On Mon, Jan 14, 2002 at 09:52:21PM +1100, Peter Moulder wrote: > On Mon, Jan 14, 2002 at 11:06:22AM +0100, Wichert Akkerman wrote: > > Previously Peter Moulder wrote: > > > Adam Heath voices what is I believe the natural reading of current > > > policy, namely that Depends implies postinst ordering, and consequently > > > that dependency cycles aren't allowed. > > Well, they are allowed but as soon as you create a cycle it will have > > to be broken so you can't assume an exact ordering anymore. > They are allowed by dpkg, whereas current policy says that they are not > allowed, hence giving false confidence that no cycles will occur and thus > that one's Depends line will always be honoured wrt ordering of > configuration, which may turn out not to be guaranteed if a cycle exists, > even if no such cycle exists at the time one releases one's package.
This used to be documented in (I think) the packaging manual: if a cycle amongst Depends: exists, the cycle will be broken by choosing the package without a postinst (if there is one) or arbitrarily, iirc. There's still some determinism to be had, but not much. If there's no cycle when you release the package, the only way there'll ever be one is if some package you depend on decides to depend on you too, in which case they'll have needed to ensure the cyclic dependencies work out right. These things don't just happen, and if you're at all sensible, you should be talking to the other maintainers involved first if you think you need to make a dependency cycle. Cheers, aj -- Anthony Towns <[EMAIL PROTECTED]> <http://azure.humbug.org.au/~aj/> I don't speak for anyone save myself. GPG signed mail preferred. The daffodils are coming. Are you? linux.conf.au, February 2002, Brisbane, Australia --- http://linux.conf.au/
