On Tue, Mar 31, 2015 at 9:32 AM, The Wanderer <wande...@fastmail.fm> wrote:
> Repeatedly over the years - I'd almost say consistently - I've seen > aptitude report that a requested package change (install, remove, or > some combination) would result in an invalid or conflicting dependency > situation, and suggest a solution which involves _not making the change > which was requested_. > > If the requested configuration is, in fact, contradictory, then this is > of course reasonable. However, in most if not all such cases, requesting > the same change of apt-get produces a workable dependency solution > immediately. > > Sometimes (when I've bothered to stick with it long enough), telling > aptitude "no, try again" a few dozen times (and rejecting "solutions" > which would downgrade or remove dozens, if not hundreds, of packages > along the way) will eventually get it to suggest a solution which will > make that change without extraneous side effects - which may or may not > be the same as the one provided by apt-get. > > But as long as aptitude continues to take this brain-dead approach to > dependency resolution, necessitating digging through obviously-bad > suggestions before finding something as reasonable as what apt-get > provides easily, it is IMO not viable for actual use - except perhaps by > people who already know completely what they are doing and how to > override aptitude's suggestions. I've grepped debian-devel, but cannot find an email that was sent to the list some months ago about tweaks to /etc/apt/apt.conf (IIRC) to make aptitude behave more sanely. Thus, I believe there are a couple of knobs to turn to make aptitude behave more expectedly. Cheers, -m -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/caolfk3vm0kzvee-bypj6yjaqfwp-ma5ybn1k7y1w-cbev1c...@mail.gmail.com