Hi,

I have a bit of a strange error on my Travis-CI build reports and strange in 
the sense, that they are not picked up by the Bioc build report. 
https://travis-ci.com/github/FelixErnst/RNAmodR

The package "rBiopaxParser" is reported as not available during the dependency 
install step.

I investigated, what let to the dependency via the BiocPkgTools package:

> depdf <- BiocPkgTools::buildPkgDependencyDataFrame(version = "3.12")
> library(dplyr)
> depdf %>% filter(dependency == "rBiopaxParser")
# A tibble: 4 x 3
  Package           dependency    edgetype
  <chr>             <chr>         <chr>
1 AnnotationHubData rBiopaxParser Imports
2 pwOmics           rBiopaxParser Imports
3 AnnotationHub     rBiopaxParser Suggests
4 NetPathMiner      rBiopaxParser Suggests

It turns out that the RNAmodR requires RNAmodR.Data, which itself is data 
package wrapping AnnotationHub resources using the AnnotationHub and 
AnnotationHubData packages. At this point rBiopaxParser Is on the hook via 
AnnotationHubData, because "Depends", "Imports" and "LinkingTo" fields are used 
to resolve the dependencies.

When checking the rBiopaxParser build report for devel and I found consistent 
failures 
(http://bioconductor.org/checkResults/devel/bioc-LATEST/rBiopaxParser/malbec1-checksrc.html),
 because the package nem is deprecated. Nem is only suggested by rBiopaxParser, 
but this gets an error on Bioc build report. So my guess is that rBiopaxParser 
was never build without error for 3.12 and even the source package is not 
available for 3.12. I also confirmed the behavior by running 
'BiocManager::install("RNAmodR.Data")' on the bioconductor_docker:devel image.

So my questions is, why doesn't the Bioc build report fail, because 
AnnotationHubData is importing rBiopaxParser, but that shouldn't be available, 
should it? Should I worry about this? What would be the best course of action 
here?

Thanks for any help and advice.

Best regards,
Felix

PS: It seems to my the nem packages was deprecated because of conditions of 
length > 1. This was enforced in Bioc 3.11 wasn't it? Its probably a bit of a 
shame then, but probably the dev didn't respond.

        [[alternative HTML version deleted]]

_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Reply via email to