilename" to > > your /etc/make.conf on your host/build system. >=20 > Can be custom profile added to portage_overlay on all stages istead of pu= tting it to tree? Yes, it can, but it is safer to put it into your tree that you're using for your snapshots. I suggest keeping a separate portage tree from "/usr/portage" for this, so you don't have to worry about accidentally hosing it. > BTW, what is the difference between profiles/default/linux and profiles/d= efault-linux ? Well, profiles/default/linux is the new location and default-linux will be going away, so that makes it rather simple. ;] > So for now I'll try to upgrade catalyst, add portage_confdir to all stage= s, leave only > stage4/use (for additional world packages) and experiment with profiles. Sounds great! > I also have some more general questions, but I'll ask them in another thr= ead. OK. > Thank you! --=20 Chris Gianelloni Release Engineering Strategic Lead Games Developer --=-DEDB7JT8nyKRNhUezTwj Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEABECAAYFAkgMs5UACgkQkT4lNIS36YEgiACgwCUVrKcSi3Mga60pSPkDRl/L vK4An27MqEq+RWCahcHGUI5FQ3PuhYOK =laDF -----END PGP SIGNATURE----- --=-DEDB7JT8nyKRNhUezTwj-- -- gentoo-catalyst@lists.gentoo.org mailing list