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
113
return pprint.pformat(self.__dict__)
115
def complete_info(self):
116
"""This makes sure that all information is properly
117
split up, based on the assumptions that can be made
118
when information is missing.
120
from bzrlib.timestamp import unpack_highres_date
121
# Put in all of the guessable information.
122
if not self.timestamp and self.date:
123
self.timestamp, self.timezone = unpack_highres_date(self.date)
125
self.real_revisions = []
126
for rev in self.revisions:
127
if rev.timestamp is None:
128
if rev.date is not None:
129
rev.timestamp, rev.timezone = \
130
unpack_highres_date(rev.date)
132
rev.timestamp = self.timestamp
133
rev.timezone = self.timezone
134
if rev.message is None and self.message:
135
rev.message = self.message
136
if rev.committer is None and self.committer:
137
rev.committer = self.committer
138
self.real_revisions.append(rev.as_revision())
140
def get_base(self, revision):
141
revision_info = self.get_revision_info(revision.revision_id)
142
if revision_info.base_id is not None:
143
if revision_info.base_id == NULL_REVISION:
146
return revision_info.base_id
147
if len(revision.parent_ids) == 0:
148
# There is no base listed, and
149
# the lowest revision doesn't have a parent
150
# so this is probably against the empty tree
151
# and thus base truly is None
154
return revision.parent_ids[-1]
156
def _get_target(self):
157
"""Return the target revision."""
158
if len(self.real_revisions) > 0:
159
return self.real_revisions[0].revision_id
160
elif len(self.revisions) > 0:
161
return self.revisions[0].revision_id
164
target = property(_get_target, doc='The target revision id')
166
def get_revision(self, revision_id):
167
for r in self.real_revisions:
168
if r.revision_id == revision_id:
170
raise KeyError(revision_id)
172
def get_revision_info(self, revision_id):
173
for r in self.revisions:
174
if r.revision_id == revision_id:
176
raise KeyError(revision_id)
178
def revision_tree(self, repository, revision_id, base=None):
179
revision_id = osutils.safe_revision_id(revision_id)
180
revision = self.get_revision(revision_id)
181
base = self.get_base(revision)
182
assert base != revision_id
183
self._validate_references_from_repository(repository)
184
revision_info = self.get_revision_info(revision_id)
185
inventory_revision_id = revision_id
186
bundle_tree = BundleTree(repository.revision_tree(base),
187
inventory_revision_id)
188
self._update_tree(bundle_tree, revision_id)
190
inv = bundle_tree.inventory
191
self._validate_inventory(inv, revision_id)
192
self._validate_revision(inv, revision_id)
196
def _validate_references_from_repository(self, repository):
197
"""Now that we have a repository which should have some of the
198
revisions we care about, go through and validate all of them
203
def add_sha(d, revision_id, sha1):
204
if revision_id is None:
206
raise BzrError('A Null revision should always'
207
'have a null sha1 hash')
210
# This really should have been validated as part
211
# of _validate_revisions but lets do it again
212
if sha1 != d[revision_id]:
213
raise BzrError('** Revision %r referenced with 2 different'
214
' sha hashes %s != %s' % (revision_id,
215
sha1, d[revision_id]))
217
d[revision_id] = sha1
219
# All of the contained revisions were checked
220
# in _validate_revisions
222
for rev_info in self.revisions:
223
checked[rev_info.revision_id] = True
224
add_sha(rev_to_sha, rev_info.revision_id, rev_info.sha1)
226
for (rev, rev_info) in zip(self.real_revisions, self.revisions):
227
add_sha(inv_to_sha, rev_info.revision_id, rev_info.inventory_sha1)
231
for revision_id, sha1 in rev_to_sha.iteritems():
232
if repository.has_revision(revision_id):
233
testament = StrictTestament.from_revision(repository,
235
local_sha1 = self._testament_sha1_from_revision(repository,
237
if sha1 != local_sha1:
238
raise BzrError('sha1 mismatch. For revision id {%s}'
239
'local: %s, bundle: %s' % (revision_id, local_sha1, sha1))
242
elif revision_id not in checked:
243
missing[revision_id] = sha1
245
for inv_id, sha1 in inv_to_sha.iteritems():
246
if repository.has_revision(inv_id):
247
# Note: branch.get_inventory_sha1() just returns the value that
248
# is stored in the revision text, and that value may be out
249
# of date. This is bogus, because that means we aren't
250
# validating the actual text, just that we wrote and read the
251
# string. But for now, what the hell.
252
local_sha1 = repository.get_inventory_sha1(inv_id)
253
if sha1 != local_sha1:
254
raise BzrError('sha1 mismatch. For inventory id {%s}'
255
'local: %s, bundle: %s' %
256
(inv_id, local_sha1, sha1))
261
# I don't know if this is an error yet
262
warning('Not all revision hashes could be validated.'
263
' Unable validate %d hashes' % len(missing))
264
mutter('Verified %d sha hashes for the bundle.' % count)
266
def _validate_inventory(self, inv, revision_id):
267
"""At this point we should have generated the BundleTree,
268
so build up an inventory, and make sure the hashes match.
271
assert inv is not None
273
# Now we should have a complete inventory entry.
274
s = serializer_v5.write_inventory_to_string(inv)
276
# Target revision is the last entry in the real_revisions list
277
rev = self.get_revision(revision_id)
278
assert rev.revision_id == revision_id
279
if sha1 != rev.inventory_sha1:
280
open(',,bogus-inv', 'wb').write(s)
281
warning('Inventory sha hash mismatch for revision %s. %s'
282
' != %s' % (revision_id, sha1, rev.inventory_sha1))
284
def _validate_revision(self, inventory, revision_id):
285
"""Make sure all revision entries match their checksum."""
287
# This is a mapping from each revision id to it's sha hash
290
rev = self.get_revision(revision_id)
291
rev_info = self.get_revision_info(revision_id)
292
assert rev.revision_id == rev_info.revision_id
293
assert rev.revision_id == revision_id
294
sha1 = self._testament_sha1(rev, inventory)
295
if sha1 != rev_info.sha1:
296
raise TestamentMismatch(rev.revision_id, rev_info.sha1, sha1)
297
if rev.revision_id in rev_to_sha1:
298
raise BzrError('Revision {%s} given twice in the list'
300
rev_to_sha1[rev.revision_id] = sha1
302
def _update_tree(self, bundle_tree, revision_id):
303
"""This fills out a BundleTree based on the information
306
:param bundle_tree: A BundleTree to update with the new information.
309
def get_rev_id(last_changed, path, kind):
310
if last_changed is not None:
311
# last_changed will be a Unicode string because of how it was
312
# read. Convert it back to utf8.
313
changed_revision_id = osutils.safe_revision_id(last_changed,
316
changed_revision_id = revision_id
317
bundle_tree.note_last_changed(path, changed_revision_id)
318
return changed_revision_id
320
def extra_info(info, new_path):
323
for info_item in info:
325
name, value = info_item.split(':', 1)
327
raise 'Value %r has no colon' % info_item
328
if name == 'last-changed':
330
elif name == 'executable':
331
assert value in ('yes', 'no'), value
332
val = (value == 'yes')
333
bundle_tree.note_executable(new_path, val)
334
elif name == 'target':
335
bundle_tree.note_target(new_path, value)
336
elif name == 'encoding':
338
return last_changed, encoding
340
def do_patch(path, lines, encoding):
341
if encoding is not None:
342
assert encoding == 'base64'
343
patch = base64.decodestring(''.join(lines))
345
patch = ''.join(lines)
346
bundle_tree.note_patch(path, patch)
348
def renamed(kind, extra, lines):
349
info = extra.split(' // ')
351
raise BzrError('renamed action lines need both a from and to'
354
if info[1].startswith('=> '):
355
new_path = info[1][3:]
359
bundle_tree.note_rename(old_path, new_path)
360
last_modified, encoding = extra_info(info[2:], new_path)
361
revision = get_rev_id(last_modified, new_path, kind)
363
do_patch(new_path, lines, encoding)
365
def removed(kind, extra, lines):
366
info = extra.split(' // ')
368
# TODO: in the future we might allow file ids to be
369
# given for removed entries
370
raise BzrError('removed action lines should only have the path'
373
bundle_tree.note_deletion(path)
375
def added(kind, extra, lines):
376
info = extra.split(' // ')
378
raise BzrError('add action lines require the path and file id'
381
raise BzrError('add action lines have fewer than 5 entries.'
384
if not info[1].startswith('file-id:'):
385
raise BzrError('The file-id should follow the path for an add'
387
# This will be Unicode because of how the stream is read. Turn it
388
# back into a utf8 file_id
389
file_id = osutils.safe_file_id(info[1][8:], warn=False)
391
bundle_tree.note_id(file_id, path, kind)
392
# this will be overridden in extra_info if executable is specified.
393
bundle_tree.note_executable(path, False)
394
last_changed, encoding = extra_info(info[2:], path)
395
revision = get_rev_id(last_changed, path, kind)
396
if kind == 'directory':
398
do_patch(path, lines, encoding)
400
def modified(kind, extra, lines):
401
info = extra.split(' // ')
403
raise BzrError('modified action lines have at least'
404
'the path in them: %r' % extra)
407
last_modified, encoding = extra_info(info[1:], path)
408
revision = get_rev_id(last_modified, path, kind)
410
do_patch(path, lines, encoding)
418
for action_line, lines in \
419
self.get_revision_info(revision_id).tree_actions:
420
first = action_line.find(' ')
422
raise BzrError('Bogus action line'
423
' (no opening space): %r' % action_line)
424
second = action_line.find(' ', first+1)
426
raise BzrError('Bogus action line'
427
' (missing second space): %r' % action_line)
428
action = action_line[:first]
429
kind = action_line[first+1:second]
430
if kind not in ('file', 'directory', 'symlink'):
431
raise BzrError('Bogus action line'
432
' (invalid object kind %r): %r' % (kind, action_line))
433
extra = action_line[second+1:]
435
if action not in valid_actions:
436
raise BzrError('Bogus action line'
437
' (unrecognized action): %r' % action_line)
438
valid_actions[action](kind, extra, lines)
440
def install_revisions(self, target_repo):
441
"""Install revisions and return the target revision"""
442
apply_bundle.install_bundle(target_repo, self)
446
class BundleTree(Tree):
447
def __init__(self, base_tree, revision_id):
448
self.base_tree = base_tree
449
self._renamed = {} # Mapping from old_path => new_path
450
self._renamed_r = {} # new_path => old_path
451
self._new_id = {} # new_path => new_id
452
self._new_id_r = {} # new_id => new_path
453
self._kinds = {} # new_id => kind
454
self._last_changed = {} # new_id => revision_id
455
self._executable = {} # new_id => executable value
457
self._targets = {} # new path => new symlink target
459
self.contents_by_id = True
460
self.revision_id = revision_id
461
self._inventory = None
464
return pprint.pformat(self.__dict__)
466
def note_rename(self, old_path, new_path):
467
"""A file/directory has been renamed from old_path => new_path"""
468
assert new_path not in self._renamed
469
assert old_path not in self._renamed_r
470
self._renamed[new_path] = old_path
471
self._renamed_r[old_path] = new_path
473
def note_id(self, new_id, new_path, kind='file'):
474
"""Files that don't exist in base need a new id."""
475
self._new_id[new_path] = new_id
476
self._new_id_r[new_id] = new_path
477
self._kinds[new_id] = kind
479
def note_last_changed(self, file_id, revision_id):
480
if (file_id in self._last_changed
481
and self._last_changed[file_id] != revision_id):
482
raise BzrError('Mismatched last-changed revision for file_id {%s}'
483
': %s != %s' % (file_id,
484
self._last_changed[file_id],
486
self._last_changed[file_id] = revision_id
488
def note_patch(self, new_path, patch):
489
"""There is a patch for a given filename."""
490
self.patches[new_path] = patch
492
def note_target(self, new_path, target):
493
"""The symlink at the new path has the given target"""
494
self._targets[new_path] = target
496
def note_deletion(self, old_path):
497
"""The file at old_path has been deleted."""
498
self.deleted.append(old_path)
500
def note_executable(self, new_path, executable):
501
self._executable[new_path] = executable
503
def old_path(self, new_path):
504
"""Get the old_path (path in the base_tree) for the file at new_path"""
505
assert new_path[:1] not in ('\\', '/')
506
old_path = self._renamed.get(new_path)
507
if old_path is not None:
509
dirname,basename = os.path.split(new_path)
510
# dirname is not '' doesn't work, because
511
# dirname may be a unicode entry, and is
512
# requires the objects to be identical
514
old_dir = self.old_path(dirname)
518
old_path = pathjoin(old_dir, basename)
521
#If the new path wasn't in renamed, the old one shouldn't be in
523
if old_path in self._renamed_r:
527
def new_path(self, old_path):
528
"""Get the new_path (path in the target_tree) for the file at old_path
531
assert old_path[:1] not in ('\\', '/')
532
new_path = self._renamed_r.get(old_path)
533
if new_path is not None:
535
if new_path in self._renamed:
537
dirname,basename = os.path.split(old_path)
539
new_dir = self.new_path(dirname)
543
new_path = pathjoin(new_dir, basename)
546
#If the old path wasn't in renamed, the new one shouldn't be in
548
if new_path in self._renamed:
552
def path2id(self, path):
553
"""Return the id of the file present at path in the target tree."""
554
file_id = self._new_id.get(path)
555
if file_id is not None:
557
old_path = self.old_path(path)
560
if old_path in self.deleted:
562
if getattr(self.base_tree, 'path2id', None) is not None:
563
return self.base_tree.path2id(old_path)
565
return self.base_tree.inventory.path2id(old_path)
567
def id2path(self, file_id):
568
"""Return the new path in the target tree of the file with id file_id"""
569
path = self._new_id_r.get(file_id)
572
old_path = self.base_tree.id2path(file_id)
575
if old_path in self.deleted:
577
return self.new_path(old_path)
579
def old_contents_id(self, file_id):
580
"""Return the id in the base_tree for the given file_id.
581
Return None if the file did not exist in base.
583
if self.contents_by_id:
584
if self.base_tree.has_id(file_id):
588
new_path = self.id2path(file_id)
589
return self.base_tree.path2id(new_path)
591
def get_file(self, file_id):
592
"""Return a file-like object containing the new contents of the
593
file given by file_id.
595
TODO: It might be nice if this actually generated an entry
596
in the text-store, so that the file contents would
599
base_id = self.old_contents_id(file_id)
600
if (base_id is not None and
601
base_id != self.base_tree.inventory.root.file_id):
602
patch_original = self.base_tree.get_file(base_id)
604
patch_original = None
605
file_patch = self.patches.get(self.id2path(file_id))
606
if file_patch is None:
607
if (patch_original is None and
608
self.get_kind(file_id) == 'directory'):
610
assert patch_original is not None, "None: %s" % file_id
611
return patch_original
613
assert not file_patch.startswith('\\'), \
614
'Malformed patch for %s, %r' % (file_id, file_patch)
615
return patched_file(file_patch, patch_original)
617
def get_symlink_target(self, file_id):
618
new_path = self.id2path(file_id)
620
return self._targets[new_path]
622
return self.base_tree.get_symlink_target(file_id)
624
def get_kind(self, file_id):
625
if file_id in self._kinds:
626
return self._kinds[file_id]
627
return self.base_tree.inventory[file_id].kind
629
def is_executable(self, file_id):
630
path = self.id2path(file_id)
631
if path in self._executable:
632
return self._executable[path]
634
return self.base_tree.inventory[file_id].executable
636
def get_last_changed(self, file_id):
637
path = self.id2path(file_id)
638
if path in self._last_changed:
639
return self._last_changed[path]
640
return self.base_tree.inventory[file_id].revision
642
def get_size_and_sha1(self, file_id):
643
"""Return the size and sha1 hash of the given file id.
644
If the file was not locally modified, this is extracted
645
from the base_tree. Rather than re-reading the file.
647
new_path = self.id2path(file_id)
650
if new_path not in self.patches:
651
# If the entry does not have a patch, then the
652
# contents must be the same as in the base_tree
653
ie = self.base_tree.inventory[file_id]
654
if ie.text_size is None:
655
return ie.text_size, ie.text_sha1
656
return int(ie.text_size), ie.text_sha1
657
fileobj = self.get_file(file_id)
658
content = fileobj.read()
659
return len(content), sha_string(content)
661
def _get_inventory(self):
662
"""Build up the inventory entry for the BundleTree.
664
This need to be called before ever accessing self.inventory
666
from os.path import dirname, basename
668
assert self.base_tree is not None
669
base_inv = self.base_tree.inventory
670
inv = Inventory(None, self.revision_id)
672
def add_entry(file_id):
673
path = self.id2path(file_id)
679
parent_path = dirname(path)
680
parent_id = self.path2id(parent_path)
682
kind = self.get_kind(file_id)
683
revision_id = self.get_last_changed(file_id)
685
name = basename(path)
686
if kind == 'directory':
687
ie = InventoryDirectory(file_id, name, parent_id)
689
ie = InventoryFile(file_id, name, parent_id)
690
ie.executable = self.is_executable(file_id)
691
elif kind == 'symlink':
692
ie = InventoryLink(file_id, name, parent_id)
693
ie.symlink_target = self.get_symlink_target(file_id)
694
ie.revision = revision_id
696
if kind in ('directory', 'symlink'):
697
ie.text_size, ie.text_sha1 = None, None
699
ie.text_size, ie.text_sha1 = self.get_size_and_sha1(file_id)
700
if (ie.text_size is None) and (kind == 'file'):
701
raise BzrError('Got a text_size of None for file_id %r' % file_id)
704
sorted_entries = self.sorted_path_id()
705
for path, file_id in sorted_entries:
710
# Have to overload the inherited inventory property
711
# because _get_inventory is only called in the parent.
712
# Reading the docs, property() objects do not use
713
# overloading, they use the function as it was defined
715
inventory = property(_get_inventory)
718
for path, entry in self.inventory.iter_entries():
721
def sorted_path_id(self):
723
for result in self._new_id.iteritems():
725
for id in self.base_tree:
726
path = self.id2path(id)
729
paths.append((path, id))
734
def patched_file(file_patch, original):
735
"""Produce a file-like object with the patched version of a text"""
736
from bzrlib.patches import iter_patched
737
from bzrlib.iterablefile import IterableFile
739
return IterableFile(())
740
# string.splitlines(True) also splits on '\r', but the iter_patched code
741
# only expects to iterate over '\n' style lines
742
return IterableFile(iter_patched(original,
743
StringIO(file_patch).readlines()))
added
removed
bzrlib/bundle/bundle_data.py
