Feature #8073

Automated cleaning of identical issues

Added by Florian Westreicher over 11 years ago. Updated over 11 years ago.

Status:NewStart date:2011-04-05
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:Issues
Target version:-
Resolution:

Description

Our software automatically reports bugs to our Redmine install via email.
Since we have about 15 thousand different installations, pretty much every bug is reported more than once. It would be very desirable to have a feature that allows you to automatically detect identical issues (field, description etc) and remove all but one from the system.

It should be configurable which fields you want to check for duplicates (for example, select only the subject, tracker and reporter). The GUI could work just like the
current GUI for issue browsing.

It should be possible to have the job run automatically or by admin interaction. If user interaction is required, a preview of which issues will be removed would of course be necessary.

History

#1 Updated by Etienne Massip over 11 years ago

  • Category set to Issues

#2 Updated by Etienne Massip over 11 years ago

Someone came recently with the idea to have an UI help when creating a new issue ; some list of issues with similar subjects would dynamically be prompted to the user as he's typing the new issue's subject so he can know if he's actually creating a duplicate.

Alas, can't find the issue.

#3 Updated by Florian Westreicher over 11 years ago

yea but we do not create them via the web gui, we get them automagically by email.

i was just closing about 700 duplicates by hand, your guess how boring that is ;)

it would also be nice to not just close the issues but to set them as duplicate from the "first" reported bug.

#4 Updated by Etienne Massip over 11 years ago

You could add a custom field with the id of the error (should be smartly computed) and filter on this field to get all of them.

#5 Updated by Florian Westreicher over 11 years ago

that would require us to have an id ;) but i guess we could hash the text or something like that.
sounds like a plan for now, would still love to see that feature tho.

Also available in: Atom PDF