Merge branch 'index' into improvements

* index:
  Removed index test marker for custom commits as this boils down to a good way to add files to the index/remove them and make commits which are possibly customized with custom parents
  index writing now creates a sha on the content making it possible to write valid indices after manually removing or altering entriesgst
  Implemented merge/resolve handling , but realized that index writing is not yet working properly as it is sha1 checked as well. This explains what my 20 byte 'extension_data' actually is ;)
  index: Added write_tree method including test
  index.iter_blobs method added including tests ( which have been improved generally for more coverage )
  tree.traverse: Added prune functionality - previously the predciate did both, pruning and preventing to return items
  Improved testing of index against trees, tests succeed with next commit
  Initial version of merge including tests for one-way, two-way and tree-way merge
  index writing added including simple test, improved docs of IndexEntry
  improved IndexEntry type and added test for parsing of the stage
  Improved tuple access of EntryIndex class including test, stage and type access still needs to be decoded though
  Added initial version of the index reading from file - IndexEntry interface is to be improved though, writing needs to be implemented as well
  Added frame for index implementation and testing

1 files changed, 13 insertions, 11 deletions

diff --git a/TODO b/TODO
index f1b49a0c..afda659a 100644
--- a/TODO
+++ b/TODO
@@ -26,7 +26,10 @@ Object
   It would be good to improve things there as cat-file keeps all the data in a buffer
   before it writes it. Hence it does not write to a stream directly, which can be 
   bad if files are large, say 1GB :).
-
+* Effectively Objects only store hexsha's in their id attributes, so in fact 
+  it should be renamed to 'sha'. There was a time when references where allowed as 
+  well, but now objects will be 'baked' to the actual sha to assure comparisons work.
+  
 Config
 ------
 * Expand .get* methods of GitConfigParser to support default value. If it is not None, 
@@ -46,12 +49,13 @@ Overhaul docs - check examples, check looks, improve existing docs
 
 Index
 -----
-* Index class required for special handling ? Probably considering what I want 
-  to do ! Dulwich can already write the index, and read trees, although it 
-  could be improved as well and could possibly be made faster unless we want 
-  to use the c modules ( not for now )
-* Index Merge ( merge two trees into the index to quickly see conflicts ). 
-  Its possible to write it into a separate index file that can be read separately.
+* write_tree should write a tree directly, which would require ability to create
+  objects in the first place. Should be rather simple as it is 
+  "tree" bytes datablock | sha1sum and zipped.
+  Currently we use some file swapping and the git command to do it which probably 
+  is much slower. The thing is that properly writing a tree from an index involves
+  creating several tree objects, so in the end it might be slower. 
+  Hmm, probably its okay to use the command unless we go c(++)
 
 Remote
 ------
@@ -74,8 +78,6 @@ Tree
 * Should return submodules during iteration ( identifies as commit )
 * Work through test and check for test-case cleanup and completeness ( what about
   testing whether it raises on invalid input ? ). See 6dc7799d44e1e5b9b77fd19b47309df69ec01a99
-* Derive from Iterable, simple pipe it through to Commit objects and iterate using 
-  commit.tree.
   
 Testing
 -------
@@ -84,5 +86,5 @@ Testing
   as forking a shared repo in a tmp directory. In that moment, we probably want to 
   facility committing and checkouts as well.
   - Use these tests for git-remote as we need to test push
-
-
+  - Also assure that the test-case setup is a bit more consistent ( Derive from TestCase, possibly
+    make repo a class member instead of an instance member