Rev 5371: For a single page test, we parse() still takes 47us in http://bazaar.launchpad.net/~jameinel/bzr/2.3-btree-chk-leaf

John Arbash Meinel john at arbash-meinel.com
Wed Aug 4 00:46:33 BST 2010


At http://bazaar.launchpad.net/~jameinel/bzr/2.3-btree-chk-leaf

------------------------------------------------------------
revno: 5371
revision-id: john at arbash-meinel.com-20100803234618-iqbgjncawrqp1wt1
parent: john at arbash-meinel.com-20100803234101-742wpbkuyhqc9tpp
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: 2.3-btree-chk-leaf
timestamp: Tue 2010-08-03 18:46:18 -0500
message:
  For a single page test, we parse() still takes 47us
  parse().all_keys() takes 66.6us if we don't intern, and 86.9us if we do.
  (10us is hash() time) We'll need to think about whether interning is 
  worthwhile. For some cases we are optimizing for, it will bloat.
  
  However, if we only do keys that we actually return, then likely we are
  ok, and ahead to intern. Because *those* keys will then get used to
  read chk maps, or have been read from them.
  
  So I'll leave intern() in for now.
-------------- next part --------------
=== modified file 'bzrlib/_btree_serializer_pyx.pyx'
--- a/bzrlib/_btree_serializer_pyx.pyx	2010-08-03 23:41:01 +0000
+++ b/bzrlib/_btree_serializer_pyx.pyx	2010-08-03 23:46:18 +0000
@@ -455,7 +455,7 @@
     key = StaticTuple_New(1)
     Py_INCREF(hexxed)
     StaticTuple_SET_ITEM(key, 0, hexxed)
-    # key = StaticTuple_Intern(key)
+    key = StaticTuple_Intern(key)
     return key
 
 



More information about the bazaar-commits mailing list