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
31
from ..bundle_data import (
35
from ....diff import internal_diff
36
from ....revision import NULL_REVISION
37
from ....sixish import text_type
38
from ...testament import StrictTestament
39
from ....timestamp import (
42
from ....textfile import text_file
43
from ....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])
81
p_texts.append('%s:%s' % prop)
83
text.append(' // '.join(p_texts))
84
text_line = ''.join(text).encode('utf-8')
86
while len(text_line) > available:
87
to_file.write(text_line[:available])
88
text_line = text_line[available:]
89
to_file.write(b'\n... ')
90
available = 79 - len(b'... ')
91
to_file.write(text_line + b'\n')
94
class BundleSerializerV08(BundleSerializer):
97
"""Read the rest of the bundles from the supplied file.
99
:param f: The file to read from
100
:return: A list of bundles
102
return BundleReader(f).info
104
def check_compatible(self):
105
if self.source.supports_rich_root():
106
raise errors.IncompatibleBundleFormat('0.8', repr(self.source))
108
def write(self, source, revision_ids, forced_bases, f):
109
"""Write the bundless to the supplied files.
111
:param source: A source for revision information
112
:param revision_ids: The list of revision ids to serialize
113
:param forced_bases: A dict of revision -> base that overrides default
114
:param f: The file to output to
117
self.revision_ids = revision_ids
118
self.forced_bases = forced_bases
120
self.check_compatible()
121
with source.lock_read():
122
self._write_main_header()
123
with ui.ui_factory.nested_progress_bar() as pb:
124
self._write_revisions(pb)
126
def write_bundle(self, repository, revision_id, base_revision_id, out):
127
"""Helper function for translating write_bundle to write"""
128
forced_bases = {revision_id: base_revision_id}
129
if base_revision_id is NULL_REVISION:
130
base_revision_id = None
131
graph = repository.get_graph()
132
revision_ids = graph.find_unique_ancestors(revision_id,
134
revision_ids = list(repository.get_graph().iter_topo_order(
136
revision_ids.reverse()
137
self.write(repository, revision_ids, forced_bases, out)
140
def _write_main_header(self):
141
"""Write the header for the changes"""
143
f.write(_get_bundle_header('0.8'))
146
def _write(self, key, value, indent=1, trailing_space_when_empty=False):
147
"""Write out meta information, with proper indenting, etc.
149
:param trailing_space_when_empty: To work around a bug in earlier
150
bundle readers, when writing an empty property, we use "prop: \n"
151
rather than writing "prop:\n".
152
If this parameter is True, and value is the empty string, we will
153
write an extra space.
156
raise ValueError('indentation must be greater than 0')
158
f.write(b'#' + (b' ' * indent))
159
f.write(key.encode('utf-8'))
161
if trailing_space_when_empty and value == '':
165
elif isinstance(value, bytes):
169
elif isinstance(value, text_type):
171
f.write(value.encode('utf-8'))
176
f.write(b'#' + (b' ' * (indent + 2)))
177
if isinstance(entry, bytes):
180
f.write(entry.encode('utf-8'))
183
def _write_revisions(self, pb):
184
"""Write the information for all of the revisions."""
186
# Optimize for the case of revisions in order
190
i_max = len(self.revision_ids)
191
for i, rev_id in enumerate(self.revision_ids):
192
pb.update("Generating revision data", i, i_max)
193
rev = self.source.get_revision(rev_id)
194
if rev_id == last_rev_id:
195
rev_tree = last_rev_tree
197
rev_tree = self.source.revision_tree(rev_id)
198
if rev_id in self.forced_bases:
200
base_id = self.forced_bases[rev_id]
202
base_id = NULL_REVISION
204
explicit_base = False
206
base_id = rev.parent_ids[-1]
208
base_id = NULL_REVISION
210
if base_id == last_rev_id:
211
base_tree = last_rev_tree
213
base_tree = self.source.revision_tree(base_id)
214
force_binary = (i != 0)
215
self._write_revision(rev, rev_tree, base_id, base_tree,
216
explicit_base, force_binary)
218
last_rev_id = base_id
219
last_rev_tree = base_tree
221
def _testament_sha1(self, revision_id):
222
return StrictTestament.from_revision(self.source,
223
revision_id).as_sha1()
225
def _write_revision(self, rev, rev_tree, base_rev, base_tree,
226
explicit_base, force_binary):
227
"""Write out the information for a revision."""
229
self._write(key, value, indent=1)
231
w('message', rev.message.split('\n'))
232
w('committer', rev.committer)
233
w('date', format_highres_date(rev.timestamp, rev.timezone))
234
self.to_file.write(b'\n')
236
self._write_delta(rev_tree, base_tree, rev.revision_id, force_binary)
238
w('revision id', rev.revision_id)
239
w('sha1', self._testament_sha1(rev.revision_id))
240
w('inventory sha1', rev.inventory_sha1)
242
w('parent ids', rev.parent_ids)
244
w('base id', base_rev)
246
self._write('properties', None, indent=1)
247
for name, value in sorted(rev.properties.items()):
248
self._write(name, value, indent=3,
249
trailing_space_when_empty=True)
251
# Add an extra blank space at the end
252
self.to_file.write(b'\n')
254
def _write_action(self, name, parameters, properties=None):
255
if properties is None:
257
p_texts = ['%s:%s' % v for v in properties]
258
self.to_file.write(b'=== ')
259
self.to_file.write(' '.join([name] + parameters).encode('utf-8'))
260
self.to_file.write(' // '.join(p_texts).encode('utf-8'))
261
self.to_file.write(b'\n')
263
def _write_delta(self, new_tree, old_tree, default_revision_id,
265
"""Write out the changes between the trees."""
266
DEVNULL = '/dev/null'
270
def do_diff(file_id, old_path, new_path, action, force_binary):
271
def tree_lines(tree, path, require_text=False):
273
tree_file = tree.get_file(path)
274
except errors.NoSuchFile:
277
if require_text is True:
278
tree_file = text_file(tree_file)
279
return tree_file.readlines()
283
raise errors.BinaryFile()
284
old_lines = tree_lines(old_tree, old_path, require_text=True)
285
new_lines = tree_lines(new_tree, new_path, require_text=True)
286
action.write(self.to_file)
287
internal_diff(old_path, old_lines, new_path, new_lines,
289
except errors.BinaryFile:
290
old_lines = tree_lines(old_tree, old_path, require_text=False)
291
new_lines = tree_lines(new_tree, new_path, require_text=False)
292
action.add_property('encoding', 'base64')
293
action.write(self.to_file)
294
binary_diff(old_path, old_lines, new_path, new_lines,
297
def finish_action(action, file_id, kind, meta_modified, text_modified,
299
entry = new_tree.root_inventory.get_entry(file_id)
300
if entry.revision != default_revision_id:
301
action.add_utf8_property('last-changed', entry.revision)
303
action.add_bool_property('executable', entry.executable)
304
if text_modified and kind == "symlink":
305
action.add_property('target', entry.symlink_target)
306
if text_modified and kind == "file":
307
do_diff(file_id, old_path, new_path, action, force_binary)
309
action.write(self.to_file)
311
delta = new_tree.changes_from(old_tree, want_unchanged=True,
313
for change in delta.removed:
314
action = Action('removed', [change.kind[0], change.path[0]]).write(self.to_file)
316
# TODO(jelmer): Treat copied specially here?
317
for change in delta.added + delta.copied:
319
'added', [change.kind[1], change.path[1]],
320
[('file-id', change.file_id.decode('utf-8'))])
321
meta_modified = (change.kind[1] == 'file' and
322
change.executable[1])
323
finish_action(action, change.file_id, change.kind[1], meta_modified, change.changed_content,
324
DEVNULL, change.path[1])
326
for change in delta.renamed:
327
action = Action('renamed', [change.kind[1], change.path[0]], [(change.path[1],)])
328
finish_action(action, change.file_id, change.kind[1], change.meta_modified(), change.changed_content,
329
change.path[0], change.path[1])
331
for change in delta.modified:
332
action = Action('modified', [change.kind[1], change.path[1]])
333
finish_action(action, change.file_id, change.kind[1], change.meta_modified(), change.changed_content,
334
change.path[0], change.path[1])
336
for change in delta.unchanged:
337
new_rev = new_tree.get_file_revision(change.path[1])
340
old_rev = old_tree.get_file_revision(change.path[0])
341
if new_rev != old_rev:
342
action = Action('modified', [change.kind[1], change.path[1]])
343
action.add_utf8_property('last-changed', new_rev)
344
action.write(self.to_file)
347
class BundleReader(object):
348
"""This class reads in a bundle from a file, and returns
349
a Bundle object, which can then be applied against a tree.
352
def __init__(self, from_file):
353
"""Read in the bundle from the file.
355
:param from_file: A file-like object (must have iterator support).
357
object.__init__(self)
358
self.from_file = iter(from_file)
359
self._next_line = None
361
self.info = self._get_info()
362
# We put the actual inventory ids in the footer, so that the patch
363
# is easier to read for humans.
364
# Unfortunately, that means we need to read everything before we
365
# can create a proper bundle.
370
return BundleInfo08()
374
while self._next_line is not None:
375
if not self._read_revision_header():
377
if self._next_line is None:
383
"""Make sure that the information read in makes sense
384
and passes appropriate checksums.
386
# Fill in all the missing blanks for the revisions
387
# and generate the real_revisions list.
388
self.info.complete_info()
391
"""yield the next line, but secretly
392
keep 1 extra line for peeking.
394
for line in self.from_file:
395
last = self._next_line
396
self._next_line = line
398
#mutter('yielding line: %r' % last)
400
last = self._next_line
401
self._next_line = None
402
#mutter('yielding line: %r' % last)
405
def _read_revision_header(self):
406
found_something = False
407
self.info.revisions.append(RevisionInfo(None))
408
for line in self._next():
409
# The bzr header is terminated with a blank line
410
# which does not start with '#'
411
if line is None or line == b'\n':
413
if not line.startswith(b'#'):
415
found_something = True
416
self._handle_next(line)
417
if not found_something:
418
# Nothing was there, so remove the added revision
419
self.info.revisions.pop()
420
return found_something
422
def _read_next_entry(self, line, indent=1):
423
"""Read in a key-value pair
425
if not line.startswith(b'#'):
426
raise errors.MalformedHeader('Bzr header did not start with #')
427
line = line[1:-1].decode('utf-8') # Remove the '#' and '\n'
428
if line[:indent] == ' ' * indent:
431
return None, None # Ignore blank lines
433
loc = line.find(': ')
436
value = line[loc + 2:]
438
value = self._read_many(indent=indent + 2)
439
elif line[-1:] == ':':
441
value = self._read_many(indent=indent + 2)
443
raise errors.MalformedHeader('While looking for key: value pairs,'
444
' did not find the colon %r' % (line))
446
key = key.replace(' ', '_')
447
#mutter('found %s: %s' % (key, value))
450
def _handle_next(self, line):
453
key, value = self._read_next_entry(line, indent=1)
454
mutter('_handle_next %r => %r' % (key, value))
458
revision_info = self.info.revisions[-1]
459
if key in revision_info.__dict__:
460
if getattr(revision_info, key) is None:
461
if key in ('file_id', 'revision_id', 'base_id'):
462
value = value.encode('utf8')
463
elif key in ('parent_ids'):
464
value = [v.encode('utf8') for v in value]
465
elif key in ('testament_sha1', 'inventory_sha1', 'sha1'):
466
value = value.encode('ascii')
467
setattr(revision_info, key, value)
469
raise errors.MalformedHeader('Duplicated Key: %s' % key)
471
# What do we do with a key we don't recognize
472
raise errors.MalformedHeader('Unknown Key: "%s"' % key)
474
def _read_many(self, indent):
475
"""If a line ends with no entry, that means that it should be
476
followed with multiple lines of values.
478
This detects the end of the list, because it will be a line that
479
does not start properly indented.
482
start = b'#' + (b' ' * indent)
484
if self._next_line is None or not self._next_line.startswith(start):
487
for line in self._next():
488
values.append(line[len(start):-1].decode('utf-8'))
489
if self._next_line is None or not self._next_line.startswith(start):
493
def _read_one_patch(self):
494
"""Read in one patch, return the complete patch, along with
497
:return: action, lines, do_continue
499
#mutter('_read_one_patch: %r' % self._next_line)
500
# Peek and see if there are no patches
501
if self._next_line is None or self._next_line.startswith(b'#'):
502
return None, [], False
506
for line in self._next():
508
if not line.startswith(b'==='):
509
raise errors.MalformedPatches('The first line of all patches'
510
' should be a bzr meta line "==="'
512
action = line[4:-1].decode('utf-8')
513
elif line.startswith(b'... '):
514
action += line[len(b'... '):-1].decode('utf-8')
516
if (self._next_line is not None and
517
self._next_line.startswith(b'===')):
518
return action, lines, True
519
elif self._next_line is None or self._next_line.startswith(b'#'):
520
return action, lines, False
524
elif not line.startswith(b'... '):
527
return action, lines, False
529
def _read_patches(self):
531
revision_actions = []
533
action, lines, do_continue = self._read_one_patch()
534
if action is not None:
535
revision_actions.append((action, lines))
536
if self.info.revisions[-1].tree_actions is not None:
537
raise AssertionError()
538
self.info.revisions[-1].tree_actions = revision_actions
540
def _read_footer(self):
541
"""Read the rest of the meta information.
543
:param first_line: The previous step iterates past what it
544
can handle. That extra line is given here.
546
for line in self._next():
547
self._handle_next(line)
548
if self._next_line is None:
550
if not self._next_line.startswith(b'#'):
551
# Consume the trailing \n and stop processing
556
class BundleInfo08(BundleInfo):
558
def _update_tree(self, bundle_tree, revision_id):
559
bundle_tree.note_last_changed('', revision_id)
560
BundleInfo._update_tree(self, bundle_tree, revision_id)
562
def _testament_sha1_from_revision(self, repository, revision_id):
563
testament = StrictTestament.from_revision(repository, revision_id)
564
return testament.as_sha1()
566
def _testament(self, revision, tree):
567
return StrictTestament(revision, tree)