An $> eix-test-obsolete brief gives ... Not installed but in /etc/portage/package.accept_keywords: ... at my system however I do not have that directory, I do only have /etc/portage/package.keywords Reproducible: Always
(In reply to comment #0) > Not installed but in /etc/portage/package.accept_keywords: I am aware of this problem, but eix does not keep track about where exactly a mask occurs (in which file and which form): This would need a lot of additional data structures which are not implemented - partially, because of the work and partially because of the fact that they would need a lot of memory and time even for "normal" eix runs. As a result, eix-test-obsolete cannot distinguish between package.keywords and package.accept_keywords; printing only one of them is not nice, but printing both of them is not nice either (e.g. it would no longer fit into a line). Since it is documented that both are equivalent, I believe that perhaps the current solution is acceptable. Or maybe you have a better suggestion?
Maybe: "Not installed but in /etc/portage/package.*keywords:" ??
(In reply to comment #2) > Maybe: "Not installed but in /etc/portage/package.*keywords:" ?? Thank you. Good idea. Maybe /etc/portage/package.{,accept_}keywords is even clearer. (At least, it draws attention to /etc/portage/package.accept_keywords: I assume that most portage users are not aware that this is meant to replace /etc/portage/package.keywords in the long run.) This will be fixed in >=eix-0.22.6
in tree +*eix-0.22.6 (04 Jan 2011) + + 04 Jan 2011; Jeremy Olexa <darkside@gentoo.org> +eix-0.22.6.ebuild: + Version bump from upstream, minor fixes includes zsh-completion now