repodetails, and split inventory work..

[Robert Collins]

I've written a plugin to extract details for a repository; it depends on
my repository branch (patches to remove this dependency will be
accepted; I just won't be doing that myself).

I'll be uploading it shortly to
bazaar.launchpad.net/~lifeless/+junk/bzr-repodetails.

Sample output:
(my bzr repository, which has managed to getopenoffice and a few other
things into it :P)

:!./bzr repository-details ..
Commits: 21429
                     Raw    %  Compressed    %  Objects
Revisions:     11585 KiB   0%    8373 KiB   2%    21429
Inventories: 3074406 KiB  72%   43144 KiB  12%    21429
Texts:       1146601 KiB  27%  284512 KiB  83%   158564
Signatures:     3350 KiB   0%    3156 KiB   0%     8892
Total:       4235944 KiB 100%  339186 KiB 100%   210314

A mysql server - the first 200 commits in chk-inventory format:
:!../../../repository/bzr repository-details ../../../test-repos/dev3/
Commits: 365
                    Raw    %  Compressed    %  Objects
Revisions:     2433 KiB   0%     646 KiB   1%      365
Inventories: 103628 KiB  24%   52683 KiB  86%     5198
Texts:       311822 KiB  74%    7748 KiB  12%     4103
Signatures:       0 KiB   0%       0 KiB   0%        0
Total:       417883 KiB 100%   61078 KiB 100%     9666

The same content, but in btree repository format:
Commits: 365
                    Raw    %  Compressed    %  Objects
Revisions:     2418 KiB   0%     637 KiB   6%      365
Inventories: 274346 KiB  46%     988 KiB  10%      365
Texts:       311822 KiB  52%    7748 KiB  82%     4103
Signatures:       0 KiB   0%       0 KiB   0%        0
Total:       588587 KiB 100%    9373 KiB 100%     4833

Interesting things to note:
in full text terms, the CHK format is smaller (417MB vs 588MB) but in
compressed terms, the BTree format using xml inventories is smaller
(61MB vs 9MB).

There are 5200 objects to represent all the inventories and file ids in
the CHK inventory - up from 365 in xml form. Over 50% of the object
count is these objects.

Cheers, and enjoy,
-Rob

-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20081017/5fd655fb/attachment.pgp