Merge branch 'johnsca-sf-master' into 0.3

Merge all fixes from #134

Author: Byron

git/cmd.py

@@ -88,6 +88,7 @@ def __del__(self):
 # try to kill it
 try:
 os.kill(self.proc.pid, 2) # interrupt signal
+self.proc.wait() # ensure process goes away
 exceptOSError:
 pass# ignore error when process already died
 exceptAttributeError:

git/objects/commit.py

@@ -8,6 +8,7 @@
 Actor,
 Iterable,
 Stats,
+finalize_process
  )
 fromgit.diffimportDiffable
 fromtreeimportTree
@@ -65,7 +66,6 @@ def __init__(self, repo, binsha, tree=None, author=None, authored_date=None, aut
 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 @@ def _iter_from_process_or_stream(cls, repo, proc_or_stream):
 assertlen(hexsha) ==40, "Invalid line: %s"%hexsha
 yieldCommit(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
+ifhasattr(proc_or_stream, 'wait'):
+finalize_process(proc_or_stream)
 
 
 @classmethod
@@ -430,14 +434,21 @@ def _deserialize(self, stream):
 
 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()
+whilenext_line.startswith('mergetag '):
+next_line=readline()
+whilenext_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()
 whilebuf!="":
 ifbuf[0:10] =="encoding ":
 self.encoding=buf[buf.find(' ')+1:]

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")
-iflen(name) >len(name_enc):
-name=name_enc
+try:
+name_enc=name.decode("utf-8")
+exceptUnicodeDecodeError:
+pass
+else:
+iflen(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):
 returnout
 
 
+
 def_find_by_name(tree_data, name, is_dir, start_at):
 """return data entry matching the given name and tree mode
  or None.

git/objects/util.py

@@ -167,17 +167,22 @@ def parse_date(string_date):
 
 # precompiled regex
 _re_actor_epoch=re.compile(r'^.+? (.*) (\d+) ([+-]\d+).*$')
+_re_only_actor=re.compile(r'^.+? (.*)$')
 
 defparse_actor_and_date(line):
 """Parse out the actor (author or committer) info from a line like::
 
  author Tom Preston-Werner <[email protected]> 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()
+ifm:
+actor, epoch, offset=m.groups()
+else:
+m=_re_only_actor.search(line)
+actor=m.group(1) ifmelselineor''
 return (Actor._from_string(actor), int(epoch), utctz_to_altz(offset))
-
 
 #} END functions
 

git/remote.py

@@ -24,7 +24,10 @@
 TagReference
  )
 
-fromgit.utilimportjoin_path
+fromgit.utilimport (
+join_path,
+finalize_process
+ )
 fromgitdb.utilimportjoin
 
 importre
@@ -58,18 +61,6 @@ def digest_process_messages(fh, progress):
 # END while file is not done reading
 returndropped_lines
 
-deffinalize_process(proc):
-"""Wait for the process (clone, fetch, pull or push) and handle its errors accordingly"""
-try:
-proc.wait()
-exceptGitCommandError,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
-ifproc.poll() ==128:
-raise
-pass
-# END exception handling
-
 defadd_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 

git/repo/base.py

@@ -6,15 +6,17 @@
 
 fromgit.excimportInvalidGitRepositoryError, NoSuchPathError
 fromgit.cmdimportGit
-fromgit.utilimportActor
+fromgit.utilimport (
+Actor,
+finalize_process
+ )
 fromgit.refsimport*
 fromgit.indeximportIndexFile
 fromgit.objectsimport*
 fromgit.configimportGitConfigParser
 fromgit.remoteimport (
 Remote,
 digest_process_messages,
-finalize_process,
 add_progress
  )
 
@@ -540,6 +542,7 @@ def untracked_files(self):
 iffilename[0] ==filename[-1] =='"':
 filename=filename[1:-1].decode('string_escape')
 untracked_files.append(filename)
+finalize_process(proc)
 returnuntracked_files
 
 @property

git/test/lib/helper.py

@@ -227,7 +227,7 @@ class TestBase(TestCase):
  """
 
 @classmethod
-defsetUp(cls):
+defsetUpClass(cls):
 """
  Dynamically add a read-only repository to our actual type. This way 
  each test type has its own repository

git/test/test_fun.py

@@ -2,7 +2,8 @@
 fromgit.objects.funimport (
 traverse_tree_recursive,
 traverse_trees_recursive, 
-tree_to_stream
+tree_to_stream,
+tree_entries_from_data
  )
 
 fromgit.index.funimport (
@@ -249,3 +250,7 @@ def test_tree_traversal_single(self):
 entries=traverse_tree_recursive(odb, commit.tree.binsha, '')
 assertentries
 # END for each commit
+
+deftest_tree_entries_from_data(self):
+r=tree_entries_from_data(b'100644 \x9f\0aaa')
+assertr== [('aaa', 33188, '\x9f')], r

git/test/test_git.py

@@ -18,8 +18,8 @@
 classTestGit(TestBase):
 
 @classmethod
-defsetUp(cls):
-super(TestGit, cls).setUp()
+defsetUpClass(cls):
+super(TestGit, cls).setUpClass()
 cls.git=Git(cls.rorepo.working_dir)
 
 @patch.object(Git, 'execute')

git/util.py

@@ -15,6 +15,8 @@
 
 # NOTE: Some of the unused imports might be used/imported by others.
 # Handle once test-cases are back up and running.
+fromexcimportGitCommandError
+
 fromgitdb.utilimport (
 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())
 
+deffinalize_process(proc):
+"""Wait for the process (clone, fetch, pull or push) and handle its errors accordingly"""
+try:
+proc.wait()
+exceptGitCommandError,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
+ifproc.poll() ==128:
+raise
+pass
+# END exception handling
+
 #} END utilities
 
 #{Classes