1
# Copyright (C) 2005, 2006, 2009 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
"""Serializer factory for reading and writing bundles.
20
from __future__ import absolute_import
26
from breezy.bundle.serializer import (
30
from breezy.bundle.serializer import binary_diff
31
from breezy.bundle.bundle_data import (
35
from breezy.diff import internal_diff
36
from breezy.revision import NULL_REVISION
37
from breezy.sixish import text_type
38
from breezy.bzr.testament import StrictTestament
39
from breezy.timestamp import (
42
from breezy.textfile import text_file
43
from breezy.trace import mutter
45
bool_text = {True: 'yes', False: 'no'}
49
"""Represent an action"""
51
def __init__(self, name, parameters=None, properties=None):
53
if parameters is None:
56
self.parameters = parameters
57
if properties is None:
60
self.properties = properties
62
def add_utf8_property(self, name, value):
63
"""Add a property whose value is currently utf8 to the action."""
64
self.properties.append((name, value.decode('utf8')))
66
def add_property(self, name, value):
67
"""Add a property to the action"""
68
self.properties.append((name, value))
70
def add_bool_property(self, name, value):
71
"""Add a boolean property to the action"""
72
self.add_property(name, bool_text[value])
74
def write(self, to_file):
75
"""Write action as to a file"""
76
p_texts = [' '.join([self.name] + self.parameters)]
77
for prop in self.properties:
79
p_texts.append(prop[0])
82
p_texts.append('%s:%s' % prop)
86
text.append(' // '.join(p_texts))
87
text_line = ''.join(text).encode('utf-8')
89
while len(text_line) > available:
90
to_file.write(text_line[:available])
91
text_line = text_line[available:]
92
to_file.write(b'\n... ')
93
available = 79 - len(b'... ')
94
to_file.write(text_line + b'\n')
97
class BundleSerializerV08(BundleSerializer):
100
"""Read the rest of the bundles from the supplied file.
102
:param f: The file to read from
103
:return: A list of bundles
105
return BundleReader(f).info
107
def check_compatible(self):
108
if self.source.supports_rich_root():
109
raise errors.IncompatibleBundleFormat('0.8', repr(self.source))
111
def write(self, source, revision_ids, forced_bases, f):
112
"""Write the bundless to the supplied files.
114
:param source: A source for revision information
115
:param revision_ids: The list of revision ids to serialize
116
:param forced_bases: A dict of revision -> base that overrides default
117
:param f: The file to output to
120
self.revision_ids = revision_ids
121
self.forced_bases = forced_bases
123
self.check_compatible()
124
with source.lock_read():
125
self._write_main_header()
126
with ui.ui_factory.nested_progress_bar() as pb:
127
self._write_revisions(pb)
129
def write_bundle(self, repository, revision_id, base_revision_id, out):
130
"""Helper function for translating write_bundle to write"""
131
forced_bases = {revision_id: base_revision_id}
132
if base_revision_id is NULL_REVISION:
133
base_revision_id = None
134
graph = repository.get_graph()
135
revision_ids = graph.find_unique_ancestors(revision_id,
137
revision_ids = list(repository.get_graph().iter_topo_order(
139
revision_ids.reverse()
140
self.write(repository, revision_ids, forced_bases, out)
143
def _write_main_header(self):
144
"""Write the header for the changes"""
146
f.write(_get_bundle_header('0.8'))
149
def _write(self, key, value, indent=1, trailing_space_when_empty=False):
150
"""Write out meta information, with proper indenting, etc.
152
:param trailing_space_when_empty: To work around a bug in earlier
153
bundle readers, when writing an empty property, we use "prop: \n"
154
rather than writing "prop:\n".
155
If this parameter is True, and value is the empty string, we will
156
write an extra space.
159
raise ValueError('indentation must be greater than 0')
161
f.write(b'#' + (b' ' * indent))
162
f.write(key.encode('utf-8'))
164
if trailing_space_when_empty and value == '':
168
elif isinstance(value, bytes):
172
elif isinstance(value, text_type):
174
f.write(value.encode('utf-8'))
179
f.write(b'#' + (b' ' * (indent + 2)))
180
if isinstance(entry, bytes):
183
f.write(entry.encode('utf-8'))
186
def _write_revisions(self, pb):
187
"""Write the information for all of the revisions."""
189
# Optimize for the case of revisions in order
193
i_max = len(self.revision_ids)
194
for i, rev_id in enumerate(self.revision_ids):
195
pb.update("Generating revision data", i, i_max)
196
rev = self.source.get_revision(rev_id)
197
if rev_id == last_rev_id:
198
rev_tree = last_rev_tree
200
rev_tree = self.source.revision_tree(rev_id)
201
if rev_id in self.forced_bases:
203
base_id = self.forced_bases[rev_id]
205
base_id = NULL_REVISION
207
explicit_base = False
209
base_id = rev.parent_ids[-1]
211
base_id = NULL_REVISION
213
if base_id == last_rev_id:
214
base_tree = last_rev_tree
216
base_tree = self.source.revision_tree(base_id)
217
force_binary = (i != 0)
218
self._write_revision(rev, rev_tree, base_id, base_tree,
219
explicit_base, force_binary)
221
last_rev_id = base_id
222
last_rev_tree = base_tree
224
def _testament_sha1(self, revision_id):
225
return StrictTestament.from_revision(self.source,
226
revision_id).as_sha1()
228
def _write_revision(self, rev, rev_tree, base_rev, base_tree,
229
explicit_base, force_binary):
230
"""Write out the information for a revision."""
232
self._write(key, value, indent=1)
234
w('message', rev.message.split('\n'))
235
w('committer', rev.committer)
236
w('date', format_highres_date(rev.timestamp, rev.timezone))
237
self.to_file.write(b'\n')
239
self._write_delta(rev_tree, base_tree, rev.revision_id, force_binary)
241
w('revision id', rev.revision_id)
242
w('sha1', self._testament_sha1(rev.revision_id))
243
w('inventory sha1', rev.inventory_sha1)
245
w('parent ids', rev.parent_ids)
247
w('base id', base_rev)
249
self._write('properties', None, indent=1)
250
for name, value in sorted(rev.properties.items()):
251
self._write(name, value, indent=3,
252
trailing_space_when_empty=True)
254
# Add an extra blank space at the end
255
self.to_file.write(b'\n')
257
def _write_action(self, name, parameters, properties=None):
258
if properties is None:
260
p_texts = ['%s:%s' % v for v in properties]
261
self.to_file.write(b'=== ')
262
self.to_file.write(' '.join([name] + parameters).encode('utf-8'))
263
self.to_file.write(' // '.join(p_texts).encode('utf-8'))
264
self.to_file.write(b'\n')
266
def _write_delta(self, new_tree, old_tree, default_revision_id,
268
"""Write out the changes between the trees."""
269
DEVNULL = '/dev/null'
273
def do_diff(file_id, old_path, new_path, action, force_binary):
274
def tree_lines(tree, path, require_text=False):
275
if tree.has_id(file_id):
276
tree_file = tree.get_file(path)
277
if require_text is True:
278
tree_file = text_file(tree_file)
279
return tree_file.readlines()
285
raise errors.BinaryFile()
286
old_lines = tree_lines(old_tree, old_path, require_text=True)
287
new_lines = tree_lines(new_tree, new_path, require_text=True)
288
action.write(self.to_file)
289
internal_diff(old_path, old_lines, new_path, new_lines,
291
except errors.BinaryFile:
292
old_lines = tree_lines(old_tree, old_path, require_text=False)
293
new_lines = tree_lines(new_tree, new_path, require_text=False)
294
action.add_property('encoding', 'base64')
295
action.write(self.to_file)
296
binary_diff(old_path, old_lines, new_path, new_lines,
299
def finish_action(action, file_id, kind, meta_modified, text_modified,
301
entry = new_tree.root_inventory.get_entry(file_id)
302
if entry.revision != default_revision_id:
303
action.add_utf8_property('last-changed', entry.revision)
305
action.add_bool_property('executable', entry.executable)
306
if text_modified and kind == "symlink":
307
action.add_property('target', entry.symlink_target)
308
if text_modified and kind == "file":
309
do_diff(file_id, old_path, new_path, action, force_binary)
311
action.write(self.to_file)
313
delta = new_tree.changes_from(old_tree, want_unchanged=True,
315
for path, file_id, kind in delta.removed:
316
action = Action('removed', [kind, path]).write(self.to_file)
318
for path, file_id, kind in delta.added:
320
'added', [kind, path], [('file-id', file_id.decode('utf-8'))])
321
meta_modified = (kind == 'file' and
322
new_tree.is_executable(path))
323
finish_action(action, file_id, kind, meta_modified, True,
326
for (old_path, new_path, file_id, kind,
327
text_modified, meta_modified) in delta.renamed:
328
action = Action('renamed', [kind, old_path], [(new_path,)])
329
finish_action(action, file_id, kind, meta_modified, text_modified,
332
for (path, file_id, kind,
333
text_modified, meta_modified) in delta.modified:
334
action = Action('modified', [kind, path])
335
finish_action(action, file_id, kind, meta_modified, text_modified,
338
for path, file_id, kind in delta.unchanged:
339
new_rev = new_tree.get_file_revision(path)
342
old_rev = old_tree.get_file_revision(old_tree.id2path(file_id))
343
if new_rev != old_rev:
344
action = Action('modified', [new_tree.kind(path), path])
345
action.add_utf8_property('last-changed', new_rev)
346
action.write(self.to_file)
349
class BundleReader(object):
350
"""This class reads in a bundle from a file, and returns
351
a Bundle object, which can then be applied against a tree.
354
def __init__(self, from_file):
355
"""Read in the bundle from the file.
357
:param from_file: A file-like object (must have iterator support).
359
object.__init__(self)
360
self.from_file = iter(from_file)
361
self._next_line = None
363
self.info = self._get_info()
364
# We put the actual inventory ids in the footer, so that the patch
365
# is easier to read for humans.
366
# Unfortunately, that means we need to read everything before we
367
# can create a proper bundle.
372
return BundleInfo08()
376
while self._next_line is not None:
377
if not self._read_revision_header():
379
if self._next_line is None:
385
"""Make sure that the information read in makes sense
386
and passes appropriate checksums.
388
# Fill in all the missing blanks for the revisions
389
# and generate the real_revisions list.
390
self.info.complete_info()
393
"""yield the next line, but secretly
394
keep 1 extra line for peeking.
396
for line in self.from_file:
397
last = self._next_line
398
self._next_line = line
400
#mutter('yielding line: %r' % last)
402
last = self._next_line
403
self._next_line = None
404
#mutter('yielding line: %r' % last)
407
def _read_revision_header(self):
408
found_something = False
409
self.info.revisions.append(RevisionInfo(None))
410
for line in self._next():
411
# The bzr header is terminated with a blank line
412
# which does not start with '#'
413
if line is None or line == b'\n':
415
if not line.startswith(b'#'):
417
found_something = True
418
self._handle_next(line)
419
if not found_something:
420
# Nothing was there, so remove the added revision
421
self.info.revisions.pop()
422
return found_something
424
def _read_next_entry(self, line, indent=1):
425
"""Read in a key-value pair
427
if not line.startswith(b'#'):
428
raise errors.MalformedHeader('Bzr header did not start with #')
429
line = line[1:-1].decode('utf-8') # Remove the '#' and '\n'
430
if line[:indent] == ' ' * indent:
433
return None, None # Ignore blank lines
435
loc = line.find(': ')
438
value = line[loc + 2:]
440
value = self._read_many(indent=indent + 2)
441
elif line[-1:] == ':':
443
value = self._read_many(indent=indent + 2)
445
raise errors.MalformedHeader('While looking for key: value pairs,'
446
' did not find the colon %r' % (line))
448
key = key.replace(' ', '_')
449
#mutter('found %s: %s' % (key, value))
452
def _handle_next(self, line):
455
key, value = self._read_next_entry(line, indent=1)
456
mutter('_handle_next %r => %r' % (key, value))
460
revision_info = self.info.revisions[-1]
461
if key in revision_info.__dict__:
462
if getattr(revision_info, key) is None:
463
if key in ('file_id', 'revision_id', 'base_id'):
464
value = value.encode('utf8')
465
elif key in ('parent_ids'):
466
value = [v.encode('utf8') for v in value]
467
elif key in ('testament_sha1', 'inventory_sha1', 'sha1'):
468
value = value.encode('ascii')
469
setattr(revision_info, key, value)
471
raise errors.MalformedHeader('Duplicated Key: %s' % key)
473
# What do we do with a key we don't recognize
474
raise errors.MalformedHeader('Unknown Key: "%s"' % key)
476
def _read_many(self, indent):
477
"""If a line ends with no entry, that means that it should be
478
followed with multiple lines of values.
480
This detects the end of the list, because it will be a line that
481
does not start properly indented.
484
start = b'#' + (b' ' * indent)
486
if self._next_line is None or not self._next_line.startswith(start):
489
for line in self._next():
490
values.append(line[len(start):-1].decode('utf-8'))
491
if self._next_line is None or not self._next_line.startswith(start):
495
def _read_one_patch(self):
496
"""Read in one patch, return the complete patch, along with
499
:return: action, lines, do_continue
501
#mutter('_read_one_patch: %r' % self._next_line)
502
# Peek and see if there are no patches
503
if self._next_line is None or self._next_line.startswith(b'#'):
504
return None, [], False
508
for line in self._next():
510
if not line.startswith(b'==='):
511
raise errors.MalformedPatches('The first line of all patches'
512
' should be a bzr meta line "==="'
514
action = line[4:-1].decode('utf-8')
515
elif line.startswith(b'... '):
516
action += line[len(b'... '):-1].decode('utf-8')
518
if (self._next_line is not None and
519
self._next_line.startswith(b'===')):
520
return action, lines, True
521
elif self._next_line is None or self._next_line.startswith(b'#'):
522
return action, lines, False
526
elif not line.startswith(b'... '):
529
return action, lines, False
531
def _read_patches(self):
533
revision_actions = []
535
action, lines, do_continue = self._read_one_patch()
536
if action is not None:
537
revision_actions.append((action, lines))
538
if self.info.revisions[-1].tree_actions is not None:
539
raise AssertionError()
540
self.info.revisions[-1].tree_actions = revision_actions
542
def _read_footer(self):
543
"""Read the rest of the meta information.
545
:param first_line: The previous step iterates past what it
546
can handle. That extra line is given here.
548
for line in self._next():
549
self._handle_next(line)
550
if self._next_line is None:
552
if not self._next_line.startswith(b'#'):
553
# Consume the trailing \n and stop processing
558
class BundleInfo08(BundleInfo):
560
def _update_tree(self, bundle_tree, revision_id):
561
bundle_tree.note_last_changed('', revision_id)
562
BundleInfo._update_tree(self, bundle_tree, revision_id)
564
def _testament_sha1_from_revision(self, repository, revision_id):
565
testament = StrictTestament.from_revision(repository, revision_id)
566
return testament.as_sha1()
568
def _testament_sha1(self, revision, tree):
569
return StrictTestament(revision, tree).as_sha1()