4 0.7 development release
7 inbetween (0.8 release?) and more:
8 - try to find a way to perform special actions on a cache miss (like a sanity check
10 - Have an option in chkindex which let it try if there is an inode at that point
11 - if files will be saved to the same disk, make sure you reserve their space on the disk's
12 bitmap first (in case someone run chkbfs) [you currently can't recover to the same
13 partition (and it makes sense until you can select the files to recover)]
14 - have an option that recovers only certain files (like those with a specific
15 path, or file extension)
16 - try to make the Cache class easier to use -> like the one in BFS!
17 - incorporate changes from the BFS BPlusTree (BPlusTree::Find(), TreeIterator), duplicates, etc.
18 - test the B+tree code with hand-corrupted trees
19 - try to unify the BFS BPlusTree code to only maintain *one* file
20 - recreate indices on target disk (all) -> chkindex should get that functionality
25 - the name index scanner could use its own tree iterator functions, which directly
26 visit every single node and use some heuristics to determine the node type
27 - better cache buffer management in the Disk class? Reading from an image is much
28 faster than reading from a real disk currently...
29 - reduce memory overhead: use just the block_run references instead of real Inodes
30 in the hashtables (not in the missing table, of course)
31 - sort hashtables before copying after block_run