Le Thu, Jul 07, 2016 at 05:51:49PM -0500, Dirk Eddelbuettel a écrit : > > What we have works. Making a change risks having to pick up pieces of > shrapnel. I don't really have time for that --- I cannot test for every > random piece of software being present. > > But the error was in another package which I do not maintain, so I kindly ask > you to either reassign this report, or open a new one. > > I do not see this as a 'fail' in r-cran.mk.
Hi Dirk and Aaron, after reading the whole discussion, and looking at the history of the xvfb package, I opened <https://bugs.debian.org/830661>, where I ask to change the Recommends relationship on xauth to a Depends. That would prevent further build failures. The failure happened on the hppa architecture, which is not a release architecture. In addition, other hppa buildds are clean, and whithin a day, r-cran-rngtools was built successfully on "physik". This also means that one can ask the hppa porters to clean the "bell" buildd (where the failure happened), since the presence of xvfb there does not seem to be necessary. If none of this works and failures pop up on release architectures, then I think that adding an ad-hoc test in r-cran.mk would be safe, although it is not the most satisfying solution. Have a nice Sunday, Charles -- Charles Plessy Debian Med packaging team, http://www.debian.org/devel/debian-med Tsurumi, Kanagawa, Japan
