Hey Gianfranco, On Wed, Sep 18, 2019 at 04:04PM, Gianfranco Costamagna wrote: > Hello, after my ghc 8.8.1 upload targeted to the wrong suite (sorry for > that!)I tried to quickly rebootstrap ghc and I went successful, for amd64 all > i386 arm64 and ppc64el IIRC.
Thanks for trying to fix this so quickly. Let's see what else is pending. > Unfortunately armhf escaped from my jail, and went built successfully, and > the build I did failed, when I retried and uploaded a fixed one...somebody > had already triggered some binNMUs. Do we know if someone actually scheduled those binNMUs (because they saw a new GHC on unstable), or if there is an automated process in place? I was against the latter since it could cause exactly the kind of problems we are facing now. > Unfortunately the binNMU script doesn't detect them, so I don't really know > how to provide a list of affected packages... It seems that haskell-devscript has broken (again) with ghc-8.8.1. This causes packages to build with wrong provides, which as a side-effect, breaks the binNMU. I have tracked them down, and will request binNMUs for them, but please keep that in mind and do not upload ghc-8.8.1 to unstable before fixing haskell-devscript. Cheers, -- Ilias
