GitCmdStreamReader: fixed terrible bug which only kicked in if the stream was actually empty. This is a rare case that can happen during stream testing. Theoretically there shouldn't be any empty streams of course, but practically they do exist sometimes ;); fixed stream.seek implementation, which previously used seek on standard output

Improved GitCmd error handling

Author: Byron

lib/git/cmd.py

@@ -109,6 +109,12 @@ def __init__(self, size, stream):
 			self._size = size
 			self._nbr = 0			# num bytes read
 
+			# special case: if the object is empty, has null bytes, get the 
+			# final newline right away.
+			if size == 0:
+				stream.read(1)
+			# END handle empty streams
+			
 		def read(self, size=-1):
 			bytes_left = self._size - self._nbr
 			if bytes_left == 0:
@@ -127,7 +133,6 @@ def read(self, size=-1):
 			if self._size - self._nbr == 0:
 				self._stream.read(1)	# final newline
 			# END finish reading
-			
 			return data
 
 		def readline(self, size=-1):
@@ -146,7 +151,6 @@ def readline(self, size=-1):
 			self._nbr += len(data)
 
 			# handle final byte
-			# we inline everything, it must be fast !
 			if self._size - self._nbr == 0:
 				self._stream.read(1)
 			# END finish reading
@@ -185,8 +189,9 @@ def next(self):
 		def __del__(self):
 			bytes_left = self._size - self._nbr
 			if bytes_left:
-				# seek and discard
-				self._stream.seek(bytes_left + 1, os.SEEK_CUR)	# includes terminating newline
+				# read and discard - seeking is impossible within a stream
+				# includes terminating newline
+				self._stream.read(bytes_left + 1)
 			# END handle incomplete read
 
 
@@ -454,7 +459,13 @@ def _parse_object_header(self, header_line):
 		"""
 		tokens = header_line.split()
 		if len(tokens) != 3:
-			raise ValueError("SHA named %s could not be resolved, git returned: %r" % (tokens[0], header_line.strip()) )
+			if not tokens:
+				raise ValueError("SHA could not be resolved, git returned: %r" % (header_line.strip()))
+			else:
+				raise ValueError("SHA %s could not be resolved, git returned: %r" % (tokens[0], header_line.strip()))
+			# END handle actual return value
+		# END error handling
+		
 		if len(tokens[0]) != 40:
 			raise ValueError("Failed to parse header: %r" % header_line) 
 		return (tokens[0], tokens[1], int(tokens[2]))

lib/git/objects/fun.py

@@ -1,6 +1,7 @@
 """Module with functions which are supposed to be as fast as possible"""
 
-__all__ = ('tree_to_stream', 'tree_entries_from_data')
+__all__ = ('tree_to_stream', 'tree_entries_from_data', 'traverse_trees_recursive',
+			'traverse_tree_recursive')
 
 from stat import S_ISDIR
 

lib/git/repo.py

@@ -42,6 +42,7 @@ def is_git_dir(d):
 				os.readlink(headref).startswith('refs'))
 	return False
 
+
 class Repo(object):
 	"""
 	Represents a git repository and allows you to query references, 

test/testlib/helper.py

@@ -109,11 +109,14 @@ def repo_creator(self):
 					return func(self, rw_repo)
 				except:
 					print >> sys.stderr, "Keeping repo after failure: %s" % repo_dir
+					repo_dir = None
 					raise
 			finally:
 				os.chdir(prev_cwd)
 				rw_repo.git.clear_cache()
-				shutil.rmtree(repo_dir, onerror=_rmtree_onerror)
+				if repo_dir is not None:
+					shutil.rmtree(repo_dir, onerror=_rmtree_onerror)
+				# END rm test repo if possible
 			# END cleanup
 		# END rw repo creator
 		repo_creator.__name__ = func.__name__