« back to all changes in this revision
Viewing changes to read_changeset.py
- browse files at revision 0.5.55
- compare with another revision
- compare with revision 2624.2.14
- download diff
- download tarball
- view history from revision 0.5.55
- Committer: John Arbash Meinel
- Date: 2005-07-02 06:03:00 UTC
- mto: (0.5.85) (1185.82.1 bzr-w-changeset)
- mto: This revision was merged to the branch mainline in revision 1738.
- Revision ID: john@arbash-meinel.com-20050702060300-e36fa2c400796477
Lots of updates. Using a minimized annotations for changesets.
- files modified:
- common.py
added
removed
read_changeset.py
66
66
return rev
67
67
68
68
class ChangesetInfo(object):
69
"""This is the intermediate class that gets filled out as
70
the file is read.
69
"""This contains the meta information. Stuff that allows you to
70
recreate the revision or inventory XML.
71
71
"""
72
72
def __init__(self):
73
73
self.committer = None
78
78
79
79
# A list of RevisionInfo objects
80
80
self.revisions = []
81
# Tuples of (new_file_id, new_file_path)
82
self.new_file_ids = []
83
84
# This is a mapping from file_id to text_id
85
self.text_ids = {}
86
87
self.tree_root_id = None
88
self.file_ids = None
89
self.old_file_ids = None
90
91
self.actions = [] #this is the list of things that happened
81
self.text_ids = {} # file_id => text_id
82
83
self.actions = []
84
85
self.timestamp = None
86
self.timezone = None
92
87
93
88
def __str__(self):
94
89
return pprint.pformat(self.__dict__)
103
98
# is the first parent of the last revision listed
104
99
rev = self.revisions[-1]
105
100
self.base = rev.parents[0].revision_id
101
# In general, if self.base is None, self.base_sha1 should
102
# also be None
103
if self.base_sha1 is not None:
104
assert self.base_sha1 == rev.parent[0].revision_sha1
106
105
self.base_sha1 = rev.parents[0].revision_sha1
107
106
107
if not self.timestamp and self.date:
108
self.timestamp, self.timezone = common.unpack_highres_date(self.date)
109
108
110
for rev in self.revisions:
109
pass
110
111
def create_maps(self):
112
"""Go through the individual id sections, and generate the
113
id2path and path2id maps.
114
"""
115
# Rather than use an empty path, the changeset code seems
116
# to like to use "./." for the tree root.
117
self.id2path[self.tree_root_id] = './.'
118
self.path2id['./.'] = self.tree_root_id
119
self.id2parent[self.tree_root_id] = bzrlib.changeset.NULL_ID
120
self.old_id2path = self.id2path.copy()
121
self.old_path2id = self.path2id.copy()
122
self.old_id2parent = self.id2parent.copy()
123
124
if self.file_ids:
125
for info in self.file_ids:
126
path, f_id, parent_id = info.split('\t')
127
self.id2path[f_id] = path
128
self.path2id[path] = f_id
129
self.id2parent[f_id] = parent_id
130
if self.old_file_ids:
131
for info in self.old_file_ids:
132
path, f_id, parent_id = info.split('\t')
133
self.old_id2path[f_id] = path
134
self.old_path2id[path] = f_id
135
self.old_id2parent[f_id] = parent_id
136
137
def get_changeset(self):
138
"""Create a changeset from the data contained within."""
139
from bzrlib.changeset import Changeset, ChangesetEntry, \
140
PatchApply, ReplaceContents
141
cset = Changeset()
142
143
entry = ChangesetEntry(self.tree_root_id,
144
bzrlib.changeset.NULL_ID, './.')
145
cset.add_entry(entry)
146
for info, lines in self.actions:
147
parts = info.split(' ')
148
action = parts[0]
149
kind = parts[1]
150
extra = ' '.join(parts[2:])
151
if action == 'renamed':
152
old_path, new_path = extra.split(' => ')
153
old_path = _unescape(old_path)
154
new_path = _unescape(new_path)
155
156
new_id = self.path2id[new_path]
157
old_id = self.old_path2id[old_path]
158
assert old_id == new_id
159
160
new_parent = self.id2parent[new_id]
161
old_parent = self.old_id2parent[old_id]
162
163
entry = ChangesetEntry(old_id, old_parent, old_path)
164
entry.new_path = new_path
165
entry.new_parent = new_parent
166
if lines:
167
entry.contents_change = PatchApply(''.join(lines))
168
elif action == 'removed':
169
old_path = _unescape(extra)
170
old_id = self.old_path2id[old_path]
171
old_parent = self.old_id2parent[old_id]
172
entry = ChangesetEntry(old_id, old_parent, old_path)
173
entry.new_path = None
174
entry.new_parent = None
175
if lines:
176
# Technically a removed should be a ReplaceContents()
177
# Where you need to have the old contents
178
# But at most we have a remove style patch.
179
#entry.contents_change = ReplaceContents()
180
pass
181
elif action == 'added':
182
new_path = _unescape(extra)
183
new_id = self.path2id[new_path]
184
new_parent = self.id2parent[new_id]
185
entry = ChangesetEntry(new_id, new_parent, new_path)
186
entry.path = None
187
entry.parent = None
188
if lines:
189
# Technically an added should be a ReplaceContents()
190
# Where you need to have the old contents
191
# But at most we have an add style patch.
192
#entry.contents_change = ReplaceContents()
193
entry.contents_change = PatchApply(''.join(lines))
194
elif action == 'modified':
195
new_path = _unescape(extra)
196
new_id = self.path2id[new_path]
197
new_parent = self.id2parent[new_id]
198
entry = ChangesetEntry(new_id, new_parent, new_path)
199
entry.path = None
200
entry.parent = None
201
if lines:
202
# Technically an added should be a ReplaceContents()
203
# Where you need to have the old contents
204
# But at most we have an add style patch.
205
#entry.contents_change = ReplaceContents()
206
entry.contents_change = PatchApply(''.join(lines))
207
else:
208
raise BadChangeset('Unrecognized action: %r' % action)
209
cset.add_entry(entry)
210
return cset
111
if rev.timestamp is None and self.timestamp is not None:
112
rev.timestamp = self.timestamp
113
rev.timezone = self.timezone
114
if rev.message is None and self.message:
115
rev.message = self.message
116
if rev.committer is None and self.committer:
117
rev.committer = self.committer
211
118
212
119
class ChangesetReader(object):
213
120
"""This class reads in a changeset from a file, and returns
231
138
self._read_patches()
232
139
self._read_footer()
233
140
141
def get_info_and_tree(self, branch):
142
"""Return the meta information, and a Changeset tree which can
143
be used to populate the local stores and working tree, respectively.
144
"""
145
self.info.complete_info()
146
store_base_sha1 = branch.get_revision_sha1(self.info.base)
147
if store_base_sha1 != self.info.base_sha1:
148
raise BzrError('Base revision sha1 hash in store'
149
' does not match the one read in the changeset'
150
' (%s != %s)' % (store_base_sha1, self.info.base_sha1))
151
tree = ChangesetTree(branch.revision_tree(self.info.base))
152
self._update_tree(tree)
153
154
return self.info, self.tree
155
234
156
def _next(self):
235
157
"""yield the next line, but secretly
236
158
keep 1 extra line for peeking.
241
163
if last is not None:
242
164
yield last
243
165
244
def get_info(self):
245
"""Create the actual changeset object.
246
"""
247
self.info.complete_info()
248
return self.info
249
250
166
def _read_header(self):
251
167
"""Read the bzr header"""
252
168
header = common.get_header()
404
320
can handle. That extra line is given here.
405
321
"""
406
322
line = self._next().next()
407
if line != '# BEGIN BZR FOOTER\n':
408
raise MalformedFooter('Footer did not begin with BEGIN BZR FOOTER')
409
410
323
for line in self._next():
411
if line == '# END BZR FOOTER\n':
412
return
413
324
self._handle_next(line)
325
if self._next_line[:1] != '#':
326
break
327
328
def _update_tree(self, tree):
329
"""This fills out a ChangesetTree based on the information
330
that was read in.
331
332
:param tree: A ChangesetTree to update with the new information.
333
"""
334
from bzrlib.errors import BzrError
335
from common import decode
336
337
def get_text_id(info, file_id):
338
if info is not None:
339
if info[:8] != 'text-id:':
340
raise BzrError("Text ids should be prefixed with 'text-id:'"
341
': %r' % info)
342
text_id = decode(info[8:])
343
elif self.info.text_ids.has_key(file_id):
344
return self.info.text_ids[file_id]
345
else:
346
# If text_id was not explicitly supplied
347
# then it should be whatever we would guess it to be
348
# based on the base revision, and what we know about
349
# the target revision
350
text_id = common.guess_text_id(tree.base_tree,
351
file_id, self.info.base, True)
352
if (self.info.text_ids.has_key(file_id)
353
and self.info.text_ids[file_id] != text_id):
354
raise BzrError('Mismatched text_ids for file_id {%s}'
355
': %s != %s' % (file_id,
356
self.info.text_ids[file_id],
357
text_id))
358
# The Info object makes more sense for where
359
# to store something like text_id, since it is
360
# what will be used to generate stored inventory
361
# entries.
362
# The problem is that we are parsing the
363
# ChangesetTree right now, we really modifying
364
# the ChangesetInfo object
365
self.info.text_ids[file_id] = text_id
366
return text_id
367
368
def renamed(kind, extra, lines):
369
info = extra.split('\t')
370
if len(info) < 2:
371
raise BzrError('renamed action lines need both a from and to'
372
': %r' % extra)
373
old_path = decode(info[0])
374
if info[1][:3] == '=> ':
375
new_path = decode(info[1][3:])
376
else:
377
new_path = decode(info[1][3:])
378
379
file_id = tree.path2id(new_path)
380
if len(info) > 2:
381
text_id = get_text_id(info[2], file_id)
382
else:
383
text_id = get_text_id(None, file_id)
384
tree.note_rename(old_path, new_path)
385
if lines:
386
tree.note_patch(new_path, lines)
387
388
def removed(kind, extra, lines):
389
info = extra.split('\t')
390
if len(info) > 1:
391
# TODO: in the future we might allow file ids to be
392
# given for removed entries
393
raise BzrError('removed action lines should only have the path'
394
': %r' % extra)
395
path = decode(info[0])
396
tree.note_deletion(path)
397
398
def added(kind, extra, lines):
399
info = extra.split('\t')
400
if len(info) <= 1:
401
raise BzrError('add action lines require the path and file id'
402
': %r' % extra)
403
elif len(info) > 3:
404
raise BzrError('add action lines have fewer than 3 entries.'
405
': %r' % extra)
406
path = decode(info[0])
407
if info[1][:8] == 'file-id:':
408
raise BzrError('The file-id should follow the path for an add'
409
': %r' % extra)
410
file_id = decode(info[1][8:])
411
412
if len(info) > 2:
413
text_id = get_text_id(info[2], file_id)
414
else:
415
text_id = get_text_id(None, file_id)
416
tree.note_id(file_id, path)
417
tree.note_patch(path, lines)
418
419
def modified(kind, extra, lines):
420
info = extra.split('\t')
421
if len(info) < 1:
422
raise BzrError('modified action lines have at least'
423
'the path in them: %r' % extra)
424
path = decode(info[0])
425
426
file_id = tree.path2id(path)
427
if len(info) > 1:
428
text_id = get_text_id(info[1], file_id)
429
else:
430
text_id = get_text_id(None, file_id)
431
tree.note_patch(path, lines)
432
433
434
valid_actions = {
435
'renamed':renamed,
436
'removed':removed,
437
'added':added,
438
'modified':modified
439
}
440
for action_line, lines in self.info.actions:
441
first = action_line.find(' ')
442
if first == -1:
443
raise BzrError('Bogus action line'
444
' (no opening space): %r' % action_line)
445
second = action_line.find(' ', first)
446
if second == -1:
447
raise BzrError('Bogus action line'
448
' (missing second space): %r' % action_line)
449
action = action_line[:first]
450
kind = action_line[first+1:second]
451
if kind not in ('file', 'directory'):
452
raise BzrError('Bogus action line'
453
' (invalid object kind): %r' % action_line)
454
extra = action_line[second+1:]
455
456
if action not in valid_actions:
457
raise BzrError('Bogus action line'
458
' (unrecognized action): %r' % action_line)
459
valid_actions[action](kind, extra, lines)
414
460
415
461
def read_changeset(from_file):
416
462
"""Read in a changeset from a filelike object (must have "readline" support), and
424
470
class ChangesetTree:
425
471
def __init__(self, base_tree=None):
426
472
self.base_tree = base_tree
427
self._renamed = {}
428
self._renamed_r = {}
429
self._new_id = {}
430
self._new_id_r = {}
473
self._renamed = {} # Mapping from old_path => new_path
474
self._renamed_r = {} # new_path => old_path
475
self._new_id = {} # new_path => new_id
476
self._new_id_r = {} # new_id => new_path
431
477
self.patches = {}
432
478
self.deleted = []
433
479
self.contents_by_id = True
434
480
481
def __str__(self):
482
return pprint.pformat(self.__dict__)
483
435
484
def note_rename(self, old_path, new_path):
485
"""A file/directory has been renamed from old_path => new_path"""
436
486
assert not self._renamed.has_key(old_path)
437
487
assert not self._renamed_r.has_key(new_path)
438
488
self._renamed[new_path] = old_path
439
489
self._renamed_r[old_path] = new_path
440
490
441
491
def note_id(self, new_id, new_path):
492
"""Files that don't exist in base need a new id."""
442
493
self._new_id[new_path] = new_id
443
494
self._new_id_r[new_id] = new_path
444
495
445
496
def note_patch(self, new_path, patch):
497
"""There is a patch for a given filename."""
446
498
self.patches[new_path] = patch
447
499
448
500
def note_deletion(self, old_path):
501
"""The file at old_path has been deleted."""
449
502
self.deleted.append(old_path)
450
503
451
504
def old_path(self, new_path):
505
"""Get the old_path (path in the base_tree) for the file at new_path"""
452
506
import os.path
453
507
old_path = self._renamed.get(new_path)
454
508
if old_path is not None:
470
524
471
525
472
526
def new_path(self, old_path):
527
"""Get the new_path (path in the target_tree) for the file at old_path
528
in the base tree.
529
"""
473
530
import os.path
474
531
new_path = self._renamed_r.get(old_path)
475
532
if new_path is not None:
492
549
return new_path
493
550
494
551
def path2id(self, path):
552
"""Return the id of the file present at path in the target tree."""
495
553
file_id = self._new_id.get(path)
496
554
if file_id is not None:
497
555
return file_id
503
561
return self.base_tree.path2id(old_path)
504
562
505
563
def id2path(self, file_id):
564
"""Return the new path in the target tree of the file with id file_id"""
506
565
path = self._new_id_r.get(file_id)
507
566
if path is not None:
508
567
return path
514
573
return self.new_path(old_path)
515
574
516
575
def old_contents_id(self, file_id):
576
"""Return the id in the base_tree for the given file_id,
577
or None if the file did not exist in base.
578
579
FIXME: Something doesn't seem right here. It seems like this function
580
should always either return None or file_id. Even if
581
you are doing the by-path lookup, you are doing a
582
id2path lookup, just to do the reverse path2id lookup.
583
"""
517
584
if self.contents_by_id:
518
585
if self.base_tree.has_id(file_id):
519
586
return file_id
523
590
return self.base_tree.path2id(new_path)
524
591
525
592
def get_file(self, file_id):
593
"""Return a file-like object containing the new contents of the
594
file given by file_id.
595
596
TODO: It might be nice if this actually generated an entry
597
in the text-store, so that the file contents would
598
then be cached.
599
"""
526
600
base_id = self.old_contents_id(file_id)
527
601
if base_id is not None:
528
602
patch_original = self.base_tree.get_file(base_id)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1