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 (BundleSerializer,
29
from breezy.bundle.serializer import binary_diff
30
from breezy.bundle.bundle_data import (RevisionInfo, BundleInfo)
31
from breezy.diff import internal_diff
32
from breezy.revision import NULL_REVISION
33
from breezy.testament import StrictTestament
34
from breezy.timestamp import (
37
from breezy.textfile import text_file
38
from breezy.trace import mutter
40
bool_text = {True: 'yes', False: 'no'}
44
"""Represent an action"""
46
def __init__(self, name, parameters=None, properties=None):
48
if parameters is None:
51
self.parameters = parameters
52
if properties is None:
55
self.properties = properties
57
def add_utf8_property(self, name, value):
58
"""Add a property whose value is currently utf8 to the action."""
59
self.properties.append((name, value.decode('utf8')))
61
def add_property(self, name, value):
62
"""Add a property to the action"""
63
self.properties.append((name, value))
65
def add_bool_property(self, name, value):
66
"""Add a boolean property to the action"""
67
self.add_property(name, bool_text[value])
69
def write(self, to_file):
70
"""Write action as to a file"""
71
p_texts = [' '.join([self.name]+self.parameters)]
72
for prop in self.properties:
74
p_texts.append(prop[0])
77
p_texts.append('%s:%s' % prop)
81
text.append(' // '.join(p_texts))
82
text_line = ''.join(text).encode('utf-8')
84
while len(text_line) > available:
85
to_file.write(text_line[:available])
86
text_line = text_line[available:]
87
to_file.write('\n... ')
88
available = 79 - len('... ')
89
to_file.write(text_line+'\n')
92
class BundleSerializerV08(BundleSerializer):
94
"""Read the rest of the bundles from the supplied file.
96
:param f: The file to read from
97
:return: A list of bundles
99
return BundleReader(f).info
101
def check_compatible(self):
102
if self.source.supports_rich_root():
103
raise errors.IncompatibleBundleFormat('0.8', repr(self.source))
105
def write(self, source, revision_ids, forced_bases, f):
106
"""Write the bundless to the supplied files.
108
:param source: A source for revision information
109
:param revision_ids: The list of revision ids to serialize
110
:param forced_bases: A dict of revision -> base that overrides default
111
:param f: The file to output to
114
self.revision_ids = revision_ids
115
self.forced_bases = forced_bases
117
self.check_compatible()
118
with source.lock_read():
119
self._write_main_header()
120
with ui.ui_factory.nested_progress_bar() as pb:
121
self._write_revisions(pb)
123
def write_bundle(self, repository, target, base, fileobj):
124
return self._write_bundle(repository, target, base, fileobj)
126
def _write_main_header(self):
127
"""Write the header for the changes"""
129
f.write(_get_bundle_header('0.8'))
132
def _write(self, key, value, indent=1, trailing_space_when_empty=False):
133
"""Write out meta information, with proper indenting, etc.
135
:param trailing_space_when_empty: To work around a bug in earlier
136
bundle readers, when writing an empty property, we use "prop: \n"
137
rather than writing "prop:\n".
138
If this parameter is True, and value is the empty string, we will
139
write an extra space.
142
raise ValueError('indentation must be greater than 0')
144
f.write('#' + (' ' * indent))
145
f.write(key.encode('utf-8'))
147
if trailing_space_when_empty and value == '':
151
elif isinstance(value, str):
155
elif isinstance(value, unicode):
157
f.write(value.encode('utf-8'))
162
f.write('#' + (' ' * (indent+2)))
163
if isinstance(entry, str):
166
f.write(entry.encode('utf-8'))
169
def _write_revisions(self, pb):
170
"""Write the information for all of the revisions."""
172
# Optimize for the case of revisions in order
176
i_max = len(self.revision_ids)
177
for i, rev_id in enumerate(self.revision_ids):
178
pb.update("Generating revision data", i, i_max)
179
rev = self.source.get_revision(rev_id)
180
if rev_id == last_rev_id:
181
rev_tree = last_rev_tree
183
rev_tree = self.source.revision_tree(rev_id)
184
if rev_id in self.forced_bases:
186
base_id = self.forced_bases[rev_id]
188
base_id = NULL_REVISION
190
explicit_base = False
192
base_id = rev.parent_ids[-1]
194
base_id = NULL_REVISION
196
if base_id == last_rev_id:
197
base_tree = last_rev_tree
199
base_tree = self.source.revision_tree(base_id)
200
force_binary = (i != 0)
201
self._write_revision(rev, rev_tree, base_id, base_tree,
202
explicit_base, force_binary)
204
last_rev_id = base_id
205
last_rev_tree = base_tree
207
def _testament_sha1(self, revision_id):
208
return StrictTestament.from_revision(self.source,
209
revision_id).as_sha1()
211
def _write_revision(self, rev, rev_tree, base_rev, base_tree,
212
explicit_base, force_binary):
213
"""Write out the information for a revision."""
215
self._write(key, value, indent=1)
217
w('message', rev.message.split('\n'))
218
w('committer', rev.committer)
219
w('date', format_highres_date(rev.timestamp, rev.timezone))
220
self.to_file.write('\n')
222
self._write_delta(rev_tree, base_tree, rev.revision_id, force_binary)
224
w('revision id', rev.revision_id)
225
w('sha1', self._testament_sha1(rev.revision_id))
226
w('inventory sha1', rev.inventory_sha1)
228
w('parent ids', rev.parent_ids)
230
w('base id', base_rev)
232
self._write('properties', None, indent=1)
233
for name, value in sorted(rev.properties.items()):
234
self._write(name, value, indent=3,
235
trailing_space_when_empty=True)
237
# Add an extra blank space at the end
238
self.to_file.write('\n')
240
def _write_action(self, name, parameters, properties=None):
241
if properties is None:
243
p_texts = ['%s:%s' % v for v in properties]
244
self.to_file.write('=== ')
245
self.to_file.write(' '.join([name]+parameters).encode('utf-8'))
246
self.to_file.write(' // '.join(p_texts).encode('utf-8'))
247
self.to_file.write('\n')
249
def _write_delta(self, new_tree, old_tree, default_revision_id,
251
"""Write out the changes between the trees."""
252
DEVNULL = '/dev/null'
256
def do_diff(file_id, old_path, new_path, action, force_binary):
257
def tree_lines(tree, path, require_text=False):
258
if tree.has_id(file_id):
259
tree_file = tree.get_file(path, file_id)
260
if require_text is True:
261
tree_file = text_file(tree_file)
262
return tree_file.readlines()
268
raise errors.BinaryFile()
269
old_lines = tree_lines(old_tree, old_path, require_text=True)
270
new_lines = tree_lines(new_tree, new_path, require_text=True)
271
action.write(self.to_file)
272
internal_diff(old_path, old_lines, new_path, new_lines,
274
except errors.BinaryFile:
275
old_lines = tree_lines(old_tree, old_path, require_text=False)
276
new_lines = tree_lines(new_tree, new_path, require_text=False)
277
action.add_property('encoding', 'base64')
278
action.write(self.to_file)
279
binary_diff(old_path, old_lines, new_path, new_lines,
282
def finish_action(action, file_id, kind, meta_modified, text_modified,
284
entry = new_tree.root_inventory.get_entry(file_id)
285
if entry.revision != default_revision_id:
286
action.add_utf8_property('last-changed', entry.revision)
288
action.add_bool_property('executable', entry.executable)
289
if text_modified and kind == "symlink":
290
action.add_property('target', entry.symlink_target)
291
if text_modified and kind == "file":
292
do_diff(file_id, old_path, new_path, action, force_binary)
294
action.write(self.to_file)
296
delta = new_tree.changes_from(old_tree, want_unchanged=True,
298
for path, file_id, kind in delta.removed:
299
action = Action('removed', [kind, path]).write(self.to_file)
301
for path, file_id, kind in delta.added:
302
action = Action('added', [kind, path], [('file-id', file_id)])
303
meta_modified = (kind=='file' and
304
new_tree.is_executable(path, file_id))
305
finish_action(action, file_id, kind, meta_modified, True,
308
for (old_path, new_path, file_id, kind,
309
text_modified, meta_modified) in delta.renamed:
310
action = Action('renamed', [kind, old_path], [(new_path,)])
311
finish_action(action, file_id, kind, meta_modified, text_modified,
314
for (path, file_id, kind,
315
text_modified, meta_modified) in delta.modified:
316
action = Action('modified', [kind, path])
317
finish_action(action, file_id, kind, meta_modified, text_modified,
320
for path, file_id, kind in delta.unchanged:
321
new_rev = new_tree.get_file_revision(path, file_id)
324
old_rev = old_tree.get_file_revision(old_tree.id2path(file_id), file_id)
325
if new_rev != old_rev:
326
action = Action('modified', [new_tree.kind(path, file_id),
328
action.add_utf8_property('last-changed', new_rev)
329
action.write(self.to_file)
332
class BundleReader(object):
333
"""This class reads in a bundle from a file, and returns
334
a Bundle object, which can then be applied against a tree.
336
def __init__(self, from_file):
337
"""Read in the bundle from the file.
339
:param from_file: A file-like object (must have iterator support).
341
object.__init__(self)
342
self.from_file = iter(from_file)
343
self._next_line = None
345
self.info = self._get_info()
346
# We put the actual inventory ids in the footer, so that the patch
347
# is easier to read for humans.
348
# Unfortunately, that means we need to read everything before we
349
# can create a proper bundle.
354
return BundleInfo08()
358
while self._next_line is not None:
359
if not self._read_revision_header():
361
if self._next_line is None:
367
"""Make sure that the information read in makes sense
368
and passes appropriate checksums.
370
# Fill in all the missing blanks for the revisions
371
# and generate the real_revisions list.
372
self.info.complete_info()
375
"""yield the next line, but secretly
376
keep 1 extra line for peeking.
378
for line in self.from_file:
379
last = self._next_line
380
self._next_line = line
382
#mutter('yielding line: %r' % last)
384
last = self._next_line
385
self._next_line = None
386
#mutter('yielding line: %r' % last)
389
def _read_revision_header(self):
390
found_something = False
391
self.info.revisions.append(RevisionInfo(None))
392
for line in self._next():
393
# The bzr header is terminated with a blank line
394
# which does not start with '#'
395
if line is None or line == '\n':
397
if not line.startswith('#'):
399
found_something = True
400
self._handle_next(line)
401
if not found_something:
402
# Nothing was there, so remove the added revision
403
self.info.revisions.pop()
404
return found_something
406
def _read_next_entry(self, line, indent=1):
407
"""Read in a key-value pair
409
if not line.startswith('#'):
410
raise errors.MalformedHeader('Bzr header did not start with #')
411
line = line[1:-1].decode('utf-8') # Remove the '#' and '\n'
412
if line[:indent] == ' '*indent:
415
return None, None# Ignore blank lines
417
loc = line.find(': ')
422
value = self._read_many(indent=indent+2)
423
elif line[-1:] == ':':
425
value = self._read_many(indent=indent+2)
427
raise errors.MalformedHeader('While looking for key: value pairs,'
428
' did not find the colon %r' % (line))
430
key = key.replace(' ', '_')
431
#mutter('found %s: %s' % (key, value))
434
def _handle_next(self, line):
437
key, value = self._read_next_entry(line, indent=1)
438
mutter('_handle_next %r => %r' % (key, value))
442
revision_info = self.info.revisions[-1]
443
if key in revision_info.__dict__:
444
if getattr(revision_info, key) is None:
445
if key in ('file_id', 'revision_id', 'base_id'):
446
value = value.encode('utf8')
447
elif key in ('parent_ids'):
448
value = [v.encode('utf8') for v in value]
449
setattr(revision_info, key, value)
451
raise errors.MalformedHeader('Duplicated Key: %s' % key)
453
# What do we do with a key we don't recognize
454
raise errors.MalformedHeader('Unknown Key: "%s"' % key)
456
def _read_many(self, indent):
457
"""If a line ends with no entry, that means that it should be
458
followed with multiple lines of values.
460
This detects the end of the list, because it will be a line that
461
does not start properly indented.
464
start = '#' + (' '*indent)
466
if self._next_line is None or self._next_line[:len(start)] != start:
469
for line in self._next():
470
values.append(line[len(start):-1].decode('utf-8'))
471
if self._next_line is None or self._next_line[:len(start)] != start:
475
def _read_one_patch(self):
476
"""Read in one patch, return the complete patch, along with
479
:return: action, lines, do_continue
481
#mutter('_read_one_patch: %r' % self._next_line)
482
# Peek and see if there are no patches
483
if self._next_line is None or self._next_line.startswith('#'):
484
return None, [], False
488
for line in self._next():
490
if not line.startswith('==='):
491
raise errors.MalformedPatches('The first line of all patches'
492
' should be a bzr meta line "==="'
494
action = line[4:-1].decode('utf-8')
495
elif line.startswith('... '):
496
action += line[len('... '):-1].decode('utf-8')
498
if (self._next_line is not None and
499
self._next_line.startswith('===')):
500
return action, lines, True
501
elif self._next_line is None or self._next_line.startswith('#'):
502
return action, lines, False
506
elif not line.startswith('... '):
509
return action, lines, False
511
def _read_patches(self):
513
revision_actions = []
515
action, lines, do_continue = self._read_one_patch()
516
if action is not None:
517
revision_actions.append((action, lines))
518
if self.info.revisions[-1].tree_actions is not None:
519
raise AssertionError()
520
self.info.revisions[-1].tree_actions = revision_actions
522
def _read_footer(self):
523
"""Read the rest of the meta information.
525
:param first_line: The previous step iterates past what it
526
can handle. That extra line is given here.
528
for line in self._next():
529
self._handle_next(line)
530
if self._next_line is None:
532
if not self._next_line.startswith('#'):
533
# Consume the trailing \n and stop processing
537
class BundleInfo08(BundleInfo):
539
def _update_tree(self, bundle_tree, revision_id):
540
bundle_tree.note_last_changed('', revision_id)
541
BundleInfo._update_tree(self, bundle_tree, revision_id)
543
def _testament_sha1_from_revision(self, repository, revision_id):
544
testament = StrictTestament.from_revision(repository, revision_id)
545
return testament.as_sha1()
547
def _testament_sha1(self, revision, tree):
548
return StrictTestament(revision, tree).as_sha1()