Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 Aug 2012 10:30:58 +0200
From:      Baptiste Daroussin <bapt@FreeBSD.org>
To:        Jeffrey Bouquet <jeffreybouquet@yahoo.com>
Cc:        freebsd-ports@FreeBSD.org
Subject:   Re: pkgng 1.0 release schedule [...concerns...thread...continued...]
Message-ID:  <20120823083057.GC47706@ithaqua.etoilebsd.net>
In-Reply-To: <1345639074.52612.YahooMailClassic@web111313.mail.gq1.yahoo.com>
References:  <1345639074.52612.YahooMailClassic@web111313.mail.gq1.yahoo.com>

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

--rQ2U398070+RC21q
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Aug 22, 2012 at 05:37:54AM -0700, Jeffrey Bouquet wrote:
> I view, rightly-or-wrongly, the mandatory usage of pkgng VS
> /var/db/pkg/portname-number, as somewhat of a showstopper, at least
> without more assurances... I interact daily with /var/db/pkg as
> follows...
>=20
> using the shells' tab-completion of /var/db/pkg/ to more efficiently
> pkg_create, pkg_delete
>=20
> portmaster -d -B -P -i -g /var/db/pkg/ /var/db/pkg/ /var/db/pkg/ /var/db/=
pkg /var/db/pkg/ /var/db/pkg... && yell || yell;
>=20
> #zsh# no space on the machine for pkgng
> #zsh# in /var/db/pkg # does not pkg_complete without the pwd
> #zsh# pkg_delete -f portname-number && pkg_add /mnt/portmaster-download/p=
ortname-number...
>=20
> #var/db/pkg#  ls -lac | grep py26  # upgrading to py27
> #var/db/pkg#  ls -lac | grep diff # what tools can I use
>=20
> make build-depends-list
> ls -lac /var/db/pkg/this
> ls -lac /var/db/pkg/that
> #... which port is less-recently upgraded that might fix this port...
>=20
> /usr/ports/devel/gettext
> cp -iv /var/db/pkg/gett[tab]/+REQUIRED_BY .
> #  reference for prior-to rebuild, esp., if UPDATING ...
> ...
> #  to defer or not the upgrade of a port,=20
>=20
> Any lesser-easily equivalent to these using pkgng, if it involves
> actually writing down the name-number (lacking tab completion), would
> incur a serious time cost AFAIK, not to mention RSI.
>=20
> ...
> Also, 'fails to register' in pkgng... with /var/db/pkg, at least
> one can "make -k install"  ( I should elaborate this concern more,
> but where would the "failed to register"  exist?  Files would be
> installed but ... pkgng would put a file in /work/ detailing its
> failure to register the port and why??
>=20
> I rightly-or-wrongly picture pkgng sort of as a front-end to /var/db/pkg/=
, then removal
> of the latter.  Unless clear equivalents using pkgng to the=20
> CLI and scenarios I've posted above are elucidated, put in a=20
> flowchart  somewhere (or the wiki...)... it *sort of* forces the
> use of packages rather than building from ports  ?? Or I am
> just inexperienced...  =20
>=20

Most of the things you are doing are useless now with pkgng, because pkgng =
do
provide mechanism for them natively. pkgng also provides 'pkg query' which =
is=20
designed to help scripting to gather informations about the local database.

Concerning the tab-completion, I have already gives to some user pointers to
update _portmaster completion (for zsh) function so that it works with pkgn=
g, it
is mostly one line to modify.

I will run a real zsh completion function for pkgng itself one day, given h=
ow
bad is the current completion for pkg_* tools, I don't see it as a show sto=
pper.
Plus it is not because CURRENT will be switch by default to pkgng that pkg_*
tools will be removed at the same time, there will be a cohabitation period=
 to
allow people to upgrade their scripts, learn about pkgng and new way to do
things, etc.

regards,
Bapt

--rQ2U398070+RC21q
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (FreeBSD)

iEYEARECAAYFAlA16kEACgkQ8kTtMUmk6EwQ4gCdH8EJHl5VrggzlIyhkdsVkclg
HHMAnjXRzY8jeU34tFAltNd+4h9WsXt9
=8kKp
-----END PGP SIGNATURE-----

--rQ2U398070+RC21q--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20120823083057.GC47706>