Rev 416: A discovery about the inventory cache. in http://bazaar.launchpad.net/~jameinel/loggerhead/history_db
John Arbash Meinel
john at arbash-meinel.com
Wed Apr 14 21:13:15 BST 2010
At http://bazaar.launchpad.net/~jameinel/loggerhead/history_db
------------------------------------------------------------
revno: 416
revision-id: john at arbash-meinel.com-20100414201256-aa8ey6e3xlmygu4e
parent: john at arbash-meinel.com-20100414194330-hlil7qe0opsow8r2
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: history_db
timestamp: Wed 2010-04-14 15:12:56 -0500
message:
A discovery about the inventory cache.
-------------- next part --------------
=== modified file 'loggerhead/history.py'
--- a/loggerhead/history.py 2010-04-14 19:43:30 +0000
+++ b/loggerhead/history.py 2010-04-14 20:12:56 +0000
@@ -291,6 +291,8 @@
self._branch = branch
# XXX: _inventory_cache isn't even an LRU cache... seems like a clear
# potential for memory bloat
+ # However, a given History object is only valid during a single
+ # HTTP request, so maybe that is reasonable for now...
self._inventory_cache = {}
self._branch_nick = self._branch.get_config().get_nickname()
self.log = logging.getLogger('loggerhead.%s' % (self._branch_nick,))
More information about the bazaar-commits
mailing list