On Wed, Sep 14, 2022 at 03:23:47PM +0200, Graham Inggs wrote: > On Wed, 14 Sept 2022 at 13:31, Andreas Tille <andreas@an3as.eu> wrote: > > I'm not sure whether this is an hdf5 issue or something else. > > Any idea how to fix this before asking upstream? > > Well it passes in unstable right now [1], and there's a new version of > hdf5 in unstable [2] that is blocked from migrating. > Does r-bioc-gsva miss somehow a dependency against the version of hdf5 > it was built against? If you look at the latest upstream commit[3]. Which makes changes in the code and corresponding test that is currently failing. It fixes something wrt Rmatrix 1.4-2 which I suppose is not backwards compatible with rmatrix versions. There is r-cran-matrix 1.5-1 uploaded today itself to unstable. However, both salsa CI and debci are building it against rmatrix 1.4-1-1 which is expected by debci because that's the rmatrix version in testing. Salsa CI would build it against version in unstable, but I guess src:rmatrix did not dinstall at the time that CI job ran. I added in a versioned-dep against rmatrix, and uploaded (after messsing it up in first upload :-S) and this change IMHO should fix the situation. However, I think that upstream should properly fix this in DESCRIPTION file. > [1] https://ci.debian.net/packages/r/r-bioc-gsva/unstable/amd64/ > [2] https://tracker.debian.org/pkg/hdf5 [3] https://github.com/rcastelo/GSVA/commit/b40777c510f3aff7c4bde662e466ebff1c8686c5 -- Best, Nilesh
Attachment:
signature.asc
Description: PGP signature