Hallo Joachim. Em Tue, 21 Jul 2009 00:18:05 +0200 Joachim Breitner <[email protected]> escreveu: (...) > Am Montag, den 20.07.2009, 19:03 -0300 schrieb Marco Túlio Gontijo e > Silva: > > > This is actually a neat idea to avoid the sourceful uploads that I > > > mentioned in my last mail: If the upcoming haskell-devscripts that will > > > move haddock files to -dev would depend on these five packages, it is no > > > problem if some haskell-* packages still list them as B-D-I, as they > > > have haddock as a B-D. > > > > On the other hand, ghc6-doc, haddock and hscolour will be a B-D. > > This is not a problem if .haddock file gets to -dev, but it is if it > > don't. > > not sure what you mean: B-D will be satisfied for both binary-indep and > binary-dep, i.e. always. It is only B-D-I that might be missing in some > cases (e.g. buildds). There is no Build-Depends-Dep.
Yes. The problem is that ghc6-doc, haddock and hscolour will always be satisfied, even in arch-specific buildds, which wouldn't need them. This is, of course, only in the case where .haddock is still in -doc packages. My point is that the split of B-D in B-D and B-D-I is useful, and we would lost this distinction making haskell-devscripts depends on ghc6-doc, haddock and hscolour. Auf Wiedersehen. -- marcot http://marcot.iaaeee.org/ -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
