From 8b892901a98421b9b70c131bcc7af3a5e2ce62d9 Mon Sep 17 00:00:00 2001 From: Joey Hess Date: Thu, 17 Nov 2011 18:20:06 -0400 Subject: analysis; not a bug but a feature --- doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn') diff --git a/doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn b/doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn index 322c354a2..9d1bbc4e0 100644 --- a/doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn +++ b/doc/bugs/old_data_isn__39__t_unused_after_migration.mdwn @@ -45,3 +45,11 @@ Output: unused . (checking for unused data...) (checking master...) ok The two files are hardlinked, so it's not taking up extra space, but it would be nice to be able to remove the old keys. + +> `git annex unused` checks the content of all branches, and assumes that, +> when a branch contains a file that points to a key, that key is still +> used. In this case, the migration has staged a change to the file, +> but it is not yet committed, so when it checks the master branch, it +> still finds a file referring to the old key. +> +> So, slightly surprising, but not a bug. --[[Joey]] [[done]] -- cgit v1.2.3