On Thu, Jul 19, 2018 at 12:39:58PM +0800, Drew Parsons wrote: > On Tue, 2018-07-17 at 13:20 +0200, Graham Inggs wrote: > > Hi Drew > > > > > On Tue, 17 Jul 2018, Drew Parsons wrote: > > > > The configuration of ci.debian.org is not consistent. A failing > > > > test > > > > of deal.ii is marked on > > > > https://ci.debian.net/packages/d/deal.ii/testing/amd64/ > > > > as triggered by petsc/3.9.3+dfsg1-2, but the test log shows that > ... > > From the times in the logs, it appears that the test was run before > > deal.ii's binNMU happened . > > So the delay is only temporary. > > > > As soon as the test is successful, the migration delay will > > disappear. > > > > I've requested a retry for that test, so it should happen soon. > > > Thanks Graham. When we looked closer, the underlying problem was that > petsc (both 3.8 and 3.9) were built against openmpi 3.1.1, but the CI > test used openmpi 3.1.0. PETSc objects to the backwards version. > > I've tightened up the petsc dependency in 3.9.3+dfsg1-3 which should > improve things. > > Does this still count as an inconsistency in CI configuration? The test > is triggered by a package in unstable, but CI uses testing to actually > run the test. > > Perhaps that's actually not a bad inconsistency. It gives an extra > layer of testing. If the version difference between testing and > unstable really does matter (as it does for petsc/openmpi in this case) > then that problem should be declared in the versioned dependencies. that's the exact point of the testing migration: to test whether a given package from unstable could migrate to testing, assuming the current contents of testing. If your package needs something that is not in testing yet (e.g. something that is still in unstabel only), you *must* have that proper (versioned) dependency in the package.
Description: PGP signature