Rev 3888: Change the _GCIndex._get_entries code to not build a set that we don't use. in http://bzr.arbash-meinel.com/branches/bzr/brisbane/hack3

[John Arbash Meinel]

At http://bzr.arbash-meinel.com/branches/bzr/brisbane/hack3

------------------------------------------------------------
revno: 3888
revision-id: john at arbash-meinel.com-20090323184002-yn1gmjy96ezys08c
parent: john at arbash-meinel.com-20090323173537-opqyhyax7bv2w2j1
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: hack3
timestamp: Mon 2009-03-23 13:40:02 -0500
message:
  Change the _GCIndex._get_entries code to not build a set that we don't use.
-------------- next part --------------
=== modified file 'bzrlib/groupcompress.py'
--- a/bzrlib/groupcompress.py	2009-03-23 17:35:37 +0000
+++ b/bzrlib/groupcompress.py	2009-03-23 18:40:02 +0000
@@ -556,17 +556,16 @@
         compressor = GroupCompressor()
         tstart = time.time()
         old_length = self._block._content_length
-        cur_endpoint = 0
+        end_point = 0
         for factory in self._factories:
             bytes = factory.get_bytes_as('fulltext')
-            (found_sha1, end_point, type,
+            (found_sha1, start_point, end_point, type,
              length) = compressor.compress(factory.key, bytes, factory.sha1)
             # Now update this factory with the new offsets, etc
             factory.sha1 = found_sha1
-            factory._start = cur_endpoint
+            factory._start = start_point
             factory._end = end_point
-            cur_endpoint = end_point
-        self._last_byte = cur_endpoint
+        self._last_byte = end_point
         new_block = compressor.flush()
         # TODO: Should we check that new_block really *is* smaller than the old
         #       block? It seems hard to come up with a method that it would
@@ -1703,23 +1702,25 @@
 
         :param keys: An iterable of index key tuples.
         """
-        # TODO: don't do set(keys) or build up found_keys unless we actually
-        #       use them.
-        keys = set(keys)
-        found_keys = set()
-        if self._parents:
-            for node in self._graph_index.iter_entries(keys):
-                yield node
-                found_keys.add(node[1])
-        else:
-            # adapt parentless index to the rest of the code.
-            for node in self._graph_index.iter_entries(keys):
-                yield node[0], node[1], node[2], ()
-                found_keys.add(node[1])
         if check_present:
-            missing_keys = keys.difference(found_keys)
-            if missing_keys:
-                raise RevisionNotPresent(missing_keys.pop(), self)
+            missing_keys = set(keys)
+            if self._parents:
+                for node in self._graph_index.iter_entries(keys):
+                    yield node
+                    missing_keys.discard(node[1])
+            else:
+                # adapt parentless index to the rest of the code.
+                for node in self._graph_index.iter_entries(keys):
+                    yield node[0], node[1], node[2], ()
+                    missing_keys.discard(node[1])
+        else:
+            if self._parents:
+                for node in self._graph_index.iter_entries(keys):
+                    yield node
+            else:
+                # adapt parentless index to the rest of the code.
+                for node in self._graph_index.iter_entries(keys):
+                    yield node[0], node[1], node[2], ()
 
     def get_parent_map(self, keys):
         """Get a map of the parents of keys.