Sam James

I'm just asking why you think it'd be useful. What use case would it enable? You would scan your own packages anyway? There's really not _that_ many one would need...

> Personally, I'd suggest using multiple inherit lines instead of extra variables for stylistic reasons. In terms of the check, I forget what exactly affects the check although I do...

Thinking about this a bit more after @eli-schwartz pointed out an issue with `meson-9999` and: ``` if [[ ${PV} == 9999 ]] ; then inherit ninja-utils else ... fi ```...

Further (and I don't know if this is practical), we could try to fetch the old SRC_URI and new and compare checksums, but I doubt that's worth it. I guess...

> Got another collision in dev-python/rpyc-5.2.3 ; at some point the ebuild switched from using github releases to pypi releases, which don't contain the same data. Gentoo mirrors still have...

A quick way of making a bad ebuild which breaks the tree: 1. Create an ebuild `sys-apps/foo` with `LICENSE="all-rights-reserved"` 2. Add it to `@system` in `profiles/` 3. Try to do...

> I can perform the check, by each package checking it sub packages. But, what should be the correct behavior here? In theory I see no issue with a free...

> > Maybe it could be like `RequiredUseDefaults` where we shouldn't have it be by default. This minimises the impact to users from such incidents like the linked bugs (it...

@mgorny Okay, so that's for ::gentoo. This should be useful for GURU and overlays though, right?

> Unfortunately, since Portage maintainers have historically preferred to add ugly hacks over fixing ebuilds, Portage currently pretends that "pure RDEPEND" cycles aren't there. See also https://bugs.gentoo.org/647824.