path: root/bzrlib/tests/per_pack_repository.py

# Copyright (C) 2008-2011 Canonical Ltd
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

"""Tests for pack repositories.

These tests are repeated for all pack-based repository formats.
"""

from stat import S_ISDIR

from bzrlib.btree_index import BTreeGraphIndex
from bzrlib.index import GraphIndex
from bzrlib import (
    controldir,
    errors,
    inventory,
    osutils,
    repository,
    revision as _mod_revision,
    tests,
    transport,
    ui,
    )
from bzrlib.repofmt.groupcompress_repo import RepositoryFormat2a
from bzrlib.smart import (
    client,
    )
from bzrlib.tests import (
    TestCaseWithTransport,
    TestNotApplicable,
    )
from bzrlib.transport import (
    memory,
    )
from bzrlib.tests import test_server


class TestPackRepository(TestCaseWithTransport):
    """Tests to be repeated across all pack-based formats.

    The following are populated from the test scenario:

    :ivar format_name: Registered name fo the format to test.
    :ivar format_string: On-disk format marker.
    :ivar format_supports_external_lookups: Boolean.
    """

    def get_format(self):
        return controldir.format_registry.make_bzrdir(self.format_name)

    def test_attribute__fetch_order(self):
        """Packs do not need ordered data retrieval."""
        format = self.get_format()
        repo = self.make_repository('.', format=format)
        self.assertEqual('unordered', repo._format._fetch_order)

    def test_attribute__fetch_uses_deltas(self):
        """Packs reuse deltas."""
        format = self.get_format()
        repo = self.make_repository('.', format=format)
        if isinstance(format.repository_format, RepositoryFormat2a):
            # TODO: This is currently a workaround. CHK format repositories
            #       ignore the 'deltas' flag, but during conversions, we can't
            #       do unordered delta fetches. Remove this clause once we
            #       improve the inter-format fetching.
            self.assertEqual(False, repo._format._fetch_uses_deltas)
        else:
            self.assertEqual(True, repo._format._fetch_uses_deltas)

    def test_disk_layout(self):
        format = self.get_format()
        repo = self.make_repository('.', format=format)
        # in case of side effects of locking.
        repo.lock_write()
        repo.unlock()
        t = repo.bzrdir.get_repository_transport(None)
        self.check_format(t)
        # XXX: no locks left when unlocked at the moment
        # self.assertEqualDiff('', t.get('lock').read())
        self.check_databases(t)

    def check_format(self, t):
        self.assertEqualDiff(
            self.format_string, # from scenario
            t.get('format').read())

    def assertHasNoKndx(self, t, knit_name):
        """Assert that knit_name has no index on t."""
        self.assertFalse(t.has(knit_name + '.kndx'))

    def assertHasNoKnit(self, t, knit_name):
        """Assert that knit_name exists on t."""
        # no default content
        self.assertFalse(t.has(knit_name + '.knit'))

    def check_databases(self, t):
        """check knit content for a repository."""
        # check conversion worked
        self.assertHasNoKndx(t, 'inventory')
        self.assertHasNoKnit(t, 'inventory')
        self.assertHasNoKndx(t, 'revisions')
        self.assertHasNoKnit(t, 'revisions')
        self.assertHasNoKndx(t, 'signatures')
        self.assertHasNoKnit(t, 'signatures')
        self.assertFalse(t.has('knits'))
        # revision-indexes file-container directory
        self.assertEqual([],
            list(self.index_class(t, 'pack-names', None).iter_all_entries()))
        self.assertTrue(S_ISDIR(t.stat('packs').st_mode))
        self.assertTrue(S_ISDIR(t.stat('upload').st_mode))
        self.assertTrue(S_ISDIR(t.stat('indices').st_mode))
        self.assertTrue(S_ISDIR(t.stat('obsolete_packs').st_mode))

    def test_shared_disk_layout(self):
        format = self.get_format()
        repo = self.make_repository('.', shared=True, format=format)
        # we want:
        t = repo.bzrdir.get_repository_transport(None)
        self.check_format(t)
        # XXX: no locks left when unlocked at the moment
        # self.assertEqualDiff('', t.get('lock').read())
        # We should have a 'shared-storage' marker file.
        self.assertEqualDiff('', t.get('shared-storage').read())
        self.check_databases(t)

    def test_shared_no_tree_disk_layout(self):
        format = self.get_format()
        repo = self.make_repository('.', shared=True, format=format)
        repo.set_make_working_trees(False)
        # we want:
        t = repo.bzrdir.get_repository_transport(None)
        self.check_format(t)
        # XXX: no locks left when unlocked at the moment
        # self.assertEqualDiff('', t.get('lock').read())
        # We should have a 'shared-storage' marker file.
        self.assertEqualDiff('', t.get('shared-storage').read())
        # We should have a marker for the no-working-trees flag.
        self.assertEqualDiff('', t.get('no-working-trees').read())
        # The marker should go when we toggle the setting.
        repo.set_make_working_trees(True)
        self.assertFalse(t.has('no-working-trees'))
        self.check_databases(t)

    def test_adding_revision_creates_pack_indices(self):
        format = self.get_format()
        tree = self.make_branch_and_tree('.', format=format)
        trans = tree.branch.repository.bzrdir.get_repository_transport(None)
        self.assertEqual([],
            list(self.index_class(trans, 'pack-names', None).iter_all_entries()))
        tree.commit('foobarbaz')
        index = self.index_class(trans, 'pack-names', None)
        index_nodes = list(index.iter_all_entries())
        self.assertEqual(1, len(index_nodes))
        node = index_nodes[0]
        name = node[1][0]
        # the pack sizes should be listed in the index
        pack_value = node[2]
        sizes = [int(digits) for digits in pack_value.split(' ')]
        for size, suffix in zip(sizes, ['.rix', '.iix', '.tix', '.six']):
            stat = trans.stat('indices/%s%s' % (name, suffix))
            self.assertEqual(size, stat.st_size)

    def test_pulling_nothing_leads_to_no_new_names(self):
        format = self.get_format()
        tree1 = self.make_branch_and_tree('1', format=format)
        tree2 = self.make_branch_and_tree('2', format=format)
        tree1.branch.repository.fetch(tree2.branch.repository)
        trans = tree1.branch.repository.bzrdir.get_repository_transport(None)
        self.assertEqual([],
            list(self.index_class(trans, 'pack-names', None).iter_all_entries()))

    def test_commit_across_pack_shape_boundary_autopacks(self):
        format = self.get_format()
        tree = self.make_branch_and_tree('.', format=format)
        trans = tree.branch.repository.bzrdir.get_repository_transport(None)
        # This test could be a little cheaper by replacing the packs
        # attribute on the repository to allow a different pack distribution
        # and max packs policy - so we are checking the policy is honoured
        # in the test. But for now 11 commits is not a big deal in a single
        # test.
        for x in range(9):
            tree.commit('commit %s' % x)
        # there should be 9 packs:
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(9, len(list(index.iter_all_entries())))
        # insert some files in obsolete_packs which should be removed by pack.
        trans.put_bytes('obsolete_packs/foo', '123')
        trans.put_bytes('obsolete_packs/bar', '321')
        # committing one more should coalesce to 1 of 10.
        tree.commit('commit triggering pack')
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(1, len(list(index.iter_all_entries())))
        # packing should not damage data
        tree = tree.bzrdir.open_workingtree()
        check_result = tree.branch.repository.check(
            [tree.branch.last_revision()])
        nb_files = 5 # .pack, .rix, .iix, .tix, .six
        if tree.branch.repository._format.supports_chks:
            nb_files += 1 # .cix
        # We should have 10 x nb_files files in the obsolete_packs directory.
        obsolete_files = list(trans.list_dir('obsolete_packs'))
        self.assertFalse('foo' in obsolete_files)
        self.assertFalse('bar' in obsolete_files)
        self.assertEqual(10 * nb_files, len(obsolete_files))
        # XXX: Todo check packs obsoleted correctly - old packs and indices
        # in the obsolete_packs directory.
        large_pack_name = list(index.iter_all_entries())[0][1][0]
        # finally, committing again should not touch the large pack.
        tree.commit('commit not triggering pack')
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(2, len(list(index.iter_all_entries())))
        pack_names = [node[1][0] for node in index.iter_all_entries()]
        self.assertTrue(large_pack_name in pack_names)

    def test_commit_write_group_returns_new_pack_names(self):
        # This test doesn't need real disk.
        self.vfs_transport_factory = memory.MemoryServer
        format = self.get_format()
        repo = self.make_repository('foo', format=format)
        repo.lock_write()
        try:
            # All current pack repository styles autopack at 10 revisions; and
            # autopack as well as regular commit write group needs to return
            # the new pack name. Looping is a little ugly, but we don't have a
            # clean way to test both the autopack logic and the normal code
            # path without doing this loop.
            for pos in range(10):
                revid = str(pos)
                repo.start_write_group()
                try:
                    inv = inventory.Inventory(revision_id=revid)
                    inv.root.revision = revid
                    repo.texts.add_lines((inv.root.file_id, revid), [], [])
                    rev = _mod_revision.Revision(timestamp=0, timezone=None,
                        committer="Foo Bar <foo@example.com>", message="Message",
                        revision_id=revid)
                    rev.parent_ids = ()
                    repo.add_revision(revid, rev, inv=inv)
                except:
                    repo.abort_write_group()
                    raise
                else:
                    old_names = repo._pack_collection._names.keys()
                    result = repo.commit_write_group()
                    cur_names = repo._pack_collection._names.keys()
                    new_names = list(set(cur_names) - set(old_names))
                    self.assertEqual(new_names, result)
        finally:
            repo.unlock()

    def test_fail_obsolete_deletion(self):
        # failing to delete obsolete packs is not fatal
        format = self.get_format()
        server = test_server.FakeNFSServer()
        self.start_server(server)
        t = transport.get_transport_from_url(server.get_url())
        bzrdir = self.get_format().initialize_on_transport(t)
        repo = bzrdir.create_repository()
        repo_transport = bzrdir.get_repository_transport(None)
        self.assertTrue(repo_transport.has('obsolete_packs'))
        # these files are in use by another client and typically can't be deleted
        repo_transport.put_bytes('obsolete_packs/.nfsblahblah', 'contents')
        repo._pack_collection._clear_obsolete_packs()
        self.assertTrue(repo_transport.has('obsolete_packs/.nfsblahblah'))

    def test_pack_collection_sets_sibling_indices(self):
        """The CombinedGraphIndex objects in the pack collection are all
        siblings of each other, so that search-order reorderings will be copied
        to each other.
        """
        repo = self.make_repository('repo')
        pack_coll = repo._pack_collection
        indices = set([pack_coll.revision_index, pack_coll.inventory_index,
                pack_coll.text_index, pack_coll.signature_index])
        if pack_coll.chk_index is not None:
            indices.add(pack_coll.chk_index)
        combined_indices = set(idx.combined_index for idx in indices)
        for combined_index in combined_indices:
            self.assertEqual(
                combined_indices.difference([combined_index]),
                combined_index._sibling_indices)

    def test_pack_after_two_commits_packs_everything(self):
        format = self.get_format()
        tree = self.make_branch_and_tree('.', format=format)
        trans = tree.branch.repository.bzrdir.get_repository_transport(None)
        tree.commit('start')
        tree.commit('more work')
        tree.branch.repository.pack()
        # there should be 1 pack:
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(1, len(list(index.iter_all_entries())))
        self.assertEqual(2, len(tree.branch.repository.all_revision_ids()))

    def test_pack_preserves_all_inventories(self):
        # This is related to bug:
        #   https://bugs.launchpad.net/bzr/+bug/412198
        # Stacked repositories need to keep the inventory for parents, even
        # after a pack operation. However, it is harder to test that, then just
        # test that all inventory texts are preserved.
        format = self.get_format()
        builder = self.make_branch_builder('source', format=format)
        builder.start_series()
        builder.build_snapshot('A-id', None, [
            ('add', ('', 'root-id', 'directory', None))])
        builder.build_snapshot('B-id', None, [
            ('add', ('file', 'file-id', 'file', 'B content\n'))])
        builder.build_snapshot('C-id', None, [
            ('modify', ('file-id', 'C content\n'))])
        builder.finish_series()
        b = builder.get_branch()
        b.lock_read()
        self.addCleanup(b.unlock)
        repo = self.make_repository('repo', shared=True, format=format)
        repo.lock_write()
        self.addCleanup(repo.unlock)
        repo.fetch(b.repository, revision_id='B-id')
        inv = b.repository.iter_inventories(['C-id']).next()
        repo.start_write_group()
        repo.add_inventory('C-id', inv, ['B-id'])
        repo.commit_write_group()
        self.assertEqual([('A-id',), ('B-id',), ('C-id',)],
                         sorted(repo.inventories.keys()))
        repo.pack()
        self.assertEqual([('A-id',), ('B-id',), ('C-id',)],
                         sorted(repo.inventories.keys()))
        # Content should be preserved as well
        self.assertEqual(inv, repo.iter_inventories(['C-id']).next())

    def test_pack_layout(self):
        # Test that the ordering of revisions in pack repositories is
        # tip->ancestor
        format = self.get_format()
        tree = self.make_branch_and_tree('.', format=format)
        trans = tree.branch.repository.bzrdir.get_repository_transport(None)
        tree.commit('start', rev_id='1')
        tree.commit('more work', rev_id='2')
        tree.branch.repository.pack()
        tree.lock_read()
        self.addCleanup(tree.unlock)
        pack = tree.branch.repository._pack_collection.get_pack_by_name(
            tree.branch.repository._pack_collection.names()[0])
        # revision access tends to be tip->ancestor, so ordering that way on
        # disk is a good idea.
        for _1, key, val, refs in pack.revision_index.iter_all_entries():
            if type(format.repository_format) is RepositoryFormat2a:
                # group_start, group_len, internal_start, internal_len
                pos = map(int, val.split())
            else:
                # eol_flag, start, len
                pos = int(val[1:].split()[0])
            if key == ('1',):
                pos_1 = pos
            else:
                pos_2 = pos
        self.assertTrue(pos_2 < pos_1, 'rev 1 came before rev 2 %s > %s'
                                       % (pos_1, pos_2))

    def test_pack_repositories_support_multiple_write_locks(self):
        format = self.get_format()
        self.make_repository('.', shared=True, format=format)
        r1 = repository.Repository.open('.')
        r2 = repository.Repository.open('.')
        r1.lock_write()
        self.addCleanup(r1.unlock)
        r2.lock_write()
        r2.unlock()

    def _add_text(self, repo, fileid):
        """Add a text to the repository within a write group."""
        repo.texts.add_lines((fileid, 'samplerev+'+fileid), [],
            ['smaplerev+'+fileid])

    def test_concurrent_writers_merge_new_packs(self):
        format = self.get_format()
        self.make_repository('.', shared=True, format=format)
        r1 = repository.Repository.open('.')
        r2 = repository.Repository.open('.')
        r1.lock_write()
        try:
            # access enough data to load the names list
            list(r1.all_revision_ids())
            r2.lock_write()
            try:
                # access enough data to load the names list
                list(r2.all_revision_ids())
                r1.start_write_group()
                try:
                    r2.start_write_group()
                    try:
                        self._add_text(r1, 'fileidr1')
                        self._add_text(r2, 'fileidr2')
                    except:
                        r2.abort_write_group()
                        raise
                except:
                    r1.abort_write_group()
                    raise
                # both r1 and r2 have open write groups with data in them
                # created while the other's write group was open.
                # Commit both which requires a merge to the pack-names.
                try:
                    r1.commit_write_group()
                except:
                    r1.abort_write_group()
                    r2.abort_write_group()
                    raise
                r2.commit_write_group()
                # tell r1 to reload from disk
                r1._pack_collection.reset()
                # Now both repositories should know about both names
                r1._pack_collection.ensure_loaded()
                r2._pack_collection.ensure_loaded()
                self.assertEqual(r1._pack_collection.names(), r2._pack_collection.names())
                self.assertEqual(2, len(r1._pack_collection.names()))
            finally:
                r2.unlock()
        finally:
            r1.unlock()

    def test_concurrent_writer_second_preserves_dropping_a_pack(self):
        format = self.get_format()
        self.make_repository('.', shared=True, format=format)
        r1 = repository.Repository.open('.')
        r2 = repository.Repository.open('.')
        # add a pack to drop
        r1.lock_write()
        try:
            r1.start_write_group()
            try:
                self._add_text(r1, 'fileidr1')
            except:
                r1.abort_write_group()
                raise
            else:
                r1.commit_write_group()
            r1._pack_collection.ensure_loaded()
            name_to_drop = r1._pack_collection.all_packs()[0].name
        finally:
            r1.unlock()
        r1.lock_write()
        try:
            # access enough data to load the names list
            list(r1.all_revision_ids())
            r2.lock_write()
            try:
                # access enough data to load the names list
                list(r2.all_revision_ids())
                r1._pack_collection.ensure_loaded()
                try:
                    r2.start_write_group()
                    try:
                        # in r1, drop the pack
                        r1._pack_collection._remove_pack_from_memory(
                            r1._pack_collection.get_pack_by_name(name_to_drop))
                        # in r2, add a pack
                        self._add_text(r2, 'fileidr2')
                    except:
                        r2.abort_write_group()
                        raise
                except:
                    r1._pack_collection.reset()
                    raise
                # r1 has a changed names list, and r2 an open write groups with
                # changes.
                # save r1, and then commit the r2 write group, which requires a
                # merge to the pack-names, which should not reinstate
                # name_to_drop
                try:
                    r1._pack_collection._save_pack_names()
                    r1._pack_collection.reset()
                except:
                    r2.abort_write_group()
                    raise
                try:
                    r2.commit_write_group()
                except:
                    r2.abort_write_group()
                    raise
                # Now both repositories should now about just one name.
                r1._pack_collection.ensure_loaded()
                r2._pack_collection.ensure_loaded()
                self.assertEqual(r1._pack_collection.names(), r2._pack_collection.names())
                self.assertEqual(1, len(r1._pack_collection.names()))
                self.assertFalse(name_to_drop in r1._pack_collection.names())
            finally:
                r2.unlock()
        finally:
            r1.unlock()

    def test_concurrent_pack_triggers_reload(self):
        # create 2 packs, which we will then collapse
        tree = self.make_branch_and_tree('tree')
        tree.lock_write()
        try:
            rev1 = tree.commit('one')
            rev2 = tree.commit('two')
            r2 = repository.Repository.open('tree')
            r2.lock_read()
            try:
                # Now r2 has read the pack-names file, but will need to reload
                # it after r1 has repacked
                tree.branch.repository.pack()
                self.assertEqual({rev2:(rev1,)}, r2.get_parent_map([rev2]))
            finally:
                r2.unlock()
        finally:
            tree.unlock()

    def test_concurrent_pack_during_get_record_reloads(self):
        tree = self.make_branch_and_tree('tree')
        tree.lock_write()
        try:
            rev1 = tree.commit('one')
            rev2 = tree.commit('two')
            keys = [(rev1,), (rev2,)]
            r2 = repository.Repository.open('tree')
            r2.lock_read()
            try:
                # At this point, we will start grabbing a record stream, and
                # trigger a repack mid-way
                packed = False
                result = {}
                record_stream = r2.revisions.get_record_stream(keys,
                                    'unordered', False)
                for record in record_stream:
                    result[record.key] = record
                    if not packed:
                        tree.branch.repository.pack()
                        packed = True
                # The first record will be found in the original location, but
                # after the pack, we have to reload to find the next record
                self.assertEqual(sorted(keys), sorted(result.keys()))
            finally:
                r2.unlock()
        finally:
            tree.unlock()

    def test_concurrent_pack_during_autopack(self):
        tree = self.make_branch_and_tree('tree')
        tree.lock_write()
        try:
            for i in xrange(9):
                tree.commit('rev %d' % (i,))
            r2 = repository.Repository.open('tree')
            r2.lock_write()
            try:
                # Monkey patch so that pack occurs while the other repo is
                # autopacking. This is slightly bad, but all current pack
                # repository implementations have a _pack_collection, and we
                # test that it gets triggered. So if a future format changes
                # things, the test will fail rather than succeed accidentally.
                autopack_count = [0]
                r1 = tree.branch.repository
                orig = r1._pack_collection.pack_distribution
                def trigger_during_auto(*args, **kwargs):
                    ret = orig(*args, **kwargs)
                    if not autopack_count[0]:
                        r2.pack()
                    autopack_count[0] += 1
                    return ret
                r1._pack_collection.pack_distribution = trigger_during_auto
                tree.commit('autopack-rev')
                # This triggers 2 autopacks. The first one causes r2.pack() to
                # fire, but r2 doesn't see the new pack file yet. The
                # autopack restarts and sees there are 2 files and there
                # should be only 1 for 10 commits. So it goes ahead and
                # finishes autopacking.
                self.assertEqual([2], autopack_count)
            finally:
                r2.unlock()
        finally:
            tree.unlock()

    def test_lock_write_does_not_physically_lock(self):
        repo = self.make_repository('.', format=self.get_format())
        repo.lock_write()
        self.addCleanup(repo.unlock)
        self.assertFalse(repo.get_physical_lock_status())

    def prepare_for_break_lock(self):
        # Setup the global ui factory state so that a break-lock method call
        # will find usable input in the input stream.
        ui.ui_factory = ui.CannedInputUIFactory([True])

    def test_break_lock_breaks_physical_lock(self):
        repo = self.make_repository('.', format=self.get_format())
        repo._pack_collection.lock_names()
        repo.control_files.leave_in_place()
        repo.unlock()
        repo2 = repository.Repository.open('.')
        self.assertTrue(repo.get_physical_lock_status())
        self.prepare_for_break_lock()
        repo2.break_lock()
        self.assertFalse(repo.get_physical_lock_status())

    def test_broken_physical_locks_error_on__unlock_names_lock(self):
        repo = self.make_repository('.', format=self.get_format())
        repo._pack_collection.lock_names()
        self.assertTrue(repo.get_physical_lock_status())
        repo2 = repository.Repository.open('.')
        self.prepare_for_break_lock()
        repo2.break_lock()
        self.assertRaises(errors.LockBroken, repo._pack_collection._unlock_names)

    def test_fetch_without_find_ghosts_ignores_ghosts(self):
        # we want two repositories at this point:
        # one with a revision that is a ghost in the other
        # repository.
        # 'ghost' is present in has_ghost, 'ghost' is absent in 'missing_ghost'.
        # 'references' is present in both repositories, and 'tip' is present
        # just in has_ghost.
        # has_ghost       missing_ghost
        #------------------------------
        # 'ghost'             -
        # 'references'    'references'
        # 'tip'               -
        # In this test we fetch 'tip' which should not fetch 'ghost'
        has_ghost = self.make_repository('has_ghost', format=self.get_format())
        missing_ghost = self.make_repository('missing_ghost',
            format=self.get_format())

        def add_commit(repo, revision_id, parent_ids):
            repo.lock_write()
            repo.start_write_group()
            inv = inventory.Inventory(revision_id=revision_id)
            inv.root.revision = revision_id
            root_id = inv.root.file_id
            sha1 = repo.add_inventory(revision_id, inv, [])
            repo.texts.add_lines((root_id, revision_id), [], [])
            rev = _mod_revision.Revision(timestamp=0,
                                         timezone=None,
                                         committer="Foo Bar <foo@example.com>",
                                         message="Message",
                                         inventory_sha1=sha1,
                                         revision_id=revision_id)
            rev.parent_ids = parent_ids
            repo.add_revision(revision_id, rev)
            repo.commit_write_group()
            repo.unlock()
        add_commit(has_ghost, 'ghost', [])
        add_commit(has_ghost, 'references', ['ghost'])
        add_commit(missing_ghost, 'references', ['ghost'])
        add_commit(has_ghost, 'tip', ['references'])
        missing_ghost.fetch(has_ghost, 'tip')
        # missing ghost now has tip and not ghost.
        rev = missing_ghost.get_revision('tip')
        inv = missing_ghost.get_inventory('tip')
        self.assertRaises(errors.NoSuchRevision,
            missing_ghost.get_revision, 'ghost')
        self.assertRaises(errors.NoSuchRevision,
            missing_ghost.get_inventory, 'ghost')

    def make_write_ready_repo(self):
        format = self.get_format()
        if isinstance(format.repository_format, RepositoryFormat2a):
            raise TestNotApplicable("No missing compression parents")
        repo = self.make_repository('.', format=format)
        repo.lock_write()
        self.addCleanup(repo.unlock)
        repo.start_write_group()
        self.addCleanup(repo.abort_write_group)
        return repo

    def test_missing_inventories_compression_parent_prevents_commit(self):
        repo = self.make_write_ready_repo()
        key = ('junk',)
        repo.inventories._index._missing_compression_parents.add(key)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)

    def test_missing_revisions_compression_parent_prevents_commit(self):
        repo = self.make_write_ready_repo()
        key = ('junk',)
        repo.revisions._index._missing_compression_parents.add(key)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)

    def test_missing_signatures_compression_parent_prevents_commit(self):
        repo = self.make_write_ready_repo()
        key = ('junk',)
        repo.signatures._index._missing_compression_parents.add(key)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)

    def test_missing_text_compression_parent_prevents_commit(self):
        repo = self.make_write_ready_repo()
        key = ('some', 'junk')
        repo.texts._index._missing_compression_parents.add(key)
        self.assertRaises(errors.BzrCheckError, repo.commit_write_group)
        e = self.assertRaises(errors.BzrCheckError, repo.commit_write_group)

    def test_supports_external_lookups(self):
        repo = self.make_repository('.', format=self.get_format())
        self.assertEqual(self.format_supports_external_lookups,
            repo._format.supports_external_lookups)

    def _lock_write(self, write_lockable):
        """Lock write_lockable, add a cleanup and return the result.
        
        :param write_lockable: An object with a lock_write method.
        :return: The result of write_lockable.lock_write().
        """
        result = write_lockable.lock_write()
        self.addCleanup(result.unlock)
        return result

    def test_abort_write_group_does_not_raise_when_suppressed(self):
        """Similar to per_repository.test_write_group's test of the same name.

        Also requires that the exception is logged.
        """
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        # Damage the repository on the filesystem
        self.get_transport('').rename('repo', 'foo')
        # abort_write_group will not raise an error
        self.assertEqual(None, repo.abort_write_group(suppress_errors=True))
        # But it does log an error
        log = self.get_log()
        self.assertContainsRe(log, 'abort_write_group failed')
        self.assertContainsRe(log, r'INFO  bzr: ERROR \(ignored\):')
        if token is not None:
            repo.leave_lock_in_place()

    def test_abort_write_group_does_raise_when_not_suppressed(self):
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        # Damage the repository on the filesystem
        self.get_transport('').rename('repo', 'foo')
        # abort_write_group will not raise an error
        self.assertRaises(Exception, repo.abort_write_group)
        if token is not None:
            repo.leave_lock_in_place()

    def test_suspend_write_group(self):
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        repo.texts.add_lines(('file-id', 'revid'), (), ['lines'])
        wg_tokens = repo.suspend_write_group()
        expected_pack_name = wg_tokens[0] + '.pack'
        expected_names = [wg_tokens[0] + ext for ext in
                            ('.rix', '.iix', '.tix', '.six')]
        if repo.chk_bytes is not None:
            expected_names.append(wg_tokens[0] + '.cix')
        expected_names.append(expected_pack_name)
        upload_transport = repo._pack_collection._upload_transport
        limbo_files = upload_transport.list_dir('')
        self.assertEqual(sorted(expected_names), sorted(limbo_files))
        md5 = osutils.md5(upload_transport.get_bytes(expected_pack_name))
        self.assertEqual(wg_tokens[0], md5.hexdigest())

    def test_resume_chk_bytes(self):
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        if repo.chk_bytes is None:
            raise TestNotApplicable('no chk_bytes for this repository')
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        text = 'a bit of text\n'
        key = ('sha1:' + osutils.sha_string(text),)
        repo.chk_bytes.add_lines(key, (), [text])
        wg_tokens = repo.suspend_write_group()
        same_repo = repo.bzrdir.open_repository()
        same_repo.lock_write()
        self.addCleanup(same_repo.unlock)
        same_repo.resume_write_group(wg_tokens)
        self.assertEqual([key], list(same_repo.chk_bytes.keys()))
        self.assertEqual(
            text, same_repo.chk_bytes.get_record_stream([key],
                'unordered', True).next().get_bytes_as('fulltext'))
        same_repo.abort_write_group()
        self.assertEqual([], list(same_repo.chk_bytes.keys()))

    def test_resume_write_group_then_abort(self):
        # Create a repo, start a write group, insert some data, suspend.
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        text_key = ('file-id', 'revid')
        repo.texts.add_lines(text_key, (), ['lines'])
        wg_tokens = repo.suspend_write_group()
        # Get a fresh repository object for the repo on the filesystem.
        same_repo = repo.bzrdir.open_repository()
        # Resume
        same_repo.lock_write()
        self.addCleanup(same_repo.unlock)
        same_repo.resume_write_group(wg_tokens)
        same_repo.abort_write_group()
        self.assertEqual(
            [], same_repo._pack_collection._upload_transport.list_dir(''))
        self.assertEqual(
            [], same_repo._pack_collection._pack_transport.list_dir(''))

    def test_commit_resumed_write_group(self):
        self.vfs_transport_factory = memory.MemoryServer
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        text_key = ('file-id', 'revid')
        repo.texts.add_lines(text_key, (), ['lines'])
        wg_tokens = repo.suspend_write_group()
        # Get a fresh repository object for the repo on the filesystem.
        same_repo = repo.bzrdir.open_repository()
        # Resume
        same_repo.lock_write()
        self.addCleanup(same_repo.unlock)
        same_repo.resume_write_group(wg_tokens)
        same_repo.commit_write_group()
        expected_pack_name = wg_tokens[0] + '.pack'
        expected_names = [wg_tokens[0] + ext for ext in
                            ('.rix', '.iix', '.tix', '.six')]
        if repo.chk_bytes is not None:
            expected_names.append(wg_tokens[0] + '.cix')
        self.assertEqual(
            [], same_repo._pack_collection._upload_transport.list_dir(''))
        index_names = repo._pack_collection._index_transport.list_dir('')
        self.assertEqual(sorted(expected_names), sorted(index_names))
        pack_names = repo._pack_collection._pack_transport.list_dir('')
        self.assertEqual([expected_pack_name], pack_names)

    def test_resume_malformed_token(self):
        self.vfs_transport_factory = memory.MemoryServer
        # Make a repository with a suspended write group
        repo = self.make_repository('repo', format=self.get_format())
        token = self._lock_write(repo).repository_token
        repo.start_write_group()
        text_key = ('file-id', 'revid')
        repo.texts.add_lines(text_key, (), ['lines'])
        wg_tokens = repo.suspend_write_group()
        # Make a new repository
        new_repo = self.make_repository('new_repo', format=self.get_format())
        token = self._lock_write(new_repo).repository_token
        hacked_wg_token = (
            '../../../../repo/.bzr/repository/upload/' + wg_tokens[0])
        self.assertRaises(
            errors.UnresumableWriteGroup,
            new_repo.resume_write_group, [hacked_wg_token])


class TestPackRepositoryStacking(TestCaseWithTransport):

    """Tests for stacking pack repositories"""

    def setUp(self):
        if not self.format_supports_external_lookups:
            raise TestNotApplicable("%r doesn't support stacking"
                % (self.format_name,))
        super(TestPackRepositoryStacking, self).setUp()

    def get_format(self):
        return controldir.format_registry.make_bzrdir(self.format_name)

    def test_stack_checks_rich_root_compatibility(self):
        # early versions of the packing code relied on pack internals to
        # stack, but the current version should be able to stack on any
        # format.
        #
        # TODO: Possibly this should be run per-repository-format and raise
        # TestNotApplicable on formats that don't support stacking. -- mbp
        # 20080729
        repo = self.make_repository('repo', format=self.get_format())
        if repo.supports_rich_root():
            # can only stack on repositories that have compatible internal
            # metadata
            if getattr(repo._format, 'supports_tree_reference', False):
                matching_format_name = 'pack-0.92-subtree'
            else:
                if repo._format.supports_chks:
                    matching_format_name = '2a'
                else:
                    matching_format_name = 'rich-root-pack'
            mismatching_format_name = 'pack-0.92'
        else:
            # We don't have a non-rich-root CHK format.
            if repo._format.supports_chks:
                raise AssertionError("no non-rich-root CHK formats known")
            else:
                matching_format_name = 'pack-0.92'
            mismatching_format_name = 'pack-0.92-subtree'
        base = self.make_repository('base', format=matching_format_name)
        repo.add_fallback_repository(base)
        # you can't stack on something with incompatible data
        bad_repo = self.make_repository('mismatch',
            format=mismatching_format_name)
        e = self.assertRaises(errors.IncompatibleRepositories,
            repo.add_fallback_repository, bad_repo)
        self.assertContainsRe(str(e),
            r'(?m)KnitPackRepository.*/mismatch/.*\nis not compatible with\n'
            r'.*Repository.*/repo/.*\n'
            r'different rich-root support')

    def test_stack_checks_serializers_compatibility(self):
        repo = self.make_repository('repo', format=self.get_format())
        if getattr(repo._format, 'supports_tree_reference', False):
            # can only stack on repositories that have compatible internal
            # metadata
            matching_format_name = 'pack-0.92-subtree'
            mismatching_format_name = 'rich-root-pack'
        else:
            if repo.supports_rich_root():
                if repo._format.supports_chks:
                    matching_format_name = '2a'
                else:
                    matching_format_name = 'rich-root-pack'
                mismatching_format_name = 'pack-0.92-subtree'
            else:
                raise TestNotApplicable('No formats use non-v5 serializer'
                    ' without having rich-root also set')
        base = self.make_repository('base', format=matching_format_name)
        repo.add_fallback_repository(base)
        # you can't stack on something with incompatible data
        bad_repo = self.make_repository('mismatch',
            format=mismatching_format_name)
        e = self.assertRaises(errors.IncompatibleRepositories,
            repo.add_fallback_repository, bad_repo)
        self.assertContainsRe(str(e),
            r'(?m)KnitPackRepository.*/mismatch/.*\nis not compatible with\n'
            r'.*Repository.*/repo/.*\n'
            r'different serializers')

    def test_adding_pack_does_not_record_pack_names_from_other_repositories(self):
        base = self.make_branch_and_tree('base', format=self.get_format())
        base.commit('foo')
        referencing = self.make_branch_and_tree('repo', format=self.get_format())
        referencing.branch.repository.add_fallback_repository(base.branch.repository)
        local_tree = referencing.branch.create_checkout('local')
        local_tree.commit('bar')
        new_instance = referencing.bzrdir.open_repository()
        new_instance.lock_read()
        self.addCleanup(new_instance.unlock)
        new_instance._pack_collection.ensure_loaded()
        self.assertEqual(1, len(new_instance._pack_collection.all_packs()))

    def test_autopack_only_considers_main_repo_packs(self):
        format = self.get_format()
        base = self.make_branch_and_tree('base', format=format)
        base.commit('foo')
        tree = self.make_branch_and_tree('repo', format=format)
        tree.branch.repository.add_fallback_repository(base.branch.repository)
        trans = tree.branch.repository.bzrdir.get_repository_transport(None)
        # This test could be a little cheaper by replacing the packs
        # attribute on the repository to allow a different pack distribution
        # and max packs policy - so we are checking the policy is honoured
        # in the test. But for now 11 commits is not a big deal in a single
        # test.
        local_tree = tree.branch.create_checkout('local')
        for x in range(9):
            local_tree.commit('commit %s' % x)
        # there should be 9 packs:
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(9, len(list(index.iter_all_entries())))
        # committing one more should coalesce to 1 of 10.
        local_tree.commit('commit triggering pack')
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(1, len(list(index.iter_all_entries())))
        # packing should not damage data
        tree = tree.bzrdir.open_workingtree()
        check_result = tree.branch.repository.check(
            [tree.branch.last_revision()])
        nb_files = 5 # .pack, .rix, .iix, .tix, .six
        if tree.branch.repository._format.supports_chks:
            nb_files += 1 # .cix
        # We should have 10 x nb_files files in the obsolete_packs directory.
        obsolete_files = list(trans.list_dir('obsolete_packs'))
        self.assertFalse('foo' in obsolete_files)
        self.assertFalse('bar' in obsolete_files)
        self.assertEqual(10 * nb_files, len(obsolete_files))
        # XXX: Todo check packs obsoleted correctly - old packs and indices
        # in the obsolete_packs directory.
        large_pack_name = list(index.iter_all_entries())[0][1][0]
        # finally, committing again should not touch the large pack.
        local_tree.commit('commit not triggering pack')
        index = self.index_class(trans, 'pack-names', None)
        self.assertEqual(2, len(list(index.iter_all_entries())))
        pack_names = [node[1][0] for node in index.iter_all_entries()]
        self.assertTrue(large_pack_name in pack_names)


class TestKeyDependencies(TestCaseWithTransport):

    def get_format(self):
        return controldir.format_registry.make_bzrdir(self.format_name)

    def create_source_and_target(self):
        builder = self.make_branch_builder('source', format=self.get_format())
        builder.start_series()
        builder.build_snapshot('A-id', None, [
            ('add', ('', 'root-id', 'directory', None))])
        builder.build_snapshot('B-id', ['A-id', 'ghost-id'], [])
        builder.finish_series()
        repo = self.make_repository('target', format=self.get_format())
        b = builder.get_branch()
        b.lock_read()
        self.addCleanup(b.unlock)
        repo.lock_write()
        self.addCleanup(repo.unlock)
        return b.repository, repo

    def test_key_dependencies_cleared_on_abort(self):
        source_repo, target_repo = self.create_source_and_target()
        target_repo.start_write_group()
        try:
            stream = source_repo.revisions.get_record_stream([('B-id',)],
                                                             'unordered', True)
            target_repo.revisions.insert_record_stream(stream)
            key_refs = target_repo.revisions._index._key_dependencies
            self.assertEqual([('B-id',)], sorted(key_refs.get_referrers()))
        finally:
            target_repo.abort_write_group()
        self.assertEqual([], sorted(key_refs.get_referrers()))

    def test_key_dependencies_cleared_on_suspend(self):
        source_repo, target_repo = self.create_source_and_target()
        target_repo.start_write_group()
        try:
            stream = source_repo.revisions.get_record_stream([('B-id',)],
                                                             'unordered', True)
            target_repo.revisions.insert_record_stream(stream)
            key_refs = target_repo.revisions._index._key_dependencies
            self.assertEqual([('B-id',)], sorted(key_refs.get_referrers()))
        finally:
            target_repo.suspend_write_group()
        self.assertEqual([], sorted(key_refs.get_referrers()))

    def test_key_dependencies_cleared_on_commit(self):
        source_repo, target_repo = self.create_source_and_target()
        target_repo.start_write_group()
        try:
            # Copy all texts, inventories, and chks so that nothing is missing
            # for revision B-id.
            for vf_name in ['texts', 'chk_bytes', 'inventories']:
                source_vf = getattr(source_repo, vf_name, None)
                if source_vf is None:
                    continue
                target_vf = getattr(target_repo, vf_name)
                stream = source_vf.get_record_stream(
                    source_vf.keys(), 'unordered', True)
                target_vf.insert_record_stream(stream)
            # Copy just revision B-id
            stream = source_repo.revisions.get_record_stream(
                [('B-id',)], 'unordered', True)
            target_repo.revisions.insert_record_stream(stream)
            key_refs = target_repo.revisions._index._key_dependencies
            self.assertEqual([('B-id',)], sorted(key_refs.get_referrers()))
        finally:
            target_repo.commit_write_group()
        self.assertEqual([], sorted(key_refs.get_referrers()))


class TestSmartServerAutopack(TestCaseWithTransport):

    def setUp(self):
        super(TestSmartServerAutopack, self).setUp()
        # Create a smart server that publishes whatever the backing VFS server
        # does.
        self.smart_server = test_server.SmartTCPServer_for_testing()
        self.start_server(self.smart_server, self.get_server())
        # Log all HPSS calls into self.hpss_calls.
        client._SmartClient.hooks.install_named_hook(
            'call', self.capture_hpss_call, None)
        self.hpss_calls = []

    def capture_hpss_call(self, params):
        self.hpss_calls.append(params.method)

    def get_format(self):
        return controldir.format_registry.make_bzrdir(self.format_name)

    def test_autopack_or_streaming_rpc_is_used_when_using_hpss(self):
        # Make local and remote repos
        format = self.get_format()
        tree = self.make_branch_and_tree('local', format=format)
        self.make_branch_and_tree('remote', format=format)
        remote_branch_url = self.smart_server.get_url() + 'remote'
        remote_branch = controldir.ControlDir.open(remote_branch_url).open_branch()
        # Make 9 local revisions, and push them one at a time to the remote
        # repo to produce 9 pack files.
        for x in range(9):
            tree.commit('commit %s' % x)
            tree.branch.push(remote_branch)
        # Make one more push to trigger an autopack
        self.hpss_calls = []
        tree.commit('commit triggering pack')
        tree.branch.push(remote_branch)
        autopack_calls = len([call for call in self.hpss_calls if call ==
            'PackRepository.autopack'])
        streaming_calls = len([call for call in self.hpss_calls if call in
            ('Repository.insert_stream', 'Repository.insert_stream_1.19')])
        if autopack_calls:
            # Non streaming server
            self.assertEqual(1, autopack_calls)
            self.assertEqual(0, streaming_calls)
        else:
            # Streaming was used, which autopacks on the remote end.
            self.assertEqual(0, autopack_calls)
            # NB: The 2 calls are because of the sanity check that the server
            # supports the verb (see remote.py:RemoteSink.insert_stream for
            # details).
            self.assertEqual(2, streaming_calls)


def load_tests(basic_tests, module, loader):
    # these give the bzrdir canned format name, and the repository on-disk
    # format string
    scenarios_params = [
         dict(format_name='pack-0.92',
              format_string="Bazaar pack repository format 1 (needs bzr 0.92)\n",
              format_supports_external_lookups=False,
              index_class=GraphIndex),
         dict(format_name='pack-0.92-subtree',
              format_string="Bazaar pack repository format 1 "
              "with subtree support (needs bzr 0.92)\n",
              format_supports_external_lookups=False,
              index_class=GraphIndex),
         dict(format_name='1.6',
              format_string="Bazaar RepositoryFormatKnitPack5 (bzr 1.6)\n",
              format_supports_external_lookups=True,
              index_class=GraphIndex),
         dict(format_name='1.6.1-rich-root',
              format_string="Bazaar RepositoryFormatKnitPack5RichRoot "
                  "(bzr 1.6.1)\n",
              format_supports_external_lookups=True,
              index_class=GraphIndex),
         dict(format_name='1.9',
              format_string="Bazaar RepositoryFormatKnitPack6 (bzr 1.9)\n",
              format_supports_external_lookups=True,
              index_class=BTreeGraphIndex),
         dict(format_name='1.9-rich-root',
              format_string="Bazaar RepositoryFormatKnitPack6RichRoot "
                  "(bzr 1.9)\n",
              format_supports_external_lookups=True,
              index_class=BTreeGraphIndex),
         dict(format_name='2a',
              format_string="Bazaar repository format 2a "
                "(needs bzr 1.16 or later)\n",
              format_supports_external_lookups=True,
              index_class=BTreeGraphIndex),
         ]
    # name of the scenario is the format name
    scenarios = [(s['format_name'], s) for s in scenarios_params]
    return tests.multiply_tests(basic_tests, scenarios, loader.suiteClass())