/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

ret = []

197

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

198

if name in existing_children:

199

continue

200

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

201

ret.append((c_path, None, base_inv.path2id(c_path), None))

202

if stat.S_ISDIR(mode):

203

ret.extend(remove_disappeared_children(

204

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

205

return ret

206

207

208

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

209

base_inv, parent_id, revision_id, parent_invs,

210

lookup_object, (base_mode, mode), store_updater,

211

lookup_file_id, allow_submodules=False):

212

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

213

214

:param texts: VersionedFiles object to add to

215

:param path: Path in the tree

216

:param tree: A git tree object

217

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

218

:return: Inventory delta for this subtree

219

"""

220

if base_hexsha == hexsha and base_mode == mode:

221

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

222

return [], {}

223

invdelta = []

224

file_id = lookup_file_id(path)

225

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

226

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

227

tree = lookup_object(hexsha)

228

if base_hexsha is None:

229

base_tree = None

230

old_path = None # Newly appeared here

231

else:

232

base_tree = lookup_object(base_hexsha)

233

old_path = path # Renames aren't supported yet

234

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

235

ie.revision = revision_id

236

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

237

texts.insert_record_stream([

238

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

239

# Remember for next time

240

existing_children = set()

241

child_modes = {}

242

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

243

existing_children.add(name)

244

child_path = posixpath.join(path, name)

245

if type(base_tree) is Tree:

246

try:

247

child_base_mode, child_base_hexsha = base_tree[name]

248

except KeyError:

249

child_base_hexsha = None

250

child_base_mode = 0

251

else:

252

child_base_hexsha = None

253

child_base_mode = 0

254

if stat.S_ISDIR(child_mode):

255

subinvdelta, grandchildmodes = import_git_tree(texts, mapping,

256

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

257

file_id, revision_id, parent_invs, lookup_object,

258

(child_base_mode, child_mode), store_updater, lookup_file_id,

259

allow_submodules=allow_submodules)

260

elif S_ISGITLINK(child_mode): # submodule

261

if not allow_submodules:

262

raise SubmodulesRequireSubtrees()

263

subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,

264

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

265

file_id, revision_id, parent_invs, lookup_object,

266

(child_base_mode, child_mode), store_updater, lookup_file_id)

267

else:

268

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

269

(child_base_hexsha, child_hexsha), base_inv, file_id,

270

revision_id, parent_invs, lookup_object,

271

(child_base_mode, child_mode), store_updater, lookup_file_id)

272

grandchildmodes = {}

273

child_modes.update(grandchildmodes)

274

invdelta.extend(subinvdelta)

275

if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,

276

stat.S_IFLNK, DEFAULT_FILE_MODE|0111):

277

child_modes[child_path] = child_mode

278

# Remove any children that have disappeared

279

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

280

invdelta.extend(remove_disappeared_children(base_inv, old_path,

281

base_tree, existing_children, lookup_object))

282

store_updater.add_object(tree, ie)

283

return invdelta, child_modes

284

285

286

def verify_commit_reconstruction(target_git_object_retriever, lookup_object,

287

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

288

new_unusual_modes = mapping.export_unusual_file_modes(rev)

289

if new_unusual_modes != unusual_modes:

290

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

291

unusual_modes, new_unusual_modes))

292

# Verify that we can reconstruct the commit properly

293

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

294

if rec_o != o:

295

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

296

rec_o, o))

297

diff = []

298

new_objs = {}

299

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

300

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

301

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

302

new_objs[path] = obj

303

if obj.id != old_obj_id:

304

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

305

for (path, old_obj, new_obj) in diff:

306

while (old_obj.type_name == "tree" and

307

new_obj.type_name == "tree" and

308

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

309

for name in old_obj:

310

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

311

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

312

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

313

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

314

# Found a differing child, delve deeper

315

path = posixpath.join(path, name)

316

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

317

new_obj = new_objs[path]

318

break

319

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

320

old_obj, new_obj))

321

322

323

def import_git_commit(repo, mapping, head, lookup_object,

324

target_git_object_retriever, trees_cache):

325

o = lookup_object(head)

326

rev = mapping.import_commit(o,

327

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

328

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

329

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

330

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

331

parent_trees = trees_cache.revision_trees(rev.parent_ids)

332

if parent_trees == []:

333

base_inv = Inventory(root_id=None)

334

base_tree = None

335

base_mode = None

336

else:

337

base_inv = parent_trees[0].inventory

338

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

339

base_mode = stat.S_IFDIR

340

store_updater = target_git_object_retriever._get_updater(rev)

341

store_updater.add_object(o, None)

342

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

343

inv_delta, unusual_modes = import_git_tree(repo.texts,

344

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

345

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

346

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

347

lookup_file_id,

348

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

349

store_updater.finish()

350

if unusual_modes != {}:

351

for path, mode in unusual_modes.iteritems():

352

warn_unusual_mode(rev.foreign_revid, path, mode)

353

mapping.import_unusual_file_modes(rev, unusual_modes)

354

try:

355

basis_id = rev.parent_ids[0]

356

except IndexError:

357

basis_id = NULL_REVISION

358

base_inv = None

359

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

360

inv_delta, rev.revision_id, rev.parent_ids,

361

base_inv)

362

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

363

trees_cache.add(ret_tree)

364

repo.add_revision(rev.revision_id, rev)

365

if "verify" in debug.debug_flags:

366

verify_commit_reconstruction(target_git_object_retriever,

367

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

368

unusual_modes)

369

370

371

def import_git_objects(repo, mapping, object_iter,

372

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

373

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

374

375

:param repo: Target Bazaar repository

376

:param mapping: Mapping to use

377

:param object_iter: Iterator over Git objects.

378

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

379

"""

380

def lookup_object(sha):

381

try:

382

return object_iter[sha]

383

except KeyError:

384

return target_git_object_retriever[sha]

385

graph = []

386

checked = set()

387

heads = list(set(heads))

388

trees_cache = LRUTreeCache(repo)

389

# Find and convert commit objects

390

while heads:

391

if pb is not None:

392

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

393

head = heads.pop()

394

assert isinstance(head, str)

395

try:

396

o = lookup_object(head)

397

except KeyError:

398

continue

399

if isinstance(o, Commit):

400

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

401

if repo.has_revision(rev.revision_id):

402

continue

403

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

404

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

405

elif isinstance(o, Tag):

406

if o.object[1] not in checked:

407

heads.append(o.object[1])

408

else:

409

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

410

checked.add(o.id)

411

del checked

412

# Order the revisions

413

# Create the inventory objects

414

batch_size = 1000

415

revision_ids = topo_sort(graph)

416

pack_hints = []

417

if limit is not None:

418

revision_ids = revision_ids[:limit]

419

last_imported = None

420

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

421

target_git_object_retriever.start_write_group()

422

try:

423

repo.start_write_group()

424

try:

425

for i, head in enumerate(

426

revision_ids[offset:offset+batch_size]):

427

if pb is not None:

428

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

429

len(revision_ids))

430

import_git_commit(repo, mapping, head, lookup_object,

431

target_git_object_retriever, trees_cache)

432

last_imported = head

433

except:

434

repo.abort_write_group()

435

raise

436

else:

437

hint = repo.commit_write_group()

438

if hint is not None:

439

pack_hints.extend(hint)

440

except:

441

target_git_object_retriever.abort_write_group()

442

raise

443

else:

444

target_git_object_retriever.commit_write_group()

445

return pack_hints, last_imported

446

447

448

class InterGitRepository(InterRepository):

449

450

_matching_repo_format = GitRepositoryFormat()

451

452

@staticmethod

453

def _get_repo_format_to_test():

454

return None

455

456

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

457

"""See InterRepository.copy_content."""

458

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

459

460

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

461

mapping=None, fetch_spec=None):

462

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

463

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

464

465

466

class InterGitNonGitRepository(InterGitRepository):

467

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

468

repository."""

469

470

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

471

mapping=None, fetch_spec=None):

472

if mapping is None:

473

mapping = self.source.get_mapping()

474

if revision_id is not None:

475

interesting_heads = [revision_id]

476

elif fetch_spec is not None:

477

interesting_heads = fetch_spec.heads

478

else:

479

interesting_heads = None

480

self._refs = {}

481

def determine_wants(refs):

482

self._refs = refs

483

if interesting_heads is None:

484

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

485

else:

486

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

487

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

488

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

489

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

490

self.target.pack(hint=pack_hint)

491

return self._refs

492

493

494

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

495

def report_git_progress(pb, text):

496

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

497

g = _GIT_PROGRESS_RE.match(text)

498

if g is not None:

499

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

500

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

501

else:

502

pb.update(text, 0, 0)

503

504

505

class InterRemoteGitNonGitRepository(InterGitNonGitRepository):

506

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

507

repository."""

508

509

def get_target_heads(self):

510

# FIXME: This should be more efficient

511

all_revs = self.target.all_revision_ids()

512

parent_map = self.target.get_parent_map(all_revs)

513

all_parents = set()

514

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

515

return set(all_revs) - all_parents

516

517

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

518

def progress(text):

519

report_git_progress(pb, text)

520

store = BazaarObjectStore(self.target, mapping)

521

self.target.lock_write()

522

try:

523

heads = self.get_target_heads()

524

graph_walker = store.get_graph_walker(

525

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

526

recorded_wants = []

527

528

def record_determine_wants(heads):

529

wants = determine_wants(heads)

530

recorded_wants.extend(wants)

531

return wants

532

533

create_pb = None

534

if pb is None:

535

create_pb = pb = ui.ui_factory.nested_progress_bar()

536

try:

537

objects_iter = self.source.fetch_objects(record_determine_wants,

538

graph_walker, store.get_raw, progress)

539

return import_git_objects(self.target, mapping,

540

objects_iter, store, recorded_wants, pb, limit)

541

finally:

542

if create_pb:

543

create_pb.finished()

544

finally:

545

self.target.unlock()

546

547

@staticmethod

548

def is_compatible(source, target):

549

"""Be compatible with GitRepository."""

550

return (isinstance(source, RemoteGitRepository) and

551

target.supports_rich_root() and

552

not isinstance(target, GitRepository) and

553

target.texts is not None)

554

555

556

class InterLocalGitNonGitRepository(InterGitNonGitRepository):

557

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

558

repository."""

559

560

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

561

"""Fetch objects.

562

"""

563

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

564

create_pb = None

565

if pb is None:

566

create_pb = pb = ui.ui_factory.nested_progress_bar()

567

target_git_object_retriever = BazaarObjectStore(self.target, mapping)

568

try:

569

self.target.lock_write()

570

try:

571

return import_git_objects(self.target, mapping,

572

self.source._git.object_store,

573

target_git_object_retriever, wants, pb, limit)

574

finally:

575

self.target.unlock()

576

finally:

577

if create_pb:

578

create_pb.finished()

579

580

@staticmethod

581

def is_compatible(source, target):

582

"""Be compatible with GitRepository."""

583

return (isinstance(source, LocalGitRepository) and

584

target.supports_rich_root() and

585

not isinstance(target, GitRepository) and

586

target.texts is not None)

587

588

589

class InterGitGitRepository(InterGitRepository):

590

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

591

592

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

593

def progress(text):

594

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

595

graphwalker = self.target._git.get_graph_walker()

596

if (isinstance(self.source, LocalGitRepository) and

597

isinstance(self.target, LocalGitRepository)):

598

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

599

progress)

600

elif (isinstance(self.source, LocalGitRepository) and

601

isinstance(self.target, RemoteGitRepository)):

602

raise NotImplementedError

603

elif (isinstance(self.source, RemoteGitRepository) and

604

isinstance(self.target, LocalGitRepository)):

605

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

606

try:

607

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

608

graphwalker, f.write, progress)

609

commit()

610

return refs

611

except:

612

f.close()

613

raise

614

else:

615

raise AssertionError

616

617

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

618

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

619

if mapping is None:

620

mapping = self.source.get_mapping()

621

r = self.target._git

622

if revision_id is not None:

623

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

624

elif fetch_spec is not None:

625

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

626

if branches is not None:

627

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

628

elif fetch_spec is None and revision_id is None:

629

determine_wants = r.object_store.determine_wants_all

630

else:

631

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

632

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

633

634

635

@staticmethod

636

def is_compatible(source, target):

637

"""Be compatible with GitRepository."""

638

return (isinstance(source, GitRepository) and

639

isinstance(target, GitRepository))