Summary: | portage-2.2.0_alpha74: Display of blockers broken | ||
---|---|---|---|
Product: | Portage Development | Reporter: | Dennis Schridde <dschridde+gentoobugs> |
Component: | Core - Dependencies | Assignee: | Portage team <dev-portage> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | esigra |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 385391 | ||
Bug Blocks: | 300071 |
Description
Dennis Schridde
2011-11-13 09:23:48 UTC
(In reply to comment #0) > --- > !!! The following update(s) have been skipped due to unsatisfied dependencies > !!! triggered by backtracking: > > sys-devel/clang:0 > --- P.S: It appears that this list is incomplete: gawk-4.0.0 was also obviously skipped. (In reply to comment #0) > The 2 blockers in the package list suggest that the slot conflict (which no > doubt is real, but can be resolved by skipping the upgrade, s.b.) is not the > actual reason for the aborted merge. Actually, you have multiple reasons for the aborted merge. Either the blockers or the slot conflicts would be enough to make it abort. It seems to me that emerge is behaving as designed, given the conflicts that you have. Can you suggest some way to improve it? Note that the libreoffice blockers are fixed if you sync (see bug #390271). The slot conflicts might resolve automatically if you use a higher emerge backtrack setting, such as `--backtrack=30`. (In reply to comment #2) > (In reply to comment #0) > > The 2 blockers in the package list suggest that the slot conflict (which no > > doubt is real, but can be resolved by skipping the upgrade, s.b.) is not the > > actual reason for the aborted merge. > > Actually, you have multiple reasons for the aborted merge. Either the blockers > or the slot conflicts would be enough to make it abort. The slot conflicts alone (I p.masked >=libreoffice-l10n-3.4.4, which caused the block) do not cause an abort. They both will be resolved automatically by skipping the update. So the reason for the aborted merge seems to be the blockers, only. However, portage suggests that the reason would be the slot conflicts, which appears to be wrong. My suggestion: Print the blockers as the actual reason for the abort, not the slot conflicts. It's related to this change, from bug #385391: http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=77b651300731ec007cd535a83b8ee9a898602783 (In reply to comment #3) > My suggestion: Print the blockers as the actual reason for the abort, not the > slot conflicts. The thing is, it doesn't know that the blockers are the only reason, because the blockers make it quit before it tries to find a solution for the slot conflicts. (In reply to comment #5) > (In reply to comment #3) > > My suggestion: Print the blockers as the actual reason for the abort, not the > > slot conflicts. > > The thing is, it doesn't know that the blockers are the only reason, because > the blockers make it quit before it tries to find a solution for the slot > conflicts. Could you apply the "slot conflict display" to blockers, too? And just display both? And possibly add the (short!) hint that slot conflicts can in some cases be resolved automatically by skipping upgrades ("... see manpage for more info")? Or just resolve everything, if possible, before bailing out. Or does that take too long (runtime speed)? I think we can consider this fixed since 2.2.0_alpha180, by the following commit: https://github.com/gentoo/portage/commit/f4058826bd2928fb9f12b5cac1d225d79e58c993 |