On Sat, 5 Mar 2016, Sebastiaan Couwenberg wrote: > On 05-03-16 16:44, Yaroslav Halchenko wrote: > > On Sat, 05 Mar 2016, Sebastiaan Couwenberg wrote: > > I would better seek removal of pandas from those platforms. We do know > > that pandas has issues on those and upstream doesn't support them and we > > (or I) so far have failed to provide adequate support on those. I would > > prefer not to carry it there altogether then. > > > > What do you think? > > I'm not opposed, it better reflects that those architectures are not > supported (if you set the list of supported architectures in pandas). > > It does require all pandas reverse dependencies to also be removed from > those architectures, and recursively their reverse dependencies as well.
Hello everybody. I see that this bug was downgraded to important because we wanted pandas to propagate to testing despite the build failures on some architectures. But this is not really a good scenario in the long term: https://buildd.debian.org/status/package.php?p=pandas Would the maintainer and interested parties consider the original plan instead, namely, using a custom Architecture field which explicitly list the supported architectures? Thanks.
