split inventory progress

[Martin Pool]

On Wed, Nov 12, 2008 at 9:08 PM, Robert Collins
<robertc at robertcollins.net> wrote:
>
>
>> so 78MB down to 2.5MB. I'm pulling across a bigger sample set now.
>
> revno 1000 in mysql-server
>
> btree:
> Commits: 4041
>                      Raw    %    Compressed    %  Objects
> Revisions:      10995 KiB   0%      3773 KiB   8%     4041
> Inventories:  5204462 KiB  57%     10861 KiB  23%     4041
> Texts:        3861735 KiB  42%     32452 KiB  68%    21951
> Signatures:         0 KiB   0%         0 KiB   0%        0
> Total:        9077193 KiB 100%     47087 KiB 100%    30033
>
> split inventory:
> Commits: 4041
>                      Raw    %    Compressed    %  Objects
> Revisions:      11152 KiB   0%      3878 KiB   5%     4041
> Inventories:    62388 KiB   1%     34374 KiB  48%    73628
> Texts:        3861735 KiB  98%     32452 KiB  45%    21951
> Signatures:         0 KiB   0%         0 KiB   0%        0
> Total:        3935277 KiB 100%     70704 KiB 100%    99620
>
> 5.2GB raw data down to 62MB. Woo!

Indeed.

I see the compressed size has actually grown from 10MB to 34.  I would
presume that's because at the moment the only delta compression we do
is in changed/unchanged nodes, and so the smaller similarities between
them are not getting compressed out.

-- 
Martin <http://launchpad.net/~mbp/>