smithjd added a comment.
In D11529#233550 <https://phabricator.kde.org/D11529#233550>, @michaelh wrote: > Under the premise that I still have to learn about baloo's inner workings, here are some concerns: > > - I'm not convinced, that index cleaning should be part of the dbus interface. Why should cleaning be done out of process? Do we really want any application to trigger database manipulation? Could you elaborate your rationale please? IndexCleaner fits the current implementation of checking for file changes quite well. > - Which application is using baloo dbus interface anyway and which functions? @mgallien How is elisa using it? > - Purging stale entries is by far not enough to get the index in a good shape. As soon as removable drives or network shares are involved all sorts of weird things can happen. IndexCleaner is much too simple to account for that. IndexCleaner is simple and can handle 99.9% of file deindexing requirements, and is now safe for use in combination with unmounted indexed volumes. IndexCleaner is a vital MIA part of Baloo. An additional purger could conceivably be useful in certain cicumstances, but given that the index is regeneratable data, I don't know that it's even warranted. This could be a good use for baloodb, leaving the generic automatic part to Baloo. > - Looks like indexcleaner has been dead code until now. Do we need to be concerned about Vishesh's commit message? https://cgit.kde.org/baloo.git/commit/?id=ea2afe88b0c4299d7540e5b6c8b8e46858336f0c As above, this already neatly fits into the workflow. REPOSITORY R293 Baloo REVISION DETAIL https://phabricator.kde.org/D11529 To: smithjd, #baloo, vhanda, michaelh Cc: mgallien, #frameworks, ashaposhnikov, michaelh, astippich, spoorun, nicolasfella, ngraham, alexeymin
