On Sun, 2017-11-26 at 16:22 +0100, Andreas Metzler wrote: > mariadb-10.1 1:10.1.29-6 seems to be stuck in sid. It does not > propagate to testing although > https://qa.debian.org/excuses.php?package=mariadb-10.1 lists it as > valid candidate. > > Could you please check the cause?
There's a bit of backstory, but effectively: the mariadb-test binary package in testing is built from the mariadb-10.1 source package, and has strictly versioned dependencies on other binaries built from that source package. In unstable, the mariadb-10.2 source package builds the binary package instead, but FTBFS on several architectures so is not a candidate. The net result is that when britney tries to migrate mariadb -10.1 1:10.1.29-6, the mariadb-test binary package in testing becomes uninstallable, and the migration attempt is aborted. Regards, Adam
