Rev 4488: Removing compression_parent Content objs drops us from 238MB down to 146MB. in http://bazaar.launchpad.net/~jameinel/bzr/1.17-rework-annotate

[John Arbash Meinel]

At http://bazaar.launchpad.net/~jameinel/bzr/1.17-rework-annotate

------------------------------------------------------------
revno: 4488
revision-id: john at arbash-meinel.com-20090623185734-bum18jak2qv7mfsg
parent: john at arbash-meinel.com-20090623185110-pjo35ee505ywpvd3
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: 1.17-rework-annotate
timestamp: Tue 2009-06-23 13:57:34 -0500
message:
  Removing compression_parent Content objs drops us from 238MB down to 146MB.
  
  Setting copy_base_content = False seems to cause errors ATM :(. So we need to dig deeper.
-------------- next part --------------
=== modified file 'bzrlib/knit.py'
--- a/bzrlib/knit.py	2009-06-23 18:51:10 +0000
+++ b/bzrlib/knit.py	2009-06-23 18:57:34 +0000
@@ -3413,11 +3413,11 @@
             if num == 0:
                 base_content = self._content_objects.pop(compression_parent)
                 self._num_compression_children.pop(compression_parent)
-                copy_base_content = True
+                copy_base_content = False
             else:
                 self._num_compression_children[compression_parent] = num
                 base_content = self._content_objects[compression_parent]
-                copy_base_content = False
+                copy_base_content = True
             content, _ = self._vf._factory.parse_record(
                 key, record, record_details, base_content,
                 copy_base_content=True)
@@ -3425,6 +3425,8 @@
             # Fulltext record
             content, _ = self._vf._factory.parse_record(
                 key, record, record_details, None)
+        # TODO: Only track the content when there are compression children.
+        #       Otherwise we only need the lines
         self._content_objects[key] = content
         lines = content.text()
         self._text_cache[key] = lines