On Thu, Dec 24, 2015 at 01:50:45AM +0100, Alberto Salvia Novella wrote: > Luis Felipe Tabera Alonso: > David Kalnischkies: > > It is not a good idea to perform autoremoves unattended for situations > > in which you have installed A (gui) depends B (console) depends > > C (data), but later decide that you don't like A. > > What if: > - "remove" removes a package and all the unused depends, recommends and > suggests. The problem is the word "unused" here. You have obviously a different definition (which isn't implementable) than apt has, otherwise "apt autoremove $pkg" would already do what you describe – your complain was after all that apt declares some packages you consider unused as used… > - but before removing asks the user to exclude packages from the removal, > and from that moment marks those as installed by the user. apt shows a list of packages which could be autoremoved as well as a list of packages which will be removed after pressing yes. Preventing a package to be considered for autoremove is as simple as installing it… Asking the user to confirm each and every removal would end up being pretty annoying very fast and if its something you don't want to train users to do, its is brainlessly pressing yes a hundred times as the important question nr 101 will surely be answered just as quick with yes as well – potentially followed a split-second later by a "NOOOO!" but its too late then. Also note that we talked about "unattended" here. Who should answer all those questions if the point of the exercise was to make autoremove do its thing without asking any questions… (And if you can't imagine that many removes, dist-upgrades can easy include that much and thanks to the gcc5-transition this time around it will be an especially big list) > - "autoremove" only removes depends. You will have to give slightly more details on this one as by itself, that sentence makes no sense for me. > David Kalnischkies: > > After 6 years I think I have enough 'battle' experience to say that > > even I have still ideas which look good on paper only... and its good > > that others put a stop to such ideas before those ideas have a chance > > to hurt me (and I can assure you, I implemented ideas which never > > should have been and now taunt me by their mere existence). > > Imagine that we have the perfect way to innovate. That we have decided > slowly along with other people, the change is small, we have put it on test > as prototype, and the outcome seems to be very positive. > > Would that make us get rid of complains? If you add free icecream and a cherry topping… maybe… that wasn't at all what I was trying to say through, as implementing good ideas is (comparatively) easy in free software. The hard part is figuring out the good ideas as people tend to feel very attached to their own ideas and can take it the wrong way if others tell them the idea isn't as good as they believe - even if they provide examples and hard facts. Imagine that you drive on the highway and a wrong-way driver crosses your path. Then a second. And a third. And yet another… How many wrong-way drivers does it take before you realize that maybe it is you who is driving in the wrong direction? Now read all the replies to your idea again and compare numbers. Enough from me for this year & thread now through, so: Happy "package management" days and best regards, David Kalnichkies
Attachment:
signature.asc
Description: PGP signature