On 29/06/2014 23:57, gottl...@nyu.edu wrote: > On Sun, Jun 29 2014, Alan McKinnon wrote: > >> On 29/06/2014 22:09, gottl...@nyu.edu wrote: >>> My desktop is a fully stable (empty package.accept_keywords) systemd >>> system. The profile is .../gnome/system and it boots init=systemd. >>> >>> The 3 june news asserts "all systemd users are recommended to stay with >>> sys-power/upower". >>> >>> However update world wants to uninstall upower and install >>> upower-pm-utils. >>> >>> Adding sys-power/upower-pm-utils to /etc/package.mask, "fixed" the >>> problem, but I wonder if I chose an appropriate fix. I was a little >>> surprised that a stable system would need an entry in package.mask. >>> >>> Any advice would be appreciated, >>> allan >>> >>> PS I realize that the news item concerned hibernate/suspend so is not >>> relevant, but the same issue uninstall/install occurs on my laptop. >>> Those systems have a less simple "goingstable" setup and are more >>> important to me so I prefer to first change the fully stable desktop. >> >> >> Are you still dealing with this same output you posted about on the 26th? >> >> Calculating dependencies... done! >> >> [ebuild U ] x11-wm/sawfish-1.9.1-r2 [1.9.1-r1] USE="emacs%* nls >> -xinerama" 2,556 kB >> [nomerge ] gnome-base/gnome-3.10.0:2.0 USE="bluetooth cdr classic >> cups extras -accessibility" >> [nomerge ] gnome-base/gnome-shell-3.10.4-r2 USE="bluetooth i18n >> networkmanager (-openrc-force)" PYTHON_TARGETS="python2_7" >> [nomerge ] sys-power/upower-pm-utils-0.9.23-r2 >> USE="introspection -ios" >> [blocks b ] sys-power/upower ("sys-power/upower" is blocking >> sys-power/upower-pm-utils-0.9.23-r2) >> [uninstall ] sys-power/upower-0.9.23-r3 USE="introspection -doc >> -ios" >> [ebuild N ] sys-power/upower-pm-utils-0.9.23-r2 >> USE="introspection -ios" 0 kB >> >> That is gnome-shell pulling in upower || upower-pm-utils and for some >> reason it chose the one you do not want. >> >> Using package.mask is valid (it's a documented tool and not only for >> ~arch - all it means is that you do not want the listed packages and >> there could be many reasons for that) but it does seem a bit heavy-handed. >> >> Normally, manually installing upower should be enough to satisfy the dep >> and keep upower-pm-utils off your machine, but bugs are possible I suppose. >> >> Is there any bugs on b.g.o. about this? >> Run emerge with -t and post the relevant section, let's see why the >> wrong package is being pulled in. Also the output of >> >> equery depends upower >> equery depends upower-pm-utils > > Yes it is the same general issue. But I moved from my "goingstable" > laptop to my fully stable desktop, hoping it would be clearer. > I did run the emerge world with -t on the desktop. I am now temporarily > removing the package.mask entry and running it again. The output is > > Calculating dependencies... done! > [nomerge ] gnome-base/gnome-3.10.0:2.0 USE="bluetooth cdr classic cups > extras -accessibility" > [nomerge ] gnome-base/gnome-shell-3.10.4-r2 USE="bluetooth i18n > networkmanager (-openrc-force)" PYTHON_TARGETS="python2_7" > [nomerge ] sys-power/upower-pm-utils-0.9.23-r2 USE="introspection > -ios" > [blocks b ] sys-power/upower ("sys-power/upower" is blocking > sys-power/upower-pm-utils-0.9.23-r2) > [uninstall ] sys-power/upower-0.9.23-r3 USE="introspection -doc > -ios" > [ebuild N ] sys-power/upower-pm-utils-0.9.23-r2 USE="introspection > -ios" 0 kB > Total: 1 package (1 new, 1 uninstall), Size of downloads: 0 kB > Conflict: 1 block > > I did not do this upgrade and with my package.mask in place, update > world says "nothing to merge". > > The equery's give > > allan ~ # equery depends upower > * These packages depend on upower: > app-misc/tracker-0.16.4 (laptop ? <sys-power/upower-0.99) > gnome-base/gnome-control-center-3.10.3 (<sys-power/upower-0.99) > gnome-base/gnome-session-3.10.1 (<sys-power/upower-0.99) > gnome-base/gnome-settings-daemon-3.10.2 (<sys-power/upower-0.99) > gnome-base/gnome-shell-3.10.4-r2 (<sys-power/upower-0.99[introspection]) > gnome-extra/gnome-power-manager-3.10.1 (<sys-power/upower-0.99) > net-im/telepathy-mission-control-5.14.1 (upower ? >=sys-power/upower-0.9.11) > (upower ? <sys-power/upower-0.99) > net-misc/networkmanager-0.9.8.8 (sys-power/upower) > x11-wm/mutter-3.10.4 (<sys-power/upower-0.99) > allan ~ # equery depends upower-pm-utils > * These packages depend on upower-pm-utils: > app-misc/tracker-0.16.4 (laptop ? sys-power/upower-pm-utils) > gnome-base/gnome-control-center-3.10.3 (sys-power/upower-pm-utils) > gnome-base/gnome-session-3.10.1 (sys-power/upower-pm-utils) > gnome-base/gnome-settings-daemon-3.10.2 (sys-power/upower-pm-utils) > gnome-base/gnome-shell-3.10.4-r2 (sys-power/upower-pm-utils[introspection]) > gnome-extra/gnome-power-manager-3.10.1 (sys-power/upower-pm-utils) > net-im/telepathy-mission-control-5.14.1 (sys-power/upower-pm-utils) > net-misc/networkmanager-0.9.8.8 (sys-power/upower-pm-utils) > x11-wm/mutter-3.10.4 (sys-power/upower-pm-utils) > allan ~ # > > When I started this project there were no related bugs in b.g.o. > I will investigate them > > 513842 - talks about a upower mask from the openrc perspective > 513700 - is about stabilizing upower-0.99.0-r1 > 512252 - again a problem for openrc users > 515228 - talks about upower versions and differences with different systems. > > One phrase from the last bug is "Probably we need to document this > better, for example in the wiki." > > So perhaps the package.mask is not terrible. > Would you advise I put the package.mask in a "goingstable" laptop and > try an update world? I do have two laptops whose configurations are > "close", one is my main system. I would first update world on the other > one.
Hi Allen, It's late and perhaps I'm not think too clearly, but I think I see the problem: Almost everything listed in equery depends upower is looking for this: <sys-power/upower-0.99 But look at this: $ eix -e upower * sys-power/upower Available versions: [M]0.9.23-r3 0.99.0-r1(0/2) {doc +introspection ios KERNEL="FreeBSD linux"} Homepage: http://upower.freedesktop.org/ Description: D-Bus abstraction for enumerating power devices and querying history and statistics There is no stable version of upower you can use (0.9.23-r3 is masked in the profile) so only upower-pm-utils satisfies the dep. However, the mask is reversed in the systemd profile. Are you using the systemd profile? If not, set it and do a pretend emerge run. I feel this will sort out your problem. The situation with upower does feel clunky to my mind; profiles are supposed to be for the most part entirely optional and serve as a convenience. Requiring a specific profile to get stuff to work doesn't feel right on a user level even though the maths logic driving it is correct -- Alan McKinnon alan.mckin...@gmail.com