Merge branch 'johnsca-sf-master' into 0.3

Merge all fixes from https://github.com/gitpython-developers/GitPython/pull/134

10 files changed, 62 insertions, 27 deletions

diff --git a/git/cmd.py b/git/cmd.py
index 5425d8ff..a1780de7 100644
--- a/git/cmd.py
+++ b/git/cmd.py
@@ -88,6 +88,7 @@ class Git(LazyMixin):
             # try to kill it
             try:
                 os.kill(self.proc.pid, 2)   # interrupt signal
+                self.proc.wait()    # ensure process goes away
             except OSError:
                 pass # ignore error when process already died
             except AttributeError:
diff --git a/git/objects/commit.py b/git/objects/commit.py
index 4380f472..bc437e8b 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -8,6 +8,7 @@ from git.util import        (
                             Actor,
                             Iterable,
                             Stats,
+                            finalize_process
                         )
 from git.diff import Diffable
 from tree import Tree
@@ -65,7 +66,6 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
                  message=None,  parents=None, encoding=None, gpgsig=None):
         """Instantiate a new Commit. All keyword arguments taking None as default will
         be implicitly set on first query.
-
         :param binsha: 20 byte sha1
         :param parents: tuple( Commit, ... ) 
             is a tuple of commit ids or actual Commits
@@ -252,6 +252,10 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
             assert len(hexsha) == 40, "Invalid line: %s" % hexsha
             yield Commit(repo, hex_to_bin(hexsha))
         # END for each line in stream
+        # TODO: Review this - it seems process handling got a bit out of control
+        # due to many developers trying to fix the open file handles issue
+        if hasattr(proc_or_stream, 'wait'):
+            finalize_process(proc_or_stream)
         
         
     @classmethod
@@ -430,14 +434,21 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         
         self.author, self.authored_date, self.author_tz_offset = parse_actor_and_date(next_line)
         self.committer, self.committed_date, self.committer_tz_offset = parse_actor_and_date(readline())
-        
+
+        # we might run into one or more mergetag blocks, skip those for now
+        next_line = readline()
+        while next_line.startswith('mergetag '):
+            next_line = readline()
+            while next_line.startswith(' '):
+                next_line = readline()
         
         # now we can have the encoding line, or an empty line followed by the optional
         # message.
         self.encoding = self.default_encoding
 
         # read headers
-        buf = readline().strip()
+        enc = next_line
+        buf = enc.strip()
         while buf != "":
             if buf[0:10] == "encoding ":
                 self.encoding = buf[buf.find(' ')+1:]
diff --git a/git/objects/fun.py b/git/objects/fun.py
index f73be542..66b7998e 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -70,9 +70,13 @@ def tree_entries_from_data(data):
         # default encoding for strings in git is utf8
         # Only use the respective unicode object if the byte stream was encoded
         name = data[ns:i]
-        name_enc = name.decode("utf-8")
-        if len(name) > len(name_enc):
-            name = name_enc
+        try:
+            name_enc = name.decode("utf-8")
+        except UnicodeDecodeError:
+            pass
+        else:
+            if len(name) > len(name_enc):
+                name = name_enc
         # END handle encoding
         
         # byte is NULL, get next 20
@@ -84,6 +88,7 @@ def tree_entries_from_data(data):
     return out
     
     
+
 def _find_by_name(tree_data, name, is_dir, start_at):
     """return data entry matching the given name and tree mode
     or None.
diff --git a/git/objects/util.py b/git/objects/util.py
index 2e44c9c0..cdf72bed 100644
--- a/git/objects/util.py
+++ b/git/objects/util.py
@@ -167,6 +167,7 @@ def parse_date(string_date):
     
 # precompiled regex
 _re_actor_epoch = re.compile(r'^.+? (.*) (\d+) ([+-]\d+).*$')
+_re_only_actor = re.compile(r'^.+? (.*)$')
 
 def parse_actor_and_date(line):
     """Parse out the actor (author or committer) info from a line like::
@@ -174,10 +175,14 @@ def parse_actor_and_date(line):
         author Tom Preston-Werner <tom@mojombo.com> 1191999972 -0700
     
     :return: [Actor, int_seconds_since_epoch, int_timezone_offset]"""
+    actor, epoch, offset = '', 0, 0
     m = _re_actor_epoch.search(line)
-    actor, epoch, offset = m.groups()
+    if m:
+        actor, epoch, offset = m.groups()
+    else:
+        m = _re_only_actor.search(line)
+        actor = m.group(1) if m else line or ''
     return (Actor._from_string(actor), int(epoch), utctz_to_altz(offset))
-    
 
 #} END functions
 
diff --git a/git/remote.py b/git/remote.py
index b06c0686..75e88e43 100644
--- a/git/remote.py
+++ b/git/remote.py
@@ -24,7 +24,10 @@ from refs import (
                     TagReference
                 )
 
-from git.util import join_path
+from git.util import (
+		join_path,
+		finalize_process
+	)
 from gitdb.util import join
 
 import re
@@ -58,18 +61,6 @@ def digest_process_messages(fh, progress):
     # END while file is not done reading
     return dropped_lines
 
-def finalize_process(proc):
-    """Wait for the process (clone, fetch, pull or push) and handle its errors accordingly"""
-    try:
-        proc.wait()
-    except GitCommandError,e:
-        # if a push has rejected items, the command has non-zero return status
-        # a return status of 128 indicates a connection error - reraise the previous one
-        if proc.poll() == 128:
-            raise
-        pass
-    # END exception handling
-
 def add_progress(kwargs, git, progress):
     """Add the --progress flag to the given kwargs dict if supported by the 
     git command. If the actual progress in the given progress instance is not 
diff --git a/git/repo/base.py b/git/repo/base.py
index 933c8c82..6a311f0b 100644
--- a/git/repo/base.py
+++ b/git/repo/base.py
@@ -6,7 +6,10 @@
 
 from git.exc import InvalidGitRepositoryError, NoSuchPathError
 from git.cmd import Git
-from git.util import Actor
+from git.util import (
+		Actor,
+		finalize_process
+	)
 from git.refs import *
 from git.index import IndexFile
 from git.objects import *
@@ -14,7 +17,6 @@ from git.config import GitConfigParser
 from git.remote import (
                         Remote,
                         digest_process_messages,
-                        finalize_process,
                         add_progress
                     )
 
@@ -540,6 +542,7 @@ class Repo(object):
             if filename[0] == filename[-1] == '"':
                 filename = filename[1:-1].decode('string_escape')
             untracked_files.append(filename)
+        finalize_process(proc)
         return untracked_files
 
     @property
diff --git a/git/test/lib/helper.py b/git/test/lib/helper.py
index 2634ecdb..a76f1a15 100644
--- a/git/test/lib/helper.py
+++ b/git/test/lib/helper.py
@@ -227,7 +227,7 @@ class TestBase(TestCase):
     """
     
     @classmethod
-    def setUp(cls):
+    def setUpClass(cls):
         """
         Dynamically add a read-only repository to our actual type. This way 
         each test type has its own repository
diff --git a/git/test/test_fun.py b/git/test/test_fun.py
index 530988ef..096cd368 100644
--- a/git/test/test_fun.py
+++ b/git/test/test_fun.py
@@ -2,7 +2,8 @@ from git.test.lib import *
 from git.objects.fun import (
                                 traverse_tree_recursive,
                                 traverse_trees_recursive, 
-                                tree_to_stream
+                                tree_to_stream,
+                                tree_entries_from_data
                             )
 
 from git.index.fun import (
@@ -249,3 +250,7 @@ class TestFun(TestBase):
             entries = traverse_tree_recursive(odb, commit.tree.binsha, '')
             assert entries
         # END for each commit
+
+    def test_tree_entries_from_data(self):
+        r = tree_entries_from_data(b'100644 \x9f\0aaa')
+        assert r == [('aaa', 33188, '\x9f')], r
diff --git a/git/test/test_git.py b/git/test/test_git.py
index cdea1d3e..7132aa83 100644
--- a/git/test/test_git.py
+++ b/git/test/test_git.py
@@ -18,8 +18,8 @@ from git import (   Git,
 class TestGit(TestBase):
 
     @classmethod
-    def setUp(cls):
-        super(TestGit, cls).setUp()
+    def setUpClass(cls):
+        super(TestGit, cls).setUpClass()
         cls.git = Git(cls.rorepo.working_dir)
 
     @patch.object(Git, 'execute')
diff --git a/git/util.py b/git/util.py
index 30ccfa66..955d23de 100644
--- a/git/util.py
+++ b/git/util.py
@@ -15,6 +15,8 @@ import getpass
 
 # NOTE:  Some of the unused imports might be used/imported by others.
 # Handle once test-cases are back up and running.
+from exc import GitCommandError
+
 from gitdb.util import (
     make_sha, 
     LockedFD, 
@@ -117,6 +119,18 @@ def get_user_id():
     """:return: string identifying the currently active system user as name@node"""
     return "%s@%s" % (getpass.getuser(), platform.node())
 
+def finalize_process(proc):
+	"""Wait for the process (clone, fetch, pull or push) and handle its errors accordingly"""
+	try:
+		proc.wait()
+	except GitCommandError,e:
+		# if a push has rejected items, the command has non-zero return status
+		# a return status of 128 indicates a connection error - reraise the previous one
+		if proc.poll() == 128:
+			raise
+		pass
+	# END exception handling
+
 #} END utilities
 
 #{ Classes