Rev 104: update TODO a little bit. in http://bazaar.launchpad.net/%7Ebzr/bzr-groupcompress/rabin

[John Arbash Meinel]

At http://bazaar.launchpad.net/%7Ebzr/bzr-groupcompress/rabin

------------------------------------------------------------
revno: 104
revision-id: john at arbash-meinel.com-20090304042506-zaf29b1u9jnajp2u
parent: john at arbash-meinel.com-20090303225027-dd26kj3xasgfi7bv
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: rabin
timestamp: Tue 2009-03-03 22:25:06 -0600
message:
  update TODO a little bit.
-------------- next part --------------
=== modified file 'TODO'
--- a/TODO	2009-01-19 05:46:53 +0000
+++ b/TODO	2009-03-04 04:25:06 +0000
@@ -1,6 +1,16 @@
 Todo for groupcompress:
 
  * layers - gc reader/writer and vf layers should be separate
- * mpdiff usage
+ * alternate copy/insert encoding?
  * other stuff from design
- * use byte offsets not line offsets.
+
+ * Look at splitting the 'labels' section into a header/index for all
+   compressed entries, separate from the actual content. This could potentially
+   improve compression efficiency.
+
+ * Consider a more 'binary' layout for things like sha1 sums, or length
+   indicators.
+
+ * Investigate sha1 overhead. Currently when doing 'bzr pack' we sha once on
+   extraction, and then one more time on insertion. Removing both calls saves
+   6s out of 32s for 'bzr pack' (and 8s=>5.5s for 'repository-details')