Date: Sat, 15 Aug 2009 20:33:39 +0100 From: Matthew Seaman <m.seaman@infracaninophile.co.uk> To: Thomas Backman <serenity@exscape.org> Cc: Doug Barton <dougb@freebsd.org>, Miroslav Lachman <000.fbsd@quip.cz>, freebsd-ports@freebsd.org Subject: Re: portversion and pkg_version have different opinions on current versions Message-ID: <4A870D93.3090302@infracaninophile.co.uk> In-Reply-To: <6B5B7698-CCD8-48FF-A5FB-0349D4CC1143@exscape.org> References: <B787D58E-9157-48E7-ADF3-E8D54F8AF22F@exscape.org> <4A86FF1E.1030705@quip.cz> <6B5B7698-CCD8-48FF-A5FB-0349D4CC1143@exscape.org>
next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 2440 and 3156)
--------------enig7B0D176C8FAD4EA64DD9F7FB
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: quoted-printable
Thomas Backman wrote:
> However, a new issue appeared... Kind of. I know I read something about=
=20
> portsnap and INDEX on the -current list recently, so I'm guessing this =
> is related? Maybe not, though (see later in the mail).
> libtool-1.5.26 ! Comparison failed
This is because the libtool-1.5.x port -- devel/libtool15 -- was renamed
to devel/libtool22 and the port updated to libtool-1.22.x. See the =20
20090802 entry in /usr/ports/UPDATING for special instructions on how to
deal with this.
> So... Using the index causes problems (or the opposite!). Could I be=20
> using an index for something like HEAD despite not using that ports=20
> tree? (Again, pretty new to this!)
No -- using the INDEX shows you what the INDEX is capable of showing you,=
which isn't everything you need to know when using the ports. You should=
get in the habit of casting an eye over /usr/ports/UPDATING any time you
want to do an upgrade.
There is only one ports tree, and you should be using the HEAD of it no m=
atter
what OS version you're running. There aren't any OS version specific bra=
nches
-- just tags to mark the point in time at which support for a particular =
OS
version was dropped, or to mark the specific set of ports bundled with a =
release.
HEAD is a continuously moving target -- of the order of a hundred port up=
dates
will be checked in daily largely driven by the availability of new upstea=
m
versions.
The generated INDEXes are labeled with the OS major version, out of a cho=
ice
of 5, 6, 7 or 8 -- which are the only OS versions the ports system is set=
up
to work with right now. Actually, I think 5.x support may have been drop=
ped
already. The choice of 6, 7 or 8 covers all of the supported release bra=
nches
and the bleeding edge 8-CURRENT stuff. Once the release of 8.0-STABLE ha=
ppens,
and the bleeding edge moves to 9-CURRENT there will be an INDEX-9 generat=
ed.
The reason for the difference is that certain ports are not supported on =
all OS
versions, some have variations in their dependency lists and some have a =
different
set of default options, so the INDEX comes out differently on the differe=
nt
versions.
> I don't know how the INDEX files work, but I do know (thank you DTrace)=
=20
> that INDEX-8 was the only one read during "pkg_version -vIL=3D".
Yes. You'll only use the INDEX corresponding to the major version of Fre=
eBSD=20
that you have installed. See portindex(5) for details of what the INDEX
contains.
> Oh, and my understanding is that the INDEX-8 is fetched via portsnap?=20
> Running the "fetch update" took less than 20 seconds (the cron job ran =
> about 2 hours ago, though), so I guess it cannot have been built (that =
> does take a lot of time, yes?)?
Yes. Building the INDEX on a fairly beefy machine takes 20min or more an=
d
thrashes disk IO while doing so. The index available for download by:
# cd /usr/ports
# make fetchindex
should be rebuilt approximately hourly, but it isn't unknown for problems=
to
prevent a new INDEX being available for a number of hours. I believe the=
=20
INDEXes portsnap supplies are generated similarly. In fact, depending on=
the
ports management software you use, the INDEX may be pretty much irrelevan=
t --
portmaster(1) ignores it -- or certain discrepancies may be ignored -- po=
rtupgrade(1)
uses data in the INDEX as a guide but double checks against the reality o=
f the
ports tree when working out what needs updating.
If you've made any non-standard settings in /etc/make.conf such as, eg. r=
unning
mysql-5.1.x instead of the default mysql-5.0.x or using apache13 or enabl=
ing LDAP=20
or SASL functionality then you may find it worthwhile to build your own I=
NDEX[*]
as this can help avoid different bits of the system getting conflicting i=
deas
about exactly what depends on what.
Cheers,
Matthew
[*] Modesty prevents me from mentioning ports-mgmt/p5-FreeBSD-Portindex.
--=20
Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard
Flat 3
PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate
Kent, CT11 9PW
--------------enig7B0D176C8FAD4EA64DD9F7FB
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.12 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iEYEAREIAAYFAkqHDZkACgkQ8Mjk52CukIzoDQCeN/qxMxbFBGL5a2Dnlmp2J9p7
AfQAn1ysKOB7R1HpKt2Sp/gT6VAo0gF/
=YJLh
-----END PGP SIGNATURE-----
--------------enig7B0D176C8FAD4EA64DD9F7FB--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A870D93.3090302>