Created attachment 56932 [details] A modified mergeresources.py that exposes this behavior. Version: 4.6 (using KDE 4.6.0) OS: Linux I have been playing around with the resource merging in playground: http://websvn.kde.org/trunk/playground/base/nepomuk-kde/toolbox/mergeresources.py Using the following query one finds hundreds of duplicates: select distinct ?people count(?r) where { ?r a nco:PersonContact . ?r nco:fullname ?people . } I tried to merge these and I noticed a huge improve in queries afterwards. Reproducible: Always Steps to Reproduce: Install and make use of the nepomuk email, calendar and contact feeders. Actual Results: A duplicate resource is created in nepomuk every time the feeder starts. Expected Results: Each task, contact, email, etc. should only be fed to nepomuk once. Feeders should properly check for this. I am posting a modified mergeresources.py that exposes this behavior.
The same for audio files. I think that the problem is related to a bad serialization, a bad transaction or both problems at the same time.
Although the situation has incredibly improved with the new nepomuk feeder, I am still getting some duplicates, especially with my own e-mail. I am also seeing "MERGING FAILED!" messages during indexing.
Created attachment 64212 [details] Screenshot Krunner showing multiplicated contact Is this related? I guess so.
not sure if e91113b8e0956b8ac8ae7614a2b1faa7dfbed0f0 in kdebase-runtime fixes this. Adding vhanda
After some time running the new nepomukfeeder I have no more nco:PersonContact duplicates. I still have duplicate tags, especially the "Birthday" tag is still not being merged properly.
No more duplicates in kde 4.8. Thanks