Please add geoip-database package and make it RDEPEND of dev-libs/geoip. Like https://packages.gentoo.org/package/sys-apps/hwids , but for geoip. Debian and Arch already have it: https://packages.debian.org/sid/geoip-database https://www.archlinux.org/packages/extra/any/geoip-database/ The main problem with geoipupdate* is inability to ensure installing same databases to all systems. Also files in /usr/share/ should be created by package manager, not some random scripts.
(In reply to Nikoli from comment #0) > Debian and Arch already have it: > https://packages.debian.org/sid/geoip-database > https://www.archlinux.org/packages/extra/any/geoip-database/ How does that affect Gentoo? > The main problem with geoipupdate* is inability to ensure installing same > databases to all systems. Why is that? > Also files in /usr/share/ should be created by > package manager, not some random scripts. Wrong.
> How does that affect Gentoo? It indicates that having geoip-database package in Linux distro is not something unusual. > Why is that? I mean suggested by ebuild usage of geoipupdate in every system. Of course it is possible to do geoipupdate in one system and then use 'rsync -avc --delete-after' or similar tool, but doing 'emerge geoip-database' is more correct and efficient. > Wrong. Why? Dealing with files in system partition is task of package manager. P.S. Not sure if 'dev-libs' is correct category for geoip-database.
Generally, putting the databases *back* in a package is a problem for Gentoo, since the ebuild is invalidated every time upstream updates their files. And you haven't pointed out yet what the problem is with geoipupdate* in "installing [sic] same databases to all systems".
(In reply to Jeroen Roovers from comment #3) > Generally, putting the databases *back* in a package is a problem for > Gentoo, since the ebuild is invalidated every time upstream updates their > files. > > And you haven't pointed out yet what the problem is with geoipupdate* in > "installing [sic] same databases to all systems". For my use case: installing same systems offline without access to internet in government computers, only using local mirror of packages. And when I should find and fix bugs I want to be sure that systems is exactly same, because I need method to reproduce problems on testing environment.