[Http-crcsync] crccache ready for some testing I think
Rusty Russell
rusty at rustcorp.com.au
Sun Apr 5 09:22:34 EDT 2009
On Sunday 05 April 2009 07:33:06 Alex Wulms wrote:
> Hi Rusty,
>
> Many thanks!!!
>
> I'll adapt crccache code this sunday to support the improved algorithm and
> submit all to the git repo.
>
> Hi Toby,
>
> I'll change the crccache-client to send the original file size in stead of
> the blocksize in the header. In that way, the crccache-server can calculate
> the blocksize and the tailsize by itself. It is more efficient then using two
> headers (one with blocksize, one with tailsize). I'll also adapt the client
> to put the crc of all 21 blocks (the 20 full blocks + the 1 tail block) in
> the header. I must still figure out what to do in case that the tail has a
> zero size (e.g. if filesize is exact multiple of 20) but I think that while
> writing the code, the answer will crystalize by itself.
I think Toby's draft, which suggested just leaving the trailing block as
a literal, is an interesting possibility. Worst case, it's 19 bytes.
I also think we should go to 60-bit csums (but needs new CRC algo).
Finally, I think the strong hash sent in request in Toby's draft is not useful
as the block signature + size can be used for caching in exactly the same way.
Cheers,
Rusty.
More information about the Http-crcsync
mailing list