Fixed io types to make tests work on PY2 once again.

Now it's about going through PY3 issues

Author: Byron

Diff for: git/compat.py

@@ -16,7 +16,8 @@
 
 from gitdb.utils.encoding import (
     string_types,
-    text_type
+    text_type,
+    force_bytes
 )
 
 if PY3:

Diff for: git/index/base.py

@@ -8,7 +8,7 @@
 import sys
 import subprocess
 import glob
-from io import StringIO
+from io import BytesIO
 
 from stat import S_ISLNK
 
@@ -43,6 +43,7 @@
     izip,
     xrange,
     string_types,
+    force_bytes
 )
 
 from git.util import (
@@ -562,7 +563,8 @@ def _store_path(self, filepath, fprogress):
         st = os.lstat(filepath)     # handles non-symlinks as well
         stream = None
         if S_ISLNK(st.st_mode):
-            stream = StringIO(os.readlink(filepath))
+            # in PY3, readlink is string, but we need bytes. In PY2, it's just OS encoded bytes, we assume UTF-8
+            stream = BytesIO(force_bytes(os.readlink(filepath), encoding='utf-8'))
         else:
             stream = open(filepath, 'rb')
         # END handle stream

Diff for: git/index/fun.py

@@ -12,7 +12,7 @@
 
 S_IFGITLINK = S_IFLNK | S_IFDIR     # a submodule
 
-from io import StringIO
+from io import BytesIO
 
 from git.util import IndexFileSHA1Writer
 from git.exc import UnmergedEntriesError
@@ -218,7 +218,7 @@ def write_tree_from_cache(entries, odb, sl, si=0):
     # END for each entry
 
     # finally create the tree
-    sio = StringIO()
+    sio = BytesIO()
     tree_to_stream(tree_items, sio.write)
     sio.seek(0)
 

Diff for: git/objects/commit.py

@@ -30,7 +30,7 @@
     altzone
 )
 import os
-from io import StringIO
+from io import BytesIO
 import logging
 
 log = logging.getLogger('git.objects.commit')
@@ -133,7 +133,7 @@ def _set_cache_(self, attr):
         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 @@ def create_from_tree(cls, repo, tree, message, parent_commits=None, head=False,
                          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 for: git/objects/submodule/base.py

@@ -8,7 +8,7 @@
     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 @@ def _clear_cache(self):
 
     @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 for: 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 @@ def flush_to_index(self):
         """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:

Diff for: git/refs/log.py

@@ -85,7 +85,7 @@ def from_line(cls, line):
         :param line: line without trailing newline
         :raise ValueError: If line could not be parsed"""
         try:
-            info, msg = line.split('\t', 2)
+            info, msg = line.split('\t', 1)
         except ValueError:
             raise ValueError("line is missing tab separator")
         # END handle first plit

Diff for: git/test/fixtures/git_config_global

@@ -1,3 +1,4 @@
+# just a comment
 [alias]
 	st = status
 	ci = commit

Diff for: git/test/lib/helper.py

@@ -48,8 +48,8 @@ class StringProcessAdapter(object):
     Its tailored to work with the test system only"""
 
     def __init__(self, input_string):
-        self.stdout = io.StringIO(input_string)
-        self.stderr = io.StringIO()
+        self.stdout = io.BytesIO(input_string)
+        self.stderr = io.BytesIO()
 
     def wait(self):
         return 0

Diff for: git/test/performance/test_commit.py

@@ -4,7 +4,7 @@
 # This module is part of GitPython and is released under
 # the BSD License: http://www.opensource.org/licenses/bsd-license.php
 from __future__ import print_function
-from io import StringIO
+from io import BytesIO
 from time import time
 import sys
 
@@ -93,7 +93,7 @@ def test_commit_serialization(self):
                         hc.committer, hc.committed_date, hc.committer_tz_offset,
                         str(i), parents=hc.parents, encoding=hc.encoding)
 
-            stream = StringIO()
+            stream = BytesIO()
             cm._serialize(stream)
             slen = stream.tell()
             stream.seek(0)

Diff for: git/test/test_commit.py

@@ -25,7 +25,7 @@
     text_type
 )
 
-from io import StringIO
+from io import BytesIO
 import time
 import sys
 import re
@@ -44,7 +44,7 @@ def assert_commit_serialization(rwrepo, commit_id, print_performance_info=False)
 
         # assert that we deserialize commits correctly, hence we get the same
         # sha on serialization
-        stream = StringIO()
+        stream = BytesIO()
         cm._serialize(stream)
         ns += 1
         streamlen = stream.tell()
@@ -59,7 +59,7 @@ def assert_commit_serialization(rwrepo, commit_id, print_performance_info=False)
                     cm.message, cm.parents, cm.encoding)
 
         assert nc.parents == cm.parents
-        stream = StringIO()
+        stream = BytesIO()
         nc._serialize(stream)
         ns += 1
         streamlen = stream.tell()
@@ -276,7 +276,7 @@ def test_serialization_unicode_support(self):
         cmt.author.name = "äüß".decode("utf-8")
         assert len(cmt.author.name) == 3
 
-        cstream = StringIO()
+        cstream = BytesIO()
         cmt._serialize(cstream)
         cstream.seek(0)
         assert len(cstream.getvalue())
@@ -316,7 +316,7 @@ def test_gpgsig(self):
         cmt.gpgsig = "<test\ndummy\nsig>"
         assert cmt.gpgsig != fixture_sig
 
-        cstream = StringIO()
+        cstream = BytesIO()
         cmt._serialize(cstream)
         assert re.search(r"^gpgsig <test\n dummy\n sig>$", cstream.getvalue(), re.MULTILINE)
 
@@ -326,6 +326,6 @@ def test_gpgsig(self):
         assert cmt.gpgsig == "<test\ndummy\nsig>"
 
         cmt.gpgsig = None
-        cstream = StringIO()
+        cstream = BytesIO()
         cmt._serialize(cstream)
         assert not re.search(r"^gpgsig ", cstream.getvalue(), re.MULTILINE)

Diff for: git/test/test_config.py

@@ -12,7 +12,7 @@
     GitConfigParser
 )
 from git.compat import string_types
-import StringIO
+import io
 from copy import copy
 from ConfigParser import NoSectionError
 
@@ -21,7 +21,7 @@ class TestBase(TestCase):
 
     def _to_memcache(self, file_path):
         fp = open(file_path, "r")
-        sio = StringIO.StringIO(fp.read())
+        sio = io.BytesIO(fp.read())
         sio.name = file_path
         return sio
 

Diff for: git/test/test_fun.py

@@ -24,7 +24,7 @@
 )
 
 from git.index import IndexFile
-from io import StringIO
+from io import BytesIO
 
 
 class TestFun(TestBase):
@@ -72,7 +72,7 @@ def test_aggressive_tree_merge(self):
 
     def mktree(self, odb, entries):
         """create a tree from the given tree entries and safe it to the database"""
-        sio = StringIO()
+        sio = BytesIO()
         tree_to_stream(entries, sio.write)
         sio.seek(0)
         istream = odb.store(IStream(str_tree_type, len(sio.getvalue()), sio))

Diff for: git/test/test_index.py

@@ -31,7 +31,7 @@
     ST_MODE
 )
 
-from io import StringIO
+from io import BytesIO
 from gitdb.base import IStream
 from git.objects import Blob
 from git.index.typ import (
@@ -698,9 +698,9 @@ def test_index_bare_add(self, rw_bare_repo):
         # instead of throwing the Exception we are expecting. This is
         # a quick hack to make this test fail when expected.
         rw_bare_repo._working_tree_dir = None
-        contents = 'This is a StringIO file'
+        contents = b'This is a BytesIO file'
         filesize = len(contents)
-        fileobj = StringIO(contents)
+        fileobj = BytesIO(contents)
         filename = 'my-imaginary-file'
         istream = rw_bare_repo.odb.store(
             IStream(Blob.type, filesize, fileobj))

Diff for: git/test/test_repo.py

@@ -36,7 +36,7 @@
 import sys
 import tempfile
 import shutil
-from io import StringIO
+from io import BytesIO
 
 
 class TestRepo(TestBase):
@@ -362,22 +362,22 @@ def test_comparison_and_hash(self):
     def test_git_cmd(self):
         # test CatFileContentStream, just to be very sure we have no fencepost errors
         # last \n is the terminating newline that it expects
-        l1 = "0123456789\n"
-        l2 = "abcdefghijklmnopqrstxy\n"
-        l3 = "z\n"
-        d = "%s%s%s\n" % (l1, l2, l3)
+        l1 = b"0123456789\n"
+        l2 = b"abcdefghijklmnopqrstxy\n"
+        l3 = b"z\n"
+        d = b"%s%s%s\n" % (l1, l2, l3)
 
         l1p = l1[:5]
 
         # full size
         # size is without terminating newline
         def mkfull():
-            return Git.CatFileContentStream(len(d) - 1, StringIO(d))
+            return Git.CatFileContentStream(len(d) - 1, BytesIO(d))
 
         ts = 5
 
         def mktiny():
-            return Git.CatFileContentStream(ts, StringIO(d))
+            return Git.CatFileContentStream(ts, BytesIO(d))
 
         # readlines no limit
         s = mkfull()

Diff for: git/test/test_tree.py

@@ -11,7 +11,7 @@
     Blob
 )
 
-from io import StringIO
+from io import BytesIO
 
 
 class TestTree(TestBase):
@@ -30,7 +30,7 @@ def test_serializable(self):
             orig_data = tree.data_stream.read()
             orig_cache = tree._cache
 
-            stream = StringIO()
+            stream = BytesIO()
             tree._serialize(stream)
             assert stream.getvalue() == orig_data
 
@@ -82,7 +82,7 @@ def test_serializable(self):
             mod.set_done()      # multiple times are okay
 
             # serialize, its different now
-            stream = StringIO()
+            stream = BytesIO()
             testtree._serialize(stream)
             stream.seek(0)
             assert stream.getvalue() != orig_data