Rev 4734: Minor tweak to Inventory.path2id in http://bazaar.launchpad.net/~jameinel/bzr/2.1-chk-inv-path2id

[John Arbash Meinel]

At http://bazaar.launchpad.net/~jameinel/bzr/2.1-chk-inv-path2id

------------------------------------------------------------
revno: 4734
revision-id: john at arbash-meinel.com-20091008221157-34vady0r80jpfhvk
parent: pqm at pqm.ubuntu.com-20091008172636-tygnfi5hsnn9203g
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: 2.1-chk-inv-path2id
timestamp: Thu 2009-10-08 17:11:57 -0500
message:
  Minor tweak to Inventory.path2id
  
  Name the parameter 'relpath' rather than 'name', as 'name' generally means
  just the last component of the path.
  Change the CHKInventory.path2id code to cache all of the intermediate
  values, rather than just caching the final value.
  Also change the lookup loop to use the cache if it is available.
  Note that ideally we would actually check the cache in reverse order
  first, and then build up whatever is missing afterwards...
-------------- next part --------------
=== modified file 'bzrlib/inventory.py'
--- a/bzrlib/inventory.py	2009-10-02 05:43:41 +0000
+++ b/bzrlib/inventory.py	2009-10-08 22:11:57 +0000
@@ -958,7 +958,7 @@
         descend(self.root, u'')
         return accum
 
-    def path2id(self, name):
+    def path2id(self, relpath):
         """Walk down through directories to return entry of last component.
 
         names may be either a list of path components, or a single
@@ -969,10 +969,10 @@
 
         Returns None IFF the path is not found.
         """
-        if isinstance(name, basestring):
-            name = osutils.splitpath(name)
-
-        # mutter("lookup path %r" % name)
+        if isinstance(relpath, basestring):
+            names = osutils.splitpath(relpath)
+        else:
+            names = relpath
 
         try:
             parent = self.root
@@ -981,7 +981,7 @@
             return None
         if parent is None:
             return None
-        for f in name:
+        for f in names:
             try:
                 children = getattr(parent, 'children', None)
                 if children is None:
@@ -2170,35 +2170,41 @@
             delta.append((old_path, new_path, file_id, entry))
         return delta
 
-    def path2id(self, name):
+    def path2id(self, relpath):
         """See CommonInventory.path2id()."""
         # TODO: perhaps support negative hits?
-        result = self._path_to_fileid_cache.get(name, None)
+        result = self._path_to_fileid_cache.get(relpath, None)
         if result is not None:
             return result
-        if isinstance(name, basestring):
-            names = osutils.splitpath(name)
+        if isinstance(relpath, basestring):
+            names = osutils.splitpath(relpath)
         else:
-            names = name
+            names = relpath
         current_id = self.root_id
         if current_id is None:
             return None
         parent_id_index = self.parent_id_basename_to_file_id
+        cur_path = None
         for basename in names:
-            # TODO: Cache each path we figure out in this function.
+            if cur_path is None:
+                cur_path = basename
+            else:
+                cur_path = cur_path + '/' + basename
             basename_utf8 = basename.encode('utf8')
-            key_filter = [(current_id, basename_utf8)]
-            file_id = None
-            for (parent_id, name_utf8), file_id in parent_id_index.iteritems(
-                key_filter=key_filter):
-                if parent_id != current_id or name_utf8 != basename_utf8:
-                    raise errors.BzrError("corrupt inventory lookup! "
-                        "%r %r %r %r" % (parent_id, current_id, name_utf8,
-                        basename_utf8))
+            file_id = self._path_to_fileid_cache.get(cur_path, None)
             if file_id is None:
-                return None
+                key_filter = [(current_id, basename_utf8)]
+                items = parent_id_index.iteritems(key_filter)
+                for (parent_id, name_utf8), file_id in items:
+                    if parent_id != current_id or name_utf8 != basename_utf8:
+                        raise errors.BzrError("corrupt inventory lookup! "
+                            "%r %r %r %r" % (parent_id, current_id, name_utf8,
+                            basename_utf8))
+                if file_id is None:
+                    return None
+                else:
+                    self._path_to_fileid_cache[cur_path] = file_id
             current_id = file_id
-        self._path_to_fileid_cache[name] = current_id
         return current_id
 
     def to_lines(self):