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
29
from bzrlib.bundle import apply_bundle
30
from bzrlib.errors import (TestamentMismatch, BzrError,
31
MalformedHeader, MalformedPatches, NotABundle)
32
from bzrlib.inventory import (Inventory, InventoryEntry,
33
InventoryDirectory, InventoryFile,
35
from bzrlib.osutils import sha_file, sha_string, pathjoin
36
from bzrlib.revision import Revision, NULL_REVISION
37
from bzrlib.testament import StrictTestament
38
from bzrlib.trace import mutter, warning
39
import bzrlib.transport
40
from bzrlib.tree import Tree
41
import bzrlib.urlutils
42
from bzrlib.xml5 import serializer_v5
45
class RevisionInfo(object):
46
"""Gets filled out for each revision object that is read.
48
def __init__(self, revision_id):
49
self.revision_id = revision_id
55
self.inventory_sha1 = None
57
self.parent_ids = None
60
self.properties = None
61
self.tree_actions = None
64
return pprint.pformat(self.__dict__)
66
def as_revision(self):
67
rev = Revision(revision_id=self.revision_id,
68
committer=self.committer,
69
timestamp=float(self.timestamp),
70
timezone=int(self.timezone),
71
inventory_sha1=self.inventory_sha1,
72
message='\n'.join(self.message))
75
rev.parent_ids.extend(self.parent_ids)
78
for property in self.properties:
79
key_end = property.find(': ')
81
if not property.endswith(':'):
82
raise ValueError(property)
83
key = str(property[:-1])
86
key = str(property[:key_end])
87
value = property[key_end+2:]
88
rev.properties[key] = value
93
def from_revision(revision):
94
revision_info = RevisionInfo(revision.revision_id)
95
date = timestamp.format_highres_date(revision.timestamp,
97
revision_info.date = date
98
revision_info.timezone = revision.timezone
99
revision_info.timestamp = revision.timestamp
100
revision_info.message = revision.message.split('\n')
101
revision_info.properties = [': '.join(p) for p in
102
revision.properties.iteritems()]
106
class BundleInfo(object):
107
"""This contains the meta information. Stuff that allows you to
108
recreate the revision or inventory XML.
110
def __init__(self, bundle_format=None):
111
self.bundle_format = None
112
self.committer = None
116
# A list of RevisionInfo objects
119
# The next entries are created during complete_info() and
120
# other post-read functions.
122
# A list of real Revision objects
123
self.real_revisions = []
125
self.timestamp = None
128
# Have we checked the repository yet?
129
self._validated_revisions_against_repo = False
132
return pprint.pformat(self.__dict__)
134
def complete_info(self):
135
"""This makes sure that all information is properly
136
split up, based on the assumptions that can be made
137
when information is missing.
139
from bzrlib.timestamp import unpack_highres_date
140
# Put in all of the guessable information.
141
if not self.timestamp and self.date:
142
self.timestamp, self.timezone = unpack_highres_date(self.date)
144
self.real_revisions = []
145
for rev in self.revisions:
146
if rev.timestamp is None:
147
if rev.date is not None:
148
rev.timestamp, rev.timezone = \
149
unpack_highres_date(rev.date)
151
rev.timestamp = self.timestamp
152
rev.timezone = self.timezone
153
if rev.message is None and self.message:
154
rev.message = self.message
155
if rev.committer is None and self.committer:
156
rev.committer = self.committer
157
self.real_revisions.append(rev.as_revision())
159
def get_base(self, revision):
160
revision_info = self.get_revision_info(revision.revision_id)
161
if revision_info.base_id is not None:
162
if revision_info.base_id == NULL_REVISION:
165
return revision_info.base_id
166
if len(revision.parent_ids) == 0:
167
# There is no base listed, and
168
# the lowest revision doesn't have a parent
169
# so this is probably against the empty tree
170
# and thus base truly is None
173
return revision.parent_ids[-1]
175
def _get_target(self):
176
"""Return the target revision."""
177
if len(self.real_revisions) > 0:
178
return self.real_revisions[0].revision_id
179
elif len(self.revisions) > 0:
180
return self.revisions[0].revision_id
183
target = property(_get_target, doc='The target revision id')
185
def get_revision(self, revision_id):
186
for r in self.real_revisions:
187
if r.revision_id == revision_id:
189
raise KeyError(revision_id)
191
def get_revision_info(self, revision_id):
192
for r in self.revisions:
193
if r.revision_id == revision_id:
195
raise KeyError(revision_id)
197
def revision_tree(self, repository, revision_id, base=None):
198
revision = self.get_revision(revision_id)
199
base = self.get_base(revision)
200
if base == revision_id:
201
raise AssertionError()
202
if not self._validated_revisions_against_repo:
203
self._validate_references_from_repository(repository)
204
revision_info = self.get_revision_info(revision_id)
205
inventory_revision_id = revision_id
206
bundle_tree = BundleTree(repository.revision_tree(base),
207
inventory_revision_id)
208
self._update_tree(bundle_tree, revision_id)
210
inv = bundle_tree.inventory
211
self._validate_inventory(inv, revision_id)
212
self._validate_revision(inv, revision_id)
216
def _validate_references_from_repository(self, repository):
217
"""Now that we have a repository which should have some of the
218
revisions we care about, go through and validate all of them
223
def add_sha(d, revision_id, sha1):
224
if revision_id is None:
226
raise BzrError('A Null revision should always'
227
'have a null sha1 hash')
230
# This really should have been validated as part
231
# of _validate_revisions but lets do it again
232
if sha1 != d[revision_id]:
233
raise BzrError('** Revision %r referenced with 2 different'
234
' sha hashes %s != %s' % (revision_id,
235
sha1, d[revision_id]))
237
d[revision_id] = sha1
239
# All of the contained revisions were checked
240
# in _validate_revisions
242
for rev_info in self.revisions:
243
checked[rev_info.revision_id] = True
244
add_sha(rev_to_sha, rev_info.revision_id, rev_info.sha1)
246
for (rev, rev_info) in zip(self.real_revisions, self.revisions):
247
add_sha(inv_to_sha, rev_info.revision_id, rev_info.inventory_sha1)
251
for revision_id, sha1 in rev_to_sha.iteritems():
252
if repository.has_revision(revision_id):
253
testament = StrictTestament.from_revision(repository,
255
local_sha1 = self._testament_sha1_from_revision(repository,
257
if sha1 != local_sha1:
258
raise BzrError('sha1 mismatch. For revision id {%s}'
259
'local: %s, bundle: %s' % (revision_id, local_sha1, sha1))
262
elif revision_id not in checked:
263
missing[revision_id] = sha1
265
for inv_id, sha1 in inv_to_sha.iteritems():
266
if repository.has_revision(inv_id):
267
# Note: branch.get_inventory_sha1() just returns the value that
268
# is stored in the revision text, and that value may be out
269
# of date. This is bogus, because that means we aren't
270
# validating the actual text, just that we wrote and read the
271
# string. But for now, what the hell.
272
local_sha1 = repository.get_inventory_sha1(inv_id)
273
if sha1 != local_sha1:
274
raise BzrError('sha1 mismatch. For inventory id {%s}'
275
'local: %s, bundle: %s' %
276
(inv_id, local_sha1, sha1))
281
# I don't know if this is an error yet
282
warning('Not all revision hashes could be validated.'
283
' Unable validate %d hashes' % len(missing))
284
mutter('Verified %d sha hashes for the bundle.' % count)
285
self._validated_revisions_against_repo = True
287
def _validate_inventory(self, inv, revision_id):
288
"""At this point we should have generated the BundleTree,
289
so build up an inventory, and make sure the hashes match.
291
# Now we should have a complete inventory entry.
292
s = serializer_v5.write_inventory_to_string(inv)
294
# Target revision is the last entry in the real_revisions list
295
rev = self.get_revision(revision_id)
296
if rev.revision_id != revision_id:
297
raise AssertionError()
298
if sha1 != rev.inventory_sha1:
299
open(',,bogus-inv', 'wb').write(s)
300
warning('Inventory sha hash mismatch for revision %s. %s'
301
' != %s' % (revision_id, sha1, rev.inventory_sha1))
303
def _validate_revision(self, inventory, revision_id):
304
"""Make sure all revision entries match their checksum."""
306
# This is a mapping from each revision id to it's sha hash
309
rev = self.get_revision(revision_id)
310
rev_info = self.get_revision_info(revision_id)
311
if not (rev.revision_id == rev_info.revision_id):
312
raise AssertionError()
313
if not (rev.revision_id == revision_id):
314
raise AssertionError()
315
sha1 = self._testament_sha1(rev, inventory)
316
if sha1 != rev_info.sha1:
317
raise TestamentMismatch(rev.revision_id, rev_info.sha1, sha1)
318
if rev.revision_id in rev_to_sha1:
319
raise BzrError('Revision {%s} given twice in the list'
321
rev_to_sha1[rev.revision_id] = sha1
323
def _update_tree(self, bundle_tree, revision_id):
324
"""This fills out a BundleTree based on the information
327
:param bundle_tree: A BundleTree to update with the new information.
330
def get_rev_id(last_changed, path, kind):
331
if last_changed is not None:
332
# last_changed will be a Unicode string because of how it was
333
# read. Convert it back to utf8.
334
changed_revision_id = osutils.safe_revision_id(last_changed,
337
changed_revision_id = revision_id
338
bundle_tree.note_last_changed(path, changed_revision_id)
339
return changed_revision_id
341
def extra_info(info, new_path):
344
for info_item in info:
346
name, value = info_item.split(':', 1)
348
raise 'Value %r has no colon' % info_item
349
if name == 'last-changed':
351
elif name == 'executable':
352
val = (value == 'yes')
353
bundle_tree.note_executable(new_path, val)
354
elif name == 'target':
355
bundle_tree.note_target(new_path, value)
356
elif name == 'encoding':
358
return last_changed, encoding
360
def do_patch(path, lines, encoding):
361
if encoding == 'base64':
362
patch = base64.decodestring(''.join(lines))
363
elif encoding is None:
364
patch = ''.join(lines)
366
raise ValueError(encoding)
367
bundle_tree.note_patch(path, patch)
369
def renamed(kind, extra, lines):
370
info = extra.split(' // ')
372
raise BzrError('renamed action lines need both a from and to'
375
if info[1].startswith('=> '):
376
new_path = info[1][3:]
380
bundle_tree.note_rename(old_path, new_path)
381
last_modified, encoding = extra_info(info[2:], new_path)
382
revision = get_rev_id(last_modified, new_path, kind)
384
do_patch(new_path, lines, encoding)
386
def removed(kind, extra, lines):
387
info = extra.split(' // ')
389
# TODO: in the future we might allow file ids to be
390
# given for removed entries
391
raise BzrError('removed action lines should only have the path'
394
bundle_tree.note_deletion(path)
396
def added(kind, extra, lines):
397
info = extra.split(' // ')
399
raise BzrError('add action lines require the path and file id'
402
raise BzrError('add action lines have fewer than 5 entries.'
405
if not info[1].startswith('file-id:'):
406
raise BzrError('The file-id should follow the path for an add'
408
# This will be Unicode because of how the stream is read. Turn it
409
# back into a utf8 file_id
410
file_id = osutils.safe_file_id(info[1][8:], warn=False)
412
bundle_tree.note_id(file_id, path, kind)
413
# this will be overridden in extra_info if executable is specified.
414
bundle_tree.note_executable(path, False)
415
last_changed, encoding = extra_info(info[2:], path)
416
revision = get_rev_id(last_changed, path, kind)
417
if kind == 'directory':
419
do_patch(path, lines, encoding)
421
def modified(kind, extra, lines):
422
info = extra.split(' // ')
424
raise BzrError('modified action lines have at least'
425
'the path in them: %r' % extra)
428
last_modified, encoding = extra_info(info[1:], path)
429
revision = get_rev_id(last_modified, path, kind)
431
do_patch(path, lines, encoding)
439
for action_line, lines in \
440
self.get_revision_info(revision_id).tree_actions:
441
first = action_line.find(' ')
443
raise BzrError('Bogus action line'
444
' (no opening space): %r' % action_line)
445
second = action_line.find(' ', first+1)
447
raise BzrError('Bogus action line'
448
' (missing second space): %r' % action_line)
449
action = action_line[:first]
450
kind = action_line[first+1:second]
451
if kind not in ('file', 'directory', 'symlink'):
452
raise BzrError('Bogus action line'
453
' (invalid object kind %r): %r' % (kind, action_line))
454
extra = action_line[second+1:]
456
if action not in valid_actions:
457
raise BzrError('Bogus action line'
458
' (unrecognized action): %r' % action_line)
459
valid_actions[action](kind, extra, lines)
461
def install_revisions(self, target_repo, stream_input=True):
462
"""Install revisions and return the target revision
464
:param target_repo: The repository to install into
465
:param stream_input: Ignored by this implementation.
467
apply_bundle.install_bundle(target_repo, self)
470
def get_merge_request(self, target_repo):
471
"""Provide data for performing a merge
473
Returns suggested base, suggested target, and patch verification status
475
return None, self.target, 'inapplicable'
478
class BundleTree(Tree):
479
def __init__(self, base_tree, revision_id):
480
self.base_tree = base_tree
481
self._renamed = {} # Mapping from old_path => new_path
482
self._renamed_r = {} # new_path => old_path
483
self._new_id = {} # new_path => new_id
484
self._new_id_r = {} # new_id => new_path
485
self._kinds = {} # new_id => kind
486
self._last_changed = {} # new_id => revision_id
487
self._executable = {} # new_id => executable value
489
self._targets = {} # new path => new symlink target
491
self.contents_by_id = True
492
self.revision_id = revision_id
493
self._inventory = None
496
return pprint.pformat(self.__dict__)
498
def note_rename(self, old_path, new_path):
499
"""A file/directory has been renamed from old_path => new_path"""
500
if new_path in self._renamed:
501
raise AssertionError(new_path)
502
if old_path in self._renamed_r:
503
raise AssertionError(old_path)
504
self._renamed[new_path] = old_path
505
self._renamed_r[old_path] = new_path
507
def note_id(self, new_id, new_path, kind='file'):
508
"""Files that don't exist in base need a new id."""
509
self._new_id[new_path] = new_id
510
self._new_id_r[new_id] = new_path
511
self._kinds[new_id] = kind
513
def note_last_changed(self, file_id, revision_id):
514
if (file_id in self._last_changed
515
and self._last_changed[file_id] != revision_id):
516
raise BzrError('Mismatched last-changed revision for file_id {%s}'
517
': %s != %s' % (file_id,
518
self._last_changed[file_id],
520
self._last_changed[file_id] = revision_id
522
def note_patch(self, new_path, patch):
523
"""There is a patch for a given filename."""
524
self.patches[new_path] = patch
526
def note_target(self, new_path, target):
527
"""The symlink at the new path has the given target"""
528
self._targets[new_path] = target
530
def note_deletion(self, old_path):
531
"""The file at old_path has been deleted."""
532
self.deleted.append(old_path)
534
def note_executable(self, new_path, executable):
535
self._executable[new_path] = executable
537
def old_path(self, new_path):
538
"""Get the old_path (path in the base_tree) for the file at new_path"""
539
if new_path[:1] in ('\\', '/'):
540
raise ValueError(new_path)
541
old_path = self._renamed.get(new_path)
542
if old_path is not None:
544
dirname,basename = os.path.split(new_path)
545
# dirname is not '' doesn't work, because
546
# dirname may be a unicode entry, and is
547
# requires the objects to be identical
549
old_dir = self.old_path(dirname)
553
old_path = pathjoin(old_dir, basename)
556
#If the new path wasn't in renamed, the old one shouldn't be in
558
if old_path in self._renamed_r:
562
def new_path(self, old_path):
563
"""Get the new_path (path in the target_tree) for the file at old_path
566
if old_path[:1] in ('\\', '/'):
567
raise ValueError(old_path)
568
new_path = self._renamed_r.get(old_path)
569
if new_path is not None:
571
if new_path in self._renamed:
573
dirname,basename = os.path.split(old_path)
575
new_dir = self.new_path(dirname)
579
new_path = pathjoin(new_dir, basename)
582
#If the old path wasn't in renamed, the new one shouldn't be in
584
if new_path in self._renamed:
588
def path2id(self, path):
589
"""Return the id of the file present at path in the target tree."""
590
file_id = self._new_id.get(path)
591
if file_id is not None:
593
old_path = self.old_path(path)
596
if old_path in self.deleted:
598
if getattr(self.base_tree, 'path2id', None) is not None:
599
return self.base_tree.path2id(old_path)
601
return self.base_tree.inventory.path2id(old_path)
603
def id2path(self, file_id):
604
"""Return the new path in the target tree of the file with id file_id"""
605
path = self._new_id_r.get(file_id)
608
old_path = self.base_tree.id2path(file_id)
611
if old_path in self.deleted:
613
return self.new_path(old_path)
615
def old_contents_id(self, file_id):
616
"""Return the id in the base_tree for the given file_id.
617
Return None if the file did not exist in base.
619
if self.contents_by_id:
620
if self.base_tree.has_id(file_id):
624
new_path = self.id2path(file_id)
625
return self.base_tree.path2id(new_path)
627
def get_file(self, file_id):
628
"""Return a file-like object containing the new contents of the
629
file given by file_id.
631
TODO: It might be nice if this actually generated an entry
632
in the text-store, so that the file contents would
635
base_id = self.old_contents_id(file_id)
636
if (base_id is not None and
637
base_id != self.base_tree.inventory.root.file_id):
638
patch_original = self.base_tree.get_file(base_id)
640
patch_original = None
641
file_patch = self.patches.get(self.id2path(file_id))
642
if file_patch is None:
643
if (patch_original is None and
644
self.get_kind(file_id) == 'directory'):
646
if patch_original is None:
647
raise AssertionError("None: %s" % file_id)
648
return patch_original
650
if file_patch.startswith('\\'):
652
'Malformed patch for %s, %r' % (file_id, file_patch))
653
return patched_file(file_patch, patch_original)
655
def get_symlink_target(self, file_id):
656
new_path = self.id2path(file_id)
658
return self._targets[new_path]
660
return self.base_tree.get_symlink_target(file_id)
662
def get_kind(self, file_id):
663
if file_id in self._kinds:
664
return self._kinds[file_id]
665
return self.base_tree.inventory[file_id].kind
667
def is_executable(self, file_id):
668
path = self.id2path(file_id)
669
if path in self._executable:
670
return self._executable[path]
672
return self.base_tree.inventory[file_id].executable
674
def get_last_changed(self, file_id):
675
path = self.id2path(file_id)
676
if path in self._last_changed:
677
return self._last_changed[path]
678
return self.base_tree.inventory[file_id].revision
680
def get_size_and_sha1(self, file_id):
681
"""Return the size and sha1 hash of the given file id.
682
If the file was not locally modified, this is extracted
683
from the base_tree. Rather than re-reading the file.
685
new_path = self.id2path(file_id)
688
if new_path not in self.patches:
689
# If the entry does not have a patch, then the
690
# contents must be the same as in the base_tree
691
ie = self.base_tree.inventory[file_id]
692
if ie.text_size is None:
693
return ie.text_size, ie.text_sha1
694
return int(ie.text_size), ie.text_sha1
695
fileobj = self.get_file(file_id)
696
content = fileobj.read()
697
return len(content), sha_string(content)
699
def _get_inventory(self):
700
"""Build up the inventory entry for the BundleTree.
702
This need to be called before ever accessing self.inventory
704
from os.path import dirname, basename
705
base_inv = self.base_tree.inventory
706
inv = Inventory(None, self.revision_id)
708
def add_entry(file_id):
709
path = self.id2path(file_id)
715
parent_path = dirname(path)
716
parent_id = self.path2id(parent_path)
718
kind = self.get_kind(file_id)
719
revision_id = self.get_last_changed(file_id)
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(file_id)
727
elif kind == 'symlink':
728
ie = InventoryLink(file_id, name, parent_id)
729
ie.symlink_target = self.get_symlink_target(file_id)
730
ie.revision = revision_id
732
if kind in ('directory', 'symlink'):
733
ie.text_size, ie.text_sha1 = None, None
735
ie.text_size, ie.text_sha1 = self.get_size_and_sha1(file_id)
736
if (ie.text_size is None) and (kind == 'file'):
737
raise BzrError('Got a text_size of None for file_id %r' % file_id)
740
sorted_entries = self.sorted_path_id()
741
for path, file_id in sorted_entries:
746
# Have to overload the inherited inventory property
747
# because _get_inventory is only called in the parent.
748
# Reading the docs, property() objects do not use
749
# overloading, they use the function as it was defined
751
inventory = property(_get_inventory)
754
for path, entry in self.inventory.iter_entries():
757
def sorted_path_id(self):
759
for result in self._new_id.iteritems():
761
for id in self.base_tree:
762
path = self.id2path(id)
765
paths.append((path, id))
770
def patched_file(file_patch, original):
771
"""Produce a file-like object with the patched version of a text"""
772
from bzrlib.patches import iter_patched
773
from bzrlib.iterablefile import IterableFile
775
return IterableFile(())
776
# string.splitlines(True) also splits on '\r', but the iter_patched code
777
# only expects to iterate over '\n' style lines
778
return IterableFile(iter_patched(original,
779
StringIO(file_patch).readlines()))
added
removed
bzrlib/bundle/bundle_data.py
