0.0.8 has been released

Database and synchronization dedupe has been implemented.

From this version, files are split into chunks and store with these hash.
and all entries represented as a list of a chunk.

To simplify, if chunk length is 5 chars,
If you write long text:
"supercalifragilisticexpialidocious "

will be stored as:
"super calif ragil istic expia lidoc ious " -> "ha hb hc hd he hf"

and you modify tail some letters:
"supercalifragilisticexpialidocious!"

will be stored as:
"super calif ragil istic expia lidoc ious!" -> "ha hb hc hd he hg"

And now, we only spent two (entry and hash:hg) new database items.
(Actually, a delete mark is synchronized too. In synchronization we have to transfer three items.)