[R-sig-Debian] levelplot

[Dirk Eddelbuettel]

On 18 April 2006 at 16:23, deepayan at cs.wisc.edu wrote:
| 
| > I'm running debian unstable, and have been happy with the packages
| > that Dirk provides for unstable so far. update.packages() doesn't seem to
| > update lattice to 0.13.
| 
| It's in a special directory (because it's not compatible with R 2.2) that
| update.packages doesn't know about (I think).

Ahh. And while Dirk checks daily, he only checks in $CRAN/src/contrib and
doesn't see it there.  And I just checked via update.packages() in the R
2.3.0 snapshot I am running -- Version 2.3.0 beta (2006-04-14 r37779) -- and
that doesn't point to the subdir either. 

So that's a bug, but more on the human interaction side between Deepayan,
CRAN and me :)  

Seriously, with the versioned Depends: we can easily produce versions of
packages that depends on the upcoming-but-already-in-Debian releases.  But
someone needs to tell me ....

| > Perhaps email to r-sig-debian?
| 
| Sounds like the proper way to go, so I'm CC-ing.
| 
| I realise that this is a tricky situation: there are multiple copies of
| recommended packages, one set that goes with an R release and another that
| gets updated more or less independently. I'm not sure what the best
| strategy would be from the Debian packaging perspective (perhaps looking
| in both places and choosing the one with higher version number but still
| compatible with the corresponding R version).

I am sure we can work something out.  Generally speaking, I am happy with the
general policy of 'Debian has the most recent R release in unstable, and
hence almost always also in testing' and helping R Core with pre-release in
Debian unstable prior to a new release. The prereleases never make it to
testing -- by ensuring I replace them frequently enough.

This seems to server most of us well.  We can, I think, augment this with
updates of r-recommended in unstable as well.  All we need is proper Depends
a la     Depends: r-base-core (>> 2.2.1)   which would ensure that the
package cannot slip into testing as testing has R 2.2.1 -- whereas unstable
has what I currently call 2.2.1.svn$SVNRELEASE where $SVNRELEASE is the
five-digit number.  

Does this make sense?

Dirk

| 
| Deepayan
| 
| > --
| > Edzer
| >
| > deepayan at cs.wisc.edu wrote:
| >>> Deepayan, after updating to R 2.3.0 (beta), I noticed that levelplot
| >>> now
| >>> seems to draw boxes around the coloured squares. Is this intentional?
| >>> Is
| >>> there
| >>> a way to not draw them? Need I rebuild lattice?
| >>>
| >>
| >> Well, the lattice version should be 0.13-something. Depending on how you
| >> got the R 2.3.0 sources, you may need to run ./tools/rsync-recommended.
| >> Can you check with that? I don't see any problem in my installation.
| >>
| >> Deepayan
| >>
| >>
| >>> Best regards,
| >>> --
| >>> Edzer
| >>>
| >>>  > version
| >>>                _
| >>> platform       i486-pc-linux-gnu
| >>> arch           i486
| >>> os             linux-gnu
| >>> system         i486, linux-gnu
| >>> status         alpha
| >>> major          2
| >>> minor          3.0
| >>> year           2006
| >>> month          04
| >>> day            07
| >>> svn rev        37668
| >>> language       R
| >>> version.string Version 2.3.0 alpha (2006-04-07 r37668)
| >>>  >ip = installed.packages()
| >>>  > ip[row.names(ip) == "lattice",]
| >>>         Package   LibPath                   Version   Priority
| >>> Bundle
| >>> lattice "lattice" "/usr/lib/R/library"      "0.12-11" "recommended" NA
| >>> lattice "lattice" "/usr/lib/R/site-library" "0.12-11" "recommended" NA
| >>>         Contains Depends        Suggests Imports
| >>> lattice NA       "R (>= 2.2.0)" "grid"   "grid, grDevices, graphics,
| >>> stats"
| >>> lattice NA       "R (>= 2.2.0)" "grid"   "grid, grDevices, graphics,
| >>> stats"
| >>>         Built
| >>> lattice "2.2.0"
| >>> lattice "2.2.1"
| >>>  >
| >>>
| >>>
| >>>
| >>
| >>
| >>
| >
| 
| _______________________________________________
| R-SIG-Debian mailing list
| R-SIG-Debian at r-project.org
| https://stat.ethz.ch/mailman/listinfo/r-sig-debian

-- 
Hell, there are no rules here - we're trying to accomplish something. 
                                                  -- Thomas A. Edison