test_commit works once again

Author: Byron

doc/source/changes.rst

@@ -5,6 +5,7 @@ Changelog
 0.3.4 - python 3 support
 ========================
 * Internally, hexadecimal SHA1 are treated as ascii encoded strings. Binary SHA1 are treated as bytes.
+* Id attribute of Commit objects is now `hexsha`, instead of `binsha`. The latter makes no sense in python 3 and I see no application of it anyway besides its artificial usage in test cases.
 
 0.3.3
 =====

git/cmd.py

@@ -320,6 +320,7 @@ def execute(self, command,
             always be created with a pipe due to issues with subprocess.
             This merely is a workaround as data will be copied from the
             output pipe to the given output stream directly.
+            Judging from the implementation, you shouldn't use this flag !
 
         :param subprocess_kwargs:
             Keyword arguments to be passed to subprocess.Popen. Please note that
@@ -411,9 +412,13 @@ def execute(self, command,
             else:
                 raise GitCommandError(command, status, stderr_value)
 
+
+        if isinstance(stdout_value, bytes): # could also be output_stream
+            stdout_value = stdout_value.decode(defenc)
+
         # Allow access to the command's status code
         if with_extended_output:
-            return (status, stdout_value, stderr_value)
+            return (status, stdout_value, stderr_value.decode(defenc))
         else:
             return stdout_value
 

git/config.py

@@ -201,6 +201,11 @@ def __del__(self):
                 self.write()
             except IOError:
                 log.error("Exception during destruction of GitConfigParser", exc_info=True)
+            except ReferenceError:
+                # This happens in PY3 ... and usually means that some state cannot be written
+                # as the sections dict cannot be iterated
+                # Usually when shutting down the interpreter, don'y know how to fix this
+                pass
         finally:
             self._lock._release_lock()
 
@@ -345,7 +350,7 @@ def write(self):
         # END get lock for physical files
 
         if not hasattr(fp, "seek"):
-            fp = open(self._file_or_files, "w")
+            fp = open(self._file_or_files, "wb")
             close_fp = True
         else:
             fp.seek(0)

git/diff.py

@@ -195,7 +195,7 @@ class Diff(object):
                             """, re.VERBOSE | re.MULTILINE)
     # can be used for comparisons
     NULL_HEX_SHA = "0" * 40
-    NULL_BIN_SHA = "\0" * 20
+    NULL_BIN_SHA = b"\0" * 20
 
     __slots__ = ("a_blob", "b_blob", "a_mode", "b_mode", "new_file", "deleted_file",
                  "rename_from", "rename_to", "diff")

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 @@ def __hash__(self):
 
     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 @@ def __repr__(self):
     @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):

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 @@ def _serialize(self, stream):
             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'))
 

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):

git/refs/symbolic.py

@@ -308,7 +308,7 @@ def set_reference(self, ref, logmsg=None):
 
         lfd = LockedFD(fpath)
         fd = lfd.open(write=True, stream=True)
-        fd.write(write_value)
+        fd.write(write_value.encode('ascii'))
         lfd.commit()
 
         # Adjust the reflog

git/repo/base.py

@@ -47,7 +47,10 @@
     read_gitfile,
     touch,
 )
-from git.compat import text_type
+from git.compat import (
+    text_type,
+    defenc
+)
 
 import os
 import sys
@@ -177,11 +180,11 @@ def __hash__(self):
     # Description property
     def _get_description(self):
         filename = join(self.git_dir, 'description')
-        return open(filename).read().rstrip()
+        return open(filename, 'rb').read().rstrip().decode(defenc)
 
     def _set_description(self, descr):
         filename = join(self.git_dir, 'description')
-        open(filename, 'w').write(descr + '\n')
+        open(filename, 'wb').write((descr + '\n').encode(defenc))
 
     description = property(_get_description, _set_description,
                            doc="the project's description")
@@ -464,8 +467,8 @@ def _get_alternates(self):
 
         if os.path.exists(alternates_path):
             try:
-                f = open(alternates_path)
-                alts = f.read()
+                f = open(alternates_path, 'rb')
+                alts = f.read().decode(defenc)
             finally:
                 f.close()
             return alts.strip().splitlines()
@@ -489,8 +492,8 @@ def _set_alternates(self, alts):
                 os.remove(alternates_path)
         else:
             try:
-                f = open(alternates_path, 'w')
-                f.write("\n".join(alts))
+                f = open(alternates_path, 'wb')
+                f.write("\n".join(alts).encode(defenc))
             finally:
                 f.close()
             # END file handling

git/test/test_commit.py

@@ -51,7 +51,7 @@ def assert_commit_serialization(rwrepo, commit_id, print_performance_info=False)
         stream.seek(0)
 
         istream = rwrepo.odb.store(IStream(Commit.type, streamlen, stream))
-        assert istream.hexsha == cm.hexsha
+        assert istream.hexsha == cm.hexsha.encode('ascii')
 
         nc = Commit(rwrepo, Commit.NULL_BIN_SHA, cm.tree,
                     cm.author, cm.authored_date, cm.author_tz_offset,
@@ -129,7 +129,7 @@ def check_entries(d):
 
     def test_unicode_actor(self):
         # assure we can parse unicode actors correctly
-        name = "Üäöß ÄußÉ".decode("utf-8")
+        name = u"Üäöß ÄußÉ"
         assert len(name) == 9
         special = Actor._from_string(u"%s <something@this.com>" % name)
         assert special.name == name
@@ -146,13 +146,13 @@ def test_traversal(self):
         # basic branch first, depth first
         dfirst = start.traverse(branch_first=False)
         bfirst = start.traverse(branch_first=True)
-        assert dfirst.next() == p0
-        assert dfirst.next() == p00
+        assert next(dfirst) == p0
+        assert next(dfirst) == p00
 
-        assert bfirst.next() == p0
-        assert bfirst.next() == p1
-        assert bfirst.next() == p00
-        assert bfirst.next() == p10
+        assert next(bfirst) == p0
+        assert next(bfirst) == p1
+        assert next(bfirst) == p00
+        assert next(bfirst) == p10
 
         # at some point, both iterations should stop
         assert list(bfirst)[-1] == first
@@ -161,19 +161,19 @@ def test_traversal(self):
         assert len(l[0]) == 2
 
         # ignore self
-        assert start.traverse(ignore_self=False).next() == start
+        assert next(start.traverse(ignore_self=False)) == start
 
         # depth
         assert len(list(start.traverse(ignore_self=False, depth=0))) == 1
 
         # prune
-        assert start.traverse(branch_first=1, prune=lambda i, d: i == p0).next() == p1
+        assert next(start.traverse(branch_first=1, prune=lambda i, d: i == p0)) == p1
 
         # predicate
-        assert start.traverse(branch_first=1, predicate=lambda i, d: i == p1).next() == p1
+        assert next(start.traverse(branch_first=1, predicate=lambda i, d: i == p1)) == p1
 
         # traversal should stop when the beginning is reached
-        self.failUnlessRaises(StopIteration, first.traverse().next)
+        self.failUnlessRaises(StopIteration, next, first.traverse())
 
         # parents of the first commit should be empty ( as the only parent has a null
         # sha )
@@ -210,7 +210,7 @@ def test_rev_list_bisect_all(self):
                                         first_parent=True,
                                         bisect_all=True)
 
-        commits = Commit._iter_from_process_or_stream(self.rorepo, StringProcessAdapter(revs))
+        commits = Commit._iter_from_process_or_stream(self.rorepo, StringProcessAdapter(revs.encode('ascii')))
         expected_ids = (
             '7156cece3c49544abb6bf7a0c218eb36646fad6d',
             '1f66cfbbce58b4b552b041707a12d437cc5f400a',
@@ -224,8 +224,10 @@ def test_count(self):
         assert self.rorepo.tag('refs/tags/0.1.5').commit.count() == 143
 
     def test_list(self):
+        # This doesn't work anymore, as we will either attempt getattr with bytes, or compare 20 byte string
+        # with actual 20 byte bytes. This usage makes no sense anyway
         assert isinstance(Commit.list_items(self.rorepo, '0.1.5', max_count=5)[
-                          hex_to_bin('5117c9c8a4d3af19a9958677e45cda9269de1541')], Commit)
+                          '5117c9c8a4d3af19a9958677e45cda9269de1541'], Commit)
 
     def test_str(self):
         commit = Commit(self.rorepo, Commit.NULL_BIN_SHA)
@@ -247,12 +249,12 @@ def test_iter_parents(self):
         c = self.rorepo.commit('0.1.5')
         for skip in (0, 1):
             piter = c.iter_parents(skip=skip)
-            first_parent = piter.next()
+            first_parent = next(piter)
             assert first_parent != c
             assert first_parent == c.parents[0]
         # END for each
 
-    def test_base(self):
+    def test_name_rev(self):
         name_rev = self.rorepo.head.commit.name_rev
         assert isinstance(name_rev, string_types)
 
@@ -270,10 +272,10 @@ def test_serialization_unicode_support(self):
         assert isinstance(cmt.message, text_type)     # it automatically decodes it as such
         assert isinstance(cmt.author.name, text_type)  # same here
 
-        cmt.message = "üäêèß".decode("utf-8")
+        cmt.message = u"üäêèß"
         assert len(cmt.message) == 5
 
-        cmt.author.name = "äüß".decode("utf-8")
+        cmt.author.name = u"äüß"
         assert len(cmt.author.name) == 3
 
         cstream = BytesIO()
@@ -292,7 +294,7 @@ def test_serialization_unicode_support(self):
 
     def test_gpgsig(self):
         cmt = self.rorepo.commit()
-        cmt._deserialize(open(fixture_path('commit_with_gpgsig')))
+        cmt._deserialize(open(fixture_path('commit_with_gpgsig'), 'rb'))
 
         fixture_sig = """-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.11 (GNU/Linux)
@@ -318,7 +320,7 @@ def test_gpgsig(self):
 
         cstream = BytesIO()
         cmt._serialize(cstream)
-        assert re.search(r"^gpgsig <test\n dummy\n sig>$", cstream.getvalue(), re.MULTILINE)
+        assert re.search(r"^gpgsig <test\n dummy\n sig>$", cstream.getvalue().decode('ascii'), re.MULTILINE)
 
         cstream.seek(0)
         cmt.gpgsig = None
@@ -328,4 +330,4 @@ def test_gpgsig(self):
         cmt.gpgsig = None
         cstream = BytesIO()
         cmt._serialize(cstream)
-        assert not re.search(r"^gpgsig ", cstream.getvalue(), re.MULTILINE)
+        assert not re.search(r"^gpgsig ", cstream.getvalue().decode('ascii'), re.MULTILINE)