1
# (C) 2005 Canonical Development 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.bundle.common import get_header, header_str
23
from bzrlib.bundle.serializer import (BundleSerializer,
28
from bzrlib.bundle.serializer import binary_diff
29
from bzrlib.bundle.read_bundle import (RevisionInfo, BundleInfo, BundleTree)
30
from bzrlib.delta import compare_trees
31
from bzrlib.diff import internal_diff
32
import bzrlib.errors as errors
33
from bzrlib.osutils import pathjoin
34
from bzrlib.progress import DummyProgress
35
from bzrlib.revision import NULL_REVISION
36
from bzrlib.rio import RioWriter, read_stanzas
38
from bzrlib.testament import StrictTestament
39
from bzrlib.textfile import text_file
40
from bzrlib.trace import mutter
42
bool_text = {True: 'yes', False: 'no'}
46
"""Represent an action"""
48
def __init__(self, name, parameters=None, properties=None):
50
if parameters is None:
53
self.parameters = parameters
54
if properties is None:
57
self.properties = properties
59
def add_property(self, name, value):
60
"""Add a property to the action"""
61
self.properties.append((name, value))
63
def add_bool_property(self, name, value):
64
"""Add a boolean property to the action"""
65
self.add_property(name, bool_text[value])
67
def write(self, to_file):
68
"""Write action as to a file"""
69
p_texts = [' '.join([self.name]+self.parameters)]
70
for prop in self.properties:
72
p_texts.append(prop[0])
75
p_texts.append('%s:%s' % prop)
79
text.append(' // '.join(p_texts))
80
text_line = ''.join(text).encode('utf-8')
82
while len(text_line) > available:
83
to_file.write(text_line[:available])
84
text_line = text_line[available:]
85
to_file.write('\n... ')
86
available = 79 - len('... ')
87
to_file.write(text_line+'\n')
90
class BundleSerializerV07(BundleSerializer):
92
"""Read the rest of the bundles from the supplied file.
94
:param f: The file to read from
95
:return: A list of bundles
97
return BundleReader(f).info
99
def write(self, source, revision_ids, forced_bases, f):
100
"""Write the bundless to the supplied files.
102
:param source: A source for revision information
103
:param revision_ids: The list of revision ids to serialize
104
:param forced_bases: A dict of revision -> base that overrides default
105
:param f: The file to output to
108
self.revision_ids = revision_ids
109
self.forced_bases = forced_bases
113
self._write_main_header()
116
self._write_revisions(pb)
123
def _write_main_header(self):
124
"""Write the header for the changes"""
126
f.write(BUNDLE_HEADER)
130
def _write(self, key, value, indent=1):
131
"""Write out meta information, with proper indenting, etc"""
132
assert indent > 0, 'indentation must be greater than 0'
134
f.write('#' + (' ' * indent))
135
f.write(key.encode('utf-8'))
138
elif isinstance(value, basestring):
140
f.write(value.encode('utf-8'))
145
f.write('#' + (' ' * (indent+2)))
146
f.write(entry.encode('utf-8'))
149
def _write_revisions(self, pb):
150
"""Write the information for all of the revisions."""
152
# Optimize for the case of revisions in order
156
i_max = len(self.revision_ids)
157
for i, rev_id in enumerate(self.revision_ids):
158
pb.update("Generating revsion data", i, i_max)
159
rev = self.source.get_revision(rev_id)
160
if rev_id == last_rev_id:
161
rev_tree = last_rev_tree
163
base_tree = self.source.revision_tree(rev_id)
164
rev_tree = self.source.revision_tree(rev_id)
165
if rev_id in self.forced_bases:
167
base_id = self.forced_bases[rev_id]
169
base_id = NULL_REVISION
171
explicit_base = False
173
base_id = rev.parent_ids[-1]
175
base_id = NULL_REVISION
177
if base_id == last_rev_id:
178
base_tree = last_rev_tree
180
base_tree = self.source.revision_tree(base_id)
181
force_binary = (i != 0)
182
self._write_revision(rev, rev_tree, base_id, base_tree,
183
explicit_base, force_binary)
185
last_rev_id = base_id
186
last_rev_tree = base_tree
188
def _write_revision(self, rev, rev_tree, base_rev, base_tree,
189
explicit_base, force_binary):
190
"""Write out the information for a revision."""
192
self._write(key, value, indent=1)
194
w('message', rev.message.split('\n'))
195
w('committer', rev.committer)
196
w('date', format_highres_date(rev.timestamp, rev.timezone))
197
self.to_file.write('\n')
199
self._write_delta(rev_tree, base_tree, rev.revision_id, force_binary)
201
w('revision id', rev.revision_id)
202
w('sha1', StrictTestament.from_revision(self.source,
203
rev.revision_id).as_sha1())
204
w('inventory sha1', rev.inventory_sha1)
206
w('parent ids', rev.parent_ids)
208
w('base id', base_rev)
210
self._write('properties', None, indent=1)
211
for name, value in rev.properties.items():
212
self._write(name, value, indent=3)
214
# Add an extra blank space at the end
215
self.to_file.write('\n')
217
def _write_action(self, name, parameters, properties=None):
218
if properties is None:
220
p_texts = ['%s:%s' % v for v in properties]
221
self.to_file.write('=== ')
222
self.to_file.write(' '.join([name]+parameters).encode('utf-8'))
223
self.to_file.write(' // '.join(p_texts).encode('utf-8'))
224
self.to_file.write('\n')
226
def _write_delta(self, new_tree, old_tree, default_revision_id,
228
"""Write out the changes between the trees."""
229
DEVNULL = '/dev/null'
233
def do_diff(file_id, old_path, new_path, action, force_binary):
234
def tree_lines(tree, require_text=False):
236
tree_file = tree.get_file(file_id)
237
if require_text is True:
238
tree_file = text_file(tree_file)
239
return tree_file.readlines()
245
raise errors.BinaryFile()
246
old_lines = tree_lines(old_tree, require_text=True)
247
new_lines = tree_lines(new_tree, require_text=True)
248
action.write(self.to_file)
249
internal_diff(old_path, old_lines, new_path, new_lines,
251
except errors.BinaryFile:
252
old_lines = tree_lines(old_tree, require_text=False)
253
new_lines = tree_lines(new_tree, require_text=False)
254
action.add_property('encoding', 'base64')
255
action.write(self.to_file)
256
binary_diff(old_path, old_lines, new_path, new_lines,
259
def finish_action(action, file_id, kind, meta_modified, text_modified,
261
entry = new_tree.inventory[file_id]
262
if entry.revision != default_revision_id:
263
action.add_property('last-changed', entry.revision)
265
action.add_bool_property('executable', entry.executable)
266
if text_modified and kind == "symlink":
267
action.add_property('target', entry.symlink_target)
268
if text_modified and kind == "file":
269
do_diff(file_id, old_path, new_path, action, force_binary)
271
action.write(self.to_file)
273
delta = compare_trees(old_tree, new_tree, want_unchanged=True)
274
for path, file_id, kind in delta.removed:
275
action = Action('removed', [kind, path]).write(self.to_file)
277
for path, file_id, kind in delta.added:
278
action = Action('added', [kind, path], [('file-id', file_id)])
279
meta_modified = (kind=='file' and
280
new_tree.is_executable(file_id))
281
finish_action(action, file_id, kind, meta_modified, True,
284
for (old_path, new_path, file_id, kind,
285
text_modified, meta_modified) in delta.renamed:
286
action = Action('renamed', [kind, old_path], [(new_path,)])
287
finish_action(action, file_id, kind, meta_modified, text_modified,
290
for (path, file_id, kind,
291
text_modified, meta_modified) in delta.modified:
292
action = Action('modified', [kind, path])
293
finish_action(action, file_id, kind, meta_modified, text_modified,
296
for path, file_id, kind in delta.unchanged:
297
ie = new_tree.inventory[file_id]
298
new_rev = getattr(ie, 'revision', None)
301
old_rev = getattr(old_tree.inventory[ie.file_id], 'revision', None)
302
if new_rev != old_rev:
303
action = Action('modified', [ie.kind,
304
new_tree.id2path(ie.file_id)])
305
action.add_property('last-changed', ie.revision)
306
action.write(self.to_file)
309
class BundleReader(object):
310
"""This class reads in a bundle from a file, and returns
311
a Bundle object, which can then be applied against a tree.
313
def __init__(self, from_file):
314
"""Read in the bundle from the file.
316
:param from_file: A file-like object (must have iterator support).
318
object.__init__(self)
319
self.from_file = iter(from_file)
320
self._next_line = None
322
self.info = BundleInfo()
323
# We put the actual inventory ids in the footer, so that the patch
324
# is easier to read for humans.
325
# Unfortunately, that means we need to read everything before we
326
# can create a proper bundle.
332
while self._next_line is not None:
333
self._read_revision_header()
334
if self._next_line is None:
340
"""Make sure that the information read in makes sense
341
and passes appropriate checksums.
343
# Fill in all the missing blanks for the revisions
344
# and generate the real_revisions list.
345
self.info.complete_info()
347
def get_bundle(self, repository):
348
"""Return the meta information, and a Bundle tree which can
349
be used to populate the local stores and working tree, respectively.
351
return self.info, self.revision_tree(repository, self.info.target)
353
def revision_tree(self, repository, revision_id, base=None):
354
return self.info.revision_tree(repository, revision_id, base)
357
"""yield the next line, but secretly
358
keep 1 extra line for peeking.
360
for line in self.from_file:
361
last = self._next_line
362
self._next_line = line
364
#mutter('yielding line: %r' % last)
366
last = self._next_line
367
self._next_line = None
368
#mutter('yielding line: %r' % last)
371
def _read_header(self):
372
"""Read the bzr header"""
373
header = get_header()
375
for line in self._next():
377
# not all mailers will keep trailing whitespace
380
if (not line.startswith('# ') or not line.endswith('\n')
381
or line[2:-1].decode('utf-8') != header[0]):
382
raise MalformedHeader('Found a header, but it'
383
' was improperly formatted')
384
header.pop(0) # We read this line.
386
break # We found everything.
387
elif (line.startswith('#') and line.endswith('\n')):
388
line = line[1:-1].strip().decode('utf-8')
389
if line[:len(header_str)] == header_str:
390
if line == header[0]:
393
raise MalformedHeader('Found what looks like'
394
' a header, but did not match')
397
raise NotABundle('Did not find an opening header')
399
def _read_revision_header(self):
400
self.info.revisions.append(RevisionInfo(None))
401
for line in self._next():
402
# The bzr header is terminated with a blank line
403
# which does not start with '#'
404
if line is None or line == '\n':
406
self._handle_next(line)
408
def _read_next_entry(self, line, indent=1):
409
"""Read in a key-value pair
411
if not line.startswith('#'):
412
raise MalformedHeader('Bzr header did not start with #')
413
line = line[1:-1].decode('utf-8') # Remove the '#' and '\n'
414
if line[:indent] == ' '*indent:
417
return None, None# Ignore blank lines
419
loc = line.find(': ')
424
value = self._read_many(indent=indent+2)
425
elif line[-1:] == ':':
427
value = self._read_many(indent=indent+2)
429
raise MalformedHeader('While looking for key: value pairs,'
430
' did not find the colon %r' % (line))
432
key = key.replace(' ', '_')
433
#mutter('found %s: %s' % (key, value))
436
def _handle_next(self, line):
439
key, value = self._read_next_entry(line, indent=1)
440
mutter('_handle_next %r => %r' % (key, value))
444
revision_info = self.info.revisions[-1]
445
if hasattr(revision_info, key):
446
if getattr(revision_info, key) is None:
447
setattr(revision_info, key, value)
449
raise MalformedHeader('Duplicated Key: %s' % key)
451
# What do we do with a key we don't recognize
452
raise MalformedHeader('Unknown Key: "%s"' % key)
454
def _read_many(self, indent):
455
"""If a line ends with no entry, that means that it should be
456
followed with multiple lines of values.
458
This detects the end of the list, because it will be a line that
459
does not start properly indented.
462
start = '#' + (' '*indent)
464
if self._next_line is None or self._next_line[:len(start)] != start:
467
for line in self._next():
468
values.append(line[len(start):-1].decode('utf-8'))
469
if self._next_line is None or self._next_line[:len(start)] != start:
473
def _read_one_patch(self):
474
"""Read in one patch, return the complete patch, along with
477
:return: action, lines, do_continue
479
#mutter('_read_one_patch: %r' % self._next_line)
480
# Peek and see if there are no patches
481
if self._next_line is None or self._next_line.startswith('#'):
482
return None, [], False
486
for line in self._next():
488
if not line.startswith('==='):
489
raise MalformedPatches('The first line of all patches'
490
' should be a bzr meta line "==="'
492
action = line[4:-1].decode('utf-8')
493
elif line.startswith('... '):
494
action += line[len('... '):-1].decode('utf-8')
496
if (self._next_line is not None and
497
self._next_line.startswith('===')):
498
return action, lines, True
499
elif self._next_line is None or self._next_line.startswith('#'):
500
return action, lines, False
504
elif not line.startswith('... '):
507
return action, lines, False
509
def _read_patches(self):
511
revision_actions = []
513
action, lines, do_continue = self._read_one_patch()
514
if action is not None:
515
revision_actions.append((action, lines))
516
assert self.info.revisions[-1].tree_actions is None
517
self.info.revisions[-1].tree_actions = revision_actions
519
def _read_footer(self):
520
"""Read the rest of the meta information.
522
:param first_line: The previous step iterates past what it
523
can handle. That extra line is given here.
525
for line in self._next():
526
self._handle_next(line)
527
if not self._next_line.startswith('#'):
530
if self._next_line is None:
added
removed
bzrlib/bundle/serializer/v07.py
