Created attachment 303563 [details] emerge --info portage emerge --pretend --update --deep --newuse @system just decided that virtual/pager's || rdep isn't filled despite me having both sys-apps/most and sys-apps/util-linux[ncurses] installed. ($>> indicates bash prompt.) $>>emerge --pretend --update --deep --newuse @system These are the packages that would be merged, in order: Calculating dependencies ... done! [ebuild N ] sys-apps/less-445-r1 USE="pcre unicode" $>>equery d less * These packages depend on less: virtual/pager-0 (sys-apps/less) $>>equery l most util-linux * Searching for most ... [IP-] [ ] sys-apps/most-5.0.0a:0 * Searching for util-linux ... [IP-] [ ] sys-apps/util-linux-2.21:0 $>>equery -N u util-linux | grep ncurses + + ncurses : Adds ncurses support (console display library) - - slang : Adds support for the slang text display library (it's like ncurses, but different) $>> From the virtual pager ebuild: RDEPEND="|| ( sys-apps/less sys-apps/more sys-apps/most sys-apps/util-linux[ncurses] app-text/lv app-editors/vim[vim-pager] )" According to the changelog, the last virtual/pager change was back in January and I've updated several times since then without portage pulling in less. so it would appear to be a regression in portage itself. According to the dates on the binpkgs, I updated to portage-2.2.0_alpha89 on Feb 23, and before that to alpha87 on Feb 20, so the regression would be in 88 or 89 as the last update, on the 23rd with alpha87, didn't try to pull in less. At a casual glance bug #280336 (|| regression back with 2.2_rc35) looks like it might be related. emerge --info portage attached.
This is odd, because I don't recall any recent changes in the relevant dependency resolution code. Will this command try to remove it? emerge -pv --depclean sys-apps/less
(In reply to comment #1) > This is odd, because I don't recall any recent changes in the relevant > dependency resolution code. I've thought a bit more and believe it must be a change in the tree, not portage, but what? When I see portage in the --pretend update list, I check its changelog and manually update it first. So if it were a change in portage, after I updated portage back on the 23rd and reran the @system/@world updates to catch the rest, it would have shown up then. So it can't be a change in portage, and must instead be a change in the tree (or overlays, kde, x11, mozilla, nothing kooky). But that's still a portage bug since other tree changes shouldn't change Now's when I really miss a git-based tree, as doing a bisect would sure be nice! > Will this command try to remove it? > > emerge -pv --depclean sys-apps/less I hadn't actually done the merge yet, but I just did it so I could try depclean, and no, depclean doesn't want to remove it either. Yet virtual/pager is still the only thing pulling it in, according to equery. And equary says nothing (well, no package) is pulling in virtual/pager, so that must be the direct @system dep only, there. So what sort of tree or apparently unrelated portage change could affect a virtual with no inherits and that's only pulled in by @system, that way? Maybe a profile change to pull in less directly, instead of virtual/pager? Sounds strange but let me look... ARGH!! profiles/base/packages! $>>grep -r sys-apps/less /p/profiles /p/profiles/base/packages:*sys-apps/less /p/profiles/base/ChangeLog: adding sys-apps/less temporarly to resolve bug 398295. /p/profiles/default/bsd/fbsd/packages.build:sys-apps/less /p/profiles/default/linux/packages.build:sys-apps/less /p/profiles/use.local.desc:sys-apps/portage:less - Pull in sys-apps/less as default pager for etc-update. The base-profile changelog entry tells the story... 26 Feb 2012; Mike Doty <kingtaco@gentoo.org> packages: adding sys-apps/less temporarly to resolve bug 398295. So it's not a portage problem after all, and the proper resolution here is as a dup of 398295, it would seem. Unfortunately, the chances of further dups could be pretty high, so keep an eye out for bugs due to pulling in less. =:^\ Thanks for the fast response and sorry for the bother. Duncan *** This bug has been marked as a duplicate of bug 398295 ***
