On Mon, 15 Jun 2009 16:48:03 +0100 AllenJB <gentoo-li...@allenjb.me.uk> wrote:
> https://bugs.gentoo.org/show_bug.cgi?id=274197 > > The above bug brings up 2 issues: > > First, hplip says one thing, but does another with qt3 and qt4 > use-based dependencies. This is obviously a bug that needs to be > fixed. > > As a user, the second issue it brings up for me is what is the policy > applied to the rest of the tree with regards to packages that can use > one or more of several options (eg. qt3 or qt4) and have both / all > flags specified? > > Do packages that can use both/all always use both/all? Probably. > When both/all flags are specified, which one takes preferences? > Always the newer? When this issue arose up for www-client/opera because of upcoming Qt4 builds of (binary only) Opera, the qt team advised to only provide a qt3 USE flag and (thereby) make qt4 the default. So only if USE=qt3, then you get a Qt3 build of Opera[1]. Contrarily, net-print/hplip-3.9.4b appears to make Qt3 the default, while it could do entirely without IUSE=qt4, just like www-client/opera. Kind regards, jer [1] Technically speaking, you sometimes do get a Qt3 build in preview releases, but this shouldn't happen with stable releases of Opera.