Package: release.debian.org Severity: grave Hello everybody,
the update of R to version 3.3.3 unexpectedly breaks some R packages, which need to be reinstalled from source. In Debian terms, given how we package R packages, this can be solved by a binNMU. A quick look at ci.debian.net shows that a large number of Bioconductor packages (r-bioc-*) started to fail their tests at the end of February following the upload of r-base version 3.3.2.20170227-1, which was a release candidate of 3.3.3. Some of the failures are direct, and some may be a ripple effect. On Bioconductor's support forum (https://support.bioconductor.org/p/93695/#93719), it was suggested to reinstall the packages "BiocGenerics", "S4Vectors", "IRanges", and "GenomicRanges". I tried this on my computer (using sbuild --make-binNMU) and their regression tests restarted to pass. These packages are very central in Bioconductor's dependency graph. If time remains, how about starting with these four and see if the failures in the the other packages dissapear ? nmu r-bioc-biocgenerics_0.20.0-1 . all . -m "Rebuild for R 3.3.3." nmu r-bioc-s4vectors_0.12.1-2 . ANY . -m "Rebuild for R 3.3.3." nmu r-bioc-iranges_2.8.1-1-m . ANY . "Rebuild for R 3.3.3." nmu r-bioc-genomicranges_1.26.2-1 . ANY -m "Rebuild for R 3.3.3." Otherwise I can prepare a more extensive list of packages to rebuild. Cheers, -- Charles Plessy Debian Med packaging team, http://www.debian.org/devel/debian-med Tsurumi, Kanagawa, Japan
