Some linked bug reports are private duplicates

Bug #1086754 reported by Matthew Paul Thomas
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Apport
Triaged
Medium
Martin Pitt
Errors
New
Undecided
Unassigned

Bug Description

Right now, the #1 error listed on errors.ubuntu.com -- /usr/bin/update-manager:libqt4-test:_inline_callbacks:commit:_inline_callbacks:_run_in_dialog for update-manager -- is linked to bug 979278.

Bug 979278 is a private duplicate of public bug 975320. The problem is that unless you are in the the ubuntu-bugcontrol team, you can't follow the chain through the private report to find the public report.

To fix this, a cron job should take the existing bug in the duplicates database and update to the one it is a duplicate of.

Also, errors.ubuntu.com should tag any bug reports that it synthetically creates.

The cron job should check this tag and, if it is present, prefer a report created by crashdigger over one created by errors.ubuntu.com.

Fixing this is also necessary to tell users whether there is a fix for a crash with a particular signature.

description: updated
Revision history for this message
Martin Pitt (pitti) wrote :

We need to revive the old concept of "crash db consolidation" and run that daily, in addition to updating the bug status when we run into the next master bug. We should implement this in a more clever way than iterating through all existing master bugs, such as querying for all "apport-crash" and "daisy-created" tagged bugs which changed since the last run (modified_since in the searchTasks argument).

Changed in apport:
status: New → Triaged
assignee: nobody → Martin Pitt (pitti)
importance: Undecided → Medium
Revision history for this message
Brian Murray (brian-murray) wrote :

I previously reported the tagging issue as https://bugs.launchpad.net/errors/+bug/1049972.

Revision history for this message
Matthew Paul Thomas (mpt) wrote :

Martin, are you sure this is a duplicate of bug 910981? This bug is about errors.ubuntu.com, whereas that bug is about Apport's pre-release "the bug report displayed in the web browser" behavior.

Revision history for this message
Martin Pitt (pitti) wrote :

Yes, it comes down to the same root cause: the duplicate db that the Launchpad retracers export have master IDs which are not true "master" bugs, but itself duplicates of the real master bug.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.