Hi, On Sun, May 28, 2017 at 07:54:33PM +0200, Ivo De Decker wrote: > On Sun, May 28, 2017 at 02:39:00PM +0200, Jan Niehusmann wrote: > > May I suggest only removing runit-init from the runit source package, if > > the bug can't be fixed in time for the stretch release? > > Sure. That just needs someone to upload such an NMU very soon. I suggested > this already at the end of message #49: > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=861536#49
But who decides that this is the right way to go? Just removing a package is not usually something which should be done with an NMU. Especially as this bug only affects systems switching to runit. There may be people happily running runit as their init system. Just leaving them with an unmaintained init system doesn't seem to be a good idea, either. In fact, that may be less user friendly than this bug. How could we handle future security issues in that package, for example? As runit-init is part of current stable, we need to provide some upgrade path if we want to remove it. BTW, according to the initial reporter, "The same procedure was required to return to systemd as PID1." So systemd seems to be broken in the same way. Therefore, simply replacing runit-init with a transitional package depending on systemd wouldn't work either. > If this bug can be fixed by removing runit-init, the removal of the other > packages isn't necessary, but please note that this would need to happen very > soon. Removing runit-init from the init package may be worse than keeping it, see above. Of course, removing runit-init, runit and all reverse-dependencies would be even worse, so it may be a workaround in case the release team decides that keeping runit-init in its current form is absolutely not acceptable. Jan
