1
# Copyright (C) 2005, 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
"""Serializer factory for reading and writing bundles.
22
from bzrlib import errors
23
from bzrlib.bundle.serializer import (BundleSerializer,
26
from bzrlib.bundle.serializer import binary_diff
27
from bzrlib.bundle.bundle_data import (RevisionInfo, BundleInfo, BundleTree)
28
from bzrlib.diff import internal_diff
29
from bzrlib.osutils import pathjoin
30
from bzrlib.progress import DummyProgress
31
from bzrlib.revision import NULL_REVISION
33
from bzrlib.testament import StrictTestament
34
from bzrlib.timestamp import (
38
from bzrlib.textfile import text_file
39
from bzrlib.trace import mutter
41
bool_text = {True: 'yes', False: 'no'}
45
"""Represent an action"""
47
def __init__(self, name, parameters=None, properties=None):
49
if parameters is None:
52
self.parameters = parameters
53
if properties is None:
56
self.properties = properties
58
def add_utf8_property(self, name, value):
59
"""Add a property whose value is currently utf8 to the action."""
60
self.properties.append((name, value.decode('utf8')))
62
def add_property(self, name, value):
63
"""Add a property to the action"""
64
self.properties.append((name, value))
66
def add_bool_property(self, name, value):
67
"""Add a boolean property to the action"""
68
self.add_property(name, bool_text[value])
70
def write(self, to_file):
71
"""Write action as to a file"""
72
p_texts = [' '.join([self.name]+self.parameters)]
73
for prop in self.properties:
75
p_texts.append(prop[0])
78
p_texts.append('%s:%s' % prop)
82
text.append(' // '.join(p_texts))
83
text_line = ''.join(text).encode('utf-8')
85
while len(text_line) > available:
86
to_file.write(text_line[:available])
87
text_line = text_line[available:]
88
to_file.write('\n... ')
89
available = 79 - len('... ')
90
to_file.write(text_line+'\n')
93
class BundleSerializerV08(BundleSerializer):
95
"""Read the rest of the bundles from the supplied file.
97
:param f: The file to read from
98
:return: A list of bundles
100
return BundleReader(f).info
102
def check_compatible(self):
103
if self.source.supports_rich_root():
104
raise errors.IncompatibleBundleFormat('0.8', repr(self.source))
106
def write(self, source, revision_ids, forced_bases, f):
107
"""Write the bundless to the supplied files.
109
:param source: A source for revision information
110
:param revision_ids: The list of revision ids to serialize
111
:param forced_bases: A dict of revision -> base that overrides default
112
:param f: The file to output to
115
self.revision_ids = revision_ids
116
self.forced_bases = forced_bases
118
self.check_compatible()
121
self._write_main_header()
124
self._write_revisions(pb)
131
def _write_main_header(self):
132
"""Write the header for the changes"""
134
f.write(BUNDLE_HEADER)
138
def _write(self, key, value, indent=1, trailing_space_when_empty=False):
139
"""Write out meta information, with proper indenting, etc.
141
:param trailing_space_when_empty: To work around a bug in earlier
142
bundle readers, when writing an empty property, we use "prop: \n"
143
rather than writing "prop:\n".
144
If this parameter is True, and value is the empty string, we will
145
write an extra space.
147
assert indent > 0, 'indentation must be greater than 0'
149
f.write('#' + (' ' * indent))
150
f.write(key.encode('utf-8'))
152
if trailing_space_when_empty and value == '':
156
elif isinstance(value, str):
160
elif isinstance(value, unicode):
162
f.write(value.encode('utf-8'))
167
f.write('#' + (' ' * (indent+2)))
168
if isinstance(entry, str):
171
f.write(entry.encode('utf-8'))
174
def _write_revisions(self, pb):
175
"""Write the information for all of the revisions."""
177
# Optimize for the case of revisions in order
181
i_max = len(self.revision_ids)
182
for i, rev_id in enumerate(self.revision_ids):
183
pb.update("Generating revsion data", i, i_max)
184
rev = self.source.get_revision(rev_id)
185
if rev_id == last_rev_id:
186
rev_tree = last_rev_tree
188
rev_tree = self.source.revision_tree(rev_id)
189
if rev_id in self.forced_bases:
191
base_id = self.forced_bases[rev_id]
193
base_id = NULL_REVISION
195
explicit_base = False
197
base_id = rev.parent_ids[-1]
199
base_id = NULL_REVISION
201
if base_id == last_rev_id:
202
base_tree = last_rev_tree
204
base_tree = self.source.revision_tree(base_id)
205
force_binary = (i != 0)
206
self._write_revision(rev, rev_tree, base_id, base_tree,
207
explicit_base, force_binary)
209
last_rev_id = base_id
210
last_rev_tree = base_tree
212
def _testament_sha1(self, revision_id):
213
return StrictTestament.from_revision(self.source,
214
revision_id).as_sha1()
216
def _write_revision(self, rev, rev_tree, base_rev, base_tree,
217
explicit_base, force_binary):
218
"""Write out the information for a revision."""
220
self._write(key, value, indent=1)
222
w('message', rev.message.split('\n'))
223
w('committer', rev.committer)
224
w('date', format_highres_date(rev.timestamp, rev.timezone))
225
self.to_file.write('\n')
227
self._write_delta(rev_tree, base_tree, rev.revision_id, force_binary)
229
w('revision id', rev.revision_id)
230
w('sha1', self._testament_sha1(rev.revision_id))
231
w('inventory sha1', rev.inventory_sha1)
233
w('parent ids', rev.parent_ids)
235
w('base id', base_rev)
237
self._write('properties', None, indent=1)
238
for name, value in sorted(rev.properties.items()):
239
self._write(name, value, indent=3,
240
trailing_space_when_empty=True)
242
# Add an extra blank space at the end
243
self.to_file.write('\n')
245
def _write_action(self, name, parameters, properties=None):
246
if properties is None:
248
p_texts = ['%s:%s' % v for v in properties]
249
self.to_file.write('=== ')
250
self.to_file.write(' '.join([name]+parameters).encode('utf-8'))
251
self.to_file.write(' // '.join(p_texts).encode('utf-8'))
252
self.to_file.write('\n')
254
def _write_delta(self, new_tree, old_tree, default_revision_id,
256
"""Write out the changes between the trees."""
257
DEVNULL = '/dev/null'
261
def do_diff(file_id, old_path, new_path, action, force_binary):
262
def tree_lines(tree, require_text=False):
264
tree_file = tree.get_file(file_id)
265
if require_text is True:
266
tree_file = text_file(tree_file)
267
return tree_file.readlines()
273
raise errors.BinaryFile()
274
old_lines = tree_lines(old_tree, require_text=True)
275
new_lines = tree_lines(new_tree, require_text=True)
276
action.write(self.to_file)
277
internal_diff(old_path, old_lines, new_path, new_lines,
279
except errors.BinaryFile:
280
old_lines = tree_lines(old_tree, require_text=False)
281
new_lines = tree_lines(new_tree, require_text=False)
282
action.add_property('encoding', 'base64')
283
action.write(self.to_file)
284
binary_diff(old_path, old_lines, new_path, new_lines,
287
def finish_action(action, file_id, kind, meta_modified, text_modified,
289
entry = new_tree.inventory[file_id]
290
if entry.revision != default_revision_id:
291
action.add_utf8_property('last-changed', entry.revision)
293
action.add_bool_property('executable', entry.executable)
294
if text_modified and kind == "symlink":
295
action.add_property('target', entry.symlink_target)
296
if text_modified and kind == "file":
297
do_diff(file_id, old_path, new_path, action, force_binary)
299
action.write(self.to_file)
301
delta = new_tree.changes_from(old_tree, want_unchanged=True,
303
for path, file_id, kind in delta.removed:
304
action = Action('removed', [kind, path]).write(self.to_file)
306
for path, file_id, kind in delta.added:
307
action = Action('added', [kind, path], [('file-id', file_id)])
308
meta_modified = (kind=='file' and
309
new_tree.is_executable(file_id))
310
finish_action(action, file_id, kind, meta_modified, True,
313
for (old_path, new_path, file_id, kind,
314
text_modified, meta_modified) in delta.renamed:
315
action = Action('renamed', [kind, old_path], [(new_path,)])
316
finish_action(action, file_id, kind, meta_modified, text_modified,
319
for (path, file_id, kind,
320
text_modified, meta_modified) in delta.modified:
321
action = Action('modified', [kind, path])
322
finish_action(action, file_id, kind, meta_modified, text_modified,
325
for path, file_id, kind in delta.unchanged:
326
ie = new_tree.inventory[file_id]
327
new_rev = getattr(ie, 'revision', None)
330
old_rev = getattr(old_tree.inventory[ie.file_id], 'revision', None)
331
if new_rev != old_rev:
332
action = Action('modified', [ie.kind,
333
new_tree.id2path(ie.file_id)])
334
action.add_utf8_property('last-changed', ie.revision)
335
action.write(self.to_file)
338
class BundleReader(object):
339
"""This class reads in a bundle from a file, and returns
340
a Bundle object, which can then be applied against a tree.
342
def __init__(self, from_file):
343
"""Read in the bundle from the file.
345
:param from_file: A file-like object (must have iterator support).
347
object.__init__(self)
348
self.from_file = iter(from_file)
349
self._next_line = None
351
self.info = self._get_info()
352
# We put the actual inventory ids in the footer, so that the patch
353
# is easier to read for humans.
354
# Unfortunately, that means we need to read everything before we
355
# can create a proper bundle.
360
return BundleInfo08()
364
while self._next_line is not None:
365
if not self._read_revision_header():
367
if self._next_line is None:
373
"""Make sure that the information read in makes sense
374
and passes appropriate checksums.
376
# Fill in all the missing blanks for the revisions
377
# and generate the real_revisions list.
378
self.info.complete_info()
381
"""yield the next line, but secretly
382
keep 1 extra line for peeking.
384
for line in self.from_file:
385
last = self._next_line
386
self._next_line = line
388
#mutter('yielding line: %r' % last)
390
last = self._next_line
391
self._next_line = None
392
#mutter('yielding line: %r' % last)
395
def _read_revision_header(self):
396
found_something = False
397
self.info.revisions.append(RevisionInfo(None))
398
for line in self._next():
399
# The bzr header is terminated with a blank line
400
# which does not start with '#'
401
if line is None or line == '\n':
403
if not line.startswith('#'):
405
found_something = True
406
self._handle_next(line)
407
if not found_something:
408
# Nothing was there, so remove the added revision
409
self.info.revisions.pop()
410
return found_something
412
def _read_next_entry(self, line, indent=1):
413
"""Read in a key-value pair
415
if not line.startswith('#'):
416
raise errors.MalformedHeader('Bzr header did not start with #')
417
line = line[1:-1].decode('utf-8') # Remove the '#' and '\n'
418
if line[:indent] == ' '*indent:
421
return None, None# Ignore blank lines
423
loc = line.find(': ')
428
value = self._read_many(indent=indent+2)
429
elif line[-1:] == ':':
431
value = self._read_many(indent=indent+2)
433
raise errors.MalformedHeader('While looking for key: value pairs,'
434
' did not find the colon %r' % (line))
436
key = key.replace(' ', '_')
437
#mutter('found %s: %s' % (key, value))
440
def _handle_next(self, line):
443
key, value = self._read_next_entry(line, indent=1)
444
mutter('_handle_next %r => %r' % (key, value))
448
revision_info = self.info.revisions[-1]
449
if key in revision_info.__dict__:
450
if getattr(revision_info, key) is None:
451
if key in ('file_id', 'revision_id', 'base_id'):
452
value = value.encode('utf8')
453
elif key in ('parent_ids'):
454
value = [v.encode('utf8') for v in value]
455
setattr(revision_info, key, value)
457
raise errors.MalformedHeader('Duplicated Key: %s' % key)
459
# What do we do with a key we don't recognize
460
raise errors.MalformedHeader('Unknown Key: "%s"' % key)
462
def _read_many(self, indent):
463
"""If a line ends with no entry, that means that it should be
464
followed with multiple lines of values.
466
This detects the end of the list, because it will be a line that
467
does not start properly indented.
470
start = '#' + (' '*indent)
472
if self._next_line is None or self._next_line[:len(start)] != start:
475
for line in self._next():
476
values.append(line[len(start):-1].decode('utf-8'))
477
if self._next_line is None or self._next_line[:len(start)] != start:
481
def _read_one_patch(self):
482
"""Read in one patch, return the complete patch, along with
485
:return: action, lines, do_continue
487
#mutter('_read_one_patch: %r' % self._next_line)
488
# Peek and see if there are no patches
489
if self._next_line is None or self._next_line.startswith('#'):
490
return None, [], False
494
for line in self._next():
496
if not line.startswith('==='):
497
raise errors.MalformedPatches('The first line of all patches'
498
' should be a bzr meta line "==="'
500
action = line[4:-1].decode('utf-8')
501
elif line.startswith('... '):
502
action += line[len('... '):-1].decode('utf-8')
504
if (self._next_line is not None and
505
self._next_line.startswith('===')):
506
return action, lines, True
507
elif self._next_line is None or self._next_line.startswith('#'):
508
return action, lines, False
512
elif not line.startswith('... '):
515
return action, lines, False
517
def _read_patches(self):
519
revision_actions = []
521
action, lines, do_continue = self._read_one_patch()
522
if action is not None:
523
revision_actions.append((action, lines))
524
assert self.info.revisions[-1].tree_actions is None
525
self.info.revisions[-1].tree_actions = revision_actions
527
def _read_footer(self):
528
"""Read the rest of the meta information.
530
:param first_line: The previous step iterates past what it
531
can handle. That extra line is given here.
533
for line in self._next():
534
self._handle_next(line)
535
if self._next_line is None:
537
if not self._next_line.startswith('#'):
538
# Consume the trailing \n and stop processing
542
class BundleInfo08(BundleInfo):
544
def _update_tree(self, bundle_tree, revision_id):
545
bundle_tree.note_last_changed('', revision_id)
546
BundleInfo._update_tree(self, bundle_tree, revision_id)
548
def _testament_sha1_from_revision(self, repository, revision_id):
549
testament = StrictTestament.from_revision(repository, revision_id)
550
return testament.as_sha1()
552
def _testament_sha1(self, revision, inventory):
553
return StrictTestament(revision, inventory).as_sha1()
added
removed
bzrlib/bundle/serializer/v08.py
