> On Friday, December 20, 2013 12:17:57 Vishesh Handa wrote: > > 1. Given a url, return a unique identifier > > 2. Given a unique identifier, return the url > > > > inodes satisfy (1), but not (2). > > inodes don't satisfy even (1). They're only unique per filesystem, > designing > Baloo around "computers with only one filesystem" seems very > limiting, well, > broken.
That's true if you're using one index for every filesystem, and I suggested to have one index per filesystem, so (1) could be satisfied ;) As for (2), I thought it would be easy to get the file path from the inode number but it seems it isn't at all. My lack of knowledge. Oh, by the way, I was still trying to figure out how things would work (Xapian + SQLite, etc.) and discovered Recoll (http://www.recoll.org/) which seemes pretty similar to what you are trying to do with Baloo. There is a english description here : http://www.lesbonscomptes.com/recoll/usermanual/RCL.INTRODUCTION.RECOLL.html - Maybe it could be smart to get in touch with the dev ? Cheers, -- François >> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
