From f6aa8d116eb33293c0a9d6d600eb7c32832758b9 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Sun, 4 Jan 2015 19:14:33 +0100
Subject: initial set of adjustments to make (most) imports work.

More to come, especially when it's about strings
---
 git/objects/__init__.py       |  7 ++++---
 git/objects/base.py           |  2 +-
 git/objects/blob.py           |  3 +--
 git/objects/commit.py         | 15 +++++++--------
 git/objects/submodule/base.py |  6 +++---
 git/objects/submodule/root.py |  7 +++++--
 git/objects/submodule/util.py |  2 +-
 git/objects/tag.py            |  6 +++---
 git/objects/tree.py           | 16 +++++++---------
 9 files changed, 32 insertions(+), 32 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/__init__.py b/git/objects/__init__.py
index 70fc52cb..ee642876 100644
--- a/git/objects/__init__.py
+++ b/git/objects/__init__.py
@@ -7,9 +7,10 @@ import inspect
 from .base import *
 # Fix import dependency - add IndexObject to the util module, so that it can be
 # imported by the submodule.base
-from .submodule import util
-util.IndexObject = IndexObject
-util.Object = Object
+from .submodule import util as smutil
+smutil.IndexObject = IndexObject
+smutil.Object = Object
+del(smutil)
 from .submodule.base import *
 from .submodule.root import *
 
diff --git a/git/objects/base.py b/git/objects/base.py
index 20147e57..1f0d5752 100644
--- a/git/objects/base.py
+++ b/git/objects/base.py
@@ -3,8 +3,8 @@
 #
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
+from .util import get_object_type_by_name
 from git.util import LazyMixin, join_path_native, stream_copy
-from util import get_object_type_by_name
 from gitdb.util import (
     bin_to_hex,
     basename
diff --git a/git/objects/blob.py b/git/objects/blob.py
index b05e5b84..322f6992 100644
--- a/git/objects/blob.py
+++ b/git/objects/blob.py
@@ -3,9 +3,8 @@
 #
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
-
 from mimetypes import guess_type
-import base
+from . import base
 
 __all__ = ('Blob', )
 
diff --git a/git/objects/commit.py b/git/objects/commit.py
index 9c733695..5b6b9a33 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -4,6 +4,8 @@
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
 
+from gitdb import IStream
+from gitdb.util import hex_to_bin
 from git.util import (
     Actor,
     Iterable,
@@ -11,26 +13,23 @@ from git.util import (
     finalize_process
 )
 from git.diff import Diffable
-from tree import Tree
-from gitdb import IStream
-from cStringIO import StringIO
 
-import base
-from gitdb.util import (
-    hex_to_bin
-)
-from util import (
+from .tree import Tree
+from . import base
+from .util import (
     Traversable,
     Serializable,
     parse_date,
     altz_to_utctz_str,
     parse_actor_and_date
 )
+
 from time import (
     time,
     altzone
 )
 import os
+from io import StringIO
 import logging
 
 log = logging.getLogger('git.objects.commit')
diff --git a/git/objects/submodule/base.py b/git/objects/submodule/base.py
index d6f8982b..5ccebd4c 100644
--- a/git/objects/submodule/base.py
+++ b/git/objects/submodule/base.py
@@ -1,5 +1,5 @@
-import util
-from util import (
+from . import util
+from .util import (
     mkhead,
     sm_name,
     sm_section,
@@ -8,7 +8,7 @@ from util import (
     find_first_remote_branch
 )
 from git.objects.util import Traversable
-from StringIO import StringIO                   # need a dict to set bloody .name field
+from io import StringIO                   # need a dict to set bloody .name field
 from git.util import (
     Iterable,
     join_path_native,
diff --git a/git/objects/submodule/root.py b/git/objects/submodule/root.py
index 708749c7..8c9afff1 100644
--- a/git/objects/submodule/root.py
+++ b/git/objects/submodule/root.py
@@ -1,5 +1,8 @@
-from base import Submodule, UpdateProgress
-from util import (
+from .base import (
+    Submodule,
+    UpdateProgress
+)
+from .util import (
     find_first_remote_branch
 )
 from git.exc import InvalidGitRepositoryError
diff --git a/git/objects/submodule/util.py b/git/objects/submodule/util.py
index 01bd03b3..cb84ccb1 100644
--- a/git/objects/submodule/util.py
+++ b/git/objects/submodule/util.py
@@ -1,7 +1,7 @@
 import git
 from git.exc import InvalidGitRepositoryError
 from git.config import GitConfigParser
-from StringIO import StringIO
+from io import StringIO
 import weakref
 
 __all__ = ('sm_section', 'sm_name', 'mkhead', 'unbare_repo', 'find_first_remote_branch',
diff --git a/git/objects/tag.py b/git/objects/tag.py
index 3c379579..5e76e230 100644
--- a/git/objects/tag.py
+++ b/git/objects/tag.py
@@ -4,12 +4,12 @@
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
 """ Module containing all object based types. """
-import base
-from gitdb.util import hex_to_bin
-from util import (
+from . import base
+from .util import (
     get_object_type_by_name,
     parse_actor_and_date
 )
+from gitdb.util import hex_to_bin
 
 __all__ = ("TagObject", )
 
diff --git a/git/objects/tree.py b/git/objects/tree.py
index c77e6056..a216322b 100644
--- a/git/objects/tree.py
+++ b/git/objects/tree.py
@@ -3,22 +3,20 @@
 #
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
-import util
-from base import IndexObject
 from git.util import join_path
-from blob import Blob
-from submodule.base import Submodule
 import git.diff as diff
+from gitdb.util import to_bin_sha
 
-from fun import (
+from . import util
+from .base import IndexObject
+from .blob import Blob
+from .submodule.base import Submodule
+
+from .fun import (
     tree_entries_from_data,
     tree_to_stream
 )
 
-from gitdb.util import (
-    to_bin_sha,
-)
-
 __all__ = ("TreeModifier", "Tree")
 
 
-- 
cgit v1.2.3


From ae2ff0f9d704dc776a1934f72a339da206a9fff4 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Sun, 4 Jan 2015 19:50:28 +0100
Subject: Dum brute force conversion of all types.

However, StringIO really is ByteIO in most cases, and py2.7 should
run but doesn't.
This should be made work first.
---
 git/objects/commit.py         | 7 ++++---
 git/objects/fun.py            | 6 +++++-
 git/objects/submodule/base.py | 3 ++-
 git/objects/tree.py           | 3 ++-
 4 files changed, 13 insertions(+), 6 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/commit.py b/git/objects/commit.py
index 5b6b9a33..c9d7ddc8 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -23,6 +23,7 @@ from .util import (
     altz_to_utctz_str,
     parse_actor_and_date
 )
+from git.compat import text_type
 
 from time import (
     time,
@@ -378,7 +379,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
 
         a = self.author
         aname = a.name
-        if isinstance(aname, unicode):
+        if isinstance(aname, text_type):
             aname = aname.encode(self.encoding)
         # END handle unicode in name
 
@@ -390,7 +391,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
 
         # encode committer
         aname = c.name
-        if isinstance(aname, unicode):
+        if isinstance(aname, text_type):
             aname = aname.encode(self.encoding)
         # END handle unicode in name
         write(fmt % ("committer", aname, c.email,
@@ -408,7 +409,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         write("\n")
 
         # write plain bytes, be sure its encoded according to our encoding
-        if isinstance(self.message, unicode):
+        if isinstance(self.message, text_type):
             write(self.message.encode(self.encoding))
         else:
             write(self.message)
diff --git a/git/objects/fun.py b/git/objects/fun.py
index 416a52e6..db2ec7c2 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -1,5 +1,9 @@
 """Module with functions which are supposed to be as fast as possible"""
 from stat import S_ISDIR
+from git.compat import (
+    xrange,
+    text_type
+)
 
 __all__ = ('tree_to_stream', 'tree_entries_from_data', 'traverse_trees_recursive',
            'traverse_tree_recursive')
@@ -28,7 +32,7 @@ def tree_to_stream(entries, write):
         # hence we must convert to an utf8 string for it to work properly.
         # According to my tests, this is exactly what git does, that is it just
         # takes the input literally, which appears to be utf8 on linux.
-        if isinstance(name, unicode):
+        if isinstance(name, text_type):
             name = name.encode("utf8")
         write("%s %s\0%s" % (mode_str, name, binsha))
     # END for each item
diff --git a/git/objects/submodule/base.py b/git/objects/submodule/base.py
index 5ccebd4c..69bf748a 100644
--- a/git/objects/submodule/base.py
+++ b/git/objects/submodule/base.py
@@ -22,6 +22,7 @@ from git.exc import (
     InvalidGitRepositoryError,
     NoSuchPathError
 )
+from git.compat import string_types
 
 import stat
 import git
@@ -93,7 +94,7 @@ class Submodule(util.IndexObject, Iterable, Traversable):
         if url is not None:
             self._url = url
         if branch_path is not None:
-            assert isinstance(branch_path, basestring)
+            assert isinstance(branch_path, string_types)
             self._branch_path = branch_path
         if name is not None:
             self._name = name
diff --git a/git/objects/tree.py b/git/objects/tree.py
index a216322b..6776a15e 100644
--- a/git/objects/tree.py
+++ b/git/objects/tree.py
@@ -11,6 +11,7 @@ from . import util
 from .base import IndexObject
 from .blob import Blob
 from .submodule.base import Submodule
+from git.compat import string_types
 
 from .fun import (
     tree_entries_from_data,
@@ -232,7 +233,7 @@ class Tree(IndexObject, diff.Diffable, util.Traversable, util.Serializable):
             info = self._cache[item]
             return self._map_id_to_type[info[1] >> 12](self.repo, info[0], info[1], join_path(self.path, info[2]))
 
-        if isinstance(item, basestring):
+        if isinstance(item, string_types):
             # compatability
             return self.__div__(item)
         # END index is basestring
-- 
cgit v1.2.3


From bc8c91200a7fb2140aadd283c66b5ab82f9ad61e Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 10:09:51 +0100
Subject: Fixed io types to make tests work on PY2 once again.

Now it's about going through PY3 issues
---
 git/objects/commit.py         | 6 +++---
 git/objects/submodule/base.py | 6 +++---
 git/objects/submodule/util.py | 4 ++--
 3 files changed, 8 insertions(+), 8 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/commit.py b/git/objects/commit.py
index c9d7ddc8..79d460ad 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -30,7 +30,7 @@ from time import (
     altzone
 )
 import os
-from io import StringIO
+from io import BytesIO
 import logging
 
 log = logging.getLogger('git.objects.commit')
@@ -133,7 +133,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         if attr in Commit.__slots__:
             # read the data in a chunk, its faster - then provide a file wrapper
             binsha, typename, self.size, stream = self.repo.odb.stream(self.binsha)
-            self._deserialize(StringIO(stream.read()))
+            self._deserialize(BytesIO(stream.read()))
         else:
             super(Commit, self)._set_cache_(attr)
         # END handle attrs
@@ -345,7 +345,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
                          committer, committer_time, committer_offset,
                          message, parent_commits, conf_encoding)
 
-        stream = StringIO()
+        stream = BytesIO()
         new_commit._serialize(stream)
         streamlen = stream.tell()
         stream.seek(0)
diff --git a/git/objects/submodule/base.py b/git/objects/submodule/base.py
index 69bf748a..0fb3f35f 100644
--- a/git/objects/submodule/base.py
+++ b/git/objects/submodule/base.py
@@ -8,7 +8,7 @@ from .util import (
     find_first_remote_branch
 )
 from git.objects.util import Traversable
-from io import StringIO                   # need a dict to set bloody .name field
+from io import BytesIO                   # need a dict to set bloody .name field
 from git.util import (
     Iterable,
     join_path_native,
@@ -187,8 +187,8 @@ class Submodule(util.IndexObject, Iterable, Traversable):
 
     @classmethod
     def _sio_modules(cls, parent_commit):
-        """:return: Configuration file as StringIO - we only access it through the respective blob's data"""
-        sio = StringIO(parent_commit.tree[cls.k_modules_file].data_stream.read())
+        """:return: Configuration file as BytesIO - we only access it through the respective blob's data"""
+        sio = BytesIO(parent_commit.tree[cls.k_modules_file].data_stream.read())
         sio.name = cls.k_modules_file
         return sio
 
diff --git a/git/objects/submodule/util.py b/git/objects/submodule/util.py
index cb84ccb1..5604dec7 100644
--- a/git/objects/submodule/util.py
+++ b/git/objects/submodule/util.py
@@ -1,7 +1,7 @@
 import git
 from git.exc import InvalidGitRepositoryError
 from git.config import GitConfigParser
-from io import StringIO
+from io import BytesIO
 import weakref
 
 __all__ = ('sm_section', 'sm_name', 'mkhead', 'unbare_repo', 'find_first_remote_branch',
@@ -83,7 +83,7 @@ class SubmoduleConfigParser(GitConfigParser):
         """Flush changes in our configuration file to the index"""
         assert self._smref is not None
         # should always have a file here
-        assert not isinstance(self._file_or_files, StringIO)
+        assert not isinstance(self._file_or_files, BytesIO)
 
         sm = self._smref()
         if sm is not None:
-- 
cgit v1.2.3


From 04357d0d46fee938a618b64daed1716606e05ca5 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 15:53:46 +0100
Subject: Intermediate commit: test_config and test_actor works

Kind of tackling the tasks step by step, picking low-hanging fruit first,
or the ones that everyone depends on
---
 git/objects/util.py | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/util.py b/git/objects/util.py
index fdf9622b..cefef862 100644
--- a/git/objects/util.py
+++ b/git/objects/util.py
@@ -46,17 +46,17 @@ def get_object_type_by_name(object_type_name):
     :param object_type_name: Member of TYPES
 
     :raise ValueError: In case object_type_name is unknown"""
-    if object_type_name == "commit":
-        import commit
+    if object_type_name == b"commit":
+        from . import commit
         return commit.Commit
-    elif object_type_name == "tag":
-        import tag
+    elif object_type_name == b"tag":
+        from . import tag
         return tag.TagObject
-    elif object_type_name == "blob":
-        import blob
+    elif object_type_name == b"blob":
+        from . import blob
         return blob.Blob
-    elif object_type_name == "tree":
-        import tree
+    elif object_type_name == b"tree":
+        from . import tree
         return tree.Tree
     else:
         raise ValueError("Cannot handle unknown object type: %s" % object_type_name)
-- 
cgit v1.2.3


From 3d0556a31916a709e9da3eafb92fc6b8bf69896c Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 16:10:50 +0100
Subject: Added test of #147 to verify it works.

Applied a few more fixes to commit implementation, possibly not the last
---
 git/objects/commit.py | 25 +++++++++----------------
 1 file changed, 9 insertions(+), 16 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/commit.py b/git/objects/commit.py
index 79d460ad..5ad7902b 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -373,40 +373,33 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
 
     def _serialize(self, stream):
         write = stream.write
-        write("tree %s\n" % self.tree)
+        write(("tree %s\n" % self.tree).encode('ascii'))
         for p in self.parents:
-            write("parent %s\n" % p)
+            write(("parent %s\n" % p).encode('ascii'))
 
         a = self.author
         aname = a.name
-        if isinstance(aname, text_type):
-            aname = aname.encode(self.encoding)
-        # END handle unicode in name
-
         c = self.committer
         fmt = "%s %s <%s> %s %s\n"
-        write(fmt % ("author", aname, a.email,
+        write((fmt % ("author", aname, a.email,
                      self.authored_date,
-                     altz_to_utctz_str(self.author_tz_offset)))
+                     altz_to_utctz_str(self.author_tz_offset))).encode(self.encoding))
 
         # encode committer
         aname = c.name
-        if isinstance(aname, text_type):
-            aname = aname.encode(self.encoding)
-        # END handle unicode in name
-        write(fmt % ("committer", aname, c.email,
+        write((fmt % ("committer", aname, c.email,
                      self.committed_date,
-                     altz_to_utctz_str(self.committer_tz_offset)))
+                     altz_to_utctz_str(self.committer_tz_offset))).encode(self.encoding))
 
         if self.encoding != self.default_encoding:
-            write("encoding %s\n" % self.encoding)
+            write(("encoding %s\n" % self.encoding).encode('ascii'))
 
         if self.gpgsig:
             write("gpgsig")
             for sigline in self.gpgsig.rstrip("\n").split("\n"):
-                write(" " + sigline + "\n")
+                write((" " + sigline + "\n").encode('ascii'))
 
-        write("\n")
+        write(b"\n")
 
         # write plain bytes, be sure its encoded according to our encoding
         if isinstance(self.message, text_type):
-- 
cgit v1.2.3


From 8a308613467a1510f8dac514624abae4e10c0779 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 16:44:54 +0100
Subject: Fixes test_blob and improved commit writing/reading

---
 git/objects/commit.py | 44 ++++++++++++++++++++++++--------------------
 git/objects/fun.py    | 26 +++++++++++++-------------
 2 files changed, 37 insertions(+), 33 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/commit.py b/git/objects/commit.py
index 5ad7902b..f8b5c969 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -419,23 +419,25 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         next_line = None
         while True:
             parent_line = readline()
-            if not parent_line.startswith('parent'):
+            if not parent_line.startswith(b'parent'):
                 next_line = parent_line
                 break
             # END abort reading parents
-            self.parents.append(type(self)(self.repo, hex_to_bin(parent_line.split()[-1])))
+            self.parents.append(type(self)(self.repo, hex_to_bin(parent_line.split()[-1].decode('ascii'))))
         # END for each parent line
         self.parents = tuple(self.parents)
 
-        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 don't know actual author encoding before we have parsed it, so keep the lines around
+        author_line = next_line
+        committer_line = readline()
 
         # we might run into one or more mergetag blocks, skip those for now
         next_line = readline()
-        while next_line.startswith('mergetag '):
+        while next_line.startswith(b'mergetag '):
             next_line = readline()
             while next_line.startswith(' '):
                 next_line = readline()
+        # end skip mergetags
 
         # now we can have the encoding line, or an empty line followed by the optional
         # message.
@@ -444,39 +446,40 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         # read headers
         enc = next_line
         buf = enc.strip()
-        while buf != "":
-            if buf[0:10] == "encoding ":
-                self.encoding = buf[buf.find(' ') + 1:]
-            elif buf[0:7] == "gpgsig ":
-                sig = buf[buf.find(' ') + 1:] + "\n"
+        while buf:
+            if buf[0:10] == b"encoding ":
+                self.encoding = buf[buf.find(' ') + 1:].decode('ascii')
+            elif buf[0:7] == b"gpgsig ":
+                sig = buf[buf.find(b' ') + 1:] + b"\n"
                 is_next_header = False
                 while True:
                     sigbuf = readline()
-                    if sigbuf == "":
+                    if not sigbuf:
                         break
-                    if sigbuf[0:1] != " ":
+                    if sigbuf[0:1] != b" ":
                         buf = sigbuf.strip()
                         is_next_header = True
                         break
                     sig += sigbuf[1:]
-                self.gpgsig = sig.rstrip("\n")
+                # end read all signature
+                self.gpgsig = sig.rstrip(b"\n").decode('ascii')
                 if is_next_header:
                     continue
             buf = readline().strip()
-
         # decode the authors name
+
         try:
-            self.author.name = self.author.name.decode(self.encoding)
+            self.author, self.authored_date, self.author_tz_offset = \
+                parse_actor_and_date(author_line.decode(self.encoding))
         except UnicodeDecodeError:
-            log.error("Failed to decode author name '%s' using encoding %s", self.author.name, self.encoding,
+            log.error("Failed to decode author line '%s' using encoding %s", author_line, self.encoding,
                       exc_info=True)
-        # END handle author's encoding
 
-        # decode committer name
         try:
-            self.committer.name = self.committer.name.decode(self.encoding)
+            self.committer, self.committed_date, self.committer_tz_offset = \
+                parse_actor_and_date(committer_line.decode(self.encoding))
         except UnicodeDecodeError:
-            log.error("Failed to decode committer name '%s' using encoding %s", self.committer.name, self.encoding,
+            log.error("Failed to decode committer line '%s' using encoding %s", committer_line, self.encoding,
                       exc_info=True)
         # END handle author's encoding
 
@@ -488,6 +491,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         except UnicodeDecodeError:
             log.error("Failed to decode message '%s' using encoding %s", self.message, self.encoding, exc_info=True)
         # END exception handling
+
         return self
 
     #} END serializable implementation
diff --git a/git/objects/fun.py b/git/objects/fun.py
index db2ec7c2..f92a4c06 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -1,6 +1,9 @@
 """Module with functions which are supposed to be as fast as possible"""
 from stat import S_ISDIR
 from git.compat import (
+    byte_ord,
+    force_bytes,
+    defenc,
     xrange,
     text_type
 )
@@ -17,13 +20,13 @@ def tree_to_stream(entries, write):
     bit_mask = 7            # 3 bits set
 
     for binsha, mode, name in entries:
-        mode_str = ''
+        mode_str = b''
         for i in xrange(6):
             mode_str = chr(((mode >> (i * 3)) & bit_mask) + ord_zero) + mode_str
         # END for each 8 octal value
 
         # git slices away the first octal if its zero
-        if mode_str[0] == '0':
+        if byte_ord(mode_str[0]) == ord_zero:
             mode_str = mode_str[1:]
         # END save a byte
 
@@ -33,16 +36,16 @@ def tree_to_stream(entries, write):
         # According to my tests, this is exactly what git does, that is it just
         # takes the input literally, which appears to be utf8 on linux.
         if isinstance(name, text_type):
-            name = name.encode("utf8")
-        write("%s %s\0%s" % (mode_str, name, binsha))
+            name = name.encode(defenc)
+        write(b''.join(mode_str, b' ', name, b'\0', binsha))
     # END for each item
 
-
 def tree_entries_from_data(data):
     """Reads the binary representation of a tree and returns tuples of Tree items
-    :param data: data block with tree data
+    :param data: data block with tree data (as bytes)
     :return: list(tuple(binsha, mode, tree_relative_path), ...)"""
     ord_zero = ord('0')
+    space_ord = ord(' ')
     len_data = len(data)
     i = 0
     out = list()
@@ -52,10 +55,10 @@ def tree_entries_from_data(data):
         # read mode
         # Some git versions truncate the leading 0, some don't
         # The type will be extracted from the mode later
-        while data[i] != ' ':
+        while byte_ord(data[i]) != space_ord:
             # move existing mode integer up one level being 3 bits
             # and add the actual ordinal value of the character
-            mode = (mode << 3) + (ord(data[i]) - ord_zero)
+            mode = (mode << 3) + (byte_ord(data[i]) - ord_zero)
             i += 1
         # END while reading mode
 
@@ -65,7 +68,7 @@ def tree_entries_from_data(data):
         # parse name, it is NULL separated
 
         ns = i
-        while data[i] != '\0':
+        while byte_ord(data[i]) != 0:
             i += 1
         # END while not reached NULL
 
@@ -73,12 +76,9 @@ def tree_entries_from_data(data):
         # Only use the respective unicode object if the byte stream was encoded
         name = data[ns:i]
         try:
-            name_enc = name.decode("utf-8")
+            name = name.decode(defenc)
         except UnicodeDecodeError:
             pass
-        else:
-            if len(name) > len(name_enc):
-                name = name_enc
         # END handle encoding
 
         # byte is NULL, get next 20
-- 
cgit v1.2.3


From e1060a2a8c90c0730c3541811df8f906dac510a7 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 17:59:22 +0100
Subject: test_commit works once again

---
 git/objects/base.py   | 7 ++++---
 git/objects/commit.py | 4 ++--
 git/objects/fun.py    | 3 +--
 3 files changed, 7 insertions(+), 7 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/base.py b/git/objects/base.py
index 1f0d5752..004e3981 100644
--- a/git/objects/base.py
+++ b/git/objects/base.py
@@ -21,7 +21,7 @@ class Object(LazyMixin):
 
     """Implements an Object which may be Blobs, Trees, Commits and Tags"""
     NULL_HEX_SHA = '0' * 40
-    NULL_BIN_SHA = '\0' * 20
+    NULL_BIN_SHA = b'\0' * 20
 
     TYPES = (dbtyp.str_blob_type, dbtyp.str_tree_type, dbtyp.str_commit_type, dbtyp.str_tag_type)
     __slots__ = ("repo", "binsha", "size")
@@ -94,7 +94,7 @@ class Object(LazyMixin):
 
     def __str__(self):
         """:return: string of our SHA1 as understood by all git commands"""
-        return bin_to_hex(self.binsha)
+        return self.hexsha
 
     def __repr__(self):
         """:return: string with pythonic representation of our object"""
@@ -103,7 +103,8 @@ class Object(LazyMixin):
     @property
     def hexsha(self):
         """:return: 40 byte hex version of our 20 byte binary sha"""
-        return bin_to_hex(self.binsha)
+        # b2a_hex produces bytes
+        return bin_to_hex(self.binsha).decode('ascii')
 
     @property
     def data_stream(self):
diff --git a/git/objects/commit.py b/git/objects/commit.py
index f8b5c969..53af22cd 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -62,7 +62,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
                  "author", "authored_date", "author_tz_offset",
                  "committer", "committed_date", "committer_tz_offset",
                  "message", "parents", "encoding", "gpgsig")
-    _id_attribute_ = "binsha"
+    _id_attribute_ = "hexsha"
 
     def __init__(self, repo, binsha, tree=None, author=None, authored_date=None, author_tz_offset=None,
                  committer=None, committed_date=None, committer_tz_offset=None,
@@ -395,7 +395,7 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
             write(("encoding %s\n" % self.encoding).encode('ascii'))
 
         if self.gpgsig:
-            write("gpgsig")
+            write(b"gpgsig")
             for sigline in self.gpgsig.rstrip("\n").split("\n"):
                 write((" " + sigline + "\n").encode('ascii'))
 
diff --git a/git/objects/fun.py b/git/objects/fun.py
index f92a4c06..610bdb5c 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -2,7 +2,6 @@
 from stat import S_ISDIR
 from git.compat import (
     byte_ord,
-    force_bytes,
     defenc,
     xrange,
     text_type
@@ -37,7 +36,7 @@ def tree_to_stream(entries, write):
         # takes the input literally, which appears to be utf8 on linux.
         if isinstance(name, text_type):
             name = name.encode(defenc)
-        write(b''.join(mode_str, b' ', name, b'\0', binsha))
+        write(b''.join((mode_str, b' ', name, b'\0', binsha)))
     # END for each item
 
 def tree_entries_from_data(data):
-- 
cgit v1.2.3


From 4a67e4e49c4e7b82e416067df69c72656213e886 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Mon, 5 Jan 2015 18:21:49 +0100
Subject: test_fun works

---
 git/objects/fun.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/fun.py b/git/objects/fun.py
index 610bdb5c..ba8dbcf4 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -4,7 +4,8 @@ from git.compat import (
     byte_ord,
     defenc,
     xrange,
-    text_type
+    text_type,
+    bchr
 )
 
 __all__ = ('tree_to_stream', 'tree_entries_from_data', 'traverse_trees_recursive',
@@ -21,7 +22,7 @@ def tree_to_stream(entries, write):
     for binsha, mode, name in entries:
         mode_str = b''
         for i in xrange(6):
-            mode_str = chr(((mode >> (i * 3)) & bit_mask) + ord_zero) + mode_str
+            mode_str = bchr(((mode >> (i * 3)) & bit_mask) + ord_zero) + mode_str
         # END for each 8 octal value
 
         # git slices away the first octal if its zero
-- 
cgit v1.2.3


From a83eee5bcee64eeb000dd413ab55aa98dcc8c7f2 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Tue, 6 Jan 2015 11:14:19 +0100
Subject: test_refs works

---
 git/objects/tag.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/tag.py b/git/objects/tag.py
index 5e76e230..c8684447 100644
--- a/git/objects/tag.py
+++ b/git/objects/tag.py
@@ -10,6 +10,7 @@ from .util import (
     parse_actor_and_date
 )
 from gitdb.util import hex_to_bin
+from git.compat import defenc
 
 __all__ = ("TagObject", )
 
@@ -52,11 +53,12 @@ class TagObject(base.Object):
         """Cache all our attributes at once"""
         if attr in TagObject.__slots__:
             ostream = self.repo.odb.stream(self.binsha)
-            lines = ostream.read().splitlines()
+            lines = ostream.read().decode(defenc).splitlines()
 
             obj, hexsha = lines[0].split(" ")       # object <hexsha>
             type_token, type_name = lines[1].split(" ")  # type <type_name>
-            self.object = get_object_type_by_name(type_name)(self.repo, hex_to_bin(hexsha))
+            self.object = \
+                get_object_type_by_name(type_name.encode('ascii'))(self.repo, hex_to_bin(hexsha))
 
             self.tag = lines[2][4:]  # tag <tag name>
 
-- 
cgit v1.2.3


From a05e49d2419d65c59c65adf5cd8c05f276550e1d Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Tue, 6 Jan 2015 14:05:30 +0100
Subject: test_repo works

---
 git/objects/submodule/base.py | 31 +++++++++++++++++++++++--------
 1 file changed, 23 insertions(+), 8 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/submodule/base.py b/git/objects/submodule/base.py
index 0fb3f35f..0ec6f656 100644
--- a/git/objects/submodule/base.py
+++ b/git/objects/submodule/base.py
@@ -17,7 +17,10 @@ from git.util import (
     rmtree
 )
 
-from git.config import SectionConstraint
+from git.config import (
+    SectionConstraint,
+    cp
+)
 from git.exc import (
     InvalidGitRepositoryError,
     NoSuchPathError
@@ -302,6 +305,7 @@ class Submodule(util.IndexObject, Iterable, Traversable):
             writer.set_value(cls.k_head_option, br.path)
             sm._branch_path = br.path
         # END handle path
+        writer.release()
         del(writer)
 
         # we deliberatly assume that our head matches our index !
@@ -419,7 +423,9 @@ class Submodule(util.IndexObject, Iterable, Traversable):
                 # the default implementation will be offended and not update the repository
                 # Maybe this is a good way to assure it doesn't get into our way, but
                 # we want to stay backwards compatible too ... . Its so redundant !
-                self.repo.config_writer().set_value(sm_section(self.name), 'url', self.url)
+                writer = self.repo.config_writer()
+                writer.set_value(sm_section(self.name), 'url', self.url)
+                writer.release()
             # END handle dry_run
         # END handle initalization
 
@@ -576,6 +582,7 @@ class Submodule(util.IndexObject, Iterable, Traversable):
                 writer = self.config_writer(index=index)        # auto-write
                 writer.set_value('path', module_path)
                 self.path = module_path
+                writer.release()
                 del(writer)
             # END handle configuration flag
         except Exception:
@@ -700,8 +707,12 @@ class Submodule(util.IndexObject, Iterable, Traversable):
 
             # now git config - need the config intact, otherwise we can't query
             # inforamtion anymore
-            self.repo.config_writer().remove_section(sm_section(self.name))
-            self.config_writer().remove_section()
+            writer = self.repo.config_writer()
+            writer.remove_section(sm_section(self.name))
+            writer.release()
+            writer = self.config_writer()
+            writer.remove_section()
+            writer.release()
         # END delete configuration
 
         # void our data not to delay invalid access
@@ -800,14 +811,18 @@ class Submodule(util.IndexObject, Iterable, Traversable):
         """
         :return: True if the submodule exists, False otherwise. Please note that
             a submodule may exist (in the .gitmodules file) even though its module
-            doesn't exist"""
+            doesn't exist on disk"""
         # keep attributes for later, and restore them if we have no valid data
         # this way we do not actually alter the state of the object
         loc = locals()
         for attr in self._cache_attrs:
-            if hasattr(self, attr):
-                loc[attr] = getattr(self, attr)
-            # END if we have the attribute cache
+            try:
+                if hasattr(self, attr):
+                    loc[attr] = getattr(self, attr)
+                # END if we have the attribute cache
+            except cp.NoSectionError:
+                # on PY3, this can happen apparently ... don't know why this doesn't happen on PY2
+                pass
         # END for each attr
         self._clear_cache()
 
-- 
cgit v1.2.3


From 4cfc682ff87d3629b31e0196004d1954810b206c Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Tue, 6 Jan 2015 14:11:48 +0100
Subject: test_submodule works

---
 git/objects/base.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'git/objects')

diff --git a/git/objects/base.py b/git/objects/base.py
index 004e3981..3f595d9d 100644
--- a/git/objects/base.py
+++ b/git/objects/base.py
@@ -60,7 +60,7 @@ class Object(LazyMixin):
         :param sha1: 20 byte binary sha1"""
         if sha1 == cls.NULL_BIN_SHA:
             # the NULL binsha is always the root commit
-            return get_object_type_by_name('commit')(repo, sha1)
+            return get_object_type_by_name(b'commit')(repo, sha1)
         # END handle special case
         oinfo = repo.odb.info(sha1)
         inst = get_object_type_by_name(oinfo.type)(repo, oinfo.binsha)
-- 
cgit v1.2.3


From 95bb489a50f6c254f49ba4562d423dbf2201554f Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Tue, 6 Jan 2015 14:20:58 +0100
Subject: test_tree works

---
 git/objects/tree.py | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/tree.py b/git/objects/tree.py
index 6776a15e..f9bee01e 100644
--- a/git/objects/tree.py
+++ b/git/objects/tree.py
@@ -159,7 +159,7 @@ class Tree(IndexObject, diff.Diffable, util.Traversable, util.Serializable):
                 raise TypeError("Unknown mode %o found in tree data for path '%s'" % (mode, path))
         # END for each item
 
-    def __div__(self, file):
+    def join(self, file):
         """Find the named object in this tree's contents
         :return: ``git.Blob`` or ``git.Tree`` or ``git.Submodule``
 
@@ -192,6 +192,14 @@ class Tree(IndexObject, diff.Diffable, util.Traversable, util.Serializable):
             raise KeyError(msg % file)
         # END handle long paths
 
+    def __div__(self, file):
+        """For PY2 only"""
+        return self.join(file)
+
+    def __truediv__(self, file):
+        """For PY3 only"""
+        return self.join(file)
+
     @property
     def trees(self):
         """:return: list(Tree, ...) list of trees directly below this tree"""
@@ -235,7 +243,7 @@ class Tree(IndexObject, diff.Diffable, util.Traversable, util.Serializable):
 
         if isinstance(item, string_types):
             # compatability
-            return self.__div__(item)
+            return self.join(item)
         # END index is basestring
 
         raise TypeError("Invalid index type: %r" % item)
-- 
cgit v1.2.3


From e0c65d6638698f4e3a9e726efca8c0bcf466cd62 Mon Sep 17 00:00:00 2001
From: Sebastian Thiel <byronimo@gmail.com>
Date: Tue, 6 Jan 2015 15:38:20 +0100
Subject: Make flake8 happy

---
 git/objects/commit.py | 8 ++++----
 git/objects/fun.py    | 1 +
 2 files changed, 5 insertions(+), 4 deletions(-)

(limited to 'git/objects')

diff --git a/git/objects/commit.py b/git/objects/commit.py
index 53af22cd..8f93d1b9 100644
--- a/git/objects/commit.py
+++ b/git/objects/commit.py
@@ -382,14 +382,14 @@ class Commit(base.Object, Iterable, Diffable, Traversable, Serializable):
         c = self.committer
         fmt = "%s %s <%s> %s %s\n"
         write((fmt % ("author", aname, a.email,
-                     self.authored_date,
-                     altz_to_utctz_str(self.author_tz_offset))).encode(self.encoding))
+                      self.authored_date,
+                      altz_to_utctz_str(self.author_tz_offset))).encode(self.encoding))
 
         # encode committer
         aname = c.name
         write((fmt % ("committer", aname, c.email,
-                     self.committed_date,
-                     altz_to_utctz_str(self.committer_tz_offset))).encode(self.encoding))
+                      self.committed_date,
+                      altz_to_utctz_str(self.committer_tz_offset))).encode(self.encoding))
 
         if self.encoding != self.default_encoding:
             write(("encoding %s\n" % self.encoding).encode('ascii'))
diff --git a/git/objects/fun.py b/git/objects/fun.py
index ba8dbcf4..c04f80b5 100644
--- a/git/objects/fun.py
+++ b/git/objects/fun.py
@@ -40,6 +40,7 @@ def tree_to_stream(entries, write):
         write(b''.join((mode_str, b' ', name, b'\0', binsha)))
     # END for each item
 
+
 def tree_entries_from_data(data):
     """Reads the binary representation of a tree and returns tuples of Tree items
     :param data: data block with tree data (as bytes)
-- 
cgit v1.2.3