Obsolete packages

[Panagiotis Christias]

On 4/25/06, Kris Kennaway <kris at obsecurity.org> wrote:
> On Mon, Apr 24, 2006 at 01:41:51PM -0500, Eric Schuele wrote:
> > Lena at lena.kiev.ua wrote:
> > >Hi,
> > >
> > >A new version of a port (www/firefox) was released on April 14.
> > >
> > ># portversion -v firefox
> > >firefox-1.5.0.1,1           <  needs updating (port has 1.5.0.2,1)
> > >
> > >But packages still (on April 24) are of previous version:
> > >
> > >$ ftp ftp://ftp.FreeBSD.org/pub/FreeBSD/ports/i386/
> > >ftp> dir packages-5-stable/All/firefox-1*
> > >-rw-r--r--    1 110      0        11188636 Apr 01 16:29
> > >firefox-1.5.0.1_2,1.tbz
> > >ftp> dir packages-6-stable/All/firefox-1*
> > >-rw-r--r--    1 110      0        11511879 Apr 02 10:21
> > >firefox-1.5.0.1_2,1.tbz
> > >ftp> dir packages-7-current/All/firefox-1*
> > >-rw-r--r--    1 110      0        11511428 Apr 03 04:40
> > >firefox-1.5.0.1_2,1.tbz
> > >
> > >Is something broken or is there insufficient computing power for
> > >building new packages more often?
> >
> > It's my understanding that packages are built "when possible".  They
> > often lag that which is in ports.  There are only so many cycles in a
> > day (per cpu and per person).  I would assume that there is some logical
> > order in which the packages are built (most used first? Though not sure
> > how that would be determined)
>
> I continuously rebuild packages using a method that only builds
> "changed" packages (new, updated to new version or with a dependency
> that was changed).  This typically gives a turnaround time on i386 of
> less than a day to several days for packages becoming available, but
> as I said in another reply I'm not uploading them now because of the
> looming release cycle.

With no intention to criticize your way of thinking or your work,
release cycles sometimes could take a bit more time than scheduled.
You, the developers and maintainers, know that better than us, the
users. In the mean time there is a whole community of (end?) users
that could benefit from the prompt availability of latest ports in
packages. I'm referring mostly to desktop or workstation users, since
the most of us build our ports from the sources for our servers.
Although, I'm eager to use the "portupgrade -P" option more often for
our (less critical) ports.

Is there a chance that you, along with the release engineering team,
reconsider your policy?

Regards,
Panagiotis