Removes the just implemented improvments.

TODO

@@ -5,39 +5,6 @@ This is the list of TODO items for SnapRAID.
 
 - Next
 
-* Improves the fix -e behaviour. At now it reads too mutch data,
-as it needs to verify the full files.
-This is done just to have confirmation that the file is OK,
-and then restore the file timestamp. But other approaches are possible.
-Indicatively we can do:
-- Read/Write only the marked bad error blocks
-- For all the files, owning this block:
-  - If the file is present and it has the original timestamp
-    - If yes, Write any fix to it, and keep the timestamp.
-    - If no, don't write the fix, and assume it as update.
-The issue could be that if the error is in a modified file
-after sync, it won't be fixed. But at now we detect only silent error,
-and they CANNOT be in modified files.
-So, if a file is modified, we just don't write it, at the next scrub
-the bad flag will be removed.
-
-Note that this way of checking the timestamp before writing fixes
-may have sense also for a generic fix.
-We already have the -m, --missing flag. It's already working in this way.
---missing fixes only deleted/missing files.
-Like --stable fixes only synced files, but unsure of the use case.
-
-We have now:
-- Lost disk. fix -d DISK
-- Deleted files. fix -m / fix -f
-- Silent errors on scrub. fix -e
-Possible new:
-- Disk with failed read, that can be rewritten.
-  - Scrub could put them in error list, and fix could rewrite the sectors to make them right
-    This won't work for modified files, in that case the error will remain.
-    But scrub can report it in the screen and log about the file.
-  - Sync could proceed anyway, just skip such blocks index, and mark the sector as bad.
-
 * Avoid to do optimized move operation if more files with
 same timestamp/size are present.