On Fri, Oct 29, 2010 at 06:45:15AM +0200, Martin Baehr wrote:
> tried that, but still found some issues:
> you received my system model email so i won't reproduce it here,
> the points that i am concerned about are the following steps to match
> the model.
>
> Erase foresight-contexts(:config)=0.1-4-1
> Erase sugar-artwork(:data :lib)=0.89.4-1-1
> Erase wqy-zenhei(:config :data)=0.8.38-2-1
> Erase factory-base-packagecreator(:recipe)=1.0-4-1
>
> those i knowingly installed manually, but it looks like genmodel did not
> detect that. (it detected other sugar packages, which is puzzling)
It has a list of components that are more likely to have been the
result of manual installs. config, data, and lib were expected to
have been mostly the result of dependency closure, and we were
trying to avoid introducing the same sort of confusion that shows
up with "conary updateall" due to PK's use of --keep-required
creating "zombie" toplevel packages.
It would be possible to augment that list on your system. Search for
"componentPriorities" under "initialForesightModel" and add one
line for each component type that you don't want to be ignored.
See what that does to improve the model for you, and let me know.
Thanks!
> the resulting model and the sync steps attached. the puzzling part is that
> this model
> causes conary to attempt to install a second version of firefox,
> conflicting with an already existing version. screen too.
...
> search 'group-world=foresight.rpath....@fl:2-qa/2.3.4.99_2010.08.31-0.1-2[is:
> x86]'
> search
> 'group-world=foresight.rpath....@fl:2-qa/2.3.4.99_2010.08.31-0.1-2[~!bootstrap,~buildtests,~!cross,desktop,~!dom0,~!domU,~!gcc.core,~grub.static,ipv6,~kernel.debugdata,krb,ldap,pam,~!pie,readline,ssl,tcl,tk,~!vmware,~!xen
> is: x86(i486,i586,i686,sse,sse2) x86_64]'
> search group-kde-dist=foresight.rpath....@fl:2-qa/2.4.8.9+2010.09.09-0.1-1
Your group-world and group-kde-dist lines are out of sync versions.
I'm quite certain that this is the source of the confusion. You end
up with group-gnome-dist and group-kde-dist with different versions
of shared components.
Unfortunately, we don't yet have the ability to print out what happens
per line (as you can see with dumpAll when building a GroupSetRecipe).
We do have the ability to give you a graphical view of the
relationships between the lines in your model and how Conary
processes them:
conary sync --info --model-graph=/tmp/model.dot
dot -Tsvg /tmp/model.dot > /tmp/model.svg
eog /tmp/model.svg
That won't answer the question you had here, but might still be
interesting. :)
_______________________________________________
Foresight-devel mailing list
Foresight-devel@lists.rpath.org
http://lists.rpath.org/mailman/listinfo/foresight-devel