On 2016-05-19 23:52, Mattia Rizzolo wrote: > The question is: shouldn't the update have tried to update both? Do we > actually try support partial upgrades?
Even if a "regular" upgrade would upgrade both packages, there is no guarantee on ordering. So I'm trying to exercise worst-case upgrade paths that are valid (i.e. not violating package relationships). I'm not even sure that this falls under "partial" upgrade. If myspell-es would get removed from testing (it has no rdepends), this bug (in huspell-es) would get more exposure. Andreas
