[R-pkg-devel] Package fails to build on CRAN Windows server: rtracklayer version too new

[Tim Stuart]

Hi Henrik,

Thanks for your explanation, that does indeed make sense. I will try
to track down the bioc package dependency that's setting the explicit
rtracklayer version dependency and go from there...

The fact that Bioconductor allows packages to be released without
successfully building on Windows does create some interesting
downstream issues for CRAN packages with bioc dependencies, and raises
a larger question for me about whether I should really host this
package on CRAN or if I should try to move it to Bioconductor.

Anyway, thanks again for your help (also I'm a big fan of your future package)

Cheers,
Tim


On Mon, Apr 19, 2021 at 9:47 PM Henrik Bengtsson
<henrik.bengtsson using gmail.com> wrote:
>
> Not that it solves your problem per se, but here's what I think CRAN
> only has rtracklayer 1.48.0 installed on WIndows, i.e. why you get
> error "namespace 'rtracklayer' 1.48.0 is being loaded, but >= 1.51.5
> is required.":
>
> * The current Bioconductor "release" is version 3.12
> * As you say, Bioconductor 3.12 provides rtracklayer 1.50.0
> (https://www.bioconductor.org/packages/3.12/bioc/html/rtracklayer.html)
> * However, if you look at that page, for Windows it only provides
> rtracklayer 1.49.5 (this is because rtracklayer never successfully
> checked on Bioconductor 3.12)
> * Since the minor version component of rtracklayer 1.49.5 is _odd_, it
> means it's from a Bioconductor 'devel' branch, i.e. when Bioconductor
> 3.12 was in devel.
> * Thus, the most recent Bioconductor "release" where rtracklayer was
> successfully distributed on Windows is Bioconductor 3.11.
> * Bioconductor 3.11 provides rtracklayer 1.48.0
> (https://www.bioconductor.org/packages/3.11/bioc/html/rtracklayer.html)
>
> Clear as mud?
>
> I suspect you're using Suggest: rtracklayer
> (https://github.com/timoast/signac/blob/develop/DESCRIPTION) in your
> new version.  If so, that's good.  I also don't think you're doing
> anything wrong here.  Instead, I think it's one of your other
> Bioconductor dependencies under 'Imports:', or possibly under
> 'Suggests:', that depend on 'rtracklayer'.  The fact that error says
> "namespace 'rtracklayer' 1.48.0 is being loaded, but >= 1.51.5 is
> required.", means that one of those Bioconductor dependencies, or
> their dependencies, has an explicit 'rtracklayer (>= 1.51.5)' under
> either Depends: or Imports:.  My recommendation is that you start
> identifying that Bioconductor package, and in what way your 'Signac'
> depends on that package.  That will help you decide on what to do
> next.
>
> /Henrik
>
> PS. This can happen because Bioconductor allows a package to propagate
> to a new "release" branch even if it fails on one of the operating
> systems.
>
> On Fri, Apr 16, 2021 at 1:59 PM Tim Stuart <tstuart using nygenome.org> wrote:
> >
> > Hi all,
> >
> > I am trying to submit an update to my package
> > (https://cran.r-project.org/package=Signac) to CRAN, but it is
> > currently failing to install on the Windows server. I see the
> > following installation error: "namespace 'rtracklayer' 1.48.0 is being
> > loaded, but >= 1.51.5 is required."
> >
> > The current release version for rtracklayer is 1.50.0
> > (https://www.bioconductor.org/packages/release/bioc/html/rtracklayer.html).
> > My package does not set a version requirement for rtracklayer, so this
> > installation error must be caused by having a pre-release version of
> > one of my bioconductor dependencies installed on the windows server.
> >
> > Does anyone know how I might be able to solve this issue? I have tried
> > contacting the CRAN team a few times by reply-all to the CRAN emails,
> > but have not had any response from them. This seems like an issue that
> > could only be solved by fixing the installation of packages on the
> > CRAN Windows server.
> >
> > Incoming test results:
> > https://win-builder.r-project.org/incoming_pretest/Signac_1.1.2_20210415_225235/
> >
> > Thanks,
> > Tim
> >
> > --
> >
> > This message is for the recipient’s use only, and may contain
> > confidential, privileged or protected information. Any unauthorized use or
> > dissemination of this communication is prohibited. If you received this
> > message in error, please immediately notify the sender and destroy all
> > copies of this message. The recipient should check this email and any
> > attachments for the presence of viruses, as we accept no liability for any
> > damage caused by any virus transmitted by this email.
> >
> > ______________________________________________
> > R-package-devel using r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-package-devel

-- 

This message is for the recipient’s use only, and may contain 
confidential, privileged or protected information. Any unauthorized use or 
dissemination of this communication is prohibited. If you received this 
message in error, please immediately notify the sender and destroy all 
copies of this message. The recipient should check this email and any 
attachments for the presence of viruses, as we accept no liability for any 
damage caused by any virus transmitted by this email.