[HEADS-UP] Announcing the end of port CVS

Wed Sep 12 14:17:35 UTC 2012

--- On Wed, 9/12/12, Jeffrey Bouquet <jeffreybouquet at yahoo.com> wrote:

> From: Jeffrey Bouquet <jeffreybouquet at yahoo.com>
> Subject: Re: [HEADS-UP] Announcing the end of port CVS
> To: freebsd-ports at FreeBSD.org, "Beat Gaetzi" <beat at FreeBSD.org>
> Date: Wednesday, September 12, 2012, 6:31 AM
> 
> 
> --- On Fri, 9/7/12, Beat Gaetzi <beat at FreeBSD.org>
> wrote:
> 
> > From: Beat Gaetzi <beat at FreeBSD.org>
> > Subject: [HEADS-UP] Announcing the end of port CVS
> > To: freebsd-ports at FreeBSD.org
> > Date: Friday, September 7, 2012, 5:36 AM
> > The development of FreeBSD ports is
> > done in Subversion nowadays.
> > For the sake of compatibility a Subversion to CVS
> exporter
> > is
> > in place which has some limitations. For CVSup
> mirroring
> > cvsup
> > based on Ezm3 is used which breaks regularly especially
> on
> > amd64
> > and with Clang and becomes more and more
> unmaintainable.
> > 
> > For those reasons by February 28th 2013 the FreeBSD
> ports
> > tree will
> > no longer be exported to CVS. Therefore ports tree
> updates
> > via CVS
> > or CVSup will no longer available after that date. All
> users
> > who use
> > CVS or CVSup to update the ports tree are encouraged
> to
> > switch to
> > portsnap(8) [1] or for users which need more control
> over
> > their ports
> > collection checkout use Subversion directly:
> > 
> > % svn co https://svn0.us-west.FreeBSD.org/ports/head /usr/ports
> > 
> > and update a checked out repository using:
> > 
> > % cd /usr/ports && svn update
> > 
> > Advanced users, or larger sites, might consider setting
> up a
> > local
> > svn mirror. Both for people doing direct checkouts and
> for
> > people
> > wanting to use a local mirror, they can access one of
> the
> > public
> > subversion servers [2].
> > 
> > How to set up a Subversion mirror using svnsync(1) is
> > described in
> > the FreeBSD Committers Guide [3]. Initial seeds to set
> up a
> > svnsync
> > mirror are provided on the FreeBSD FTP mirror sites
> under
> > /pub/FreeBSD/development/subversion/.
> > 
> > Binary packages for pkg_install are still provided via
> the
> > FTP mirror
> > network. There is also pkgng which is a feature rich
> > replacement tool
> > for pkg_install available in the ports tree under
> > ports/ports-mgmt/pkg.
> > Packages for pkgng are available on pkg.FreeBSD.org.
> > 
> > To use pkg.FreeBSD.org at least pkgng 1.0 RC6 is needed
> and
> > can be
> > enabled in pkg.conf like this (where ${ABI} is
> dependent on
> > your
> > system):
> > PACKAGESITE         : http://pkg.freebsd.org/${ABI}/latest
> > SRV_MIRRORS         : YES
> > 
> > With pkgng 1.0 SRV_MIRRORS is enabled by default and
> no
> > longer needs
> > to be set explicitly. If pkgng prior to 1.0 RC6 is
> used
> > http://pkgbeta.FreeBSD.org can be used as packagesite
> > instead.
> > 
> > Please keep im mind that the pkgng infrastructure is
> still
> > considered
> > as beta. More information about pkgng can be found at
> > http://wiki.FreeBSD.org/pkgng and https://github.com/pkgng/pkgng.
> > 
> > Beat, on behalf of portmgr@
> > 
> > [1] http://www.FreeBSD.org/doc/handbook/updating-upgrading-portsnap.html
> > [2] http://www.FreeBSD.org/doc/handbook/mirrors-svn.html
> > [3]
> > http://www.FreeBSD.org/doc/en_US.ISO8859-1/articles/committers-guide/subversion-primer.html
> > _______________________________________________
> > freebsd-ports at freebsd.org
> > mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> > To unsubscribe, send any mail to "freebsd-ports-unsubscribe at freebsd.org"
> > 
> [1] Should not this go in UPDATING now for persons who have
> it
> set in cron and do not read this list?  Thus they would
> have time
> to prepare adequately or to ask questions at the minimum.
> 
> [2] Any URL of sites which would be portsnap or svn updated,
> yet
> export via a cvs server for persons to continue using
> csup/cvsup?
> 
> I had a random thought that this change could be delayed one
> release
> so that csup could depend upon a new .so. "on purpose" in
> v10 that
> would notify the user somehow that it is deprecated in
> v11... but
> that neglects cvsup... 
> 
> J. Bouquet
> _______________________________________________
> freebsd-ports at freebsd.org
> mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "freebsd-ports-unsubscribe at freebsd.org"
> 

As an example of a problem they may encounter, I am stuck at
crafting a solution (a .svn or portsnap ports tree, which I 
understand may not allow (at least without specific
commands to "version" them...) local log files, local
Makefile.local, etc...) 

And a ports tree which includes many
of the latter. 

It seems it would be somewhat of a three-way merge
which at many points would be not adequately scripted, since
copies from the newer (.svn, portsnap) may not be adequately mirrored to the
more-files-included (as above) "final working ports tree", etc, unless it
specifically copied/gcp'd/rsync'd different types of directories
within the tree (find -depth, /files/, /src/, etc etc...) so as
to remove as well as add files depending upon the location of
the directory within the port's PWD. 
It seems doable, but at the expense of many many hours of
scripting.  Unless someone's done it already and put up a guide
somewhere. 

Or I'm approaching it wrong, and simply want a ports tree server
on the network... or some near-equivalent.

Apologies if there is some simple step/solution that I read and
have not comprehended in a man page already, or a wiki page, or
equivalent.

Again, I see this as impacting more than just a few of some admin's who
use csup/cvsup as a matter of course, and may have a few 
critical FreeBSD machines in a large infrastructure, or many
in a smaller infrastructure, and their updating procedure
may thus stall irrevocably (even worse, a remote upgrade one
version to another, or an early version 9 to a later version 9,
and a few hours of downtime to cvsup, say 2-3 am and some may
discover the command no longer works... maybe breaking some
process of site (or router) (or firewall) (or network...) ...)

J. Bouquet