On 15/08/14 01:14 PM, Daniel Baumann wrote: > (without having tried^diffed the two outputs) the idea would be that you > would only see a line: > > W: automatically excluding packages: libdb5.1 > > (or something like that) in the current case, and if that package is > gone from testing eventually, you wouldn't see anything anymore at all.
It's deeper than that. If our goal is to make the inclusion of standard "like what happens when you install standard with d-i", then we need to emulate what tasksel does, which is to omit all this material (and that's considerably more than just libdb5.1, it's any package in a lib section with Priority: standard). Now, I don't know if those libs would be included *anyway* because they're depended on by something else that is standard, as I haven't tried it yet, but the difference between the two grep-dctrl's is certainly a lot more than just libdb5.1. e.g. let's take a hypothetical libfoo that is Priority: standard but is *not* in the middle of a transition/in a broken state in the archive, and which no other Priority: standard package depends on. Without my patch, it will be uselessly included in debian-live images, but d-i images will not install it. > imho lb should get rid of all embedded magics in lb, and when we keep > one, it should 'show' itself when it's active. OK, but for this to work, you would have to run the resolver on the whole set of packages to be included in the image to get the real difference between the two methods. Otherwise you get something that looks like this: $ grep-aptavail -n -sPackage -FPriority standard -a -! -e -F Section '^(lib|/)' | sort > magic $ grep-aptavail -n -sPackage -FPriority standard | sort > nomagic $ diff -u nomagic magic | diffstat magic | 103 ------------------------------------------------------------------ 1 file changed, 103 deletions(-) That's a significant reduction in packages, most of which won't be "real" omissions because they're included by some other Priority: standard thing. Ben -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org