bugs get unreasonably locked in fixreleased

Bug #887875 reported by Martin Pool
38
This bug affects 6 people
Affects Status Importance Assigned to Milestone
Launchpad itself
Triaged
Low
Unassigned

Bug Description

In bug 664096 Launchpad changed so that fixreleased bugtasks can only be reopened by the original reporter or by the pillar's bug supervisor. The stated reason is that some bugs that are either contentious or have broad symptoms like "black screen on boot" cause noise by being reopened, which probably is a real problem.

However, this solution causes trouble because many bugs are in fact not contentious, and this means that affected users (who aren't the original reporter) or developers (who aren't the bug supervisor) are unable to reopen them, even if the bug was incorrectly closed.

For instance https://bugs.launchpad.net/bzr/+bug/485601

There are secondary problems:
 * the control looks editable but all the options are greyed out
 * there's no indication why it's not editable

This doesn't seem like a very complete solution to noise edits anyhow.

Martin Pool (mbp)
Changed in launchpad:
status: New → Triaged
importance: Undecided → Low
Revision history for this message
Timothy Arceri (t-fridey) wrote :

I agree totally this is a very poor solution. All that is going to happen is people will be forced to created new bug reports meaning all the previous comments patch etc will be lost, this will create much worse noise then the reopening of a couple of bugs here and there.

I to are trying to reopen a bug: https://bugs.launchpad.net/hundredpapercuts/+bug/429041 that was closed by a maintainer without being tested. What is the point of a community based bug system if you are limiting the abilities of the community?

Why not give the package maintainer the ability to mark the problem bugs that keep getting reopened as "only reopen-able by maintainer/reporter" rather then implementing a blanket ban on reopening any bug.

Revision history for this message
Martin Pool (mbp) wrote : Re: [Bug 887875] Re: bugs get unreasonably locked in fixreleased

I think it could work well to just have a generic "lock status" that
stops it being changed by anyone but bug supervisors. That would
handle this case for contentious bugs, and also others that keep
getting flipped from invalid, wontfix, or other statuses.

Revision history for this message
Sebastien Bacher (seb128) wrote :

The locking is rather a good thing imho, we used to often get random users reopening bugs closed for years because they "get the same issue", while the issue could seem similar to users in 95% of the cases it's not the same issue and even if it is, it's better to have a clean recent bug with a mention to a closed one than to add "noise" to a bug which already got forth and back comments, not to mention that the "old bug" submitter might not want to be spammed about other users issues

Revision history for this message
Bryce Harrington (bryce) wrote :

Agreed with Sebastien that locking is a good thing. The types of bugs that get improperly re-opened tend to have lots of comments and activity, so as a developer when you hit a bug that's been incorrectly reopened it takes several minutes of study before you realize what's going on - time that would have been better spent on a real bug. I agree with seb that for many reasons even if it is the same bug, it's often better to have a clean new report to work from in these cases.

I don't think we should have another knob to have to turn to lock a fixed bug. In Ubuntu a lot of bugs get set to Fix Released automatically via the changelog, and so that would need to be made to always lock (thus risking unreasonable locks), or always unlock (leaving us back at the original problem of incorrect re-opens).

A few other ideas for solving this:

  * A 'Nominate to reopen' function analogous to series nominations. Perhaps re-open once a nomination is seconded by someone else

  * Per-project setting of which states are 'auto-locked', so bzr can have it one way and ubuntu another

  * Allow bugs to be reopened but denote them in some fashion so it's more evident to triagers/developers it's a re-open

  * Have the user file a NEW bug, but establish a link / relationship to the closed bug they think this is a regression of

Revision history for this message
Sebastien Bacher (seb128) wrote :

it's especially annoying when users "hijack" a closed bug and reopen it and it turns out after a few comments that their issue is not the same one, it leads to spamming of an old bug, of its submitter and to confusion for everybody reading the bug, we are much better set encouraging users to file new bugs

Revision history for this message
Martin Pool (mbp) wrote :

I do find noise edits really annoying too, and they still happen on
other bug statuses.

Perhaps locking is not the answer.

Perhaps we can do something better to make it obvious to less-expert
users what is the right way to express their intention. For instance
adding a clear "here's the information you asked for" or "i'm still
having this bug" link, along the lines of Answers.

Revision history for this message
Richard Hansen (rhansen) wrote :

How is one supposed to reopen a bug if a new user inadvertently changes the status to 'Fix Released'? (Or if a malicious user purposely changes the status.) See bug #798414 for example.

If you need special permissions to change the status from 'Fix Released', it seems to me that you should need those same special permissions to change the status to 'Fix Released' in the first place.

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

Other bug subscribers

Related questions

Remote bug watches

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