Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 12 Sep 2012 13:04:20 -0700 (PDT)
From:      Jeffrey Bouquet <jeffreybouquet@yahoo.com>
To:        freebsd-ports@freebsd.org
Subject:   Re: [HEADS-UP] Announcing the end of port CVS
Message-ID:  <1347480260.41942.YahooMailClassic@web111301.mail.gq1.yahoo.com>
In-Reply-To: <CADLo838sde6F5E3GKd92bCjbDtG%2BjTRBXy%2BXSqkXmdJAPBD9pg@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
=0A=0A--- On Wed, 9/12/12, Chris Rees <crees@FreeBSD.org> wrote:=0A=0A> Fro=
m: Chris Rees <crees@FreeBSD.org>=0A> Subject: Re: [HEADS-UP] Announcing th=
e end of port CVS=0A> To: "Jeffrey Bouquet" <jeffreybouquet@yahoo.com>=0A> =
Cc: "Beat Gaetzi" <beat@freebsd.org>, freebsd-ports@freebsd.org=0A> Date: W=
ednesday, September 12, 2012, 12:07 PM=0A> On 12 September 2012 15:14, Jeff=
rey=0A> Bouquet <jeffreybouquet@yahoo.com>=0A> wrote:=0A> >=0A> > --- On We=
d, 9/12/12, Jeffrey Bouquet <jeffreybouquet@yahoo.com>=0A> wrote:=0A> >=0A>=
 >> From: Jeffrey Bouquet <jeffreybouquet@yahoo.com>=0A> >> Subject: Re: [H=
EADS-UP] Announcing the end of port=0A> CVS=0A> >> To: freebsd-ports@FreeBS=
D.org,=0A> "Beat Gaetzi" <beat@FreeBSD.org>=0A> >> Date: Wednesday, Septemb=
er 12, 2012, 6:31 AM=0A> >>=0A> >>=0A> >> --- On Fri, 9/7/12, Beat Gaetzi <=
beat@FreeBSD.org>=0A> >> wrote:=0A> >>=0A> >> > From: Beat Gaetzi <beat@Fre=
eBSD.org>=0A> >> > Subject: [HEADS-UP] Announcing the end of port=0A> CVS=
=0A> >> > To: freebsd-ports@FreeBSD.org=0A> >> > Date: Friday, September 7,=
 2012, 5:36 AM=0A> >> > The development of FreeBSD ports is=0A> >> > done i=
n Subversion nowadays.=0A> >> > For the sake of compatibility a Subversion =
to=0A> CVS=0A> >> exporter=0A> >> > is=0A> >> > in place which has some lim=
itations. For=0A> CVSup=0A> >> mirroring=0A> >> > cvsup=0A> >> > based on E=
zm3 is used which breaks regularly=0A> especially=0A> >> on=0A> >> > amd64=
=0A> >> > and with Clang and becomes more and more=0A> >> unmaintainable.=
=0A> >> >=0A> >> > For those reasons by February 28th 2013 the=0A> FreeBSD=
=0A> >> ports=0A> >> > tree will=0A> >> > no longer be exported to CVS. The=
refore ports=0A> tree=0A> >> updates=0A> >> > via CVS=0A> >> > or CVSup wil=
l no longer available after that=0A> date. All=0A> >> users=0A> >> > who us=
e=0A> >> > CVS or CVSup to update the ports tree are=0A> encouraged=0A> >> =
to=0A> >> > switch to=0A> >> > portsnap(8) [1] or for users which need more=
=0A> control=0A> >> over=0A> >> > their ports=0A> >> > collection checkout =
use Subversion directly:=0A> >> >=0A> >> > % svn co https://svn0.us-west.Fr=
eeBSD.org/ports/head /usr/ports=0A> >> >=0A> >> > and update a checked out =
repository using:=0A> >> >=0A> >> > % cd /usr/ports && svn update=0A> >> >=
=0A> >> > Advanced users, or larger sites, might=0A> consider setting=0A> >=
> up a=0A> >> > local=0A> >> > svn mirror. Both for people doing direct=0A>=
 checkouts and=0A> >> for=0A> >> > people=0A> >> > wanting to use a local m=
irror, they can access=0A> one of=0A> >> the=0A> >> > public=0A> >> > subve=
rsion servers [2].=0A> >> >=0A> >> > How to set up a Subversion mirror usin=
g=0A> svnsync(1) is=0A> >> > described in=0A> >> > the FreeBSD Committers G=
uide [3]. Initial=0A> seeds to set=0A> >> up a=0A> >> > svnsync=0A> >> > mi=
rror are provided on the FreeBSD FTP mirror=0A> sites=0A> >> under=0A> >> >=
 /pub/FreeBSD/development/subversion/.=0A> >> >=0A> >> > Binary packages fo=
r pkg_install are still=0A> provided via=0A> >> the=0A> >> > FTP mirror=0A>=
 >> > network. There is also pkgng which is a=0A> feature rich=0A> >> > rep=
lacement tool=0A> >> > for pkg_install available in the ports tree=0A> unde=
r=0A> >> > ports/ports-mgmt/pkg.=0A> >> > Packages for pkgng are available =
on=0A> pkg.FreeBSD.org.=0A> >> >=0A> >> > To use pkg.FreeBSD.org at least p=
kgng 1.0 RC6=0A> is needed=0A> >> and=0A> >> > can be=0A> >> > enabled in p=
kg.conf like this (where ${ABI}=0A> is=0A> >> dependent on=0A> >> > your=0A=
> >> > system):=0A> >> > PACKAGESITE=A0 =A0 =A0=0A> =A0=A0=A0: http://pkg.f=
reebsd.org/${ABI}/latest=0A> >> > SRV_MIRRORS=A0 =A0 =A0=0A> =A0=A0=A0: YES=
=0A> >> >=0A> >> > With pkgng 1.0 SRV_MIRRORS is enabled by=0A> default and=
=0A> >> no=0A> >> > longer needs=0A> >> > to be set explicitly. If pkgng pr=
ior to 1.0=0A> RC6 is=0A> >> used=0A> >> > http://pkgbeta.FreeBSD.org can b=
e used as=0A> packagesite=0A> >> > instead.=0A> >> >=0A> >> > Please keep i=
m mind that the pkgng=0A> infrastructure is=0A> >> still=0A> >> > considere=
d=0A> >> > as beta. More information about pkgng can be=0A> found at=0A> >>=
 > http://wiki.FreeBSD.org/pkgng and https://github.com/pkgng/pkgng.=0A>; >>=
 >=0A> >> > Beat, on behalf of portmgr@=0A> >> >=0A> >> > [1] http://www.Fr=
eeBSD.org/doc/handbook/updating-upgrading-portsnap.html=0A> >> > [2] http:/=
/www.FreeBSD.org/doc/handbook/mirrors-svn.html=0A> >> > [3]=0A> >> > http:/=
/www.FreeBSD.org/doc/en_US.ISO8859-1/articles/committers-guide/subversion-p=
rimer.html=0A> >> >=0A> _______________________________________________=0A>=
 >> > freebsd-ports@freebsd.org=0A> >> > mailing list=0A> >> > http://lists=
.freebsd.org/mailman/listinfo/freebsd-ports=0A> >> > To unsubscribe, send a=
ny mail to "freebsd-ports-unsubscribe@freebsd.org"=0A> >> >=0A> >> [1] Shou=
ld not this go in UPDATING now for persons=0A> who have=0A> >> it=0A> >> se=
t in cron and do not read this list?=A0 Thus=0A> they would=0A> >> have tim=
e=0A> >> to prepare adequately or to ask questions at the=0A> minimum.=0A> =
>>=0A> >> [2] Any URL of sites which would be portsnap or svn=0A> updated,=
=0A> >> yet=0A> >> export via a cvs server for persons to continue=0A> usin=
g=0A> >> csup/cvsup?=0A> >>=0A> >> I had a random thought that this change =
could be=0A> delayed one=0A> >> release=0A> >> so that csup could depend up=
on a new .so. "on=0A> purpose" in=0A> >> v10 that=0A> >> would notify the u=
ser somehow that it is deprecated=0A> in=0A> >> v11... but=0A> >> that negl=
ects cvsup...=0A> >>=0A> >> J. Bouquet=0A> >> _____________________________=
__________________=0A> >> freebsd-ports@freebsd.org=0A> >> mailing list=0A>=
 >> http://lists.freebsd.org/mailman/listinfo/freebsd-ports=0A>; >> To unsub=
scribe, send any mail to "freebsd-ports-unsubscribe@freebsd.org"=0A> >>=0A>=
 >=0A> > As an example of a problem they may encounter, I am=0A> stuck at=
=0A> > crafting a solution (a .svn or portsnap ports tree,=0A> which I=0A> =
> understand may not allow (at least without specific=0A> > commands to "ve=
rsion" them...) local log files, local=0A> > Makefile.local, etc...)=0A> =
=0A> You don't understand correctly. You've been told it's fine=0A> before;=
 svn=0A> ignores any files it doesn't know about.=0A> =0A> > And a ports tr=
ee which includes many=0A> > of the latter.=0A> >=0A> > It seems it would b=
e somewhat of a three-way merge=0A> > which at many points would be not ade=
quately scripted,=0A> since=0A> > copies from the newer (.svn, portsnap) ma=
y not be=0A> adequately mirrored to the=0A> > more-files-included (as above=
) "final working ports=0A> tree", etc, unless it=0A> > specifically copied/=
gcp'd/rsync'd different types of=0A> directories=0A> > within the tree (fin=
d -depth, /files/, /src/, etc=0A> etc...) so as=0A> > to remove as well as =
add files depending upon the=0A> location of=0A> > the directory within the=
 port's PWD.=0A> > It seems doable, but at the expense of many many hours=
=0A> of=0A> > scripting.=A0 Unless someone's done it already and=0A> put up=
 a guide=0A> > somewhere.=0A> >=0A> > Or I'm approaching it wrong, and simp=
ly want a ports=0A> tree server=0A> > on the network... or some near-equiva=
lent.=0A> =0A> NFS?=0A> =0A> > Apologies if there is some simple step/solut=
ion that I=0A> read and=0A> > have not comprehended in a man page already, =
or a wiki=0A> page, or=0A> > equivalent.=0A> =0A> If you made clear exactly=
 what you want, someone could point=0A> you=0A> towards documentation.=0A> =
=0A> >From what I can make of above, you want a three-way=0A> merge of port=
snap and svn?=0A> =0A> > Again, I see this as impacting more than just a fe=
w of=0A> some admin's who=0A> > use csup/cvsup as a matter of course, and m=
ay have a=0A> few=0A> > critical FreeBSD machines in a large infrastructure=
, or=0A> many=0A> > in a smaller infrastructure, and their updating=0A> pro=
cedure=0A> > may thus stall irrevocably (even worse, a remote=0A> upgrade o=
ne=0A> > version to another, or an early version 9 to a later=0A> version 9=
,=0A> > and a few hours of downtime to cvsup, say 2-3 am and=0A> some may=
=0A> > discover the command no longer works... maybe breaking=0A> some=0A> =
> process of site (or router) (or firewall) (or=0A> network...) ...)=0A> =
=0A> This is why over six months' notice has been given.=0A> =0A> Chris=0A>=
 _______________________________________________=0A> freebsd-ports@freebsd.=
org=0A> mailing list=0A> http://lists.freebsd.org/mailman/listinfo/freebsd-=
ports=0A> To unsubscribe, send any mail to "freebsd-ports-unsubscribe@freeb=
sd.org"=0A> =0A=0AAlright. I've not tried the precise older-files-rsync-ove=
r .svn yet.=0AThanks for the information, I've put it on the list of stuff =
to=0Atry if/when I get the time very soon.=0A=0A(Workable probably from a t=
humbdrive for some machines, some of=0Awhich do not have space for a .svn r=
epository on that filesystem)=0A=0ASounds maybe more workable than the port=
snap (man portsnap > update=0Asuggests local files will be gone.  In the in=
stance I tried,=0A"/usr/ports was not created by portsnap" although it was.=
.. maybe=0Adue to post-portsnap addition of files.)=0A...=0AI've not seen a=
 long web page with screenshots of .svn=0Aupdating equivalent to csup/cvsup=
... but that sort of ignores my=0Aconcerns about the points of failure.  Gr=
anted, .svn should=0Anot be relied upon by those unfamiliar with its use an=
d the=0Aports tree, but a beginner's guide in that regard maybe should=0Abe=
 in place before .svn officially entirely replaces csup/csvup...=0ACsup/cvs=
up is a 'binary' failure, usually, it works or it doesn't.=0AMaybe a .svn-a=
nd-ports wiki webpage where the (five?) most common=0Aequivalents, and (fiv=
e?) most common errors, could be expounded=0Aupon....=0A=0A=0AI also though=
t of today a scripted scheme using portmaster's=0Aportmaster -L --index-onl=
y...=0A=0AHowsoever, if it is this complex a workaround(s) (hours setup at=
=0Aeach machine here, and subsequent testing, only one of which I=0Ause day=
-to-day) (and still dread the pkgng changeover), I suspect=0Athat there are=
 a at least some CPU running FreeBSD where the =0Aadmins/users *don't* read=
 the ports list, forum, as closely as they=0Awould a notice in UPDATING (af=
ter all, pages and pages of forum search reveal csup/cvsup solely, with a s=
ide mention of the portsnap man page).  =0A  That is why I thought that=0At=
he notice to this list, should go in UPDATING right away, even=0Aif only te=
ntative.  Then maybe other persons also trying=0Athe switchover right away =
could write more threads or posts =0Aregarding any issues or concerns... =
=0A=0AJ. Bouquet=0A



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1347480260.41942.YahooMailClassic>