Created attachment 773291 [details] output from emerge including errors Hello, A conflict of dependencies apparently involving kde and qt is preventing any upgrades in my installation of Gentoo. The situation seems to have arisen with the upgrade of qt to 5.15.3 and nearly concurrent upgrade of kde-plasma to 5.24. I have tried emerge -auD @world, emerge -a @world, and emerge -au @world. All of them fail. I checked the listed dependencies in the error message to the results of equery depgraph for the package. These results conflict. But if a package requires >= packagex-5.15.2, does that mean that 5.15.3 is OK? Attached is a log of regular and error messages from emerge. I also tried masking qtwayland5.15.2 and unmerging it, but neither allowed the emerge to complete. What's the solution? Thanks, Bob
Bugs require emerge --info output always. The information is not complete without you providing your full emerge command. Regardless, please try with --deep and --backtrack=30 in addition. > $ grep -R dev-qt /etc/portage ^ cleanup any resulting entries inside package.mask, update any entries in package.accept_keywords or, better, take a really close look if they are actually needed. > $ grep dev-qt /var/lib/portage/world ^ cleanup any output that may yield.
Created attachment 773441 [details] output from emerge --info
Created attachment 773447 [details] output from grep -R dev-qt /etc/portage
The attached file emergelog was created with the following command: emerge -auD --backtrack=30 @world &>emergelog I attached the output of grep -R dev-qt /etc/portage. Please define what I need to clean up. Also attached output from grep dev-qt /var/lib/portage/world. What needs to be cleaned up in it? The masking of dev-qt/qtwayland-5.15.2-r17 had no effect.
This is looking like a problem in kde. the ebuilds for kde-plasma/plasma-workspace-5.24.4 and kde-plasma/layer-shell-qt-5.24.4 show dependencies on >=dev-qt/qtwayland-${QTMIN}:5= This leads to emerge insisting that there is no match for qtwayland. Can someone from the KDE team verify that '=' is necessary? Masking qtwayland-5.15.2 doesn't help. I see that the '=' is there in 5.23.5-r2 ebuild. But it may be causing a problem due to the new version of qt. Since qtwayland-5.15.2 and all other dev-qt 5.15.2 packages have been removed, it appears that we're stuck. I see that masking kde latest version is not going to help either. Best Wishes, Bob
(In reply to rjgleits from comment #5) > This is looking like a problem in kde. the ebuilds for > kde-plasma/plasma-workspace-5.24.4 and > kde-plasma/layer-shell-qt-5.24.4 show dependencies on > >=dev-qt/qtwayland-${QTMIN}:5= > This leads to emerge insisting that there is no match for qtwayland. Can > someone from the KDE team verify that '=' is necessary? Masking > qtwayland-5.15.2 doesn't help. I see that the '=' is there in 5.23.5-r2 > ebuild. But it may be causing a problem due to the new version of qt. > Yes, the = is necessary -- it means "rebuild if (sub)slot changes", not "exactly this version". Please drop all masks for Qt related software, KDE, etc. > Since qtwayland-5.15.2 and all other dev-qt 5.15.2 packages have been > removed, it appears that we're stuck. I see that masking kde latest version > is not going to help either. > Can you run: emerge -epv --backtrack=0 @world and share the output as an attachment? > Best Wishes, > > Bob
Created attachment 774564 [details] emerge -epv @world output All masks on qt and kde elements were removed.
(In reply to rjgleits from comment #7) > Created attachment 774564 [details] > emerge -epv @world output > > All masks on qt and kde elements were removed. Please emerge --deselect dev-qt/assistant:5 dev-qt/qdbusviewer:5. You also need to emerge --deselect app-shells/bashdb given it's incompatible with Bash 5.1. (You should not have things you don't personally want/use in your @world file, hence dropping Qt stuff, which your -pev @world output shows is unresolvable/gone anyway.) Then try -epv @world --backtrack=0 again (don't run it without -p), then share emerge -p -uvDU @world --backtrack=200.
Created attachment 774630 [details] output from emerge -epv --backtrack=0 @world Deselections you recommended were performed.
Created attachment 774756 [details] Output from emerge -p -uvDU @world --backtrack=200 It looks like the problem has been resolved, but I'll wait for your opinion before going ahead with the actual emerge.
(In reply to rjgleits from comment #10) > Created attachment 774756 [details] > Output from emerge -p -uvDU @world --backtrack=200 > > It looks like the problem has been resolved, but I'll wait for your opinion > before going ahead with the actual emerge. Go for launch!