1
# Copyright (C) 2006 Canonical Ltd
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
"""Read in a bundle stream, and process it into a BundleReader object."""
20
from cStringIO import StringIO
28
from bzrlib.bundle import apply_bundle
29
from bzrlib.errors import (TestamentMismatch, BzrError,
30
MalformedHeader, MalformedPatches, NotABundle)
31
from bzrlib.inventory import (Inventory, InventoryEntry,
32
InventoryDirectory, InventoryFile,
34
from bzrlib.osutils import sha_file, sha_string, pathjoin
35
from bzrlib.revision import Revision, NULL_REVISION
36
from bzrlib.testament import StrictTestament
37
from bzrlib.trace import mutter, warning
38
import bzrlib.transport
39
from bzrlib.tree import Tree
40
import bzrlib.urlutils
41
from bzrlib.xml5 import serializer_v5
44
class RevisionInfo(object):
45
"""Gets filled out for each revision object that is read.
47
def __init__(self, revision_id):
48
self.revision_id = revision_id
54
self.inventory_sha1 = None
56
self.parent_ids = None
59
self.properties = None
60
self.tree_actions = None
63
return pprint.pformat(self.__dict__)
65
def as_revision(self):
66
rev = Revision(revision_id=self.revision_id,
67
committer=self.committer,
68
timestamp=float(self.timestamp),
69
timezone=int(self.timezone),
70
inventory_sha1=self.inventory_sha1,
71
message='\n'.join(self.message))
74
rev.parent_ids.extend(self.parent_ids)
77
for property in self.properties:
78
key_end = property.find(': ')
80
assert property.endswith(':')
81
key = str(property[:-1])
84
key = str(property[:key_end])
85
value = property[key_end+2:]
86
rev.properties[key] = value
91
class BundleInfo(object):
92
"""This contains the meta information. Stuff that allows you to
93
recreate the revision or inventory XML.
100
# A list of RevisionInfo objects
103
# The next entries are created during complete_info() and
104
# other post-read functions.
106
# A list of real Revision objects
107
self.real_revisions = []
109
self.timestamp = None
112
# Have we checked the repository yet?
113
self._validated_revisions_against_repo = False
116
return pprint.pformat(self.__dict__)
118
def complete_info(self):
119
"""This makes sure that all information is properly
120
split up, based on the assumptions that can be made
121
when information is missing.
123
from bzrlib.timestamp import unpack_highres_date
124
# Put in all of the guessable information.
125
if not self.timestamp and self.date:
126
self.timestamp, self.timezone = unpack_highres_date(self.date)
128
self.real_revisions = []
129
for rev in self.revisions:
130
if rev.timestamp is None:
131
if rev.date is not None:
132
rev.timestamp, rev.timezone = \
133
unpack_highres_date(rev.date)
135
rev.timestamp = self.timestamp
136
rev.timezone = self.timezone
137
if rev.message is None and self.message:
138
rev.message = self.message
139
if rev.committer is None and self.committer:
140
rev.committer = self.committer
141
self.real_revisions.append(rev.as_revision())
143
def get_base(self, revision):
144
revision_info = self.get_revision_info(revision.revision_id)
145
if revision_info.base_id is not None:
146
if revision_info.base_id == NULL_REVISION:
149
return revision_info.base_id
150
if len(revision.parent_ids) == 0:
151
# There is no base listed, and
152
# the lowest revision doesn't have a parent
153
# so this is probably against the empty tree
154
# and thus base truly is None
157
return revision.parent_ids[-1]
159
def _get_target(self):
160
"""Return the target revision."""
161
if len(self.real_revisions) > 0:
162
return self.real_revisions[0].revision_id
163
elif len(self.revisions) > 0:
164
return self.revisions[0].revision_id
167
target = property(_get_target, doc='The target revision id')
169
def get_revision(self, revision_id):
170
for r in self.real_revisions:
171
if r.revision_id == revision_id:
173
raise KeyError(revision_id)
175
def get_revision_info(self, revision_id):
176
for r in self.revisions:
177
if r.revision_id == revision_id:
179
raise KeyError(revision_id)
181
def revision_tree(self, repository, revision_id, base=None):
182
revision_id = osutils.safe_revision_id(revision_id)
183
revision = self.get_revision(revision_id)
184
base = self.get_base(revision)
185
assert base != revision_id
186
if not self._validated_revisions_against_repo:
187
self._validate_references_from_repository(repository)
188
revision_info = self.get_revision_info(revision_id)
189
inventory_revision_id = revision_id
190
bundle_tree = BundleTree(repository.revision_tree(base),
191
inventory_revision_id)
192
self._update_tree(bundle_tree, revision_id)
194
inv = bundle_tree.inventory
195
self._validate_inventory(inv, revision_id)
196
self._validate_revision(inv, revision_id)
200
def _validate_references_from_repository(self, repository):
201
"""Now that we have a repository which should have some of the
202
revisions we care about, go through and validate all of them
207
def add_sha(d, revision_id, sha1):
208
if revision_id is None:
210
raise BzrError('A Null revision should always'
211
'have a null sha1 hash')
214
# This really should have been validated as part
215
# of _validate_revisions but lets do it again
216
if sha1 != d[revision_id]:
217
raise BzrError('** Revision %r referenced with 2 different'
218
' sha hashes %s != %s' % (revision_id,
219
sha1, d[revision_id]))
221
d[revision_id] = sha1
223
# All of the contained revisions were checked
224
# in _validate_revisions
226
for rev_info in self.revisions:
227
checked[rev_info.revision_id] = True
228
add_sha(rev_to_sha, rev_info.revision_id, rev_info.sha1)
230
for (rev, rev_info) in zip(self.real_revisions, self.revisions):
231
add_sha(inv_to_sha, rev_info.revision_id, rev_info.inventory_sha1)
235
for revision_id, sha1 in rev_to_sha.iteritems():
236
if repository.has_revision(revision_id):
237
testament = StrictTestament.from_revision(repository,
239
local_sha1 = self._testament_sha1_from_revision(repository,
241
if sha1 != local_sha1:
242
raise BzrError('sha1 mismatch. For revision id {%s}'
243
'local: %s, bundle: %s' % (revision_id, local_sha1, sha1))
246
elif revision_id not in checked:
247
missing[revision_id] = sha1
249
for inv_id, sha1 in inv_to_sha.iteritems():
250
if repository.has_revision(inv_id):
251
# Note: branch.get_inventory_sha1() just returns the value that
252
# is stored in the revision text, and that value may be out
253
# of date. This is bogus, because that means we aren't
254
# validating the actual text, just that we wrote and read the
255
# string. But for now, what the hell.
256
local_sha1 = repository.get_inventory_sha1(inv_id)
257
if sha1 != local_sha1:
258
raise BzrError('sha1 mismatch. For inventory id {%s}'
259
'local: %s, bundle: %s' %
260
(inv_id, local_sha1, sha1))
265
# I don't know if this is an error yet
266
warning('Not all revision hashes could be validated.'
267
' Unable validate %d hashes' % len(missing))
268
mutter('Verified %d sha hashes for the bundle.' % count)
269
self._validated_revisions_against_repo = True
271
def _validate_inventory(self, inv, revision_id):
272
"""At this point we should have generated the BundleTree,
273
so build up an inventory, and make sure the hashes match.
276
assert inv is not None
278
# Now we should have a complete inventory entry.
279
s = serializer_v5.write_inventory_to_string(inv)
281
# Target revision is the last entry in the real_revisions list
282
rev = self.get_revision(revision_id)
283
assert rev.revision_id == revision_id
284
if sha1 != rev.inventory_sha1:
285
open(',,bogus-inv', 'wb').write(s)
286
warning('Inventory sha hash mismatch for revision %s. %s'
287
' != %s' % (revision_id, sha1, rev.inventory_sha1))
289
def _validate_revision(self, inventory, revision_id):
290
"""Make sure all revision entries match their checksum."""
292
# This is a mapping from each revision id to it's sha hash
295
rev = self.get_revision(revision_id)
296
rev_info = self.get_revision_info(revision_id)
297
assert rev.revision_id == rev_info.revision_id
298
assert rev.revision_id == revision_id
299
sha1 = self._testament_sha1(rev, inventory)
300
if sha1 != rev_info.sha1:
301
raise TestamentMismatch(rev.revision_id, rev_info.sha1, sha1)
302
if rev.revision_id in rev_to_sha1:
303
raise BzrError('Revision {%s} given twice in the list'
305
rev_to_sha1[rev.revision_id] = sha1
307
def _update_tree(self, bundle_tree, revision_id):
308
"""This fills out a BundleTree based on the information
311
:param bundle_tree: A BundleTree to update with the new information.
314
def get_rev_id(last_changed, path, kind):
315
if last_changed is not None:
316
# last_changed will be a Unicode string because of how it was
317
# read. Convert it back to utf8.
318
changed_revision_id = osutils.safe_revision_id(last_changed,
321
changed_revision_id = revision_id
322
bundle_tree.note_last_changed(path, changed_revision_id)
323
return changed_revision_id
325
def extra_info(info, new_path):
328
for info_item in info:
330
name, value = info_item.split(':', 1)
332
raise 'Value %r has no colon' % info_item
333
if name == 'last-changed':
335
elif name == 'executable':
336
assert value in ('yes', 'no'), value
337
val = (value == 'yes')
338
bundle_tree.note_executable(new_path, val)
339
elif name == 'target':
340
bundle_tree.note_target(new_path, value)
341
elif name == 'encoding':
343
return last_changed, encoding
345
def do_patch(path, lines, encoding):
346
if encoding is not None:
347
assert encoding == 'base64'
348
patch = base64.decodestring(''.join(lines))
350
patch = ''.join(lines)
351
bundle_tree.note_patch(path, patch)
353
def renamed(kind, extra, lines):
354
info = extra.split(' // ')
356
raise BzrError('renamed action lines need both a from and to'
359
if info[1].startswith('=> '):
360
new_path = info[1][3:]
364
bundle_tree.note_rename(old_path, new_path)
365
last_modified, encoding = extra_info(info[2:], new_path)
366
revision = get_rev_id(last_modified, new_path, kind)
368
do_patch(new_path, lines, encoding)
370
def removed(kind, extra, lines):
371
info = extra.split(' // ')
373
# TODO: in the future we might allow file ids to be
374
# given for removed entries
375
raise BzrError('removed action lines should only have the path'
378
bundle_tree.note_deletion(path)
380
def added(kind, extra, lines):
381
info = extra.split(' // ')
383
raise BzrError('add action lines require the path and file id'
386
raise BzrError('add action lines have fewer than 5 entries.'
389
if not info[1].startswith('file-id:'):
390
raise BzrError('The file-id should follow the path for an add'
392
# This will be Unicode because of how the stream is read. Turn it
393
# back into a utf8 file_id
394
file_id = osutils.safe_file_id(info[1][8:], warn=False)
396
bundle_tree.note_id(file_id, path, kind)
397
# this will be overridden in extra_info if executable is specified.
398
bundle_tree.note_executable(path, False)
399
last_changed, encoding = extra_info(info[2:], path)
400
revision = get_rev_id(last_changed, path, kind)
401
if kind == 'directory':
403
do_patch(path, lines, encoding)
405
def modified(kind, extra, lines):
406
info = extra.split(' // ')
408
raise BzrError('modified action lines have at least'
409
'the path in them: %r' % extra)
412
last_modified, encoding = extra_info(info[1:], path)
413
revision = get_rev_id(last_modified, path, kind)
415
do_patch(path, lines, encoding)
423
for action_line, lines in \
424
self.get_revision_info(revision_id).tree_actions:
425
first = action_line.find(' ')
427
raise BzrError('Bogus action line'
428
' (no opening space): %r' % action_line)
429
second = action_line.find(' ', first+1)
431
raise BzrError('Bogus action line'
432
' (missing second space): %r' % action_line)
433
action = action_line[:first]
434
kind = action_line[first+1:second]
435
if kind not in ('file', 'directory', 'symlink'):
436
raise BzrError('Bogus action line'
437
' (invalid object kind %r): %r' % (kind, action_line))
438
extra = action_line[second+1:]
440
if action not in valid_actions:
441
raise BzrError('Bogus action line'
442
' (unrecognized action): %r' % action_line)
443
valid_actions[action](kind, extra, lines)
445
def install_revisions(self, target_repo):
446
"""Install revisions and return the target revision"""
447
apply_bundle.install_bundle(target_repo, self)
451
class BundleTree(Tree):
452
def __init__(self, base_tree, revision_id):
453
self.base_tree = base_tree
454
self._renamed = {} # Mapping from old_path => new_path
455
self._renamed_r = {} # new_path => old_path
456
self._new_id = {} # new_path => new_id
457
self._new_id_r = {} # new_id => new_path
458
self._kinds = {} # new_id => kind
459
self._last_changed = {} # new_id => revision_id
460
self._executable = {} # new_id => executable value
462
self._targets = {} # new path => new symlink target
464
self.contents_by_id = True
465
self.revision_id = revision_id
466
self._inventory = None
469
return pprint.pformat(self.__dict__)
471
def note_rename(self, old_path, new_path):
472
"""A file/directory has been renamed from old_path => new_path"""
473
assert new_path not in self._renamed
474
assert old_path not in self._renamed_r
475
self._renamed[new_path] = old_path
476
self._renamed_r[old_path] = new_path
478
def note_id(self, new_id, new_path, kind='file'):
479
"""Files that don't exist in base need a new id."""
480
self._new_id[new_path] = new_id
481
self._new_id_r[new_id] = new_path
482
self._kinds[new_id] = kind
484
def note_last_changed(self, file_id, revision_id):
485
if (file_id in self._last_changed
486
and self._last_changed[file_id] != revision_id):
487
raise BzrError('Mismatched last-changed revision for file_id {%s}'
488
': %s != %s' % (file_id,
489
self._last_changed[file_id],
491
self._last_changed[file_id] = revision_id
493
def note_patch(self, new_path, patch):
494
"""There is a patch for a given filename."""
495
self.patches[new_path] = patch
497
def note_target(self, new_path, target):
498
"""The symlink at the new path has the given target"""
499
self._targets[new_path] = target
501
def note_deletion(self, old_path):
502
"""The file at old_path has been deleted."""
503
self.deleted.append(old_path)
505
def note_executable(self, new_path, executable):
506
self._executable[new_path] = executable
508
def old_path(self, new_path):
509
"""Get the old_path (path in the base_tree) for the file at new_path"""
510
assert new_path[:1] not in ('\\', '/')
511
old_path = self._renamed.get(new_path)
512
if old_path is not None:
514
dirname,basename = os.path.split(new_path)
515
# dirname is not '' doesn't work, because
516
# dirname may be a unicode entry, and is
517
# requires the objects to be identical
519
old_dir = self.old_path(dirname)
523
old_path = pathjoin(old_dir, basename)
526
#If the new path wasn't in renamed, the old one shouldn't be in
528
if old_path in self._renamed_r:
532
def new_path(self, old_path):
533
"""Get the new_path (path in the target_tree) for the file at old_path
536
assert old_path[:1] not in ('\\', '/')
537
new_path = self._renamed_r.get(old_path)
538
if new_path is not None:
540
if new_path in self._renamed:
542
dirname,basename = os.path.split(old_path)
544
new_dir = self.new_path(dirname)
548
new_path = pathjoin(new_dir, basename)
551
#If the old path wasn't in renamed, the new one shouldn't be in
553
if new_path in self._renamed:
557
def path2id(self, path):
558
"""Return the id of the file present at path in the target tree."""
559
file_id = self._new_id.get(path)
560
if file_id is not None:
562
old_path = self.old_path(path)
565
if old_path in self.deleted:
567
if getattr(self.base_tree, 'path2id', None) is not None:
568
return self.base_tree.path2id(old_path)
570
return self.base_tree.inventory.path2id(old_path)
572
def id2path(self, file_id):
573
"""Return the new path in the target tree of the file with id file_id"""
574
path = self._new_id_r.get(file_id)
577
old_path = self.base_tree.id2path(file_id)
580
if old_path in self.deleted:
582
return self.new_path(old_path)
584
def old_contents_id(self, file_id):
585
"""Return the id in the base_tree for the given file_id.
586
Return None if the file did not exist in base.
588
if self.contents_by_id:
589
if self.base_tree.has_id(file_id):
593
new_path = self.id2path(file_id)
594
return self.base_tree.path2id(new_path)
596
def get_file(self, file_id):
597
"""Return a file-like object containing the new contents of the
598
file given by file_id.
600
TODO: It might be nice if this actually generated an entry
601
in the text-store, so that the file contents would
604
base_id = self.old_contents_id(file_id)
605
if (base_id is not None and
606
base_id != self.base_tree.inventory.root.file_id):
607
patch_original = self.base_tree.get_file(base_id)
609
patch_original = None
610
file_patch = self.patches.get(self.id2path(file_id))
611
if file_patch is None:
612
if (patch_original is None and
613
self.get_kind(file_id) == 'directory'):
615
assert patch_original is not None, "None: %s" % file_id
616
return patch_original
618
assert not file_patch.startswith('\\'), \
619
'Malformed patch for %s, %r' % (file_id, file_patch)
620
return patched_file(file_patch, patch_original)
622
def get_symlink_target(self, file_id):
623
new_path = self.id2path(file_id)
625
return self._targets[new_path]
627
return self.base_tree.get_symlink_target(file_id)
629
def get_kind(self, file_id):
630
if file_id in self._kinds:
631
return self._kinds[file_id]
632
return self.base_tree.inventory[file_id].kind
634
def is_executable(self, file_id):
635
path = self.id2path(file_id)
636
if path in self._executable:
637
return self._executable[path]
639
return self.base_tree.inventory[file_id].executable
641
def get_last_changed(self, file_id):
642
path = self.id2path(file_id)
643
if path in self._last_changed:
644
return self._last_changed[path]
645
return self.base_tree.inventory[file_id].revision
647
def get_size_and_sha1(self, file_id):
648
"""Return the size and sha1 hash of the given file id.
649
If the file was not locally modified, this is extracted
650
from the base_tree. Rather than re-reading the file.
652
new_path = self.id2path(file_id)
655
if new_path not in self.patches:
656
# If the entry does not have a patch, then the
657
# contents must be the same as in the base_tree
658
ie = self.base_tree.inventory[file_id]
659
if ie.text_size is None:
660
return ie.text_size, ie.text_sha1
661
return int(ie.text_size), ie.text_sha1
662
fileobj = self.get_file(file_id)
663
content = fileobj.read()
664
return len(content), sha_string(content)
666
def _get_inventory(self):
667
"""Build up the inventory entry for the BundleTree.
669
This need to be called before ever accessing self.inventory
671
from os.path import dirname, basename
673
assert self.base_tree is not None
674
base_inv = self.base_tree.inventory
675
inv = Inventory(None, self.revision_id)
677
def add_entry(file_id):
678
path = self.id2path(file_id)
684
parent_path = dirname(path)
685
parent_id = self.path2id(parent_path)
687
kind = self.get_kind(file_id)
688
revision_id = self.get_last_changed(file_id)
690
name = basename(path)
691
if kind == 'directory':
692
ie = InventoryDirectory(file_id, name, parent_id)
694
ie = InventoryFile(file_id, name, parent_id)
695
ie.executable = self.is_executable(file_id)
696
elif kind == 'symlink':
697
ie = InventoryLink(file_id, name, parent_id)
698
ie.symlink_target = self.get_symlink_target(file_id)
699
ie.revision = revision_id
701
if kind in ('directory', 'symlink'):
702
ie.text_size, ie.text_sha1 = None, None
704
ie.text_size, ie.text_sha1 = self.get_size_and_sha1(file_id)
705
if (ie.text_size is None) and (kind == 'file'):
706
raise BzrError('Got a text_size of None for file_id %r' % file_id)
709
sorted_entries = self.sorted_path_id()
710
for path, file_id in sorted_entries:
715
# Have to overload the inherited inventory property
716
# because _get_inventory is only called in the parent.
717
# Reading the docs, property() objects do not use
718
# overloading, they use the function as it was defined
720
inventory = property(_get_inventory)
723
for path, entry in self.inventory.iter_entries():
726
def sorted_path_id(self):
728
for result in self._new_id.iteritems():
730
for id in self.base_tree:
731
path = self.id2path(id)
734
paths.append((path, id))
739
def patched_file(file_patch, original):
740
"""Produce a file-like object with the patched version of a text"""
741
from bzrlib.patches import iter_patched
742
from bzrlib.iterablefile import IterableFile
744
return IterableFile(())
745
# string.splitlines(True) also splits on '\r', but the iter_patched code
746
# only expects to iterate over '\n' style lines
747
return IterableFile(iter_patched(original,
748
StringIO(file_patch).readlines()))
added
removed
bzrlib/bundle/bundle_data.py
