Hi Thomas,

the question was how the DataStore can store changes of different file 
versions. And that's not possible cause it doesn't know the previous version to 
determine the changes.

Kindly regards, Robert

-----Ursprüngliche Nachricht-----
Von: Thomas Mueller [mailto:[email protected]] 
Gesendet: Mittwoch, 13. Juli 2011 10:57
An: [email protected]
Betreff: Re: AW: Incremental/deduplicating versioning

Hi,

>The DataStore can't do such a thing, cause it knows nothing about
>versions - just take a look at the DataStore interface.

The DataStore doesn't necessarily need to know about the versions to be
able to better compress the data. For example, if a data store
implementation splits up the binaries into smaller blocks, it could re-use
blocks. Or the implementation could detect where to best split (using a
rolling checksum similar to rsync). In my view, there is no need to change
the data store API to better compress the data.

>If you need such a thing, it has to be implemented on a higher level -
>within your application

That's also an option.

Regards,
Thomas

Reply via email to