1 # object_store.py -- Object store for git objects
2 # Copyright (C) 2008-2009 Jelmer Vernooij <jelmer@samba.org>
4 # This program is free software; you can redistribute it and/or
5 # modify it under the terms of the GNU General Public License
6 # as published by the Free Software Foundation; either version 2
7 # or (at your option) a later version of the License.
9 # This program is distributed in the hope that it will be useful,
10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 # GNU General Public License for more details.
14 # You should have received a copy of the GNU General Public License
15 # along with this program; if not, write to the Free Software
16 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
20 """Git object store interfaces and implementation."""
30 from dulwich.errors import (
33 from dulwich.file import GitFile
34 from dulwich.objects import (
43 from dulwich.pack import (
56 class BaseObjectStore(object):
57 """Object store interface."""
59 def determine_wants_all(self, refs):
60 return [sha for (ref, sha) in refs.iteritems() if not sha in self and not ref.endswith("^{}")]
62 def iter_shas(self, shas):
63 """Iterate over the objects for the specified shas.
65 :param shas: Iterable object with SHAs
66 :return: Object iterator
68 return ObjectStoreIterator(self, shas)
70 def contains_loose(self, sha):
71 """Check if a particular object is present by SHA1 and is loose."""
72 raise NotImplementedError(self.contains_loose)
74 def contains_packed(self, sha):
75 """Check if a particular object is present by SHA1 and is packed."""
76 raise NotImplementedError(self.contains_packed)
78 def __contains__(self, sha):
79 """Check if a particular object is present by SHA1.
81 This method makes no distinction between loose and packed objects.
83 return self.contains_packed(sha) or self.contains_loose(sha)
87 """Iterable of pack objects."""
88 raise NotImplementedError
90 def get_raw(self, name):
91 """Obtain the raw text for an object.
93 :param name: sha for the object.
94 :return: tuple with object type and object contents.
96 raise NotImplementedError(self.get_raw)
98 def __getitem__(self, sha):
99 """Obtain an object by SHA1."""
100 type, uncomp = self.get_raw(sha)
101 return ShaFile.from_raw_string(type, uncomp)
104 """Iterate over the SHAs that are present in this store."""
105 raise NotImplementedError(self.__iter__)
107 def add_object(self, obj):
108 """Add a single object to this object store.
111 raise NotImplementedError(self.add_object)
113 def add_objects(self, objects):
114 """Add a set of objects to this object store.
116 :param objects: Iterable over a list of objects.
118 raise NotImplementedError(self.add_objects)
120 def tree_changes(self, source, target, want_unchanged=False):
121 """Find the differences between the contents of two trees
123 :param object_store: Object store to use for retrieving tree contents
124 :param tree: SHA1 of the root tree
125 :param want_unchanged: Whether unchanged files should be reported
126 :return: Iterator over tuples with (oldpath, newpath), (oldmode, newmode), (oldsha, newsha)
128 todo = set([(source, target, "")])
130 (sid, tid, path) = todo.pop()
139 for name, oldmode, oldhexsha in stree.iteritems():
143 oldchildpath = "%s/%s" % (path, name)
145 (newmode, newhexsha) = ttree[name]
146 newchildpath = oldchildpath
151 if (want_unchanged or oldmode != newmode or
152 oldhexsha != newhexsha):
153 if stat.S_ISDIR(oldmode):
154 if newmode is None or stat.S_ISDIR(newmode):
155 todo.add((oldhexsha, newhexsha, oldchildpath))
157 # entry became a file
158 todo.add((oldhexsha, None, oldchildpath))
159 yield ((None, newchildpath), (None, newmode), (None, newhexsha))
161 if newmode is not None and stat.S_ISDIR(newmode):
163 yield ((oldchildpath, None), (oldmode, None), (oldhexsha, None))
164 todo.add((None, newhexsha, newchildpath))
166 yield ((oldchildpath, newchildpath), (oldmode, newmode), (oldhexsha, newhexsha))
168 for name, newmode, newhexsha in ttree.iteritems():
172 childpath = "%s/%s" % (path, name)
173 if not name in stree:
174 if not stat.S_ISDIR(newmode):
175 yield ((None, childpath), (None, newmode), (None, newhexsha))
177 todo.add((None, newhexsha, childpath))
179 def iter_tree_contents(self, tree):
180 """Yield (path, mode, hexsha) tuples for all non-Tree objects in a tree.
182 :param tree: SHA1 of the root of the tree
184 todo = set([(tree, "")])
186 (tid, tpath) = todo.pop()
188 for name, mode, hexsha in tree.iteritems():
192 path = "%s/%s" % (tpath, name)
193 if stat.S_ISDIR(mode):
194 todo.add((hexsha, path))
196 yield path, mode, hexsha
198 def find_missing_objects(self, haves, wants, progress=None,
200 """Find the missing objects required for a set of revisions.
202 :param haves: Iterable over SHAs already in common.
203 :param wants: Iterable over SHAs of objects to fetch.
204 :param progress: Simple progress function that will be called with
205 updated progress strings.
206 :param get_tagged: Function that returns a dict of pointed-to sha -> tag
207 sha for including tags.
208 :return: Iterator over (sha, path) pairs.
210 finder = MissingObjectFinder(self, haves, wants, progress, get_tagged)
211 return iter(finder.next, None)
213 def find_common_revisions(self, graphwalker):
214 """Find which revisions this store has in common using graphwalker.
216 :param graphwalker: A graphwalker object.
217 :return: List of SHAs that are in common
220 sha = graphwalker.next()
225 sha = graphwalker.next()
228 def get_graph_walker(self, heads):
229 """Obtain a graph walker for this object store.
231 :param heads: Local heads to start search with
232 :return: GraphWalker object
234 return ObjectStoreGraphWalker(heads, lambda sha: self[sha].parents)
236 def generate_pack_contents(self, have, want):
237 """Iterate over the contents of a pack file.
239 :param have: List of SHA1s of objects that should not be sent
240 :param want: List of SHA1s of objects that should be sent
242 return self.iter_shas(self.find_missing_objects(have, want))
245 class PackBasedObjectStore(BaseObjectStore):
248 self._pack_cache = None
250 def contains_packed(self, sha):
251 """Check if a particular object is present by SHA1 and is packed."""
252 for pack in self.packs:
257 def _load_packs(self):
258 raise NotImplementedError(self._load_packs)
260 def _pack_cache_stale(self):
261 """Check whether the pack cache is stale."""
262 raise NotImplementedError(self._pack_cache_stale)
264 def _add_known_pack(self, pack):
265 """Add a newly appeared pack to the cache by path.
268 if self._pack_cache is not None:
269 self._pack_cache.append(pack)
273 """List with pack objects."""
274 if self._pack_cache is None or self._pack_cache_stale():
275 self._pack_cache = self._load_packs()
276 return self._pack_cache
278 def _iter_loose_objects(self):
279 raise NotImplementedError(self._iter_loose_objects)
281 def _get_loose_object(self, sha):
282 raise NotImplementedError(self._get_loose_object)
285 """Iterate over the SHAs that are present in this store."""
286 iterables = self.packs + [self._iter_loose_objects()]
287 return itertools.chain(*iterables)
289 def contains_loose(self, sha):
290 """Check if a particular object is present by SHA1 and is loose."""
291 return self._get_loose_object(sha) is not None
293 def get_raw(self, name):
294 """Obtain the raw text for an object.
296 :param name: sha for the object.
297 :return: tuple with object type and object contents.
300 sha = hex_to_sha(name)
302 elif len(name) == 20:
307 for pack in self.packs:
309 return pack.get_raw(sha)
313 hexsha = sha_to_hex(name)
314 ret = self._get_loose_object(hexsha)
316 return ret.type, ret.as_raw_string()
317 raise KeyError(hexsha)
319 def add_objects(self, objects):
320 """Add a set of objects to this object store.
322 :param objects: Iterable over objects, should support __len__.
324 if len(objects) == 0:
325 # Don't bother writing an empty pack file
327 f, commit = self.add_pack()
328 write_pack_data(f, objects, len(objects))
332 class DiskObjectStore(PackBasedObjectStore):
333 """Git-style object store that exists on disk."""
335 def __init__(self, path):
336 """Open an object store.
338 :param path: Path of the object store.
340 super(DiskObjectStore, self).__init__()
342 self.pack_dir = os.path.join(self.path, PACKDIR)
343 if not os.path.isdir(self.pack_dir):
344 os.makedirs(self.pack_dir)
345 self._pack_cache_time = 0
347 def _load_packs(self):
350 self._pack_cache_time = os.stat(self.pack_dir).st_mtime
351 pack_dir_contents = os.listdir(self.pack_dir)
352 for name in pack_dir_contents:
353 # TODO: verify that idx exists first
354 if name.startswith("pack-") and name.endswith(".pack"):
355 filename = os.path.join(self.pack_dir, name)
356 pack_files.append((os.stat(filename).st_mtime, filename))
358 if e.errno == errno.ENOENT:
361 pack_files.sort(reverse=True)
362 suffix_len = len(".pack")
363 return [Pack(f[:-suffix_len]) for _, f in pack_files]
365 def _pack_cache_stale(self):
367 return os.stat(self.pack_dir).st_mtime > self._pack_cache_time
369 if e.errno == errno.ENOENT:
373 def _get_shafile_path(self, sha):
376 # Check from object dir
377 return os.path.join(self.path, dir, file)
379 def _iter_loose_objects(self):
380 for base in os.listdir(self.path):
383 for rest in os.listdir(os.path.join(self.path, base)):
386 def _get_loose_object(self, sha):
387 path = self._get_shafile_path(sha)
389 return ShaFile.from_file(path)
391 if e.errno == errno.ENOENT:
395 def move_in_thin_pack(self, path):
396 """Move a specific file containing a pack into the pack directory.
398 :note: The file should be on the same file system as the
401 :param path: Path to the pack file.
403 data = PackData(path)
405 # Write index for the thin pack (do we really need this?)
406 temppath = os.path.join(self.pack_dir,
407 sha_to_hex(urllib2.randombytes(20))+".tempidx")
408 data.create_index_v2(temppath, self.get_raw)
409 p = Pack.from_objects(data, load_pack_index(temppath))
411 # Write a full pack version
412 temppath = os.path.join(self.pack_dir,
413 sha_to_hex(urllib2.randombytes(20))+".temppack")
414 write_pack(temppath, ((o, None) for o in p.iterobjects(self.get_raw)),
416 pack_sha = load_pack_index(temppath+".idx").objects_sha1()
417 newbasename = os.path.join(self.pack_dir, "pack-%s" % pack_sha)
418 os.rename(temppath+".pack", newbasename+".pack")
419 os.rename(temppath+".idx", newbasename+".idx")
420 self._add_known_pack(Pack(newbasename))
422 def move_in_pack(self, path):
423 """Move a specific file containing a pack into the pack directory.
425 :note: The file should be on the same file system as the
428 :param path: Path to the pack file.
431 entries = p.sorted_entries()
432 basename = os.path.join(self.pack_dir,
433 "pack-%s" % iter_sha1(entry[0] for entry in entries))
434 write_pack_index_v2(basename+".idx", entries, p.get_stored_checksum())
436 os.rename(path, basename + ".pack")
437 self._add_known_pack(Pack(basename))
439 def add_thin_pack(self):
440 """Add a new thin pack to this object store.
442 Thin packs are packs that contain deltas with parents that exist
445 fd, path = tempfile.mkstemp(dir=self.pack_dir, suffix=".pack")
446 f = os.fdopen(fd, 'wb')
450 if os.path.getsize(path) > 0:
451 self.move_in_thin_pack(path)
455 """Add a new pack to this object store.
457 :return: Fileobject to write to and a commit function to
458 call when the pack is finished.
460 fd, path = tempfile.mkstemp(dir=self.pack_dir, suffix=".pack")
461 f = os.fdopen(fd, 'wb')
465 if os.path.getsize(path) > 0:
466 self.move_in_pack(path)
469 def add_object(self, obj):
470 """Add a single object to this object store.
472 :param obj: Object to add
474 dir = os.path.join(self.path, obj.id[:2])
475 if not os.path.isdir(dir):
477 path = os.path.join(dir, obj.id[2:])
478 if os.path.exists(path):
479 return # Already there, no need to write again
480 f = GitFile(path, 'wb')
482 f.write(obj.as_legacy_object())
487 class MemoryObjectStore(BaseObjectStore):
488 """Object store that keeps all objects in memory."""
491 super(MemoryObjectStore, self).__init__()
494 def contains_loose(self, sha):
495 """Check if a particular object is present by SHA1 and is loose."""
496 return sha in self._data
498 def contains_packed(self, sha):
499 """Check if a particular object is present by SHA1 and is packed."""
503 """Iterate over the SHAs that are present in this store."""
504 return self._data.iterkeys()
508 """List with pack objects."""
511 def get_raw(self, name):
512 """Obtain the raw text for an object.
514 :param name: sha for the object.
515 :return: tuple with object type and object contents.
517 return self[name].as_raw_string()
519 def __getitem__(self, name):
520 return self._data[name]
522 def add_object(self, obj):
523 """Add a single object to this object store.
526 self._data[obj.id] = obj
528 def add_objects(self, objects):
529 """Add a set of objects to this object store.
531 :param objects: Iterable over a list of objects.
533 for obj, path in objects:
534 self._data[obj.id] = obj
537 class ObjectImporter(object):
538 """Interface for importing objects."""
540 def __init__(self, count):
541 """Create a new ObjectImporter.
543 :param count: Number of objects that's going to be imported.
547 def add_object(self, object):
549 raise NotImplementedError(self.add_object)
551 def finish(self, object):
552 """Finish the imoprt and write objects to disk."""
553 raise NotImplementedError(self.finish)
556 class ObjectIterator(object):
557 """Interface for iterating over objects."""
559 def iterobjects(self):
560 raise NotImplementedError(self.iterobjects)
563 class ObjectStoreIterator(ObjectIterator):
564 """ObjectIterator that works on top of an ObjectStore."""
566 def __init__(self, store, sha_iter):
567 """Create a new ObjectIterator.
569 :param store: Object store to retrieve from
570 :param sha_iter: Iterator over (sha, path) tuples
573 self.sha_iter = sha_iter
577 """Yield tuple with next object and path."""
578 for sha, path in self.itershas():
579 yield self.store[sha], path
581 def iterobjects(self):
582 """Iterate over just the objects."""
587 """Iterate over the SHAs."""
588 for sha in self._shas:
590 for sha in self.sha_iter:
591 self._shas.append(sha)
594 def __contains__(self, needle):
595 """Check if an object is present.
597 :note: This checks if the object is present in
598 the underlying object store, not if it would
599 be yielded by the iterator.
601 :param needle: SHA1 of the object to check for
603 return needle in self.store
605 def __getitem__(self, key):
606 """Find an object by SHA1.
608 :note: This retrieves the object from the underlying
609 object store. It will also succeed if the object would
610 not be returned by the iterator.
612 return self.store[key]
615 """Return the number of objects."""
616 return len(list(self.itershas()))
619 def tree_lookup_path(lookup_obj, root_sha, path):
620 """Lookup an object in a Git tree.
622 :param lookup_obj: Callback for retrieving object by SHA1
623 :param root_sha: SHA1 of the root tree
624 :param path: Path to lookup
626 parts = path.split("/")
629 obj = lookup_obj(sha)
630 if type(obj) is not Tree:
631 raise NotTreeError(sha)
638 class MissingObjectFinder(object):
639 """Find the objects missing from another object store.
641 :param object_store: Object store containing at least all objects to be
643 :param haves: SHA1s of commits not to send (already present in target)
644 :param wants: SHA1s of commits to send
645 :param progress: Optional function to report progress to.
646 :param get_tagged: Function that returns a dict of pointed-to sha -> tag
647 sha for including tags.
648 :param tagged: dict of pointed-to sha -> tag sha for including tags
651 def __init__(self, object_store, haves, wants, progress=None,
653 self.sha_done = set(haves)
654 self.objects_to_send = set([(w, None, False) for w in wants if w not in haves])
655 self.object_store = object_store
657 self.progress = lambda x: None
659 self.progress = progress
660 self._tagged = get_tagged and get_tagged() or {}
662 def add_todo(self, entries):
663 self.objects_to_send.update([e for e in entries if not e[0] in self.sha_done])
665 def parse_tree(self, tree):
666 self.add_todo([(sha, name, not stat.S_ISDIR(mode)) for (mode, name, sha) in tree.entries() if not S_ISGITLINK(mode)])
668 def parse_commit(self, commit):
669 self.add_todo([(commit.tree, "", False)])
670 self.add_todo([(p, None, False) for p in commit.parents])
672 def parse_tag(self, tag):
673 self.add_todo([(tag.object[1], None, False)])
676 if not self.objects_to_send:
678 (sha, name, leaf) = self.objects_to_send.pop()
680 o = self.object_store[sha]
681 if isinstance(o, Commit):
683 elif isinstance(o, Tree):
685 elif isinstance(o, Tag):
687 if sha in self._tagged:
688 self.add_todo([(self._tagged[sha], None, True)])
689 self.sha_done.add(sha)
690 self.progress("counting objects: %d\r" % len(self.sha_done))
694 class ObjectStoreGraphWalker(object):
695 """Graph walker that finds out what commits are missing from an object store."""
697 def __init__(self, local_heads, get_parents):
698 """Create a new instance.
700 :param local_heads: Heads to start search with
701 :param get_parents: Function for finding the parents of a SHA1.
703 self.heads = set(local_heads)
704 self.get_parents = get_parents
708 """Ack that a particular revision and its ancestors are present in the source."""
709 if sha in self.heads:
710 self.heads.remove(sha)
711 if sha in self.parents:
712 for p in self.parents[sha]:
716 """Iterate over ancestors of heads in the target."""
718 ret = self.heads.pop()
719 ps = self.get_parents(ret)
720 self.parents[ret] = ps
721 self.heads.update(ps)