Portage wants to install tcl here because of sys-libs/db, even though the use flag isn't set, also the additional nomerge line in the --tree output is rather weird. Downgraded to 2.1.2.2 - same problem. Did not test other versions yet. emerge man dev-perl/Net-IP jdepend -1pv Calculating dependencies... done! [ebuild U ] sys-apps/man-1.6e-r1 [1.6d] USE="nls" 0 kB [ebuild U ] dev-perl/Net-IP-1.25-r1 [1.25] 0 kB [ebuild U ] dev-java/jdepend-2.9-r3 [2.8.1] USE="doc -source (-jikes%)" 296 kB emerge man sys-libs/db dev-perl/Net-IP jdepend -1pv Calculating dependencies... done! [ebuild N ] dev-lang/tcl-8.4.9 USE="threads" 0 kB [ebuild U ] sys-apps/man-1.6e-r1 [1.6d] USE="nls" 0 kB [ebuild U ] dev-perl/Net-IP-1.25-r1 [1.25] 0 kB [ebuild U ] dev-java/jdepend-2.9-r3 [2.8.1] USE="doc -source (-jikes%)" 296 kB [ebuild NS ] sys-libs/db-4.3.29-r2 USE="doc test -bootstrap -java -nocxx -tcl" 5,961 kB emerge man sys-libs/db dev-perl/Net-IP jdepend -1pvt Calculating dependencies... done! [ebuild U ] dev-java/jdepend-2.9-r3 [2.8.1] USE="doc -source (-jikes%)" 296 kB [ebuild U ] dev-perl/Net-IP-1.25-r1 [1.25] 0 kB [ebuild NS ] sys-libs/db-4.3.29-r2 USE="doc test -bootstrap -java -nocxx -tcl" 5,961 kB [ebuild U ] sys-apps/man-1.6e-r1 [1.6d] USE="nls" 0 kB [nomerge ] sys-libs/db-4.3.29-r2 USE="doc test -bootstrap -java -nocxx -tcl" [ebuild N ] dev-lang/tcl-8.4.9 USE="threads" 0 kB
Here's an idea, look at db.eclass dependencies for USE=test ;)
(In reply to comment #0) > the additional nomerge line in the --tree output is rather weird. Lots of people think it's weird but it solves bug 158100.
(In reply to comment #1) > Here's an idea, look at db.eclass dependencies for USE=test ;) > Got me, fine. ;) I was rather puzzled, not having seen such a strange /bug/ in a long time. As you know the portage bugs better - is there a feature request regarding some transparency towards the user on such "overloading" issues and what's happening within eclasses in generel? Especially blockers within eclasses I have on my mind.
(In reply to comment #3) > Got me, fine. ;) I was rather puzzled, not having seen such a strange /bug/ in > a long time. As you know the portage bugs better - is there a feature request > regarding some transparency towards the user on such "overloading" issues and > what's happening within eclasses in generel? Especially blockers within > eclasses I have on my mind. We have several bug related to increasing transparency: bug 162173, bug 163684, and bug 164457. None of those are about showing eclass relationships though. Currently, the metadata generation phase doesn't return any information about which eclasses provide what attributes. An automated way to visualize that would certainly be useful though...