On Fri, 01 Jun 2012 23:13:58 Thomas Kahle wrote:
> On 07:26 Sat 02 Jun 2012, Francois Bissey wrote:
> > On Fri, 01 Jun 2012 11:52:24 Thomas Kahle wrote:
> > > Hi,
> > >
> > > now that I have sageongentoo installed I can't do world updates with
> > > the D option anymore. Looks like a portage quirk to me.
> > >
> > > There seem to be at least two problems:
> > >
> > > I have scimathematics/singular in my world file and the D option
> > > forces portage to think that I insist on the latest version. Therefore
> > > it collides with sageongentoo insisting on 3.1.3:
> > >
> > > scimathematics/singular:0
> > >
> > > (scimathematics/singular3.1.3.3::gentoo, installed) pulled in by
> > >
> > > ~scimathematics/singular3.1.3.3 required by
> > >
> > > (scimathematics/sage5.0r1::sageongentoo, installed)
> > >
> > > (scimathematics/singular3.1.4r1::gentoo, ebuild scheduled for
> > > merge)
> > >
> > > pulled in by (no parents that aren't satisfied by other packages in this
> > > slot)
> > >
> > >
> > > There also is an eselect issue. I don't know why, but eselectblas and
> > > eselectlapack are pulled in, although blocked by the new gsl:
> > >
> > > [blocks B ] appadmin/eselectblas ("appadmin/eselectblas" is
> > > blocking scilibs/gsl1.15r2, scilibs/cblasreference20110218)
> > > [blocks B
> > >
> > > ] appadmin/eselectlapack ("appadmin/eselectlapack" is blocking
> > >
> > > scilibs/gsl1.15r2, scilibs/cblasreference20110218)
> > >
> > > However, without the D option I just get a warning:
> > >
> > > WARNING: One or more updates have been skipped due to a dependency
> > > conflict:
> > >
> > > scimathematics/singular:0
> > >
> > > (scimathematics/singular3.1.4r1::gentoo, ebuild scheduled for
> > > merge)
> > >
> > > conflicts with ~scimathematics/singular3.1.3.3 required by
> > > (scimathematics/sage5.0r1::sageongentoo, installed)
> > >
> > >
> > > Any ideas? Shall I carry this to the portage mailing list?
> > > Cheers,
> >
> > We need to know what pulls these eselect modules. What is the portage
> > output when you add the "t" option. It should show what wants what.
>
> Goes back to devlang/R and scilibs/superlu. My setup has superlu
> stable, but R keyworded ~amd64 the entire slot.
>
> [ebuild U ~] scimathematics/singular3.1.4r1 [3.1.3.3] USE="boost doc
> emacs python%* readline examples test" 0 kB [nomerge ]
> devlang/R2.15.0 USE="X bashcompletion cairo doc java jpeg lapack nls
> openmp perl png readline tiff tk icu minimal profile staticlibs"
> [nomerge ] virtual/lapack3.1
> [nomerge ] scilibs/lapackreference3.1.1r1 USE="doc"
> [ebuild N ] appadmin/eselectlapack0.1 0 kB
> [nomerge ] scimathematics/sage5.0r1::sageongentoo USE="latex
> testsuite" [nomerge ] scilibs/scipy0.9.0r1 USE="doc umfpack"
> [nomerge ] scilibs/superlu4.1r1 USE="doc examples staticlibs
> test" [nomerge ] virtual/blas1.0
> [nomerge ] scilibs/blasreference20070226 USE="doc"
> [ebuild N ] appadmin/eselectblas0.1 0 kB
> [nomerge ] scimathematics/sage5.0r1::sageongentoo USE="latex
> testsuite" [nomerge ] devpython/sympy0.7.1 USE="gtk latex opengl
> pdf png doc examples imaging ipython mathml pyglet test texmacs"
> [ebuild U ] devpython/pexpect2.4::gentoo [2.0r1::sageongentoo]
> USE="doc% examples%" 0 kB [blocks B ] appadmin/eselectlapack
> ("appadmin/eselectlapack" is blocking scilibs/gsl1.15r2,
> scilibs/cblasreference20110218) [blocks B ] appadmin/eselectblas
> ("appadmin/eselectblas" is blocking scilibs/gsl1.15r2,
> scilibs/cblasreference20110218)
>
You have the virtual/{blas,lapack} from the main tree not from the science
overlay! Now I am not sure why eselectblas blocks gsl which gives you cblas.
Anyway keyword virtual/{blas,cblas,lapack} and upgrade them. Then let's see
if you still have troubles.
Francois
