On Tue, Jul 1, 2014 at 8:41 AM, Patrick Lauer <patr...@gentoo.org> wrote: > On 06/30/14 22:15, Jeroen Roovers wrote: >> On Mon, 30 Jun 2014 09:25:27 -0400 >> Rich Freeman <ri...@gentoo.org> wrote: >> >>> Agree 100%. I'm taking about masking things that HAVEN'T BEEN TESTED >>> AT ALL. The maintainer knows that they compile, and that is it. >> >> Developers who "HAVEN'T [...] TESTED AT ALL" and still commit their >> changes to the tree should immediately hand in their toys and leave >> the project. >> > > I usually avoid overlays (best way to make things hard to find), so when > there's stuff that upstream says is experimental (e.g. perl6/rakudo with > the MoarVM backend) I have no issue with adding it as un-keyworded > ebuilds to the tree. That way it's easy to test, and once there's a bit > more confidence that it works well enough it's trivial to keyword. >
If the goal is to reduce clutter in the profiles then this could be a good alternative. Nothing would prevent a maintainer from sticking a comment in the ebuild as well. Hate to derail this, but another option would be to migrate package.mask to a directory (eventually) and manage masks by project or by date. Projects could create standing files when needed, and misc masks would go into a file named by year/quarter. Then anybody looking in the directory can spot projects that are dead, or files which are old. Either would be easier to clean up. Obviously restructuring the profiles entirely as has been suggested will help, though not for masks like these. Rich
