Skip to content

[R-pkg-devel] dependency on a bioconductor package

4 messages · m@i@lb@ m@ili@g off @lum@i@uv@es, William Dunlap, Martin Morgan

#
(don't forget to 'reply all', so that others may benefit / contribute to 
the conversation).

You or one of the packages your package depends on must Depend: or 
Import: BiocInstaller. If it is your package, then you should no longer 
do so. If it is one of the packages your package depends on, then the 
responsible thing to do is to contact the author of that package and 
indicate that they should update their package to no longer depend on 
BiocInstaller.

I do not know how the CRAN team will view this warning; I would guess 
that, if it is not directly under your control, then they will not be 
overly concerned about it.

Martin
On 10/5/18 1:35 PM, maialba at alumni.uv.es wrote:
#
(sorry about the replay, my mistake)

ok, then I guess that is the problem...I do not depend on BiocInstaller so it must be one of my dependencies.


Thank you very much!

Maribel
#
It looks like affy is the culprit.
recursive=TRUE)$CAMERA
function(x)"BiocInstaller"%in%x, FUN.VALUE=NA) ]
[1] "affy"
[1] "BiocGenerics"   "Biobase"        "affyio"         "BiocInstaller"
"graphics"       "grDevices"
 [7] "methods"        "preprocessCore" "stats"          "utils"
"zlibbioc"
CRAN
                         BioCsoft
                         "https://cloud.r-project.org"    "
http://www.bioconductor.org/packages/release/bioc"
                                               BioCann
                          BioCexp
"http://bioconductor.org/packages/3.5/data/annotation" "
http://bioconductor.org/packages/3.5/data/experiment"


Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Fri, Oct 5, 2018 at 10:58 AM, Morgan, Martin <mtmorgan.bioc at gmail.com>
wrote:

  
  
#
Bioconductor has a release and a devel branch, with the idea that new 
features are introduced into 'devel'. The transition to use BiocManager 
is a new feature introduced into devel, and BiocInstaller has a 
deprecation message added in devel only. It follows that win-builder is 
using bioc-devel (this is reasonable; our next release is at the end of 
this month), and that we'd look for use of BiocInstaller in devel 
packages. affy is itself a 'core team' package, and we have removed use 
of BiocInstaller from that.

But let's see... the 'release' branch has

 > dput(repos)
c(BioCsoft = "https://bioconductor.org/packages/3.7/bioc", BioCann = 
"https://bioconductor.org/packages/3.7/data/annotation",
BioCexp = "https://bioconductor.org/packages/3.7/data/experiment",
BioCworkflows = "https://bioconductor.org/packages/3.7/workflows",
CRAN = "https://cran.rstudio.com")
 > db = available.packages(repos=repos)
 > revdeps = tools::package_dependencies("LipidMS", db, recursive=TRUE)
 > "BiocInstaller" %in% unlist(revdeps)
[1] TRUE

and indeed BiocInstaller is problematic (but BiocInstaller in the 
release branch doesn't generate the warning...) (also, I'm not sure 
about the '3.5' components of your annotation / experiment repos, these 
should be the same as the main repository).

On the other hand in our devel branch

 > repos = sub("3.7", "3.8", repos)
 > db = available.packages(repos=repos)
 > revdeps = tools::package_dependencies("LipidMS", db, recursive=TRUE)
 > "BiocInstaller" %in% unlist(revdeps)
[1] FALSE

So at least for packages that are current in CRAN and in the devel 
branch of Bioconductor there is no use of BiocInstaller.

I'm not exactly sure where this leaves us...

Martin
On 10/5/18 2:26 PM, William Dunlap wrote: