1
# Copyright (C) 2009 Jelmer Vernooij <jelmer@samba.org>
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Map from Git sha's to Bazaar objects."""
19
from dulwich.objects import (
24
from dulwich.object_store import (
35
from bzrlib.revision import (
39
from bzrlib.plugins.git.mapping import (
42
extract_unusual_modes,
46
from bzrlib.plugins.git.cache import (
47
from_repository as cache_from_repository,
54
def get_object_store(repo, mapping=None):
55
git = getattr(repo, "_git", None)
57
return git.object_store
58
return BazaarObjectStore(repo, mapping)
61
MAX_TREE_CACHE_SIZE = 50 * 1024 * 1024
64
class LRUTreeCache(object):
66
def __init__(self, repository):
67
def approx_tree_size(tree):
68
# Very rough estimate, 1k per inventory entry
69
return len(tree.inventory) * 1024
70
self.repository = repository
71
self._cache = lru_cache.LRUSizeCache(max_size=MAX_TREE_CACHE_SIZE,
72
after_cleanup_size=None, compute_size=approx_tree_size)
74
def revision_tree(self, revid):
76
tree = self._cache[revid]
78
tree = self.repository.revision_tree(revid)
80
assert tree.get_revision_id() == tree.inventory.revision_id
83
def iter_revision_trees(self, revids):
88
tree = self._cache[revid]
92
assert tree.get_revision_id() == revid
93
assert tree.inventory.revision_id == revid
95
for tree in self.repository.revision_trees(todo):
96
trees[tree.get_revision_id()] = tree
98
return (trees[r] for r in revids)
100
def revision_trees(self, revids):
101
return list(self.iter_revision_trees(revids))
104
self._cache.add(tree.get_revision_id(), tree)
107
def _find_missing_bzr_revids(get_parent_map, want, have):
108
"""Find the revisions that have to be pushed.
110
:param get_parent_map: Function that returns the parents for a sequence
112
:param want: Revisions the target wants
113
:param have: Revisions the target already has
114
:return: Set of revisions to fetch
116
pending = want - have
120
processed.update(pending)
121
next_map = get_parent_map(pending)
123
for item in next_map.iteritems():
127
next_pending.update(p for p in item[1] if p not in processed)
128
pending = next_pending
129
if NULL_REVISION in todo:
130
todo.remove(NULL_REVISION)
134
def _check_expected_sha(expected_sha, object):
135
"""Check whether an object matches an expected SHA.
137
:param expected_sha: None or expected SHA as either binary or as hex digest
138
:param object: Object to verify
140
if expected_sha is None:
142
if len(expected_sha) == 40:
143
if expected_sha != object.sha().hexdigest():
144
raise AssertionError("Invalid sha for %r: %s" % (object,
146
elif len(expected_sha) == 20:
147
if expected_sha != object.sha().digest():
148
raise AssertionError("Invalid sha for %r: %s" % (object,
149
sha_to_hex(expected_sha)))
151
raise AssertionError("Unknown length %d for %r" % (len(expected_sha),
155
def _tree_to_objects(tree, parent_trees, idmap, unusual_modes,
156
dummy_file_name=None):
157
"""Iterate over the objects that were introduced in a revision.
160
:param parent_trees: Parent revision trees
161
:param unusual_modes: Unusual file modes dictionary
162
:param dummy_file_name: File name to use for dummy files
163
in empty directories. None to skip empty directories
164
:return: Yields (path, object, ie) entries
170
base_tree = parent_trees[0]
171
other_parent_trees = parent_trees[1:]
173
base_tree = tree._repository.revision_tree(NULL_REVISION)
174
other_parent_trees = []
175
def find_unchanged_parent_ie(ie, parent_trees):
176
assert ie.kind in ("symlink", "file")
177
for ptree in parent_trees:
179
pie = ptree.inventory[ie.file_id]
180
except errors.NoSuchId:
183
if (pie.text_sha1 == ie.text_sha1 and
184
pie.kind == ie.kind and
185
pie.symlink_target == ie.symlink_target):
189
# Find all the changed blobs
190
for (file_id, path, changed_content, versioned, parent, name, kind,
191
executable) in tree.iter_changes(base_tree):
192
if kind[1] == "file":
193
ie = tree.inventory[file_id]
196
pie = find_unchanged_parent_ie(ie, other_parent_trees)
201
shamap[ie.file_id] = idmap.lookup_blob_id(
202
pie.file_id, pie.revision)
206
blob.data = tree.get_file_text(ie.file_id)
207
shamap[ie.file_id] = blob.id
208
if not file_id in shamap:
209
new_blobs.append((path[1], ie))
210
new_trees[posixpath.dirname(path[1])] = parent[1]
211
elif kind[1] == "symlink":
212
ie = tree.inventory[file_id]
214
blob = symlink_to_blob(ie)
215
shamap[file_id] = blob.id
217
find_unchanged_parent_ie(ie, other_parent_trees)
219
yield path[1], blob, ie
220
new_trees[posixpath.dirname(path[1])] = parent[1]
221
elif kind[1] not in (None, "directory"):
222
raise AssertionError(kind[1])
223
if path[0] is not None and parent[0] in tree.inventory and tree.inventory[parent[0]].kind == "directory":
225
new_trees[posixpath.dirname(path[0])] = parent[0]
227
# Fetch contents of the blobs that were changed
228
for (path, ie), chunks in tree.iter_files_bytes(
229
[(ie.file_id, (path, ie)) for (path, ie) in new_blobs]):
233
shamap[ie.file_id] = obj.id
235
for path in unusual_modes:
236
parent_path = posixpath.dirname(path)
237
new_trees[parent_path] = tree.path2id(parent_path)
241
items = new_trees.items()
243
for path, file_id in items:
244
parent_id = tree.inventory[file_id].parent_id
245
if parent_id is not None:
246
parent_path = urlutils.dirname(path)
247
new_trees[parent_path] = parent_id
248
trees[path] = file_id
250
def ie_to_hexsha(ie):
252
return shamap[ie.file_id]
254
# FIXME: Should be the same as in parent
255
if ie.kind in ("file", "symlink"):
257
return idmap.lookup_blob_id(ie.file_id, ie.revision)
261
blob.data = tree.get_file_text(ie.file_id)
263
elif ie.kind == "directory":
264
# Not all cache backends store the tree information,
265
# calculate again from scratch
266
ret = directory_to_tree(ie, ie_to_hexsha, unusual_modes,
274
for path in sorted(trees.keys(), reverse=True):
275
ie = tree.inventory[trees[path]]
276
assert ie.kind == "directory"
277
obj = directory_to_tree(ie, ie_to_hexsha, unusual_modes,
281
shamap[ie.file_id] = obj.id
284
class BazaarObjectStore(BaseObjectStore):
285
"""A Git-style object store backed onto a Bazaar repository."""
287
def __init__(self, repository, mapping=None):
288
self.repository = repository
290
self.mapping = default_mapping
292
self.mapping = mapping
293
self._cache = cache_from_repository(repository)
294
self._content_cache_types = ("tree")
295
self.start_write_group = self._cache.idmap.start_write_group
296
self.abort_write_group = self._cache.idmap.abort_write_group
297
self.commit_write_group = self._cache.idmap.commit_write_group
298
self.tree_cache = LRUTreeCache(self.repository)
300
def _update_sha_map(self, stop_revision=None):
301
graph = self.repository.get_graph()
302
if stop_revision is None:
303
heads = graph.heads(self.repository.all_revision_ids())
305
heads = set([stop_revision])
306
missing_revids = self._cache.idmap.missing_revisions(heads)
308
parents = graph.get_parent_map(heads)
310
for p in parents.values():
311
todo.update([x for x in p if x not in missing_revids])
312
heads = self._cache.idmap.missing_revisions(todo)
313
missing_revids.update(heads)
314
if NULL_REVISION in missing_revids:
315
missing_revids.remove(NULL_REVISION)
316
missing_revids = self.repository.has_revisions(missing_revids)
317
if not missing_revids:
319
self.start_write_group()
321
pb = ui.ui_factory.nested_progress_bar()
323
for i, revid in enumerate(graph.iter_topo_order(missing_revids)):
324
trace.mutter('processing %r', revid)
325
pb.update("updating git map", i, len(missing_revids))
326
self._update_sha_map_revision(revid)
330
self.abort_write_group()
333
self.commit_write_group()
336
self._update_sha_map()
337
return iter(self._cache.idmap.sha1s())
339
def _reconstruct_commit(self, rev, tree_sha, roundtrip):
340
def parent_lookup(revid):
342
return self._lookup_revision_sha1(revid)
343
except errors.NoSuchRevision:
345
return self.mapping.export_commit(rev, tree_sha, parent_lookup,
348
def _create_fileid_map_blob(self, inv):
349
# FIXME: This can probably be a lot more efficient,
350
# not all files necessarily have to be processed.
352
for (path, ie) in inv.iter_entries():
353
if self.mapping.generate_file_id(path) != ie.file_id:
354
file_ids[path] = ie.file_id
355
return self.mapping.export_fileid_map(file_ids)
357
def _revision_to_objects(self, rev, tree, roundtrip):
358
"""Convert a revision to a set of git objects.
360
:param rev: Bazaar revision object
361
:param tree: Bazaar revision tree
362
:param roundtrip: Whether to roundtrip all Bazaar revision data
364
unusual_modes = extract_unusual_modes(rev)
365
present_parents = self.repository.has_revisions(rev.parent_ids)
366
parent_trees = self.tree_cache.revision_trees(
367
[p for p in rev.parent_ids if p in present_parents])
369
for path, obj, ie in _tree_to_objects(tree, parent_trees,
370
self._cache.idmap, unusual_modes, self.mapping.BZR_DUMMY_FILE):
374
# Don't yield just yet
377
if root_tree is None:
378
# Pointless commit - get the tree sha elsewhere
379
if not rev.parent_ids:
382
base_sha1 = self._lookup_revision_sha1(rev.parent_ids[0])
383
root_tree = self[self[base_sha1].tree]
384
root_ie = tree.inventory.root
385
if roundtrip and self.mapping.BZR_FILE_IDS_FILE is not None:
386
b = self._create_fileid_map_blob(tree.inventory)
388
root_tree[self.mapping.BZR_FILE_IDS_FILE] = ((stat.S_IFREG | 0644), b.id)
389
yield self.mapping.BZR_FILE_IDS_FILE, b, None
390
yield "", root_tree, root_ie
391
commit_obj = self._reconstruct_commit(rev, root_tree.id,
394
foreign_revid, mapping = mapping_registry.parse_revision_id(
396
except errors.InvalidRevisionId:
399
_check_expected_sha(foreign_revid, commit_obj)
400
yield None, commit_obj, None
402
def _get_updater(self, rev):
403
return self._cache.get_updater(rev)
405
def _update_sha_map_revision(self, revid):
406
rev = self.repository.get_revision(revid)
407
tree = self.tree_cache.revision_tree(rev.revision_id)
408
updater = self._get_updater(rev)
409
for path, obj, ie in self._revision_to_objects(rev, tree,
411
updater.add_object(obj, ie, path)
412
commit_obj = updater.finish()
415
def _reconstruct_blobs(self, keys):
416
"""Return a Git Blob object from a fileid and revision stored in bzr.
418
:param fileid: File id of the text
419
:param revision: Revision of the text
421
stream = self.repository.iter_files_bytes(
422
((key[0], key[1], key) for key in keys))
423
for (fileid, revision, expected_sha), chunks in stream:
425
blob.chunked = chunks
426
if blob.id != expected_sha and blob.data == "":
427
# Perhaps it's a symlink ?
428
tree = self.tree_cache.revision_tree(revision)
429
entry = tree.inventory[fileid]
430
if entry.kind == 'symlink':
431
blob = symlink_to_blob(entry)
432
_check_expected_sha(expected_sha, blob)
435
def _reconstruct_tree(self, fileid, revid, inv, unusual_modes,
437
"""Return a Git Tree object from a file id and a revision stored in bzr.
439
:param fileid: fileid in the tree.
440
:param revision: Revision of the tree.
442
def get_ie_sha1(entry):
443
if entry.kind == "directory":
445
return self._cache.idmap.lookup_tree_id(entry.file_id,
447
except (NotImplementedError, KeyError):
448
obj = self._reconstruct_tree(entry.file_id, revid, inv,
454
elif entry.kind in ("file", "symlink"):
456
return self._cache.idmap.lookup_blob_id(entry.file_id,
460
return self._reconstruct_blobs(
461
[(entry.file_id, entry.revision, None)]).next().id
463
raise AssertionError("unknown entry kind '%s'" % entry.kind)
464
tree = directory_to_tree(inv[fileid], get_ie_sha1, unusual_modes,
465
self.mapping.BZR_DUMMY_FILE)
466
if (inv.root.file_id == fileid and
467
self.mapping.BZR_FILE_IDS_FILE is not None):
468
b = self._create_fileid_map_blob(inv)
469
# If this is the root tree, add the file ids
470
tree[self.mapping.BZR_FILE_IDS_FILE] = ((stat.S_IFREG | 0644), b.id)
471
_check_expected_sha(expected_sha, tree)
474
def get_parents(self, sha):
475
"""Retrieve the parents of a Git commit by SHA1.
477
:param sha: SHA1 of the commit
478
:raises: KeyError, NotCommitError
480
return self[sha].parents
482
def _lookup_revision_sha1(self, revid):
483
"""Return the SHA1 matching a Bazaar revision."""
484
from dulwich.protocol import ZERO_SHA
485
if revid == NULL_REVISION:
488
return self._cache.idmap.lookup_commit(revid)
491
return mapping_registry.parse_revision_id(revid)[0]
492
except errors.InvalidRevisionId:
493
self.repository.lock_read()
495
self._update_sha_map(revid)
497
self.repository.unlock()
498
return self._cache.idmap.lookup_commit(revid)
500
def get_raw(self, sha):
501
"""Get the raw representation of a Git object by SHA1.
503
:param sha: SHA1 of the git object
506
return (obj.type, obj.as_raw_string())
508
def __contains__(self, sha):
509
# See if sha is in map
511
(type, type_data) = self.lookup_git_sha(sha)
513
return self.repository.has_revision(type_data[0])
515
return self.repository.texts.has_key(type_data)
517
return self.repository.has_revision(type_data[1])
519
raise AssertionError("Unknown object type '%s'" % type)
523
def lookup_git_shas(self, shas, update_map=True):
524
from dulwich.protocol import ZERO_SHA
528
ret[sha] = ("commit", (NULL_REVISION, None))
531
ret[sha] = self._cache.idmap.lookup_git_sha(sha)
534
# if not, see if there are any unconverted revisions and add
535
# them to the map, search for sha in map again
536
self._update_sha_map()
539
ret[sha] = self._cache.idmap.lookup_git_sha(sha)
544
def lookup_git_sha(self, sha, update_map=True):
545
return self.lookup_git_shas([sha], update_map=update_map)[sha]
547
def __getitem__(self, sha):
548
if self._cache.content_cache is not None:
550
return self._cache.content_cache[sha]
553
(type, type_data) = self.lookup_git_sha(sha)
554
# convert object to git object
556
(revid, tree_sha) = type_data
558
rev = self.repository.get_revision(revid)
559
except errors.NoSuchRevision:
560
trace.mutter('entry for %s %s in shamap: %r, but not found in '
561
'repository', type, sha, type_data)
563
commit = self._reconstruct_commit(rev, tree_sha, roundtrip=True)
564
_check_expected_sha(sha, commit)
567
(fileid, revision) = type_data
568
return self._reconstruct_blobs([(fileid, revision, sha)]).next()
570
(fileid, revid) = type_data
572
tree = self.tree_cache.revision_tree(revid)
573
rev = self.repository.get_revision(revid)
574
except errors.NoSuchRevision:
575
trace.mutter('entry for %s %s in shamap: %r, but not found in repository', type, sha, type_data)
577
unusual_modes = extract_unusual_modes(rev)
579
return self._reconstruct_tree(fileid, revid, tree.inventory,
580
unusual_modes, expected_sha=sha)
581
except errors.NoSuchRevision:
584
raise AssertionError("Unknown object type '%s'" % type)
586
def generate_lossy_pack_contents(self, have, want, progress=None,
588
return self.generate_pack_contents(have, want, progress, get_tagged,
591
def generate_pack_contents(self, have, want, progress=None,
592
get_tagged=None, lossy=False):
593
"""Iterate over the contents of a pack file.
595
:param have: List of SHA1s of objects that should not be sent
596
:param want: List of SHA1s of objects that should be sent
599
ret = self.lookup_git_shas(have + want)
600
for commit_sha in have:
602
(type, (revid, tree_sha)) = ret[commit_sha]
606
assert type == "commit"
609
for commit_sha in want:
610
if commit_sha in have:
613
(type, (revid, tree_sha)) = ret[commit_sha]
617
assert type == "commit"
620
todo = _find_missing_bzr_revids(self.repository.get_parent_map,
622
trace.mutter('sending revisions %r', todo)
624
pb = ui.ui_factory.nested_progress_bar()
626
for i, revid in enumerate(todo):
627
pb.update("generating git objects", i, len(todo))
628
rev = self.repository.get_revision(revid)
629
tree = self.tree_cache.revision_tree(revid)
630
for path, obj, ie in self._revision_to_objects(rev, tree,
631
roundtrip=not lossy):
632
ret.append((obj, path))
637
def add_thin_pack(self):
640
fd, path = tempfile.mkstemp(suffix=".pack")
641
f = os.fdopen(fd, 'wb')
643
from dulwich.pack import PackData, Pack
644
from bzrlib.plugins.git.fetch import import_git_objects
647
if os.path.getsize(path) == 0:
650
pd.create_index_v2(path[:-5]+".idx", self.object_store.get_raw)
653
self.repository.lock_write()
655
self.repository.start_write_group()
657
import_git_objects(self.repository, self.mapping,
658
p.iterobjects(get_raw=self.get_raw),
661
self.repository.abort_write_group()
664
self.repository.commit_write_group()
666
self.repository.unlock()
669
# The pack isn't kept around anyway, so no point
670
# in treating full packs different from thin packs
671
add_pack = add_thin_pack
added
removed
object_store.py
