1
# Copyright (C) 2005-2010 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
"""Read in a bundle stream, and process it into a BundleReader object."""
19
from __future__ import absolute_import
22
from io import BytesIO
31
from . import apply_bundle
32
from ..errors import (
36
from ..bzr.inventory import (
42
from ..osutils import sha_string, pathjoin
43
from ..revision import Revision, NULL_REVISION
44
from ..sixish import (
47
from ..bzr.testament import StrictTestament
48
from ..trace import mutter, warning
49
from ..tree import Tree
50
from ..bzr.xml5 import serializer_v5
53
class RevisionInfo(object):
54
"""Gets filled out for each revision object that is read.
57
def __init__(self, revision_id):
58
self.revision_id = revision_id
64
self.inventory_sha1 = None
66
self.parent_ids = None
69
self.properties = None
70
self.tree_actions = None
73
return pprint.pformat(self.__dict__)
75
def as_revision(self):
76
rev = Revision(revision_id=self.revision_id,
77
committer=self.committer,
78
timestamp=float(self.timestamp),
79
timezone=int(self.timezone),
80
inventory_sha1=self.inventory_sha1,
81
message='\n'.join(self.message))
84
rev.parent_ids.extend(self.parent_ids)
87
for property in self.properties:
88
key_end = property.find(': ')
90
if not property.endswith(':'):
91
raise ValueError(property)
92
key = str(property[:-1])
95
key = str(property[:key_end])
96
value = property[key_end + 2:]
97
rev.properties[key] = value
102
def from_revision(revision):
103
revision_info = RevisionInfo(revision.revision_id)
104
date = timestamp.format_highres_date(revision.timestamp,
106
revision_info.date = date
107
revision_info.timezone = revision.timezone
108
revision_info.timestamp = revision.timestamp
109
revision_info.message = revision.message.split('\n')
110
revision_info.properties = [': '.join(p) for p in
111
viewitems(revision.properties)]
115
class BundleInfo(object):
116
"""This contains the meta information. Stuff that allows you to
117
recreate the revision or inventory XML.
120
def __init__(self, bundle_format=None):
121
self.bundle_format = None
122
self.committer = None
126
# A list of RevisionInfo objects
129
# The next entries are created during complete_info() and
130
# other post-read functions.
132
# A list of real Revision objects
133
self.real_revisions = []
135
self.timestamp = None
138
# Have we checked the repository yet?
139
self._validated_revisions_against_repo = False
142
return pprint.pformat(self.__dict__)
144
def complete_info(self):
145
"""This makes sure that all information is properly
146
split up, based on the assumptions that can be made
147
when information is missing.
149
from breezy.timestamp import unpack_highres_date
150
# Put in all of the guessable information.
151
if not self.timestamp and self.date:
152
self.timestamp, self.timezone = unpack_highres_date(self.date)
154
self.real_revisions = []
155
for rev in self.revisions:
156
if rev.timestamp is None:
157
if rev.date is not None:
158
rev.timestamp, rev.timezone = \
159
unpack_highres_date(rev.date)
161
rev.timestamp = self.timestamp
162
rev.timezone = self.timezone
163
if rev.message is None and self.message:
164
rev.message = self.message
165
if rev.committer is None and self.committer:
166
rev.committer = self.committer
167
self.real_revisions.append(rev.as_revision())
169
def get_base(self, revision):
170
revision_info = self.get_revision_info(revision.revision_id)
171
if revision_info.base_id is not None:
172
return revision_info.base_id
173
if len(revision.parent_ids) == 0:
174
# There is no base listed, and
175
# the lowest revision doesn't have a parent
176
# so this is probably against the empty tree
177
# and thus base truly is NULL_REVISION
180
return revision.parent_ids[-1]
182
def _get_target(self):
183
"""Return the target revision."""
184
if len(self.real_revisions) > 0:
185
return self.real_revisions[0].revision_id
186
elif len(self.revisions) > 0:
187
return self.revisions[0].revision_id
190
target = property(_get_target, doc='The target revision id')
192
def get_revision(self, revision_id):
193
for r in self.real_revisions:
194
if r.revision_id == revision_id:
196
raise KeyError(revision_id)
198
def get_revision_info(self, revision_id):
199
for r in self.revisions:
200
if r.revision_id == revision_id:
202
raise KeyError(revision_id)
204
def revision_tree(self, repository, revision_id, base=None):
205
revision = self.get_revision(revision_id)
206
base = self.get_base(revision)
207
if base == revision_id:
208
raise AssertionError()
209
if not self._validated_revisions_against_repo:
210
self._validate_references_from_repository(repository)
211
revision_info = self.get_revision_info(revision_id)
212
inventory_revision_id = revision_id
213
bundle_tree = BundleTree(repository.revision_tree(base),
214
inventory_revision_id)
215
self._update_tree(bundle_tree, revision_id)
217
inv = bundle_tree.inventory
218
self._validate_inventory(inv, revision_id)
219
self._validate_revision(bundle_tree, revision_id)
223
def _validate_references_from_repository(self, repository):
224
"""Now that we have a repository which should have some of the
225
revisions we care about, go through and validate all of them
231
def add_sha(d, revision_id, sha1):
232
if revision_id is None:
234
raise BzrError('A Null revision should always'
235
'have a null sha1 hash')
238
# This really should have been validated as part
239
# of _validate_revisions but lets do it again
240
if sha1 != d[revision_id]:
241
raise BzrError('** Revision %r referenced with 2 different'
242
' sha hashes %s != %s' % (revision_id,
243
sha1, d[revision_id]))
245
d[revision_id] = sha1
247
# All of the contained revisions were checked
248
# in _validate_revisions
250
for rev_info in self.revisions:
251
checked[rev_info.revision_id] = True
252
add_sha(rev_to_sha, rev_info.revision_id, rev_info.sha1)
254
for (rev, rev_info) in zip(self.real_revisions, self.revisions):
255
add_sha(inv_to_sha, rev_info.revision_id, rev_info.inventory_sha1)
259
for revision_id, sha1 in viewitems(rev_to_sha):
260
if repository.has_revision(revision_id):
261
testament = StrictTestament.from_revision(repository,
263
local_sha1 = self._testament_sha1_from_revision(repository,
265
if sha1 != local_sha1:
266
raise BzrError('sha1 mismatch. For revision id {%s}'
267
'local: %s, bundle: %s' % (revision_id, local_sha1, sha1))
270
elif revision_id not in checked:
271
missing[revision_id] = sha1
274
# I don't know if this is an error yet
275
warning('Not all revision hashes could be validated.'
276
' Unable validate %d hashes' % len(missing))
277
mutter('Verified %d sha hashes for the bundle.' % count)
278
self._validated_revisions_against_repo = True
280
def _validate_inventory(self, inv, revision_id):
281
"""At this point we should have generated the BundleTree,
282
so build up an inventory, and make sure the hashes match.
284
# Now we should have a complete inventory entry.
285
s = serializer_v5.write_inventory_to_string(inv)
287
# Target revision is the last entry in the real_revisions list
288
rev = self.get_revision(revision_id)
289
if rev.revision_id != revision_id:
290
raise AssertionError()
291
if sha1 != rev.inventory_sha1:
292
with open(',,bogus-inv', 'wb') as f:
294
warning('Inventory sha hash mismatch for revision %s. %s'
295
' != %s' % (revision_id, sha1, rev.inventory_sha1))
297
def _validate_revision(self, tree, revision_id):
298
"""Make sure all revision entries match their checksum."""
300
# This is a mapping from each revision id to its sha hash
303
rev = self.get_revision(revision_id)
304
rev_info = self.get_revision_info(revision_id)
305
if not (rev.revision_id == rev_info.revision_id):
306
raise AssertionError()
307
if not (rev.revision_id == revision_id):
308
raise AssertionError()
309
sha1 = self._testament_sha1(rev, tree)
310
if sha1 != rev_info.sha1:
311
raise TestamentMismatch(rev.revision_id, rev_info.sha1, sha1)
312
if rev.revision_id in rev_to_sha1:
313
raise BzrError('Revision {%s} given twice in the list'
315
rev_to_sha1[rev.revision_id] = sha1
317
def _update_tree(self, bundle_tree, revision_id):
318
"""This fills out a BundleTree based on the information
321
:param bundle_tree: A BundleTree to update with the new information.
324
def get_rev_id(last_changed, path, kind):
325
if last_changed is not None:
326
# last_changed will be a Unicode string because of how it was
327
# read. Convert it back to utf8.
328
changed_revision_id = cache_utf8.encode(last_changed)
330
changed_revision_id = revision_id
331
bundle_tree.note_last_changed(path, changed_revision_id)
332
return changed_revision_id
334
def extra_info(info, new_path):
337
for info_item in info:
339
name, value = info_item.split(':', 1)
341
raise ValueError('Value %r has no colon' % info_item)
342
if name == 'last-changed':
344
elif name == 'executable':
345
val = (value == 'yes')
346
bundle_tree.note_executable(new_path, val)
347
elif name == 'target':
348
bundle_tree.note_target(new_path, value)
349
elif name == 'encoding':
351
return last_changed, encoding
353
def do_patch(path, lines, encoding):
354
if encoding == 'base64':
355
patch = base64.b64decode(b''.join(lines))
356
elif encoding is None:
357
patch = b''.join(lines)
359
raise ValueError(encoding)
360
bundle_tree.note_patch(path, patch)
362
def renamed(kind, extra, lines):
363
info = extra.split(' // ')
365
raise BzrError('renamed action lines need both a from and to'
368
if info[1].startswith('=> '):
369
new_path = info[1][3:]
373
bundle_tree.note_rename(old_path, new_path)
374
last_modified, encoding = extra_info(info[2:], new_path)
375
revision = get_rev_id(last_modified, new_path, kind)
377
do_patch(new_path, lines, encoding)
379
def removed(kind, extra, lines):
380
info = extra.split(' // ')
382
# TODO: in the future we might allow file ids to be
383
# given for removed entries
384
raise BzrError('removed action lines should only have the path'
387
bundle_tree.note_deletion(path)
389
def added(kind, extra, lines):
390
info = extra.split(' // ')
392
raise BzrError('add action lines require the path and file id'
395
raise BzrError('add action lines have fewer than 5 entries.'
398
if not info[1].startswith('file-id:'):
399
raise BzrError('The file-id should follow the path for an add'
401
# This will be Unicode because of how the stream is read. Turn it
402
# back into a utf8 file_id
403
file_id = cache_utf8.encode(info[1][8:])
405
bundle_tree.note_id(file_id, path, kind)
406
# this will be overridden in extra_info if executable is specified.
407
bundle_tree.note_executable(path, False)
408
last_changed, encoding = extra_info(info[2:], path)
409
revision = get_rev_id(last_changed, path, kind)
410
if kind == 'directory':
412
do_patch(path, lines, encoding)
414
def modified(kind, extra, lines):
415
info = extra.split(' // ')
417
raise BzrError('modified action lines have at least'
418
'the path in them: %r' % extra)
421
last_modified, encoding = extra_info(info[1:], path)
422
revision = get_rev_id(last_modified, path, kind)
424
do_patch(path, lines, encoding)
432
for action_line, lines in \
433
self.get_revision_info(revision_id).tree_actions:
434
first = action_line.find(' ')
436
raise BzrError('Bogus action line'
437
' (no opening space): %r' % action_line)
438
second = action_line.find(' ', first + 1)
440
raise BzrError('Bogus action line'
441
' (missing second space): %r' % action_line)
442
action = action_line[:first]
443
kind = action_line[first + 1:second]
444
if kind not in ('file', 'directory', 'symlink'):
445
raise BzrError('Bogus action line'
446
' (invalid object kind %r): %r' % (kind, action_line))
447
extra = action_line[second + 1:]
449
if action not in valid_actions:
450
raise BzrError('Bogus action line'
451
' (unrecognized action): %r' % action_line)
452
valid_actions[action](kind, extra, lines)
454
def install_revisions(self, target_repo, stream_input=True):
455
"""Install revisions and return the target revision
457
:param target_repo: The repository to install into
458
:param stream_input: Ignored by this implementation.
460
apply_bundle.install_bundle(target_repo, self)
463
def get_merge_request(self, target_repo):
464
"""Provide data for performing a merge
466
Returns suggested base, suggested target, and patch verification status
468
return None, self.target, 'inapplicable'
471
class BundleTree(Tree):
473
def __init__(self, base_tree, revision_id):
474
self.base_tree = base_tree
475
self._renamed = {} # Mapping from old_path => new_path
476
self._renamed_r = {} # new_path => old_path
477
self._new_id = {} # new_path => new_id
478
self._new_id_r = {} # new_id => new_path
479
self._kinds = {} # new_path => kind
480
self._last_changed = {} # new_id => revision_id
481
self._executable = {} # new_id => executable value
483
self._targets = {} # new path => new symlink target
485
self.contents_by_id = True
486
self.revision_id = revision_id
487
self._inventory = None
490
return pprint.pformat(self.__dict__)
492
def note_rename(self, old_path, new_path):
493
"""A file/directory has been renamed from old_path => new_path"""
494
if new_path in self._renamed:
495
raise AssertionError(new_path)
496
if old_path in self._renamed_r:
497
raise AssertionError(old_path)
498
self._renamed[new_path] = old_path
499
self._renamed_r[old_path] = new_path
501
def note_id(self, new_id, new_path, kind='file'):
502
"""Files that don't exist in base need a new id."""
503
self._new_id[new_path] = new_id
504
self._new_id_r[new_id] = new_path
505
self._kinds[new_path] = kind
507
def note_last_changed(self, file_id, revision_id):
508
if (file_id in self._last_changed
509
and self._last_changed[file_id] != revision_id):
510
raise BzrError('Mismatched last-changed revision for file_id {%s}'
511
': %s != %s' % (file_id,
512
self._last_changed[file_id],
514
self._last_changed[file_id] = revision_id
516
def note_patch(self, new_path, patch):
517
"""There is a patch for a given filename."""
518
self.patches[new_path] = patch
520
def note_target(self, new_path, target):
521
"""The symlink at the new path has the given target"""
522
self._targets[new_path] = target
524
def note_deletion(self, old_path):
525
"""The file at old_path has been deleted."""
526
self.deleted.append(old_path)
528
def note_executable(self, new_path, executable):
529
self._executable[new_path] = executable
531
def old_path(self, new_path):
532
"""Get the old_path (path in the base_tree) for the file at new_path"""
533
if new_path[:1] in ('\\', '/'):
534
raise ValueError(new_path)
535
old_path = self._renamed.get(new_path)
536
if old_path is not None:
538
dirname, basename = os.path.split(new_path)
539
# dirname is not '' doesn't work, because
540
# dirname may be a unicode entry, and is
541
# requires the objects to be identical
543
old_dir = self.old_path(dirname)
547
old_path = pathjoin(old_dir, basename)
550
# If the new path wasn't in renamed, the old one shouldn't be in
552
if old_path in self._renamed_r:
556
def new_path(self, old_path):
557
"""Get the new_path (path in the target_tree) for the file at old_path
560
if old_path[:1] in ('\\', '/'):
561
raise ValueError(old_path)
562
new_path = self._renamed_r.get(old_path)
563
if new_path is not None:
565
if new_path in self._renamed:
567
dirname, basename = os.path.split(old_path)
569
new_dir = self.new_path(dirname)
573
new_path = pathjoin(new_dir, basename)
576
# If the old path wasn't in renamed, the new one shouldn't be in
578
if new_path in self._renamed:
582
def get_root_id(self):
583
return self.path2id('')
585
def path2id(self, path):
586
"""Return the id of the file present at path in the target tree."""
587
file_id = self._new_id.get(path)
588
if file_id is not None:
590
old_path = self.old_path(path)
593
if old_path in self.deleted:
595
return self.base_tree.path2id(old_path)
597
def id2path(self, file_id):
598
"""Return the new path in the target tree of the file with id file_id"""
599
path = self._new_id_r.get(file_id)
602
old_path = self.base_tree.id2path(file_id)
605
if old_path in self.deleted:
607
return self.new_path(old_path)
609
def old_contents_id(self, file_id):
610
"""Return the id in the base_tree for the given file_id.
611
Return None if the file did not exist in base.
613
if self.contents_by_id:
614
if self.base_tree.has_id(file_id):
618
new_path = self.id2path(file_id)
619
return self.base_tree.path2id(new_path)
621
def get_file(self, path):
622
"""Return a file-like object containing the new contents of the
623
file given by file_id.
625
TODO: It might be nice if this actually generated an entry
626
in the text-store, so that the file contents would
629
file_id = self.path2id(path)
630
base_id = self.old_contents_id(file_id)
631
if (base_id is not None and
632
base_id != self.base_tree.get_root_id()):
633
old_path = self.base_tree.id2path(base_id)
634
patch_original = self.base_tree.get_file(old_path)
636
patch_original = None
637
file_patch = self.patches.get(path)
638
if file_patch is None:
639
if (patch_original is None and
640
self.kind(path) == 'directory'):
642
if patch_original is None:
643
raise AssertionError("None: %s" % file_id)
644
return patch_original
646
if file_patch.startswith(b'\\'):
648
'Malformed patch for %s, %r' % (file_id, file_patch))
649
return patched_file(file_patch, patch_original)
651
def get_symlink_target(self, path):
653
return self._targets[path]
655
old_path = self.old_path(path)
656
return self.base_tree.get_symlink_target(old_path)
658
def kind(self, path):
660
return self._kinds[path]
662
old_path = self.old_path(path)
663
return self.base_tree.kind(old_path)
665
def get_file_revision(self, path):
666
if path in self._last_changed:
667
return self._last_changed[path]
669
old_path = self.old_path(path)
670
return self.base_tree.get_file_revision(old_path)
672
def is_executable(self, path):
673
if path in self._executable:
674
return self._executable[path]
676
old_path = self.old_path(path)
677
return self.base_tree.is_executable(old_path)
679
def get_last_changed(self, path):
680
if path in self._last_changed:
681
return self._last_changed[path]
682
old_path = self.old_path(path)
683
return self.base_tree.get_file_revision(old_path)
685
def get_size_and_sha1(self, new_path):
686
"""Return the size and sha1 hash of the given file id.
687
If the file was not locally modified, this is extracted
688
from the base_tree. Rather than re-reading the file.
692
if new_path not in self.patches:
693
# If the entry does not have a patch, then the
694
# contents must be the same as in the base_tree
695
base_path = self.old_path(new_path)
696
text_size = self.base_tree.get_file_size(base_path)
697
text_sha1 = self.base_tree.get_file_sha1(base_path)
698
return text_size, text_sha1
699
fileobj = self.get_file(new_path)
700
content = fileobj.read()
701
return len(content), sha_string(content)
703
def _get_inventory(self):
704
"""Build up the inventory entry for the BundleTree.
706
This need to be called before ever accessing self.inventory
708
from os.path import dirname, basename
709
inv = Inventory(None, self.revision_id)
711
def add_entry(path, file_id):
715
parent_path = dirname(path)
716
parent_id = self.path2id(parent_path)
718
kind = self.kind(path)
719
revision_id = self.get_last_changed(path)
721
name = basename(path)
722
if kind == 'directory':
723
ie = InventoryDirectory(file_id, name, parent_id)
725
ie = InventoryFile(file_id, name, parent_id)
726
ie.executable = self.is_executable(path)
727
elif kind == 'symlink':
728
ie = InventoryLink(file_id, name, parent_id)
729
ie.symlink_target = self.get_symlink_target(path)
730
ie.revision = revision_id
733
ie.text_size, ie.text_sha1 = self.get_size_and_sha1(path)
734
if ie.text_size is None:
736
'Got a text_size of None for file_id %r' % file_id)
739
sorted_entries = self.sorted_path_id()
740
for path, file_id in sorted_entries:
741
add_entry(path, file_id)
745
# Have to overload the inherited inventory property
746
# because _get_inventory is only called in the parent.
747
# Reading the docs, property() objects do not use
748
# overloading, they use the function as it was defined
750
inventory = property(_get_inventory)
752
root_inventory = property(_get_inventory)
754
def all_file_ids(self):
755
return {entry.file_id for path, entry in self.inventory.iter_entries()}
757
def all_versioned_paths(self):
758
return {path for path, entry in self.inventory.iter_entries()}
760
def list_files(self, include_root=False, from_dir=None, recursive=True):
761
# The only files returned by this are those from the version
766
from_dir_id = inv.path2id(from_dir)
767
if from_dir_id is None:
768
# Directory not versioned
770
entries = inv.iter_entries(from_dir=from_dir_id, recursive=recursive)
771
if inv.root is not None and not include_root and from_dir is None:
772
# skip the root for compatibility with the current apis.
774
for path, entry in entries:
775
yield path, 'V', entry.kind, entry
777
def sorted_path_id(self):
779
for result in viewitems(self._new_id):
781
for id in self.base_tree.all_file_ids():
782
path = self.id2path(id)
785
paths.append((path, id))
790
def patched_file(file_patch, original):
791
"""Produce a file-like object with the patched version of a text"""
792
from breezy.patches import iter_patched
793
from breezy.iterablefile import IterableFile
794
if file_patch == b"":
795
return IterableFile(())
796
# string.splitlines(True) also splits on '\r', but the iter_patched code
797
# only expects to iterate over '\n' style lines
798
return IterableFile(iter_patched(original,
799
BytesIO(file_patch).readlines()))
added
removed
breezy/bundle/bundle_data.py
