/brz/remove-bazaar : revision 0.200.1

1

2

#

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.

7

#

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.

12

#

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

16

17

from dulwich.objects import (

18

Commit,

19

Tag,

20

Tree,

21

S_ISGITLINK,

22

)

23

from dulwich.object_store import (

24

tree_lookup_path,

25

)

26

from itertools import (

27

imap,

28

)

29

import posixpath

30

import re

31

import stat

32

33

from bzrlib import (

34

debug,

35

osutils,

36

trace,

37

ui,

38

)

39

from bzrlib.errors import (

40

BzrError,

41

NoSuchId,

42

)

43

from bzrlib.inventory import (

44

Inventory,

45

InventoryDirectory,

46

InventoryFile,

47

InventoryLink,

48

TreeReference,

49

)

50

from bzrlib.repository import (

51

InterRepository,

52

)

53

from bzrlib.revision import (

54

NULL_REVISION,

55

)

56

from bzrlib.revisiontree import (

57

RevisionTree,

58

)

59

from bzrlib.tsort import (

60

topo_sort,

61

)

62

from bzrlib.versionedfile import (

63

ChunkedContentFactory,

64

)

65

66

from bzrlib.plugins.git.mapping import (

67

DEFAULT_FILE_MODE,

68

mode_is_executable,

69

mode_kind,

70

warn_unusual_mode,

71

)

72

from bzrlib.plugins.git.object_store import (

73

BazaarObjectStore,

74

LRUTreeCache,

75

_tree_to_objects,

76

)

77

from bzrlib.plugins.git.remote import (

78

RemoteGitRepository,

79

)

80

from bzrlib.plugins.git.repository import (

81

GitRepository,

82

GitRepositoryFormat,

83

LocalGitRepository,

84

)

85

86

87

def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),

88

base_inv, parent_id, revision_id,

89

parent_invs, lookup_object, (base_mode, mode), store_updater,

90

lookup_file_id):

91

"""Import a git blob object into a bzr repository.

92

93

:param texts: VersionedFiles to add to

94

:param path: Path in the tree

95

:param blob: A git blob

96

:return: Inventory delta for this file

97

"""

98

if mapping.is_control_file(path):

99

return []

100

if base_hexsha == hexsha and base_mode == mode:

101

# If nothing has changed since the base revision, we're done

102

return []

103

file_id = lookup_file_id(path)

104

if stat.S_ISLNK(mode):

105

cls = InventoryLink

106

else:

107

cls = InventoryFile

108

ie = cls(file_id, name.decode("utf-8"), parent_id)

109

ie.executable = mode_is_executable(mode)

110

if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):

111

base_ie = base_inv[base_inv.path2id(path)]

112

ie.text_size = base_ie.text_size

113

ie.text_sha1 = base_ie.text_sha1

114

ie.symlink_target = base_ie.symlink_target

115

if ie.executable == base_ie.executable:

116

ie.revision = base_ie.revision

117

else:

118

blob = lookup_object(hexsha)

119

else:

120

blob = lookup_object(hexsha)

121

if ie.kind == "symlink":

122

ie.revision = None

123

ie.symlink_target = blob.data

124

ie.text_size = None

125

ie.text_sha1 = None

126

else:

127

ie.text_size = sum(imap(len, blob.chunked))

128

ie.text_sha1 = osutils.sha_strings(blob.chunked)

129

# Check what revision we should store

130

parent_keys = []

131

for pinv in parent_invs:

132

try:

133

pie = pinv[file_id]

134

except NoSuchId:

135

continue

136

if (pie.text_sha1 == ie.text_sha1 and

137

pie.executable == ie.executable and

138

pie.symlink_target == ie.symlink_target):

139

# found a revision in one of the parents to use

140

ie.revision = pie.revision

141

break

142

parent_key = (file_id, pie.revision)

143

if not parent_key in parent_keys:

144

parent_keys.append(parent_key)

145

if ie.revision is None:

146

# Need to store a new revision

147

ie.revision = revision_id

148

assert ie.revision is not None

149

if ie.kind == 'symlink':

150

chunks = []

151

else:

152

chunks = blob.chunked

153

texts.insert_record_stream([

154

ChunkedContentFactory((file_id, ie.revision),

155

tuple(parent_keys), ie.text_sha1, chunks)])

156

invdelta = []

157

if base_hexsha is not None:

158

old_path = path # Renames are not supported yet

159

if stat.S_ISDIR(base_mode):

160

invdelta.extend(remove_disappeared_children(base_inv, old_path,

161

lookup_object(base_hexsha), [], lookup_object))

162

else:

163

old_path = None

164

invdelta.append((old_path, path, file_id, ie))

165

if base_hexsha != hexsha:

166

store_updater.add_object(blob, ie)

167

return invdelta

168

169

170

class SubmodulesRequireSubtrees(BzrError):

171

_fmt = """The repository you are fetching from contains submodules. To continue, upgrade your Bazaar repository to a format that supports nested trees, such as 'development-subtree'."""

172

internal = False

173

174

175

def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),

176

base_inv, parent_id, revision_id, parent_invs, lookup_object,

177

(base_mode, mode), store_updater, lookup_file_id):

178

if base_hexsha == hexsha and base_mode == mode:

179

return [], {}

180

file_id = lookup_file_id(path)

181

ie = TreeReference(file_id, name.decode("utf-8"), parent_id)

182

ie.revision = revision_id

183

if base_hexsha is None:

184

oldpath = None

185

else:

186

oldpath = path

187

ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)

188

texts.insert_record_stream([

189

ChunkedContentFactory((file_id, ie.revision), (), None, [])])

190

invdelta = [(oldpath, path, file_id, ie)]

191

return invdelta, {}

192

193

194

def remove_disappeared_children(base_inv, path, base_tree, existing_children,

195

lookup_object):

196

"""Generate an inventory delta for removed children.

197

198

:param base_inv: Base inventory against which to generate the

199

inventory delta.

200

:param path: Path to process

201

:param base_tree: Git Tree base object

202

:param existing_children: Children that still exist

203

:param lookup_object: Lookup a git object by its SHA1

204

:return: Inventory delta, as list

205

"""

206

ret = []

207

for name, mode, hexsha in base_tree.iteritems():

208

if name in existing_children:

209

continue

210

c_path = posixpath.join(path, name.decode("utf-8"))

211

file_id = base_inv.path2id(c_path)

212

assert file_id is not None

213

ret.append((c_path, None, file_id, None))

214

if stat.S_ISDIR(mode):

215

ret.extend(remove_disappeared_children(

216

base_inv, c_path, lookup_object(hexsha), [], lookup_object))

217

return ret

218

219

220

def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),

221

base_inv, parent_id, revision_id, parent_invs,

222

lookup_object, (base_mode, mode), store_updater,

223

lookup_file_id, allow_submodules=False):

224

"""Import a git tree object into a bzr repository.

225

226

:param texts: VersionedFiles object to add to

227

:param path: Path in the tree

228

:param tree: A git tree object

229

:param base_inv: Base inventory against which to return inventory delta

230

:return: Inventory delta for this subtree

231

"""

232

if base_hexsha == hexsha and base_mode == mode:

233

# If nothing has changed since the base revision, we're done

234

return [], {}

235

invdelta = []

236

file_id = lookup_file_id(path)

237

# We just have to hope this is indeed utf-8:

238

ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)

239

tree = lookup_object(hexsha)

240

if base_hexsha is None:

241

base_tree = None

242

old_path = None # Newly appeared here

243

else:

244

base_tree = lookup_object(base_hexsha)

245

old_path = path # Renames aren't supported yet

246

if base_tree is None or type(base_tree) is not Tree:

247

ie.revision = revision_id

248

invdelta.append((old_path, path, ie.file_id, ie))

249

texts.insert_record_stream([

250

ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])

251

# Remember for next time

252

existing_children = set()

253

child_modes = {}

254

for child_mode, name, child_hexsha in tree.entries():

255

existing_children.add(name)

256

child_path = posixpath.join(path, name)

257

if type(base_tree) is Tree:

258

try:

259

child_base_mode, child_base_hexsha = base_tree[name]

260

except KeyError:

261

child_base_hexsha = None

262

child_base_mode = 0

263

else:

264

child_base_hexsha = None

265

child_base_mode = 0

266

if stat.S_ISDIR(child_mode):

267

subinvdelta, grandchildmodes = import_git_tree(texts, mapping,

268

child_path, name, (child_base_hexsha, child_hexsha), base_inv,

269

file_id, revision_id, parent_invs, lookup_object,

270

(child_base_mode, child_mode), store_updater, lookup_file_id,

271

allow_submodules=allow_submodules)

272

elif S_ISGITLINK(child_mode): # submodule

273

if not allow_submodules:

274

raise SubmodulesRequireSubtrees()

275

subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,

276

child_path, name, (child_base_hexsha, child_hexsha), base_inv,

277

file_id, revision_id, parent_invs, lookup_object,

278

(child_base_mode, child_mode), store_updater, lookup_file_id)

279

else:

280

subinvdelta = import_git_blob(texts, mapping, child_path, name,

281

(child_base_hexsha, child_hexsha), base_inv, file_id,

282

revision_id, parent_invs, lookup_object,

283

(child_base_mode, child_mode), store_updater, lookup_file_id)

284

grandchildmodes = {}

285

child_modes.update(grandchildmodes)

286

invdelta.extend(subinvdelta)

287

if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,

288

stat.S_IFLNK, DEFAULT_FILE_MODE|0111):

289

child_modes[child_path] = child_mode

290

# Remove any children that have disappeared

291

if base_tree is not None and type(base_tree) is Tree:

292

invdelta.extend(remove_disappeared_children(base_inv, old_path,

293

base_tree, existing_children, lookup_object))

294

store_updater.add_object(tree, ie)

295

return invdelta, child_modes

296

297

298

def verify_commit_reconstruction(target_git_object_retriever, lookup_object,

299

o, rev, ret_tree, parent_trees, mapping, unusual_modes):

300

new_unusual_modes = mapping.export_unusual_file_modes(rev)

301

if new_unusual_modes != unusual_modes:

302

raise AssertionError("unusual modes don't match: %r != %r" % (

303

unusual_modes, new_unusual_modes))

304

# Verify that we can reconstruct the commit properly

305

rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree, True)

306

if rec_o != o:

307

raise AssertionError("Reconstructed commit differs: %r != %r" % (

308

rec_o, o))

309

diff = []

310

new_objs = {}

311

for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,

312

target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):

313

old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]

314

new_objs[path] = obj

315

if obj.id != old_obj_id:

316

diff.append((path, lookup_object(old_obj_id), obj))

317

for (path, old_obj, new_obj) in diff:

318

while (old_obj.type_name == "tree" and

319

new_obj.type_name == "tree" and

320

sorted(old_obj) == sorted(new_obj)):

321

for name in old_obj:

322

if old_obj[name][0] != new_obj[name][0]:

323

raise AssertionError("Modes for %s differ: %o != %o" %

324

(path, old_obj[name][0], new_obj[name][0]))

325

if old_obj[name][1] != new_obj[name][1]:

326

# Found a differing child, delve deeper

327

path = posixpath.join(path, name)

328

old_obj = lookup_object(old_obj[name][1])

329

new_obj = new_objs[path]

330

break

331

raise AssertionError("objects differ for %s: %r != %r" % (path,

332

old_obj, new_obj))

333

334

335

def import_git_commit(repo, mapping, head, lookup_object,

336

target_git_object_retriever, trees_cache):

337

o = lookup_object(head)

338

rev = mapping.import_commit(o,

339

lambda x: target_git_object_retriever.lookup_git_sha(x)[1][0])

340

# We have to do this here, since we have to walk the tree and

341

# we need to make sure to import the blobs / trees with the right

342

# path; this may involve adding them more than once.

343

parent_trees = trees_cache.revision_trees(rev.parent_ids)

344

if parent_trees == []:

345

base_inv = Inventory(root_id=None)

346

base_tree = None

347

base_mode = None

348

else:

349

base_inv = parent_trees[0].inventory

350

base_tree = lookup_object(o.parents[0]).tree

351

base_mode = stat.S_IFDIR

352

store_updater = target_git_object_retriever._get_updater(rev)

353

store_updater.add_object(o, None)

354

lookup_file_id = mapping.get_fileid_map(lookup_object, o.tree).lookup_file_id

355

inv_delta, unusual_modes = import_git_tree(repo.texts,

356

mapping, "", u"", (base_tree, o.tree), base_inv,

357

None, rev.revision_id, [p.inventory for p in parent_trees],

358

lookup_object, (base_mode, stat.S_IFDIR), store_updater,

359

lookup_file_id,

360

allow_submodules=getattr(repo._format, "supports_tree_reference", False))

361

store_updater.finish()

362

if unusual_modes != {}:

363

for path, mode in unusual_modes.iteritems():

364

warn_unusual_mode(rev.foreign_revid, path, mode)

365

mapping.import_unusual_file_modes(rev, unusual_modes)

366

try:

367

basis_id = rev.parent_ids[0]

368

except IndexError:

369

basis_id = NULL_REVISION

370

base_inv = None

371

rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,

372

inv_delta, rev.revision_id, rev.parent_ids,

373

base_inv)

374

ret_tree = RevisionTree(repo, inv, rev.revision_id)

375

trees_cache.add(ret_tree)

376

repo.add_revision(rev.revision_id, rev)

377

if "verify" in debug.debug_flags:

378

verify_commit_reconstruction(target_git_object_retriever,

379

lookup_object, o, rev, ret_tree, parent_trees, mapping,

380

unusual_modes)

381

382

383

def import_git_objects(repo, mapping, object_iter,

384

target_git_object_retriever, heads, pb=None, limit=None):

385

"""Import a set of git objects into a bzr repository.

386

387

:param repo: Target Bazaar repository

388

:param mapping: Mapping to use

389

:param object_iter: Iterator over Git objects.

390

:return: Tuple with pack hints and last imported revision id

391

"""

392

def lookup_object(sha):

393

try:

394

return object_iter[sha]

395

except KeyError:

396

return target_git_object_retriever[sha]

397

graph = []

398

checked = set()

399

heads = list(set(heads))

400

trees_cache = LRUTreeCache(repo)

401

# Find and convert commit objects

402

while heads:

403

if pb is not None:

404

pb.update("finding revisions to fetch", len(graph), None)

405

head = heads.pop()

406

assert isinstance(head, str)

407

try:

408

o = lookup_object(head)

409

except KeyError:

410

continue

411

if isinstance(o, Commit):

412

rev = mapping.import_commit(o, lambda x: None)

413

if repo.has_revision(rev.revision_id):

414

continue

415

graph.append((o.id, o.parents))

416

heads.extend([p for p in o.parents if p not in checked])

417

elif isinstance(o, Tag):

418

if o.object[1] not in checked:

419

heads.append(o.object[1])

420

else:

421

trace.warning("Unable to import head object %r" % o)

422

checked.add(o.id)

423

del checked

424

# Order the revisions

425

# Create the inventory objects

426

batch_size = 1000

427

revision_ids = topo_sort(graph)

428

pack_hints = []

429

if limit is not None:

430

revision_ids = revision_ids[:limit]

431

last_imported = None

432

for offset in range(0, len(revision_ids), batch_size):

433

target_git_object_retriever.start_write_group()

434

try:

435

repo.start_write_group()

436

try:

437

for i, head in enumerate(

438

revision_ids[offset:offset+batch_size]):

439

if pb is not None:

440

pb.update("fetching revisions", offset+i,

441

len(revision_ids))

442

import_git_commit(repo, mapping, head, lookup_object,

443

target_git_object_retriever, trees_cache)

444

last_imported = head

445

except:

446

repo.abort_write_group()

447

raise

448

else:

449

hint = repo.commit_write_group()

450

if hint is not None:

451

pack_hints.extend(hint)

452

except:

453

target_git_object_retriever.abort_write_group()

454

raise

455

else:

456

target_git_object_retriever.commit_write_group()

457

return pack_hints, last_imported

458

459

460

class InterGitRepository(InterRepository):

461

462

_matching_repo_format = GitRepositoryFormat()

463

464

@staticmethod

465

def _get_repo_format_to_test():

466

return None

467

468

def copy_content(self, revision_id=None, pb=None):

469

"""See InterRepository.copy_content."""

470

self.fetch(revision_id, pb, find_ghosts=False)

471

472

def fetch(self, revision_id=None, pb=None, find_ghosts=False,

473

mapping=None, fetch_spec=None):

474

self.fetch_refs(revision_id=revision_id, pb=pb,

475

find_ghosts=find_ghosts, mapping=mapping, fetch_spec=fetch_spec)

476

477

478

class InterGitNonGitRepository(InterGitRepository):

479

"""Base InterRepository that copies revisions from a Git into a non-Git

480

repository."""

481

482

def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,

483

mapping=None, fetch_spec=None):

484

if mapping is None:

485

mapping = self.source.get_mapping()

486

if revision_id is not None:

487

interesting_heads = [revision_id]

488

elif fetch_spec is not None:

489

interesting_heads = fetch_spec.heads

490

else:

491

interesting_heads = None

492

self._refs = {}

493

def determine_wants(refs):

494

self._refs = refs

495

if interesting_heads is None:

496

ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]

497

else:

498

ret = [self.source.lookup_bzr_revision_id(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]

499

return [rev for rev in ret if not self.target.has_revision(self.source.lookup_foreign_revision_id(rev))]

500

(pack_hint, _) = self.fetch_objects(determine_wants, mapping, pb)

501

if pack_hint is not None and self.target._format.pack_compresses:

502

self.target.pack(hint=pack_hint)

503

return self._refs

504

505

506

_GIT_PROGRESS_RE = re.compile(r"(.*?): +(\d+)% \((\d+)/(\d+)\)")

507

def report_git_progress(pb, text):

508

text = text.rstrip("\r\n")

509

g = _GIT_PROGRESS_RE.match(text)

510

if g is not None:

511

(text, pct, current, total) = g.groups()

512

pb.update(text, int(current), int(total))

513

else:

514

pb.update(text, 0, 0)

515

516

517

class InterRemoteGitNonGitRepository(InterGitNonGitRepository):

518

"""InterRepository that copies revisions from a remote Git into a non-Git

519

repository."""

520

521

def get_target_heads(self):

522

# FIXME: This should be more efficient

523

all_revs = self.target.all_revision_ids()

524

parent_map = self.target.get_parent_map(all_revs)

525

all_parents = set()

526

map(all_parents.update, parent_map.itervalues())

527

return set(all_revs) - all_parents

528

529

def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):

530

def progress(text):

531

report_git_progress(pb, text)

532

store = BazaarObjectStore(self.target, mapping)

533

self.target.lock_write()

534

try:

535

heads = self.get_target_heads()

536

graph_walker = store.get_graph_walker(

537

[store._lookup_revision_sha1(head) for head in heads])

538

recorded_wants = []

539

540

def record_determine_wants(heads):

541

wants = determine_wants(heads)

542

recorded_wants.extend(wants)

543

return wants

544

545

create_pb = None

546

if pb is None:

547

create_pb = pb = ui.ui_factory.nested_progress_bar()

548

try:

549

objects_iter = self.source.fetch_objects(record_determine_wants,

550

graph_walker, store.get_raw, progress)

551

return import_git_objects(self.target, mapping,

552

objects_iter, store, recorded_wants, pb, limit)

553

finally:

554

if create_pb:

555

create_pb.finished()

556

finally:

557

self.target.unlock()

558

559

@staticmethod

560

def is_compatible(source, target):

561

"""Be compatible with GitRepository."""

562

return (isinstance(source, RemoteGitRepository) and

563

target.supports_rich_root() and

564

not isinstance(target, GitRepository) and

565

target.texts is not None)

566

567

568

class InterLocalGitNonGitRepository(InterGitNonGitRepository):

569

"""InterRepository that copies revisions from a local Git into a non-Git

570

repository."""

571

572

def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):

573

"""Fetch objects.

574

"""

575

wants = determine_wants(self.source._git.get_refs())

576

create_pb = None

577

if pb is None:

578

create_pb = pb = ui.ui_factory.nested_progress_bar()

579

target_git_object_retriever = BazaarObjectStore(self.target, mapping)

580

try:

581

self.target.lock_write()

582

try:

583

return import_git_objects(self.target, mapping,

584

self.source._git.object_store,

585

target_git_object_retriever, wants, pb, limit)

586

finally:

587

self.target.unlock()

588

finally:

589

if create_pb:

590

create_pb.finished()

591

592

@staticmethod

593

def is_compatible(source, target):

594

"""Be compatible with GitRepository."""

595

return (isinstance(source, LocalGitRepository) and

596

target.supports_rich_root() and

597

not isinstance(target, GitRepository) and

598

target.texts is not None)

599

600

601

class InterGitGitRepository(InterGitRepository):

602

"""InterRepository that copies between Git repositories."""

603

604

def fetch_objects(self, determine_wants, mapping, pb=None):

605

def progress(text):

606

trace.note("git: %s", text)

607

graphwalker = self.target._git.get_graph_walker()

608

if (isinstance(self.source, LocalGitRepository) and

609

isinstance(self.target, LocalGitRepository)):

610

return self.source._git.fetch(self.target._git, determine_wants,

611

progress)

612

elif (isinstance(self.source, LocalGitRepository) and

613

isinstance(self.target, RemoteGitRepository)):

614

raise NotImplementedError

615

elif (isinstance(self.source, RemoteGitRepository) and

616

isinstance(self.target, LocalGitRepository)):

617

f, commit = self.target._git.object_store.add_thin_pack()

618

try:

619

refs = self.source._git.fetch_pack(determine_wants,

620

graphwalker, f.write, progress)

621

commit()

622

return refs

623

except:

624

f.close()

625

raise

626

else:

627

raise AssertionError

628

629

def fetch_refs(self, revision_id=None, pb=None, find_ghosts=False,

630

mapping=None, fetch_spec=None, branches=None):

631

if mapping is None:

632

mapping = self.source.get_mapping()

633

r = self.target._git

634

if revision_id is not None:

635

args = [mapping.revision_id_bzr_to_foreign(revision_id)[0]]

636

elif fetch_spec is not None:

637

args = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in fetch_spec.heads]

638

if branches is not None:

639

determine_wants = lambda x: [x[y] for y in branches if not x[y] in r.object_store]

640

elif fetch_spec is None and revision_id is None:

641

determine_wants = r.object_store.determine_wants_all

642

else:

643

determine_wants = lambda x: [y for y in args if not y in r.object_store]

644

return self.fetch_objects(determine_wants, mapping)[0]

645

646

647

@staticmethod

648

def is_compatible(source, target):

649

"""Be compatible with GitRepository."""

650

return (isinstance(source, GitRepository) and

651

isinstance(target, GitRepository))