Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 Jul 2014 10:46:17 +0200
From:      Baptiste Daroussin <bapt@FreeBSD.org>
To:        Andrey Chernov <ache@freebsd.org>
Cc:        svn-ports-head@freebsd.org, svn-ports-all@freebsd.org, ports-committers@freebsd.org
Subject:   Re: svn commit: r362996 - head/ports-mgmt/pkg
Message-ID:  <20140728084617.GM50802@ivaldir.etoilebsd.net>
In-Reply-To: <53D46F7F.8020609@freebsd.org>
References:  <201407262308.s6QN85Od018020@svn.freebsd.org> <53D46F7F.8020609@freebsd.org>

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

--+2GlJm56SCtLHYlr
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Jul 27, 2014 at 07:18:23AM +0400, Andrey Chernov wrote:
> On 27.07.2014 3:08, Baptiste Daroussin wrote:
> > Author: bapt
> > Date: Sat Jul 26 23:08:04 2014
> > New Revision: 362996
> > URL: http://svnweb.freebsd.org/changeset/ports/362996
> > QAT: https://qat.redports.org/buildarchive/r362996/
> >=20
> > Log:
> >   Update to 1.3.1
> >   Changes:
>=20
> Of course I notice that pkg becomes worse and worse by each version.
> F.e. 1.3.0 makes portupgrade to be slower a lot comparing to 1.2.*
> But this version surpass any degradation by just causing segmentation
> fault and dropping core in pkg-static in recent 10-stable i386. Now pkg
> can't be installed:
>=20

First you might have missed the release of pkg 1.2 because it was way worse.
Back at the time I have been asking for help to improve testing, testing and
testing.

pkg 1.3 beside the visible bug is in a way better shape that pkg 1.2.0 was =
and
=66rom very far.

if you think you can do better, then come and join the party, help us, pkg-=
devel
have been in the ports tree for a while, the said segmentation fault is due=
 to a
bad pkg.conf file a lot of users have been using (pkg never install pkg.conf
file, but sometime in the lifetime of pkg 1.2 we did provide a bad sample a=
nd
lots of users converted it to regular pkg.conf)

This bug in pkg has been fixed in the meantime.
Note that if people were using tested pre releases of pkg and reporting bug=
 the
regression in reading pkg.conf introduced in devel branch that went in pkg =
1.3.0
would have been figure out leading to the next step which is exposing the b=
ug
that has hit users with a very corner case situation in pkg.conf experiment=
ing a
segfault).

We have very limited resources and a huge technical debt we are trying to
reduce. Hence tons of changes in every single major release, we would like =
less
changes but that would mean having a decent package manager on by 2042...

pkg is quite a complex tools where a simple regression test framework is not
enough in particular for the solver and other parts like this.

We have been asking for help for all these areas for very very long and got
very few people to step in.

I would like to conclude this mail to thanks all the people that have been
testing a lot pkg during the beta/rc phases, thanks to them the stress I
experiment after each .0 release of pkg has moved from a magnitude of 10/10=
 for all
previous releases of pkg to a magnitude of 8/10 for this one.

I would love that 1.3 will only be 5/10 stressing, to please all step in, t=
est,
report bugs, provide patches, help us improving our documentation, help us
having a decent testing framework for all that can be tested automatically.

regards,
Bapt

--+2GlJm56SCtLHYlr
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlPWDdkACgkQ8kTtMUmk6ExYHgCdFWWTnv+0q02nBhT5yqwWcI3t
HHIAoK0B6k/M9OPO1hRu5CdYMksapMe2
=OqQB
-----END PGP SIGNATURE-----

--+2GlJm56SCtLHYlr--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140728084617.GM50802>