This is my current TODO list. The db4o branch is taking much longer than I had initially hoped, but I'm reasonably confident we can get something great out of it (in terms of resource usage, local performance). This is something the users want so it is important for 0.8 - perhaps not as important as FOAF and other things that may deliver massive network level performance gains, but it should get us more users and allow us to keep more users.
1. Catch up with code review process. 2. Implement regular testing code. 3. Sort out seednodes etc. 4. Db4o: Implement Bloom filters optimisation, eliminate pendingKeys code. Should eliminate the big, long-lived spike in CPU usage, disk, memory, after starting a big splitfile request (when it reaches the metadata). 5. Sort out FOAF. Option 4, 6 or 7 should make it sufficiently secure combined with current changes (an improvement in many cases); we also need the testing code, but then we can enable it. There is some routing impact but only on the first hop, so this shouldn't be a problem. 6. Db4o: Test, debug, optimise inserts. 7. Db4o: Test, debug, optimise freesite inserts. 8. Db4o: Sort out build process/freenet-ext.jar. If necessary upgrade to db4o 7.2. 9. Db4o: Garbage collection (structures not deleted from database when request deleted, no way to cleanup temp buckets), any other issues remaining. 10. Db4o: Public (easy) testing, stabilisation, and merge. Fortunately other people are handling some of the other 0.8 priorities e.g. saces is working on plugin updating over Freenet, and of course sdiz's ongoing work on the new store. Nextgens has written the FOAF code and may be able to complete it, but only after I have done the testing code. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080731/28be7d15/attachment.pgp>
