Summary: | sys-apps/portage-2.2.26: emerge @world does not pick up [rR] slot-operator dependency scheduled for re-emerging | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Anton Bolshakov <anton.bugs> |
Component: | Current packages | Assignee: | Portage team <dev-portage> |
Status: | RESOLVED OBSOLETE | ||
Severity: | normal | CC: | esigra, kripton, pacho, zerochaos |
Priority: | Highest | ||
Version: | 10.1 | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 155723 | ||
Attachments: |
config.tar.xz
debug.log.xz successful dependency calculation (>=dev-python/ipython-4 masked) |
Description
Anton Bolshakov
2016-04-22 02:14:01 UTC
portage subslot dep calculation failure? Somehow, with "world" portage does not detect that few deps need to be rebuilt too. Here is what I have on gnuradio upgrade specifically: bash# emerge -pv gnuradio These are the packages that would be merged, in reverse order: Calculating dependencies... done! [ebuild rR *] net-wireless/gr-air-modes-9999::pentoo USE="gui -fgfs -rtl -uhd" PYTHON_TARGETS="python2_7" 0 KiB [ebuild rR ~] net-wireless/gqrx-2.5::gentoo USE="-pulseaudio" 0 KiB [ebuild rR ~] net-wireless/gr-osmosdr-0.1.4_p20150730:0/0.1.4_p20150730::gentoo USE="fcd hackrf iqbalance python rtlsdr uhd -bladerf (-mirisdr)" PYTHON_TARGETS="python2_7" 0 KiB [ebuild rR ~] net-wireless/gr-iqbal-0.37.2:0/0.37.2::gentoo PYTHON_TARGETS="python2_7" 0 KiB [ebuild r U ~] net-wireless/gnuradio-3.7.9.2:0/3.7.9.2::gentoo [3.7.9:0/3.7.9::gentoo] USE="alsa analog audio channels digital dtv fcd fec filter grc noaa qt4 sdl trellis uhd utils wavelet wxwidgets -atsc -doc -examples -jack -log -oss -pager -performance-counters -portaudio {-test} -vocoder -zeromq" PYTHON_TARGETS="python2_7" 0 KiB Total: 5 packages (1 upgrade, 4 reinstalls), Size of downloads: 0 KiB The following packages are causing rebuilds: (net-wireless/gnuradio-3.7.9.2:0/3.7.9.2::gentoo, ebuild scheduled for merge) causes rebuilds for: (net-wireless/gr-air-modes-9999:0/0::pentoo, ebuild scheduled for merge) (net-wireless/gr-osmosdr-0.1.4_p20150730:0/0.1.4_p20150730::gentoo, ebuild scheduled for merge) (net-wireless/gr-iqbal-0.37.2:0/0.37.2::gentoo, ebuild scheduled for merge) (net-wireless/gqrx-2.5:0/0::gentoo, ebuild scheduled for merge) Created attachment 432032 [details]
config.tar.xz
the output of the following command:
tar -cJf /tmp/config.tar.xz -C / ./etc/portage ./var/db/pkg ./var/lib/portage/world --exclude CONTENTS --exclude environment.bz2 --exclude 'NEEDED*' --exclude '*.ebuild'
Since there is no progress so far and the bug might not be fix soonest, I'm going to upgrade gnuradio manually. I'm attaching some configs (copied from bug #566024) with a hope that it will help to reproduce the problem. Created attachment 432034 [details]
debug.log.xz
output of the following: emerge -DNUupv --with-bdeps=y --debug world > debug.log 2>&1 xz -9 debug.log
There is one more package got blocked with the similar problem after recent sync: app-text/poppler:0 (app-text/poppler-0.42.0:0/59::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (app-text/poppler-0.32.0:0/51::gentoo, installed) pulled in by app-text/poppler:0/51 required by (app-office/libreoffice-bin-5.0.5.2-r1:0/0::gentoo, installed) ^^^^^ >=app-text/poppler-0.12.3-r3:0/51= required by (app-text/texlive-core-2014-r4:0/0::gentoo, installed) ^^^^^^ (and 2 more with the same problems) I keyworded app-office/libreoffice-bin-5.1.2.2 (it is compiled against the currently stable poppler) and the same problem show up. emerge -1 libreoffice-bin worked, but "emerge -DNu @world" didn't work. we have yet another user with the same problem, but the list of problematic packages is bigger: https://bpaste.net/show/4ed26dc3520f Does anybody look at the bug? The bug sounds very critical. (In reply to Anton Bolshakov from comment #9) > Does anybody look at the bug? The bug sounds very critical. Today I plan to see if I can use your configuration to reproduce it. As a workaround, typically you can use the --verbose-conflicts option to see which packages need to rebuilt, and you can rebuild them manually. Alternatively, you can use the --ignore-built-slot-operator-deps=y option to temporarily ignore the conflict. If you do that, then the rebuilds might trigger automatically on the next emerge run. (In reply to Anton Bolshakov from comment #8) > we have yet another user with the same problem, but the list of problematic > packages is bigger: > > https://bpaste.net/show/4ed26dc3520f Yeah, that could be the same bug. (In reply to Anton Bolshakov from comment #3) > Created attachment 432032 [details] > config.tar.xz > > the output of the following command: > tar -cJf /tmp/config.tar.xz -C / ./etc/portage ./var/db/pkg > ./var/lib/portage/world --exclude CONTENTS --exclude environment.bz2 > --exclude 'NEEDED*' --exclude '*.ebuild' With your configuration, it looks like I'm getting similar results. I'm using these repos: gentoo e272bacc12364f633ee7a3b8024aaa117011a708 pentoo 4f1c01d19c2a91193ae5b0e1270d98970aeabd54 I'm testing with portage-2.2.28, since I want to eliminate interactions from bug 566024 and bug 567686. Using emerge -DNupv world, I see what look like solvable slot conflicts for ncurses, poppler, and gnuradio. If I add the --changed-deps option, the ncurses conflict is the only one that remains. I will investigate why these conflicts are not solved. (In reply to Zac Medico from comment #11) > Using emerge -DNupv world, I see what look like solvable slot conflicts for > ncurses, poppler, and gnuradio. If I add the --changed-deps option, the > ncurses conflict is the only one that remains. I will investigate why these > conflicts are not solved. Eventually, I figured out that this blocker was the only thing preventing it from solving: [blocks B ] <dev-python/ipython-4.0.0[smp] ("<dev-python/ipython-4.0.0[smp]" is blocking dev-python/ipyparallel-5.0.1) With portage-2.2.28 it solved after putting '>=dev-python/ipython-4' in package.mask. I still need to test this configuration with portage-2.2.26, to see if bug 566024 and/or bug 567686 may be involved. (In reply to Anton Bolshakov from comment #8) > we have yet another user with the same problem, but the list of problematic > packages is bigger: > > https://bpaste.net/show/4ed26dc3520f Looking at this again, it fails due to an unsatisfied dependency: emerge: there are no ebuilds to satisfy "net-analyzer/wireshark:0/1.12.8=[lua]". (dependency required by "net-analyzer/cloudshark-1.0.4::gentoo" [installed]) The means the dependency calculation didn't finish, and the slot conflicts might have solved if the dependency calculation run to completion. I know that this is not obvious to users, so we should do something about that. Created attachment 433756 [details] successful dependency calculation (>=dev-python/ipython-4 masked) I get the same results with 2.2.26 and 2.2.28, so it doesn't seem like bug 566024 or bug 567686 are involved. Actually, maybe bug 566024 or bug 567686 played a role, but I didn't reproduce that result. It's possible for hash randomization to affect dependency resolver outcomes. It were way to many changes in portage since 2016 so I guess nobody would be able to reproduce or fix this bug. |