With sys-apps/portage-2.2_rc94, as well as sys-apps/portage-2.2_rc91, with latest tree, doing emerge -vauDN world results in python-2.6.6-r1 being a blocker for many programs depending on >=python-2.5 or =python-2*, and then portage wants to downgrade it to resolve these dependency issues. Reproducible: Always Steps to Reproduce: 1. Have hard-unmasked sys-apps/portage-2.2 and dev-lang/python-2.6.6-r1 2. emerge --sync 3. emerge -vauDN world Actual Results: emerge then says [ebuild UD] dev-lang/python-2.6.5-r103 [2.6.6-r1] USE="ncurses readline sqlite ssl threads (wide-unicode) xml -berkdb -build -doc -examples -gdbm -ipv6 -tk -wi !!! One or more updates have been skipped due to a dependency conflict: dev-lang/python:2.6 (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (net-analyzer/nmap-5.35::gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2*[threads] required by (app-editors/vim-7.3::gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (app-editors/vim-7.3::gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with >=dev-lang/python-2.6.5-r99:2.6 required by (sci-mathematics/sage-4.5.3::sage-on-gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (app-admin/webapp-config-1.50.16-r3::gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (media-gfx/exiv2-0.20::gentoo, installed) (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (media-libs/alsa-lib-1.0.23::gentoo, installed) Expected Results: Use python-2.6.6-r1 I have replaced the line of /usr/portage/net-analyzer/nmap/nmap-5.35.ebuild that said "=dev-lang/python-2.5[sqlite]" with "=dev-lang/python-2.6[sqlite]" and it got out of the blocking list. So I suppose this could be a bug with (newer) portage-2.2.
There are some differences between what you wrote and what you pasted. The summary say it wants to downgrade to python-2.5, but your merge list contain 2.6.5. (In reply to comment #0) > I have replaced the line of /usr/portage/net-analyzer/nmap/nmap-5.35.ebuild > that said "=dev-lang/python-2.5[sqlite]" with "=dev-lang/python-2.6[sqlite]" > and it got out of the blocking list. The nmap-5.35 ebuild doesn't contain "=dev-lang/python-2.5[sqlite]", only ">=dev-lang/python-2.5[sqlite]" and there are no blockers in the output you pasted.
Sorry for the double mistake. You're absolutely right about the version portage wants to downgrade to. And, as I have tested editing nmap, emerge -vauDN ; revert edit ; emerge -vauDN I have simply copied the bad one... thus with a standard portage tree this would downgrade from 2.6.6 to 2.6.5 (in my case, I have sage-on-gentoo overlay providing 2.6.5-r103, but otherwise I think it's juste 2.6.5-r3) and indeed I get (dev-lang/python-2.6.6-r1::gentoo, ebuild scheduled for merge) conflicts with =dev-lang/python-2* required by (net-analyzer/nmap-5.35::gentoo, installed) along with the others. And the incriminated part of nmap-5.35.ebuild I have is (as you said, again) gtk? ( >=x11-libs/gtk+-2.6 >=dev-python/pygtk-2.6 || ( >=dev-lang/python-2.5[sqlite] >=dev-python/pysqlite-2 ) ) which shows that some strange thing is going into portage (as it says =python-2* just above). Perhaps some inherit is going there? Well, now I found out that indeed sage-4.5.3 requires CDEPEND=">=dev-lang/python-2.6.5-r99:2.6 ~dev-lang/python-2.6.5:2.6 so perhaps this is WHY portage does not want a 2.6.6 merged. If that's the case, then this should probably go to sage maintainers (for requesting 2.6.6 also good...). But anyway, portage should not show non-blockers such as =python-2* for python-2.6.6-r1, should it? So this remains a bug?
(In reply to comment #2) > But anyway, portage should not show > non-blockers such as =python-2* for python-2.6.6-r1, should it? So this remains > a bug? > Agreed.
Please create debug.log like this and post it: emerge -vuDN world --pretend --debug &> debug.log
Created attachment 250613 [details] Result of "emerge -vuDN world --pretend --debug", gzipped
Hopefully this helps: http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=8f91b6615617af839a31e51bd3bb2048ea81ea56
For the moment I'm not labeling this as a regression because I doubt that portage-2.1.8.3 handles this any better.
This is fixed in 2.1.9.19 and 2.2_rc97.