Basically this is more of an aesthetic problem than anything. Nobody is going to die if this doesn't get implemented, although it would be nice to see. This can also provide additional functionality if implemented. The thing is that if I wanted to declare a use flag for a package of a specific version I would just go to /etc/portage/package.use and add the following: =dev-vcs/subversion-1.7.5 java So the package category/name-version are prepended with an equal sign to signal that this statement/atom will be for a target version only. Now if I wanted to let portage know that I already have provided subversion v1.7.5, I would go to /etc/portage/provided/package.provided and add the following: dev-vcs/subversion-1.7.5 Sure this works, but it doesn't have the = sign in the beginning to signal that this is just for a specific version. This in my opinion is an inconsistency. I believe it should be written with an = sign in the front just as all "this statement will be for a specific version" are: =dev-vcs/subversion-1.7.5 The benefit of doing it this way is also that if I wanted to tell portage that I want to completely manage the package, I would treat it just like any atom, and remove the version and any signs from the front: dev-vcs/subversion So now portage will not bother me about any past/present/future changes to subversion because it will think that I have provided it. This applies even more to maintaining your own kernel: =sys-kernel/vanilla-sources-3.4.4 (I'm maintaining my own kernel 3.4.4 but you can still ask to upgrade higher versions) sys-kernel/vanilla-sources (I'm maintaining my own kernel completely, do not ask for any upgrades). Of course you could put 2 statements and do the same approach, but by making the format/syntax consistent, this can be avoided. 2 statements as in: =sys-kernel/vanilla-sources-3.4.4 in package.provided and then mask all versions higher than it in package.use: >sys-kernel/vanilla-sources-3.4.4
*** This bug has been marked as a duplicate of bug 277838 ***
