To install =libreoffice 4.1.0.1 I used autounmask-write. This unmasked =dev-util/mdds-0.9.0 ~ and I ran into bug 474710. Deleting the unmask and installing mdds-0.8.1 works fine for libreoffice (“>=dev-util/mdds-0.8.1:=” in ebuild) so it wasn’t neccessary to unmask 0.9.0 Reproducible: Didn't try Steps to Reproduce: # emerge -v =app-office/libreoffice-4.1.0.1 --autounmask-write Actual Results: =dev-util-0.9.0 is getting unmasked Expected Results: shouldn’t get unmasked, because >=dev-util/mdds-0.8.1 is in 0.8.1 stable.
We have not had any similar reports. If anyone can reproduce this, then please attach a debug log created as follows: emerge [args] --pretend --debug > debug.log 2>&1
I think based on ebuilds, that =libreoffice-4.1.0.1 depends on =mdds-0.8* not a greater version. I have not exactly the same issue but the opposite case in testing branch, cause in ~amd64 the last update for mdds is 0.9.0 and it conflicts with libreoffice-4.1.0.1, so 0.9.0 needs to be masked(not unmasked as Matthias said) in order to install libreoffice-4.1.0.1.
not sure if this is relevant or not but this is what i get : emerge -uaDNv @world @system Nothing to merge; quitting. but if i add --with-bdeps=y : emerge -uaDNv @world @system --with-bdeps=y [ebuild UD] dev-util/mdds-0.7.1 [0.9.0] [ebuild N ] dev-libs/libzip-0.11.1 USE="-static-libs" [ebuild UD] dev-libs/liborcus-0.3.0 [0.5.1] USE="-static-libs" [ebuild UD] app-office/libreoffice-l10n-4.0.4.2 [4.1.0.1] USE="-offlinehelp" LINGUAS="es -af -am -ar -as -ast -be -bg -bn -bn_IN -bo -br -brx -bs -ca -ca_XV -cs -cy -da -de -dgo -dz -el -en -en_GB -en_ZA -eo -et -eu -fa -fi -fr -ga -gd -gl -gu -he -hi -hr -hu -id -is -it -ja -ka -kk -km -kn -ko -kok -ks -ku -lb -lo -lt -lv -mai -mk -ml -mn -mni -mr -my -nb -ne -nl -nn -nr -nso -oc -om -or -pa_IN -pl -pt -pt_BR -ro -ru -rw -sa_IN -sat -sd -sh -si -sk -sl -sq -sr -ss -st -sv -sw_TZ -ta -te -tg -th -tn -tr -ts -tt -ug -uk -uz -ve -vi -xh -zh_CN -zh_TW -zu" [ebuild UD] app-office/libreoffice-4.0.4.2 [4.1.0.1] USE="bluetooth branding cups dbus eds gnome gstreamer gtk opengl vba (-aqua) -debug -java -jemalloc -kde -mysql -odk -postgres -telepathy {-test} -webdav (-gtk3%*)" LIBREOFFICE_EXTENSIONS="presenter-minimizer -nlpsolver -scripting-beanshell -scripting-javascript -wiki-publisher" PYTHON_SINGLE_TARGET="python2_7 -python3_3" PYTHON_TARGETS="python2_7 -python3_3" and i have no idea why it does that. Shouldnt portage actually have blocked the update to mdds-0.9* anyways because of the LO dep ? Furthermore, portage didnt complain about Lo having to be rebuilt after the mdds-0.9.0 update, it only wanted to rebuild liborcus which i did. revdep-rebulid says everthing is fine too. sorry if this unrelated, was searching and this bug looked relevant to me. portage version: Portage 2.2.0_alpha186 (default/linux/amd64/13.0/desktop/gnome, gcc-4.7.3, glibc-2.17, 3.10.0-rc3 x86_64) thanks for your time !
(In reply to gringo from comment #3) > not sure if this is relevant or not but this is what i get : > > emerge -uaDNv @world @system > Nothing to merge; quitting. > > but if i add --with-bdeps=y : > > emerge -uaDNv @world @system --with-bdeps=y > > [ebuild UD] dev-util/mdds-0.7.1 [0.9.0] It seems unrelated to this bug report. > and i have no idea why it does that. > Shouldnt portage actually have blocked the update to mdds-0.9* anyways > because of the LO dep ? It should if you use --with-bdeps. Unless --with-bdeps=y is enabled, emerge is not supposed to account for build-time (DEPEND) only dependencies, so it's perfectly valid to upgrade to dev-util/mdds-0.9.0. > Furthermore, portage didnt complain about Lo having to be rebuilt after the > mdds-0.9.0 update, it only wanted to rebuild liborcus which i did. > revdep-rebulid says everthing is fine too. That's because libreoffice only uses dev-util/mdds at build time. At runtime, it doe not need it at all.
> It seems unrelated to this bug report. thanks, could someone split this off this bug ? or should i open a new one ( in case it really is relevant of course) ? > It should if you use --with-bdeps. Unless --with-bdeps=y is enabled, emerge > is not supposed to account for build-time (DEPEND) only dependencies, so > it's perfectly valid to upgrade to dev-util/mdds-0.9.0. shouldnt there be some logic in portage to avoid this even if we call --with-bdeps ? I mean, whats the point of having =dev-util/mdds-0.8* in the LO ebuild when --with-bdeps=y wont care about it and just update anyways. and why does it want to downgrade to mdds-0.7.1 instead of mdds-0.8.1 ? or, by extension, why does it want to pick libreoffice-4.0.4.2 instead of the already installed libreoffice-4.1.0.1 ? Wouldnt just a downgrade to mdds-0.8.1 ( & rebuild) be enough ? Im not complaining because of the LO downgrade o because i have to rebuild anything, i dont care at all. Just trying to understand the logic behind all this. TIA !
(In reply to gringo from comment #5) > > It seems unrelated to this bug report. > > thanks, could someone split this off this bug ? > or should i open a new one ( in case it really is relevant of course) ? Please go ahead and file a new bug. It may turn out to be duplicate of one of the bugs blocking bug 300071, but that's okay. > > It should if you use --with-bdeps. Unless --with-bdeps=y is enabled, emerge > > is not supposed to account for build-time (DEPEND) only dependencies, so > > it's perfectly valid to upgrade to dev-util/mdds-0.9.0. > > shouldnt there be some logic in portage to avoid this even if we call > --with-bdeps ? > I mean, whats the point of having =dev-util/mdds-0.8* in the LO ebuild when > --with-bdeps=y wont care about it and just update anyways. This specific problem has already been reported as bug 427938. > and why does it want to downgrade to mdds-0.7.1 instead of mdds-0.8.1 ? or, > by extension, why does it want to pick libreoffice-4.0.4.2 instead of the > already installed libreoffice-4.1.0.1 ? Wouldnt just a downgrade to > mdds-0.8.1 ( & rebuild) be enough ? > > Im not complaining because of the LO downgrade o because i have to rebuild > anything, i dont care at all. Just trying to understand the logic behind all > this. Apparently the dependency resolver is not able to settle on a good solution, so it's producing some bad results.
So… this is still an issue. =dev-util/mdds-0.8*:= is the atom in DEPEND on libreoffice ebuild. It's weird looking, I'm assuming the := is for slotting? I don't know but it could be breaking stuff. Another workaround could be to throw this in RDEPEND so we don't end up with deps breaking. Else, fix portage? I figured in 2 months of having this if trying to update --with-bdeps=y someone would have noticed this is not working as expected™.
[ebuild rR ] dev-libs/liborcus-0.5.1 [ebuild rR ] dev-util/mdds-0.8.1 [ebuild r U ] dev-util/mdds-0.9.0 [0.8.1] [nomerge ] x11-misc/bumblebee-3.2.1 [nomerge ] x11-base/xorg-drivers-1.14 [nomerge ] x11-drivers/xf86-video-intel-2.99.901 [ebuild U ] x11-base/xorg-server-1.14.3-r1 [1.14.3] [ebuild rR ] media-libs/mesa-9.2.0 [nomerge ] app-office/libreoffice-4.1.2.2 [ebuild U ] dev-libs/nss-3.15.2 [3.15.1-r2] [ebuild U ] dev-libs/nspr-4.10.1 [4.10] [nomerge ] net-p2p/deluge-1.3.6 [nomerge ] dev-python/twisted-web-13.0.0-r1 [nomerge ] dev-python/twisted-core-13.0.0-r3 [nomerge ] net-zope/zope-interface-4.0.5 [ebuild NS ] dev-python/python-exec-2.0 [0.3.1] [nomerge ] app-misc/freemind-0.9.0-r1 [nomerge ] dev-java/javahelp-2.0.05_p63 [nomerge ] virtual/jdk-1.7.0 [nomerge ] dev-java/icedtea-bin-7.2.4.1 [nomerge ] sys-libs/glibc-2.17 [ebuild U ] sys-libs/timezone-data-2013f [2013d] !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: dev-util/mdds:0 (dev-util/mdds-0.9.0::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (dev-util/mdds-0.8.1::gentoo, ebuild scheduled for merge) pulled in by =dev-util/mdds-0.8*:= required by (app-office/libreoffice-4.1.2.2::gentoo, installed) >=dev-util/mdds-0.7.1:0/0.8.1= required by (dev-libs/libixion-0.5.0::gentoo, installed) (and 1 more with the same problems)
(In reply to tman from comment #8) > [ebuild rR ] dev-libs/liborcus-0.5.1 > [ebuild rR ] dev-util/mdds-0.8.1 > [ebuild r U ] dev-util/mdds-0.9.0 [0.8.1] > [nomerge ] x11-misc/bumblebee-3.2.1 > [nomerge ] x11-base/xorg-drivers-1.14 > [nomerge ] x11-drivers/xf86-video-intel-2.99.901 > [ebuild U ] x11-base/xorg-server-1.14.3-r1 [1.14.3] > [ebuild rR ] media-libs/mesa-9.2.0 > [nomerge ] app-office/libreoffice-4.1.2.2 > [ebuild U ] dev-libs/nss-3.15.2 [3.15.1-r2] > [ebuild U ] dev-libs/nspr-4.10.1 [4.10] > [nomerge ] net-p2p/deluge-1.3.6 > [nomerge ] dev-python/twisted-web-13.0.0-r1 > [nomerge ] dev-python/twisted-core-13.0.0-r3 > [nomerge ] net-zope/zope-interface-4.0.5 > [ebuild NS ] dev-python/python-exec-2.0 [0.3.1] > [nomerge ] app-misc/freemind-0.9.0-r1 > [nomerge ] dev-java/javahelp-2.0.05_p63 > [nomerge ] virtual/jdk-1.7.0 > [nomerge ] dev-java/icedtea-bin-7.2.4.1 > [nomerge ] sys-libs/glibc-2.17 > [ebuild U ] sys-libs/timezone-data-2013f [2013d] > > !!! Multiple package instances within a single package slot have been pulled > !!! into the dependency graph, resulting in a slot conflict: > > dev-util/mdds:0 > > (dev-util/mdds-0.9.0::gentoo, ebuild scheduled for merge) pulled in by > (no parents that aren't satisfied by other packages in this slot) > > (dev-util/mdds-0.8.1::gentoo, ebuild scheduled for merge) pulled in by > =dev-util/mdds-0.8*:= required by > (app-office/libreoffice-4.1.2.2::gentoo, installed) > >=dev-util/mdds-0.7.1:0/0.8.1= required by > (dev-libs/libixion-0.5.0::gentoo, installed) > (and 1 more with the same problems) Same here for me
tman, still a fail but a different bug for you. https://bugs.gentoo.org/show_bug.cgi?id=486550
I have this issue as well... Seems unnecessary to upgrade mdds if nobody is using it. Or have I misunderstood the problem? I thought the slotting would allow the to co-exist.
These are the packages that would be merged, in reverse order: Calculating dependencies... done! [ebuild U ] app-office/libreoffice-4.1.2.3 [4.1.2.2-r1] USE="bluetooth branding cups dbus gnome gstreamer gtk gtk3 java kde opengl telepathy vba webdav (-aqua) -debug -eds -jemalloc -mysql -odk -postgres {-test}" LIBREOFFICE_EXTENSIONS="presenter-minimizer -nlpsolver -scripting-beanshell -scripting-javascript -wiki-publisher" PYTHON_SINGLE_TARGET="python2_7 -python3_3" PYTHON_TARGETS="python2_7 python3_3" 0 kB Total: 1 package (1 upgrade), Size of downloads: 0 kB WARNING: One or more updates have been skipped due to a dependency conflict: dev-util/mdds:0 (dev-util/mdds-0.9.0::gentoo, ebuild scheduled for merge) conflicts with =dev-util/mdds-0.8*:= required by (app-office/libreoffice-4.1.2.3::gentoo, ebuild scheduled for merge)
If this issue was due to some misbehavior in backtracking triggered by a slot conflict (multiple versions of dev-util/mdds pulled into the graph), then it may have been fixed since portage-2.2.9 by solving the conflict without backtracking: https://github.com/gentoo/portage/commit/a862cc5dd1a56114fa579c5fb01b518b243666d9
Please re-open if you can reproduce this with >=sys-apps/portage-2.2.9.
