Issue Details (XML | Word | Printable)

Key: CONTRB-33
Type: Bug Bug
Status: Resolved Resolved
Resolution: Fixed
Priority: Major Major
Assignee: Shannon Krebs
Reporter: James Mortimer
Votes: 0
Watchers: 1
Operations

If you were logged in you would be able to see more operations.
Confluence Contributors Plugin

duplicate redundant 'watches' are counted in the summary

Created: 22/Jan/07 11:54 AM   Updated: 25/Jan/07 08:44 AM
Component/s: Contributors Macro, Contributors Summary Macro
Affects Version/s: 0.8.3
Fix Version/s: 1.1

Time Tracking:
Not Specified

Labels:


 Description  « Hide
Currently, confluence has a bug that allows, in some cases, a user to have multiple watches for a given page. behaviour is that they recieve notification until they remove all the watches. If done from a page, it removes them all I beleive, but if from the watch manager, it does it one at a time.

You can simulate this by 'manually' watching a page more than once. Right-click and copy the URL from a 'watch page' icon, and then load it into your address bar a few times. Unless it's been fixed, that should give you multiple watches to a single page for a single user.

This plugin will currently count ALL watches, even if they are not from a unique user. So if James has 3 duplicated watches and Shannon has 1, it will count '4' watches. However, it will only list 'unique' watchers (James is not listed 3 times).

Perhaps the 'watches' query is missing a well-placed 'distinct' clause.

Thanks,
James.



 All   Comments   Work Log   Change History   FishEye   Crucible   Builds      Sort Order: Ascending order - Click to sort in descending order
Shannon Krebs added a comment - 22/Jan/07 02:20 PM
Is there a Jira issue for the confluence bug ?

James Mortimer added a comment - 22/Jan/07 02:22 PM
not sure. it's a minor bug and doesn't impede any confluence functions.

James Mortimer added a comment - 22/Jan/07 02:27 PM

Shannon Krebs added a comment - 25/Jan/07 03:33 AM
I tried, but can't recreate this in my 2.3 environment. But i'm only running a simple standalone install for development so it could be something that only occurs in a 'real' environment?

As part of the fix for CONTRB-32 I've added a check that the notification must be unique before the count is updated which should hopefully prevent the count being incorrect.


James Mortimer added a comment - 25/Jan/07 08:44 AM
Thanks Shannon.

I'll keep a look out for this in the next version then.

We're using an LDAP name server and Oracle database, so that might have something to do with the discrepency.