ports/150078: security/tor: Recursive dependencies on ports tor and vidalia and weird vidalia default dependencies

Fabian Keil fk at fabiankeil.de
Sun Aug 29 11:50:04 UTC 2010

The following reply was made to PR ports/150078; it has been noted by GNATS.

From: Fabian Keil <fk at fabiankeil.de>
To: bf1783 at gmail.com
Cc: bug-followup at FreeBSD.org, miwi at FreeBSD.org, Jens <jan0sch at gmx.net>
Subject: Re: ports/150078: security/tor: Recursive dependencies on ports tor
 and vidalia and weird vidalia default dependencies
Date: Sun, 29 Aug 2010 13:33:00 +0200

 Content-Type: text/plain; charset=US-ASCII
 Content-Transfer-Encoding: quoted-printable
 "b. f." <bf1783 at googlemail.com> wrote:
 > The solution is to remove the VIDALIA option, and hence any
 > dependencies on net-mgmt/vidalia, from security/tor and
 > security/tor-devel, because this option introduces circular
 > dependencies, and was only there for convenience: there is no actual
 > dependency.
 I agree that the Tor ports should drop the dependency.
 This has already been discussed in 2007 when lx@ first
 reported the problem. IIRC, at that time the Tor maintainer
 agreed as well, so I'm not sure why it hasn't happened.
 > As an aside, I should mention that the dependency handling in
 > net-mgmt/vidalia is problematic.  The use of !=3D and calls to utilities
 > like pkg_info in port Makefiles, outside of targets that are not run
 > during 'make describe'  is actively discouraged, because it makes
 > computing the INDEX more costly, and this is one of the bottlenecks in
 > Ports.  portmgr went to some trouble to remove these calls earlier
 > (see, for example:
 > http://lists.freebsd.org/pipermail/freebsd-ports/2008-July/049777.html
 > http://lists.freebsd.org/pipermail/cvs-ports/2008-July/153224.html
 > ), and they should be removed from vidalia.  Yes, the handling of
 > alternative dependencies is clumsy in Ports, but the tools that most
 > people use for dealing with this, like portmanager and portupgrade,
 > have means of dealing with this problem, and as long as the entry for
 > 'make describe'  is correct for default package builds in clean
 > sandboxes, efficiency is more important. Also, it seems to me that the
 > TOR_DEVEL and TOR OPTIONS can be combined into one option.
 The use of !=3D for the dependency handling is the result of
 a discussion on freebsd-ports@ and at that time nobody had
 any better ideas.
 Content-Type: application/pgp-signature; name=signature.asc
 Content-Disposition: attachment; filename=signature.asc
 Version: GnuPG v2.0.16 (FreeBSD)

More information about the freebsd-ports-bugs mailing list