On Fri, Feb 22, 2008 at 09:55:31PM +0100, Raphael Hertzog wrote: > On Fri, 22 Feb 2008, Mike Bird wrote: > > > I won't revert anything unless you come up with some proof that this > > > causes severe issues that will disturb the lenny release process.
> > Raphael, > > What please is the benefit of unnecessarily reordering dependencies > > and leaving everyone on tenterhooks as to whether it will change > > installation outcomes? (If this has already been explained I apologize > > for overlooking it.) > 1/ Sorting makes the output deterministic, and this means that (potential) > problems are more likely to be reproduced by everybody (whatever the > architecture, etc.) Dependencies that have been *manually* ordered by the maintainer are *also* deterministic, and the order is often set to something non-alphabetical for a reason. Otavio refers to the greedy algorithm used by apt as a "bug"; I don't agree. The greedy behavior is straightforward to understand; a full dependency resolver that tries to minimize the number of packages that have to be pulled in to satisfy the dependencies would not only be harder to understand/debug, it would also be NP-complete. > 2/ debdiff uses wdiff to show changes on field values and wdiff gives > spurious differences if the sole difference between both values is > a different order. Thus debdiff output is more useful with ordered Depends > fields. And why is that dpkg's business to try to fix? If the maintainer has changed the order of dependencies in the source package, they have some reason for doing so. Neither of these are very good reasons for wresting control of the Depends: field from the maintainer. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ [EMAIL PROTECTED] [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
