Wouter Verhelst wrote: > That's still a bit of time wasted, but it's not really bad. The really > problematic version is when a package is downloaded, build-deps are > installed, and /then/ sbuild figures out that some version isn't recent > enough. According to the intersection of the debcheck build-depends test and m86k's Needs-Build [1], this will likely affect gnomesword bibletime knoda asterisk k3d loop-aes-modules or are these already caught by the buildds?
Inaccuracies are caused (at least) by debcheck only using packages unstable (not incoming etc.) and not treating conflicts. The latter also leads to the list of packages that could be retried (i.e. failed for insatisfiable dependencies but this isn't the case anymore) [2] to be not useful yet, see e.g. gnucash. Kind regards T. 1. http://people.debian.org/~tviehmann/fail-predictions.txt (NOT automatically updated though, scripts in my ~ on p.d.o) 2. http://people.debian.org/~tviehmann/couldretry.txt -- Thomas Viehmann, http://thomas.viehmann.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
