summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorGravatar Joey Hess <joeyh@joeyh.name>2016-01-01 14:26:09 -0400
committerGravatar Joey Hess <joeyh@joeyh.name>2016-01-01 14:26:09 -0400
commit797afd300cde42488a1c64201289165c48a9208e (patch)
tree60527140af5e5f5a7eb2771a1ba69e40c1281777
parent1e74cb6ce0c0fe5ea22615ad31d8fa8746bd6510 (diff)
update
-rw-r--r--doc/todo/smudge.mdwn12
1 files changed, 10 insertions, 2 deletions
diff --git a/doc/todo/smudge.mdwn b/doc/todo/smudge.mdwn
index 4ab5053ea..f7835c563 100644
--- a/doc/todo/smudge.mdwn
+++ b/doc/todo/smudge.mdwn
@@ -2,8 +2,16 @@ git-annex should use smudge/clean filters.
### implementation todo list
-* Test suite:
- - add another pass through all the tests with files being unlocked in v6 mode
+* Test suite has a currently disabled pass that tests v6 unlocked files.
+ That pass has many failures.
+* Associated files database is not populated when a repository is cloned,
+ because the smudge filters are not set up when git checks out the work
+ tree. So, git annex get etc won't work immediately after cloning.
+ Need to make init run through the whole work index and populate the
+ associated files database.
+ (Or could update it incrementally, so git-annex get foo updates the
+ database for foo's key. But, then if bar has the same content as foo, bar
+ wouldn't be populated by get foo.)
* Reconcile staged changes into the associated files database, whenever
the database is queried. This is needed to handle eg:
git add largefile