another idea...

This commit is contained in:
anarcat 2015-06-22 22:06:30 +00:00 committed by admin
parent 293fc1b00a
commit 8410a1cccb

View file

@ -0,0 +1,39 @@
so i know about the various discussions around a `du` that follows git-annex symlinks (e.g. [[forum/__34__du__34___equivalent_on_an_annex__63__/]] or [[forum/gadu_-_git-annex_disk_usage/]]. this is not about that. :)
i believe there's a distinct use-case for a simpler `du` subcommand that will calculate the disk space used locally (in case of the default `--in here`) but could also use the tracking log to determine the space usage in *other* locations. this would be especially useful on remotes that we don't have shell access to, s3 comes to mind. i thought that `git annex info` could output that, but it seems it doesn't:
<pre>
$ git annex info . --in s3 --exclude='video/original/*'
directory: .
local annex keys: 0
local annex size: 0 bytes
annexed files in working tree: 0
size of annexed files in working tree: 0 bytes
numcopies stats:
repositories containing these files:
$ git annex info --in s3 --exclude='video/original/*'
repository mode: indirect
trusted repositories: 0
semitrusted repositories: 6
00000000-0000-0000-0000-000000000002 -- bittorrent
2d61a8de-a24e-44e3-9aa0-54f033fec1e9 -- host-mp20120507-1.mp.isuma.tv
9401d7b3-44d2-48ab-a9f1-c77fac469a1a -- s3
c510ddad-24cd-4353-b5f4-03581f6f9dca -- cs.isuma.tv [here]
d2a7d4ff-1dbf-4bfa-bb97-ae593626daf6 -- sneakernet
e747d5c8-ea47-480f-8c5d-2986ce65ed89 -- isuma.tv
untrusted repositories: 2
00000000-0000-0000-0000-000000000001 -- web
36d2cb94-e0a2-446a-87c9-02f73135b302 -- anarcat@desktop008:~/src/isuma/isuma-files
transfers in progress: none
available local disk space: 67.66 gigabytes (+1 megabyte reserved)
local annex keys: 0
local annex size: 0 bytes
annexed files in working tree: 0
size of annexed files in working tree: 0 bytes
bloom filter size: 16 mebibytes (0% full)
backend usage:
</pre>
you can see the advantage of having a separate du command above: first off, it would give directly the information we're looking for. it could also work on remote repositories, use the powerful pattern matching, and so on...
thanks! [[anarcat]]