# cache the property

        _repo = self._repo

        refs = _repo.get_refs()

        keys = [('refs/heads/', 'H'),

                ('refs/remotes/origin/', 'RH'),

                ('refs/tags/', 'T')]

        _refs = {}

        for ref, sha in refs.iteritems():

            for k, type_ in keys:

                if ref.startswith(k):

                    _key = ref[len(k):]

                    if type_ == 'T':

                        obj = _repo.get_object(sha)

                        if isinstance(obj, Tag):

                            sha = _repo.get_object(sha).object[1]

                    _refs[_key] = [sha, type_]

                    break

        return _refs

    def _heads(self, reverse=False):

        refs = self._repo.get_refs()

        heads = {}

        for key, val in refs.items():

            for ref_key in ['refs/heads/', 'refs/remotes/origin/']:

                if key.startswith(ref_key):

                    n = key[len(ref_key):]

                    if n not in ['HEAD']:

                        heads[n] = val

        return heads if reverse else dict((y, x) for x, y in heads.iteritems())

    def get_changeset(self, revision=None):

        """

        Returns ``GitChangeset`` object representing commit from git repository

        at the given revision or head (most recent commit) if None given.

        """

        if isinstance(revision, GitChangeset):

            return revision

        revision = self._get_revision(revision)

        changeset = GitChangeset(repository=self, revision=revision)

        return changeset

    def get_changesets(self, start=None, end=None, start_date=None,

           end_date=None, branch_name=None, reverse=False):

        """

        Returns iterator of ``GitChangeset`` objects from start to end (both

        are inclusive), in ascending date order (unless ``reverse`` is set).

        :param start: changeset ID, as str; first returned changeset

        :param end: changeset ID, as str; last returned changeset

        :param start_date: if specified, changesets with commit date less than

          ``start_date`` would be filtered out from returned set

        :param end_date: if specified, changesets with commit date greater than

          ``end_date`` would be filtered out from returned set

        :param branch_name: if specified, changesets not reachable from given

          branch would be filtered out from returned set

        :param reverse: if ``True``, returned generator would be reversed

          (meaning that returned changesets would have descending date order)

        :raise BranchDoesNotExistError: If given ``branch_name`` does not

            exist.

        :raise ChangesetDoesNotExistError: If changeset for given ``start`` or

          ``end`` could not be found.

        """

        if branch_name and branch_name not in self.branches:

            raise BranchDoesNotExistError("Branch '%s' not found" \

                                          % branch_name)

        # actually we should check now if it's not an empty repo to not spaw

        # subprocess commands

        if self._empty:

            raise EmptyRepositoryError("There are no changesets yet")

        # %H at format means (full) commit hash, initial hashes are retrieved

        # in ascending date order

        cmd_template = 'log --date-order --reverse --pretty=format:"%H"'

        cmd_params = {}

        if start_date:

            cmd_template += ' --since "$since"'

            cmd_params['since'] = start_date.strftime('%m/%d/%y %H:%M:%S')

        if end_date:

            cmd_template += ' --until "$until"'

            cmd_params['until'] = end_date.strftime('%m/%d/%y %H:%M:%S')

        if branch_name:

            cmd_template += ' $branch_name'

            cmd_params['branch_name'] = branch_name

        else:

            rev_filter = _git_path = settings.GIT_REV_FILTER

            cmd_template += ' %s' % (rev_filter)

        cmd = string.Template(cmd_template).safe_substitute(**cmd_params)

        revs = self.run_git_command(cmd)[0].splitlines()

        start_pos = 0

        end_pos = len(revs)

        if start:

            _start = self._get_revision(start)

            try:

                start_pos = revs.index(_start)

            except ValueError:

                pass

        if end is not None:

            _end = self._get_revision(end)

            try:

                end_pos = revs.index(_end)

            except ValueError:

                pass

        if None not in [start, end] and start_pos > end_pos:

            raise RepositoryError('start cannot be after end')

        if end_pos is not None:

            end_pos += 1

        revs = revs[start_pos:end_pos]

        if reverse:

            revs = reversed(revs)

        return CollectionGenerator(self, revs)

    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,

                 context=3):

        """

        Returns (git like) *diff*, as plain text. Shows changes introduced by

        ``rev2`` since ``rev1``.

        :param rev1: Entry point from which diff is shown. Can be

          ``self.EMPTY_CHANGESET`` - in this case, patch showing all

          the changes since empty state of the repository until ``rev2``

        :param rev2: Until which revision changes should be shown.

        :param ignore_whitespace: If set to ``True``, would not show whitespace

          changes. Defaults to ``False``.

        :param context: How many lines before/after changed lines should be

          shown. Defaults to ``3``.

        """

        flags = ['-U%s' % context, '--full-index', '--binary', '-p', '-M', '--abbrev=40']

        if ignore_whitespace:

            flags.append('-w')

        if hasattr(rev1, 'raw_id'):

            rev1 = getattr(rev1, 'raw_id')

        if hasattr(rev2, 'raw_id'):

            rev2 = getattr(rev2, 'raw_id')

        if rev1 == self.EMPTY_CHANGESET:

            rev2 = self.get_changeset(rev2).raw_id

            cmd = ' '.join(['show'] + flags + [rev2])

        else:

            rev1 = self.get_changeset(rev1).raw_id

            rev2 = self.get_changeset(rev2).raw_id

            cmd = ' '.join(['diff'] + flags + [rev1, rev2])

        if path:

            cmd += ' -- "%s"' % path

        stdout, stderr = self.run_git_command(cmd)

        # If we used 'show' command, strip first few lines (until actual diff

        # starts)

        if rev1 == self.EMPTY_CHANGESET:

            lines = stdout.splitlines()

            x = 0

            for line in lines:

                if line.startswith('diff'):

                    break

                x += 1

            # Append new line just like 'diff' command do

            stdout = '\n'.join(lines[x:]) + '\n'

        return stdout

    @LazyProperty

    def in_memory_changeset(self):

        """

        Returns ``GitInMemoryChangeset`` object for this repository.

        """

        return GitInMemoryChangeset(self)

    def clone(self, url, update_after_clone=True, bare=False):

        """

        Tries to clone changes from external location.

        :param update_after_clone: If set to ``False``, git won't checkout

          working directory

        :param bare: If set to ``True``, repository would be cloned into

          *bare* git repository (no working directory at all).

        """

        url = self._get_url(url)

        cmd = ['clone', '-q']

        if bare:

            cmd.append('--bare')

        elif not update_after_clone:

            cmd.append('--no-checkout')

        cmd = ' '.join(cmd)

        # If error occurs run_git_command raises RepositoryError already

        self.run_git_command(cmd)

    def pull(self, url):

        """

        Tries to pull changes from external location.

        """

        url = self._get_url(url)

        cmd = ['pull', "--ff-only", quote(url)]

        cmd = ' '.join(cmd)

        # If error occurs run_git_command raises RepositoryError already

        self.run_git_command(cmd)

    def fetch(self, url):

        """

        Tries to pull changes from external location.

        """

        url = self._get_url(url)

        so, se = self.run_git_command('ls-remote -h %s' % quote(url))

        refs = []

        for line in (x for x in so.splitlines()):

            sha, ref = line.split('\t')

            refs.append(ref)

        refs = ' '.join(('+%s:%s' % (r, r) for r in refs))

        cmd = '''fetch %s -- %s''' % (quote(url), refs)

        self.run_git_command(cmd)

    def _update_server_info(self):

        """

        runs gits update-server-info command in this repo instance

        """

        from dulwich.server import update_server_info

        update_server_info(self._repo)

    @LazyProperty

    def workdir(self):

        """

        Returns ``Workdir`` instance for this repository.

        """

        return GitWorkdir(self)

    def get_config_value(self, section, name, config_file=None):

        """

        Returns configuration value for a given [``section``] and ``name``.

        :param section: Section we want to retrieve value from

        :param name: Name of configuration we want to retrieve

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        if config_file is None:

            config_file = []

        elif isinstance(config_file, basestring):

            config_file = [config_file]

        def gen_configs():

            for path in config_file + self._config_files:

                try:

                    yield ConfigFile.from_path(path)

                except (IOError, OSError, ValueError):

                    continue

        for config in gen_configs():

            try:

                return config.get(section, name)

            except KeyError:

                continue

        return None

    def get_user_name(self, config_file=None):

        """

        Returns user's name from global configuration file.

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        return self.get_config_value('user', 'name', config_file)

    def get_user_email(self, config_file=None):

        """

        Returns user's email from global configuration file.

        :param config_file: A path to file which should be used to retrieve

          configuration from (might also be a list of file paths)

        """

        return self.get_config_value('user', 'email', config_file)

from __future__ import with_statement

import os

import mock

import datetime

import urllib2

from kallithea.lib.vcs.backends.git import GitRepository, GitChangeset

from kallithea.lib.vcs.exceptions import RepositoryError, VCSError, NodeDoesNotExistError

from kallithea.lib.vcs.nodes import NodeKind, FileNode, DirNode, NodeState

from kallithea.lib.vcs.utils.compat import unittest

from kallithea.tests.vcs.base import BackendTestMixin

from kallithea.tests.vcs.conf import TEST_GIT_REPO, TEST_GIT_REPO_CLONE, get_new_dir

class GitRepositoryTest(unittest.TestCase):

    def __check_for_existing_repo(self):

        if os.path.exists(TEST_GIT_REPO_CLONE):

            self.fail('Cannot test git clone repo as location %s already '

                      'exists. You should manually remove it first.'

                      % TEST_GIT_REPO_CLONE)

    def setUp(self):

        self.repo = GitRepository(TEST_GIT_REPO)

    def test_wrong_repo_path(self):

        wrong_repo_path = '/tmp/errorrepo'

        self.assertRaises(RepositoryError, GitRepository, wrong_repo_path)

    def test_git_cmd_injection(self):

        with self.assertRaises(urllib2.URLError):

        with self.assertRaises(RepositoryError):

            # Should fail on direct clone call, which as of this writing does not happen outside of class

            clone_fail_repo = GitRepository(get_new_dir('injection-repo'), create=True)

        # Repo should have been created

        self.assertFalse(successfully_cloned._repo.bare)

    def test_repo_clone(self):

        self.__check_for_existing_repo()

        repo = GitRepository(TEST_GIT_REPO)

        repo_clone = GitRepository(TEST_GIT_REPO_CLONE,

            src_url=TEST_GIT_REPO, create=True, update_after_clone=True)

        self.assertEqual(len(repo.revisions), len(repo_clone.revisions))

        # Checking hashes of changesets should be enough

        for changeset in repo.get_changesets():

            raw_id = changeset.raw_id

            self.assertEqual(raw_id, repo_clone.get_changeset(raw_id).raw_id)

    def test_repo_clone_without_create(self):

        self.assertRaises(RepositoryError, GitRepository,

            TEST_GIT_REPO_CLONE + '_wo_create', src_url=TEST_GIT_REPO)

    def test_repo_clone_with_update(self):

        repo = GitRepository(TEST_GIT_REPO)

        clone_path = TEST_GIT_REPO_CLONE + '_with_update'

        repo_clone = GitRepository(clone_path,

            create=True, src_url=TEST_GIT_REPO, update_after_clone=True)

        self.assertEqual(len(repo.revisions), len(repo_clone.revisions))

        #check if current workdir was updated

        fpath = os.path.join(clone_path, 'MANIFEST.in')

        self.assertEqual(True, os.path.isfile(fpath),

            'Repo was cloned and updated but file %s could not be found'

            % fpath)

    def test_repo_clone_without_update(self):

        repo = GitRepository(TEST_GIT_REPO)

        clone_path = TEST_GIT_REPO_CLONE + '_without_update'

        repo_clone = GitRepository(clone_path,

            create=True, src_url=TEST_GIT_REPO, update_after_clone=False)

        self.assertEqual(len(repo.revisions), len(repo_clone.revisions))

        #check if current workdir was *NOT* updated

        fpath = os.path.join(clone_path, 'MANIFEST.in')

        # Make sure it's not bare repo

        self.assertFalse(repo_clone._repo.bare)

        self.assertEqual(False, os.path.isfile(fpath),

            'Repo was cloned and updated but file %s was found'

            % fpath)

    def test_repo_clone_into_bare_repo(self):

        repo = GitRepository(TEST_GIT_REPO)

        clone_path = TEST_GIT_REPO_CLONE + '_bare.git'

        repo_clone = GitRepository(clone_path, create=True,

            src_url=repo.path, bare=True)

        self.assertTrue(repo_clone._repo.bare)

    def test_create_repo_is_not_bare_by_default(self):

        repo = GitRepository(get_new_dir('not-bare-by-default'), create=True)

        self.assertFalse(repo._repo.bare)

    def test_create_bare_repo(self):

        repo = GitRepository(get_new_dir('bare-repo'), create=True, bare=True)

        self.assertTrue(repo._repo.bare)

    def test_revisions(self):

        # there are 112 revisions (by now)

        # so we can assume they would be available from now on

        subset = set([

            'c1214f7e79e02fc37156ff215cd71275450cffc3',

            '38b5fe81f109cb111f549bfe9bb6b267e10bc557',

            'fa6600f6848800641328adbf7811fd2372c02ab2',

            '102607b09cdd60e2793929c4f90478be29f85a17',

            '49d3fd156b6f7db46313fac355dca1a0b94a0017',

            '2d1028c054665b962fa3d307adfc923ddd528038',

            'd7e0d30fbcae12c90680eb095a4f5f02505ce501',

            'ff7ca51e58c505fec0dd2491de52c622bb7a806b',

            'dd80b0f6cf5052f17cc738c2951c4f2070200d7f',

            '8430a588b43b5d6da365400117c89400326e7992',

            'd955cd312c17b02143c04fa1099a352b04368118',

            'f67b87e5c629c2ee0ba58f85197e423ff28d735b',

            'add63e382e4aabc9e1afdc4bdc24506c269b7618',

            'f298fe1189f1b69779a4423f40b48edf92a703fc',

            'bd9b619eb41994cac43d67cf4ccc8399c1125808',

            '6e125e7c890379446e98980d8ed60fba87d0f6d1',

            'd4a54db9f745dfeba6933bf5b1e79e15d0af20bd',

            '0b05e4ed56c802098dfc813cbe779b2f49e92500',

            '191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e',

            '45223f8f114c64bf4d6f853e3c35a369a6305520',

            'ca1eb7957a54bce53b12d1a51b13452f95bc7c7e',

            'f5ea29fc42ef67a2a5a7aecff10e1566699acd68',

            '27d48942240f5b91dfda77accd2caac94708cc7d',

            '622f0eb0bafd619d2560c26f80f09e3b0b0d78af',

            'e686b958768ee96af8029fe19c6050b1a8dd3b2b'])

        self.assertTrue(subset.issubset(set(self.repo.revisions)))

    def test_slicing(self):

        #4 1 5 10 95

        for sfrom, sto, size in [(0, 4, 4), (1, 2, 1), (10, 15, 5),

                                 (10, 20, 10), (5, 100, 95)]:

            revs = list(self.repo[sfrom:sto])

            self.assertEqual(len(revs), size)

            self.assertEqual(revs[0], self.repo.get_changeset(sfrom))

            self.assertEqual(revs[-1], self.repo.get_changeset(sto - 1))

    def test_branches(self):

        # TODO: Need more tests here

        # Removed (those are 'remotes' branches for cloned repo)

        #self.assertTrue('master' in self.repo.branches)

        #self.assertTrue('gittree' in self.repo.branches)

        #self.assertTrue('web-branch' in self.repo.branches)

        for name, id in self.repo.branches.items():

            self.assertTrue(isinstance(

                self.repo.get_changeset(id), GitChangeset))

    def test_tags(self):

        # TODO: Need more tests here

        self.assertTrue('v0.1.1' in self.repo.tags)

        self.assertTrue('v0.1.2' in self.repo.tags)

        for name, id in self.repo.tags.items():

            self.assertTrue(isinstance(

                self.repo.get_changeset(id), GitChangeset))

    def _test_single_changeset_cache(self, revision):

        chset = self.repo.get_changeset(revision)

        self.assertTrue(revision in self.repo.changesets)

        self.assertTrue(chset is self.repo.changesets[revision])

    def test_initial_changeset(self):

        id = self.repo.revisions[0]

        init_chset = self.repo.get_changeset(id)

        self.assertEqual(init_chset.message, 'initial import\n')

        self.assertEqual(init_chset.author,

            'Marcin Kuzminski <marcin@python-blog.com>')

        for path in ('vcs/__init__.py',

                     'vcs/backends/BaseRepository.py',

                     'vcs/backends/__init__.py'):

            self.assertTrue(isinstance(init_chset.get_node(path), FileNode))

        for path in ('', 'vcs', 'vcs/backends'):

            self.assertTrue(isinstance(init_chset.get_node(path), DirNode))

        self.assertRaises(NodeDoesNotExistError, init_chset.get_node, path='foobar')

        node = init_chset.get_node('vcs/')

        self.assertTrue(hasattr(node, 'kind'))

        self.assertEqual(node.kind, NodeKind.DIR)

        node = init_chset.get_node('vcs')

        self.assertTrue(hasattr(node, 'kind'))

        self.assertEqual(node.kind, NodeKind.DIR)

        node = init_chset.get_node('vcs/__init__.py')

        self.assertTrue(hasattr(node, 'kind'))

        self.assertEqual(node.kind, NodeKind.FILE)

    def test_not_existing_changeset(self):

        self.assertRaises(RepositoryError, self.repo.get_changeset,

            'f' * 40)

    def test_changeset10(self):

        chset10 = self.repo.get_changeset(self.repo.revisions[9])

        README = """===

VCS

===

Various Version Control System management abstraction layer for Python.

Introduction

------------

TODO: To be written...

"""

        node = chset10.get_node('README.rst')

        self.assertEqual(node.kind, NodeKind.FILE)

        self.assertEqual(node.content, README)

class GitChangesetTest(unittest.TestCase):

    def setUp(self):

        self.repo = GitRepository(TEST_GIT_REPO)

    def test_default_changeset(self):

        tip = self.repo.get_changeset()

        self.assertEqual(tip, self.repo.get_changeset(None))

        self.assertEqual(tip, self.repo.get_changeset('tip'))

    def test_root_node(self):

        tip = self.repo.get_changeset()

        self.assertTrue(tip.root is tip.get_node(''))

    def test_lazy_fetch(self):

        """

        Test if changeset's nodes expands and are cached as we walk through