Summary: | The following USE changes are necessary to proceed: >=media-gfx/graphicsmagick-1.3.20-r1 imagemagick | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | David Kredba <kredba> |
Component: | [OLD] Core system | Assignee: | Portage team <dev-portage> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | esigra, Martin.vGagern, pa.lacaze |
Priority: | Normal | Keywords: | InVCS |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://forums.gentoo.org/viewtopic-t-1050078.html | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 155723, 611328 | ||
Attachments: | Emerge debug output 7z mx9 compressed |
Description
David Kredba
2015-07-06 09:48:59 UTC
Created attachment 406238 [details]
Emerge debug output 7z mx9 compressed
(In reply to David Kredba from comment #0) > Portage enforces me to change use flag on media-gfx/graphicsmagick to > include imagemagick for no real reason IMHO. What happens if you use --autounmask=n so that it won't attempt to modify your package.use settings? > When ocatve-4.0.0 was present in system it blame ocatve. After removing > octave it not blame anything. Now with octave present it blames kopete > [latex]: That seems like a symptom of bug 419381. Hello Zac, here it is: emerge: there are no ebuilds built with USE flags to satisfy "media-gfx/graphicsmagick[imagemagick]". !!! One of the following packages is required to complete your request: - media-gfx/graphicsmagick-1.3.20-r1::gentoo (Change USE: +imagemagick) (dependency required by "media-plugins/kipi-plugins-4.10.0::gentoo[videoslideshow,imagemagick]" [installed]) (dependency required by "media-gfx/digikam-4.10.0::gentoo" [installed]) (dependency required by "@selected" [set]) (dependency required by "@world" [argument]) Thank you. The dependency resolution problem is still present, but I found a workaround: "emerge -1 imagemagick". After this I can "emerge -uDN @world". (In reply to David Kredba from comment #1) > Created attachment 406238 [details] > Emerge debug output 7z mx9 compressed Apparently the problem is triggered by this slot conflict: !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: media-gfx/imagemagick:0 (media-gfx/imagemagick-6.9.1.7:0/6.9.1.7::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (media-gfx/imagemagick-6.9.1.4:0/6.9.1.4::gentoo, installed) pulled in by media-gfx/imagemagick:0/6.9.1.4=[cxx] required by (media-gfx/inkscape-0.91:0/0::gentoo, installed) ^^^^^^^^^^^ (and 1 more with the same problem) I think we need to wait until the fixes for bug 567686 and bug 566024 are released, and then see if problem is still reproducible. If it is reproducible, then we should get a copy of the configuration created as follows: tar -cJf /tmp/config.tar.xz -C / ./etc/portage ./var/db/pkg $(cd / && echo ./var/lib/portage/world*) --exclude CONTENTS --exclude '*.ebuild' --exclude environment.bz2 --exclude 'NEEDED*' Portage-2.2.27 I just released in the tree with those bugfixes. It should be availabe for merge in the next few hours depending on your rsync server you connect to. Best is to wat a few hours and then sync, emerge 2.2.27 Same problem again. I solved it as above with emerge -1 imagemagick followd by emerge -unD world. Versions: media-gfx/imagemagick-6.9.5.10 sci-mathematics/octave-4.0.3 media-gfx/graphicsmagick-1.3.25 *** Bug 601628 has been marked as a duplicate of this bug. *** (In reply to Zac Medico from comment #5) > Apparently the problem is triggered by this slot conflict: > > !!! Multiple package instances within a single package slot have been pulled > !!! into the dependency graph, resulting in a slot conflict: > > media-gfx/imagemagick:0 > > (media-gfx/imagemagick-6.9.1.7:0/6.9.1.7::gentoo, ebuild scheduled for > merge) pulled in by > (no parents that aren't satisfied by other packages in this slot) > > (media-gfx/imagemagick-6.9.1.4:0/6.9.1.4::gentoo, installed) pulled in by > media-gfx/imagemagick:0/6.9.1.4=[cxx] required by > (media-gfx/inkscape-0.91:0/0::gentoo, installed) > ^^^^^^^^^^^ > > (and 1 more with the same problem) I suspect this code from bug 531656 is triggering the problem when it tries to avoid the above conflict: https://gitweb.gentoo.org/proj/portage.git/commit/?id=a9064d08ef4c92a5d0d1bfb3dc8a01b7850812b0 That fix was released in portage-2.2.16, a few months before this bug was reported for portage-2.2.20. Because PackageTracker.match returns packages in ascending order, depgraph._wrapped_select_pkg_highest_available_imp prefers to return the lowest version involved in a slot conflict here: https://gitweb.gentoo.org/proj/portage.git/tree/pym/_emerge/depgraph.py?h=portage-2.3.2#n6070 That will cause the conflict_dowgrade flag to be set here, because avail_pkg refers to the lowest version returned from PackageTracker.match: https://gitweb.gentoo.org/proj/portage.git/tree/pym/portage/dep/dep_check.py?h=portage-2.3.2#n374 It seems like making depgraph._wrapped_select_pkg_highest_available_imp use the highest version returned from PackageTracker.match should solve the problem. Patch posted for review: https://archives.gentoo.org/gentoo-portage-dev/message/9b808bb09aa9c2d28f1bc0f4c6a3ddaa https://github.com/gentoo/portage/pull/76 This is in the master branch: https://gitweb.gentoo.org/proj/portage.git/commit/?id=ab07ac68fa1e04ed64e2e0f6c753ff169a32d517 I'm still seeing this with portage-2.3.3, during a world update that would involve upgrading imagemagick from 6.9.6.6 to 6.9.7.0.
# emerge -uND --newrepo --ask y --keep-going --with-bdeps=y --backtrack=50 --autounmask-write --autounmask @system @world
⋮
[blocks B ] media-gfx/imagemagick ("media-gfx/imagemagick" is blocking media-gfx/graphicsmagick-1.3.25)
!!! Multiple package instances within a single package slot have been pulled
!!! into the dependency graph, resulting in a slot conflict:
media-gfx/imagemagick:0
(media-gfx/imagemagick-6.9.7.0:0/6.9.7.0::gentoo, ebuild scheduled for merge) pulled in by
(no parents that aren't satisfied by other packages in this slot)
(media-gfx/imagemagick-6.9.6.6:0/6.9.6.6::gentoo, installed) pulled in by
media-gfx/imagemagick:0/6.9.6.6=[cxx] required by (media-gfx/inkscape-0.91-r3:0/0::gentoo, installed)
^^^^^^^^^^^
(and 1 more with the same problem)
⋮
* Error: The above package list contains packages which cannot be
* installed at the same time on the same system.
(media-gfx/graphicsmagick-1.3.25:0/1.3::gentoo, ebuild scheduled for merge) pulled in by media-gfx/graphicsmagick required by @selected
media-gfx/graphicsmagick[imagemagick,png] required by (app-editors/kile-2.1.3-r2:4/4::gentoo, installed)
⋮
The following USE changes are necessary to proceed:
(see "package.use" in the portage(5) man page for more details)
# required by app-editors/kile-2.1.3-r2::gentoo[png]
# required by @selected
# required by @world (argument)
>=media-gfx/graphicsmagick-1.3.25 imagemagick
That “1 more with the same problem” is likely transcode, which has an imagemagick:= dependency as well. I've got yet another conflict, due to qgis-2.18.2-r2 being incompatible with sip-4.19. So there is a sight chance that the imagemagick issue would magically disappear if the other issue got resolved.
Can I provide any more information before I resolve this by manually upgrading imagemagick first?
(In reply to Martin von Gagern from comment #13) > So there is a sight chance that the imagemagick issue would magically > disappear if the other issue got resolved. Unfortunately, but not unexpectedly, it did not. > Can I provide any more information before I resolve this by manually > upgrading imagemagick first? I pasted the full output of an emerge --pretend @world in https://gist.github.com/gagern/a9aa117f1d3c06a7ae8ff4b6a7b1c741 and will upgrade imagemagick now, to get on with my world update. (In reply to Martin von Gagern from comment #13) > I'm still seeing this with portage-2.3.3, during a world update that would > involve upgrading imagemagick from 6.9.6.6 to 6.9.7.0. I think that the _dep_check_composite_db._visible method needs to be adjusted here, to select the higest version from the package_tracker.match call, for consistency with the change in ab07ac68fa1e04ed64e2e0f6c753ff169a32d517: https://gitweb.gentoo.org/proj/portage.git/tree/pym/_emerge/depgraph.py?h=portage-2.3.3#n9126 I'm going to try and create a unit test to reproduce this problem, and use that to test a fix. Patch posted for review: https://archives.gentoo.org/gentoo-portage-dev/message/21bf91206c165da174ea824a934cee70 https://github.com/gentoo/portage/pull/82 This is in the master branch: https://gitweb.gentoo.org/proj/portage.git/commit/?id=033477d6a4cc1adcabed966b8aa82b51b42af33c I tried the patch on top or =sys-apps/portage-2.3.3 now and it solves it! Thank you. or -> of *** Bug 610672 has been marked as a duplicate of this bug. *** Fixed in portage-2.3.5. |