On Wed, 2016-08-10 at 21:19 +0200, Wolfgang Bauer wrote:
And I think it's more the "fault" of the science maintainer who submitted hdf5 to Factory in the first place, despite it breaking netcdf (which is also maintained in the science repo). I suppose staging should have caught that though and prevent its acception, no idea what went "wrong" there.
hdf5 is a ring package and as such only went through staging - netcdf on the other hand is not in a ring and is thus not considered vital to hold back anything else from breaking it. ffmpeg at the day of submission of hdf5 did not depend on netcdf - so the fact that it now does not build was introduced after hdf5 had been tested against ffmpeg in the distro. Hope that clarifies how hdf5 got through staging - despite breaking netcdf. Cheers, Dominique