On Mon, Feb 23, 2009 at 10:19:47AM +0000, Enrico Zini <[email protected]> was heard to say: > thanks for aptitude! > > At #516254, it has been pointed out that apt-xapian-index is now getting > installed on pbuilder build systems, which is a problem. > > At first I wonderer why, since I assumed that aptitude's > apt-xapian-index functions were somehow optional functions, and that it > would only Recommend it. It has then pointed out that indeed aptitude > Depends: on apt-xapian-index, pulling it in every time. > > I am obviously happy if apt-xapian-index gets used as much as possible, > but I can think of various examples where it makes sense to have > aptitude but not apt-xapian-index. Besides pbuilder, sbuild and > friends, embedded systems also come to mind, and probably even cluster > nodes and similar kinds of server systems. > > Would aptitude actually break if apt-xapian-index were not there, or can > the dependency be safely relaxed into a Recommends?
The current experimental branch of aptitude does break if apt-xapian-index isn't available (if I remember correctly, it gets a NULL database pointer and then segfaults). It should probably be patched to fall back to another search mechanism, but I haven't gotten around to doing that yet -- it is beta software after all, and I have bigger fish to fry. What I wonder is, why is a package from experimental being installed by pbuilder? That seems wrong to me. Daniel -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
