[Aptitude-devel] Bug#406976: aptitude: source-strictness is not strict enough

[Manuel A. Fernandez Montecelo]

Control: tags -1 + moreinfo


Hi Steinar,

2007-01-15 11:49 Steinar H. Gunderson:
>Package: aptitude
>Version: 0.4.4-1
>Severity: normal
>
>Hi,
>
>I'm trying to force in newer evince (since I need it for a presentation)
>_and_ still keep GNOME:
>
>  fugl:~> LANG=C sudo aptitude install gnome evince/experimental
>  Reading package lists... Done
>  [...]
>  The following packages are BROKEN:
>    gimp libgtk2.0-0
>  The following packages are unused and will be REMOVED:
>    bsh gcj-4.1-base gij gij-4.1 gimp-data gnuplot gnuplot-nox gnuplot-x11 lapack3 libg2c0
>    libgcj-bc libgcj-common libgcj7-0 libgd2-noxpm libhsqldb-java libicu36 libjaxp1.2-java
>    libjaxp1.3-java libjline-java libmdbtools libneon26 libpoppler0c2-glib libservlet2.3-java
>    libstlport4.6c2 libufsparse libwpd8c2a libxalan2-java libxerces2-java libxt-java openoffice.org
>    openoffice.org-base openoffice.org-calc openoffice.org-common openoffice.org-core
>    openoffice.org-draw openoffice.org-impress openoffice.org-java-common openoffice.org-math
>    openoffice.org-writer python-uno refblas3 ttf-opensymbol
>  The following NEW packages will be automatically installed:
>    gimp-print gimp-svg gnome-office libwmf0.2-7
>  The following NEW packages will be installed:
>    gimp-print gimp-svg gnome gnome-office libwmf0.2-7
>  0 packages upgraded, 6 newly installed, 42 to remove and 0 not upgraded.
>  Need to get 3422kB/3448kB of archives. After unpacking 295MB will be freed.
>  The following packages have unmet dependencies:
>    gimp: Depends: gimp-data (= 2.2.13-1) but it is not installable
>          Conflicts: libgimp2.0 (>= 2.3.0) but 2.3.13-1 is installed.
>    libgtk2.0-0: Conflicts: libwmf0.2-7 (<= 0.2.8.4-2) but 0.2.8.4-2 is to be installed.
>  Resolving dependencies...
>  open: 59; closed: 34; defer: 0; conflict: 11                                                         .The following actions will resolve these dependencies:
>
>  Keep the following packages at their current version:
>  libpoppler0c2-glib [0.4.5-5 (testing, unstable, now)]
>
>  Downgrade the following packages:
>  bug-buddy [2.16.0-1 (experimental, now) -> 2.14.0-4 (testing, unstable)]
>  evince [0.6.1-1 (experimental, now) -> 0.4.0-3 (testing)]
>  gimp-data [2.3.13-1 (experimental, now) -> 2.2.13-1 (testing, unstable)]
>  gtk2-engines [1:2.8.2-2 (experimental, now) -> 1:2.8.2-1 (testing, unstable)]
>  gtk2-engines-pixbuf [2.10.7-1 (experimental, now) -> 2.8.20-4 (unstable)]
>  libgimp2.0 [2.3.13-1 (experimental, now) -> 2.2.6-1sarge1 (stable)]
>  libgnomeui-0 [2.16.1-1 (experimental, now) -> 2.14.1-2 (testing, unstable)]
>  libgnomeui-common [2.16.1-1 (experimental, now) -> 2.14.1-2 (testing, unstable)]
>  libgtk2.0-0 [2.10.7-1 (experimental, now) -> 2.8.20-4 (unstable)]
>  libgtk2.0-common [2.10.7-1 (experimental, now) -> 2.8.20-4 (unstable)]
>  librsvg2-2 [2.16.0-3 (experimental, now) -> 2.14.4-2 (testing, unstable)]
>  librsvg2-common [2.16.0-3 (experimental, now) -> 2.14.4-2 (testing, unstable)]
>
>  Score is -298
>
>Given that my Request-Strictness is set to 10000 (which is now the default,
>I believe), this score is too high; my guess is that it thinks having evince
>0.4.0-3 satisfies my request for "evince/experimental". In any case, it
>appears to be quite impossible to ask it to drop all "solutions" mentioning
>evince from testing/unstable, short of removing them from the local Packages
>files.

This has happened many years ago, before the resolver was heavily
reworked in the last few years.

I am not sure if the situation is better now due to aptitude having into
account that "experimental" is the default release for this operation,
or if it's worse because it considers that versions for experimental are
not default for the other packages that you didn't request.

Have you had some recent-ish experience with this, and remember if it
behaves in the same way?

You probably know about this, but in general, if the upgrade to
experimental is indeed possible, you can guide the solution easily by
rejecting all downgrades to testing/unstable and asking for the next
solutions.


Cheers.
-- 
Manuel A. Fernandez Montecelo <manuel.montezelo at gmail.com>