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
"""WorkingTree4 format and implementation.
19
WorkingTree4 provides the dirstate based working tree logic.
21
To get a WorkingTree, call bzrdir.open_workingtree() or
22
WorkingTree.open(dir).
25
from cStringIO import StringIO
29
from bzrlib.lazy_import import lazy_import
30
lazy_import(globals(), """
31
from bisect import bisect_left
33
from copy import deepcopy
45
conflicts as _mod_conflicts,
63
from bzrlib.transport import get_transport
67
from bzrlib import symbol_versioning
68
from bzrlib.decorators import needs_read_lock, needs_write_lock
69
from bzrlib.inventory import InventoryEntry, Inventory, ROOT_ID, entry_factory
70
from bzrlib.lockable_files import LockableFiles, TransportLock
71
from bzrlib.lockdir import LockDir
72
import bzrlib.mutabletree
73
from bzrlib.mutabletree import needs_tree_write_lock
74
from bzrlib.osutils import (
86
from bzrlib.trace import mutter, note
87
from bzrlib.transport.local import LocalTransport
88
from bzrlib.tree import InterTree
89
from bzrlib.progress import DummyProgress, ProgressPhase
90
from bzrlib.revision import NULL_REVISION, CURRENT_REVISION
91
from bzrlib.rio import RioReader, rio_file, Stanza
92
from bzrlib.symbol_versioning import (deprecated_passed,
100
from bzrlib.tree import Tree
101
from bzrlib.workingtree import WorkingTree, WorkingTree3, WorkingTreeFormat3
104
class WorkingTree4(WorkingTree3):
105
"""This is the Format 4 working tree.
107
This differs from WorkingTree3 by:
108
- having a consolidated internal dirstate.
109
- not having a regular inventory attribute.
111
This is new in bzr TODO FIXME SETMEBEFORE MERGE.
114
def __init__(self, basedir,
119
"""Construct a WorkingTree for basedir.
121
If the branch is not supplied, it is opened automatically.
122
If the branch is supplied, it must be the branch for this basedir.
123
(branch.base is not cross checked, because for remote branches that
124
would be meaningless).
126
self._format = _format
127
self.bzrdir = _bzrdir
128
from bzrlib.hashcache import HashCache
129
from bzrlib.trace import note, mutter
130
assert isinstance(basedir, basestring), \
131
"base directory %r is not a string" % basedir
132
basedir = safe_unicode(basedir)
133
mutter("opening working tree %r", basedir)
134
self._branch = branch
135
assert isinstance(self.branch, bzrlib.branch.Branch), \
136
"branch %r is not a Branch" % self.branch
137
self.basedir = realpath(basedir)
138
# if branch is at our basedir and is a format 6 or less
139
# assume all other formats have their own control files.
140
assert isinstance(_control_files, LockableFiles), \
141
"_control_files must be a LockableFiles, not %r" % _control_files
142
self._control_files = _control_files
143
# update the whole cache up front and write to disk if anything changed;
144
# in the future we might want to do this more selectively
145
# two possible ways offer themselves : in self._unlock, write the cache
146
# if needed, or, when the cache sees a change, append it to the hash
147
# cache file, and have the parser take the most recent entry for a
149
cache_filename = self.bzrdir.get_workingtree_transport(None).local_abspath('stat-cache')
150
hc = self._hashcache = HashCache(basedir, cache_filename, self._control_files._file_mode)
152
# is this scan needed ? it makes things kinda slow.
161
# during a read or write lock these objects are set, and are
162
# None the rest of the time.
163
self._dirstate = None
164
self._inventory = None
167
@needs_tree_write_lock
168
def _add(self, files, ids, kinds):
169
"""See MutableTree._add."""
170
state = self.current_dirstate()
171
for f, file_id, kind in zip(files, ids, kinds):
176
file_id = generate_ids.gen_file_id(f)
177
# deliberately add the file with no cached stat or sha1
178
# - on the first access it will be gathered, and we can
179
# always change this once tests are all passing.
180
state.add(f, file_id, kind, None, '')
183
def break_lock(self):
184
"""Break a lock if one is present from another instance.
186
Uses the ui factory to ask for confirmation if the lock may be from
189
This will probe the repository for its lock as well.
191
# if the dirstate is locked by an active process, reject the break lock
194
if self._dirstate is None:
198
state = self._current_dirstate()
199
if state._lock_token is not None:
200
# we already have it locked. sheese, cant break our own lock.
201
raise errors.LockActive(self.basedir)
204
# try for a write lock - need permission to get one anyhow
207
except errors.LockContention:
208
# oslocks fail when a process is still live: fail.
209
# TODO: get the locked lockdir info and give to the user to
210
# assist in debugging.
211
raise errors.LockActive(self.basedir)
216
self._dirstate = None
217
self._control_files.break_lock()
218
self.branch.break_lock()
220
def current_dirstate(self):
221
"""Return the current dirstate object.
223
This is not part of the tree interface and only exposed for ease of
226
:raises errors.NotWriteLocked: when not in a lock.
228
if not self._control_files._lock_count:
229
raise errors.ObjectNotLocked(self)
230
return self._current_dirstate()
232
def _current_dirstate(self):
233
"""Internal function that does not check lock status.
235
This is needed for break_lock which also needs the dirstate.
237
if self._dirstate is not None:
238
return self._dirstate
239
local_path = self.bzrdir.get_workingtree_transport(None
240
).local_abspath('dirstate')
241
self._dirstate = dirstate.DirState.on_file(local_path)
242
return self._dirstate
244
def filter_unversioned_files(self, paths):
245
"""Filter out paths that are not versioned.
247
:return: set of paths.
249
# TODO: make a generic multi-bisect routine roughly that should list
250
# the paths, then process one half at a time recursively, and feed the
251
# results of each bisect in further still
252
paths = sorted(paths)
254
state = self.current_dirstate()
255
# TODO we want a paths_to_dirblocks helper I think
257
dirname, basename = os.path.split(path.encode('utf8'))
258
_, _, _, path_is_versioned = state._get_block_entry_index(
259
dirname, basename, 0)
260
if path_is_versioned:
265
"""Write all cached data to disk."""
266
if self._control_files._lock_mode != 'w':
267
raise errors.NotWriteLocked(self)
268
self.current_dirstate().save()
269
self._inventory = None
272
def _generate_inventory(self):
273
"""Create and set self.inventory from the dirstate object.
275
This is relatively expensive: we have to walk the entire dirstate.
276
Ideally we would not, and can deprecate this function.
278
#: uncomment to trap on inventory requests.
279
# import pdb;pdb.set_trace()
280
state = self.current_dirstate()
281
state._read_dirblocks_if_needed()
282
root_key, current_entry = self._get_entry(path='')
283
current_id = root_key[2]
284
assert current_entry[0][0] == 'd' # directory
285
inv = Inventory(root_id=current_id)
286
# Turn some things into local variables
287
minikind_to_kind = dirstate.DirState._minikind_to_kind
288
factory = entry_factory
289
utf8_decode = cache_utf8._utf8_decode
291
# we could do this straight out of the dirstate; it might be fast
292
# and should be profiled - RBC 20070216
293
parent_ies = {'' : inv.root}
294
for block in state._dirblocks[1:]: # skip the root
297
parent_ie = parent_ies[dirname]
299
# all the paths in this block are not versioned in this tree
301
for key, entry in block[1]:
302
minikind, link_or_sha1, size, executable, stat = entry[0]
303
if minikind in ('a', 'r'): # absent, relocated
304
# a parent tree only entry
307
name_unicode = utf8_decode(name)[0]
309
kind = minikind_to_kind[minikind]
310
inv_entry = factory[kind](file_id, name_unicode,
313
# not strictly needed: working tree
314
#entry.executable = executable
315
#entry.text_size = size
316
#entry.text_sha1 = sha1
318
elif kind == 'directory':
319
# add this entry to the parent map.
320
parent_ies[(dirname + '/' + name).strip('/')] = inv_entry
321
# These checks cost us around 40ms on a 55k entry tree
322
assert file_id not in inv_byid, ('file_id %s already in'
323
' inventory as %s' % (file_id, inv_byid[file_id]))
324
assert name_unicode not in parent_ie.children
325
inv_byid[file_id] = inv_entry
326
parent_ie.children[name_unicode] = inv_entry
327
self._inventory = inv
329
def _get_entry(self, file_id=None, path=None):
330
"""Get the dirstate row for file_id or path.
332
If either file_id or path is supplied, it is used as the key to lookup.
333
If both are supplied, the fastest lookup is used, and an error is
334
raised if they do not both point at the same row.
336
:param file_id: An optional unicode file_id to be looked up.
337
:param path: An optional unicode path to be looked up.
338
:return: The dirstate row tuple for path/file_id, or (None, None)
340
if file_id is None and path is None:
341
raise errors.BzrError('must supply file_id or path')
342
state = self.current_dirstate()
344
path = path.encode('utf8')
345
return state._get_entry(0, fileid_utf8=file_id, path_utf8=path)
347
def get_file_sha1(self, file_id, path=None, stat_value=None):
348
# check file id is valid unconditionally.
349
key, details = self._get_entry(file_id=file_id, path=path)
350
assert key is not None, 'what error should this raise'
352
# if row stat is valid, use cached sha1, else, get a new sha1.
354
path = pathjoin(key[0], key[1]).decode('utf8')
355
return self._hashcache.get_sha1(path, stat_value)
357
def _get_inventory(self):
358
"""Get the inventory for the tree. This is only valid within a lock."""
359
if self._inventory is not None:
360
return self._inventory
361
self._generate_inventory()
362
return self._inventory
364
inventory = property(_get_inventory,
365
doc="Inventory of this Tree")
368
def get_parent_ids(self):
369
"""See Tree.get_parent_ids.
371
This implementation requests the ids list from the dirstate file.
373
return self.current_dirstate().get_parent_ids()
376
def get_root_id(self):
377
"""Return the id of this trees root"""
378
return self._get_entry(path='')[0][2]
380
def has_id(self, file_id):
381
state = self.current_dirstate()
382
file_id = osutils.safe_file_id(file_id)
383
row, parents = self._get_entry(file_id=file_id)
386
return osutils.lexists(pathjoin(
387
self.basedir, row[0].decode('utf8'), row[1].decode('utf8')))
390
def id2path(self, file_id):
391
file_id = osutils.safe_file_id(file_id)
392
state = self.current_dirstate()
393
entry = self._get_entry(file_id=file_id)
394
if entry == (None, None):
396
path_utf8 = osutils.pathjoin(entry[0][0], entry[0][1])
397
return path_utf8.decode('utf8')
401
"""Iterate through file_ids for this tree.
403
file_ids are in a WorkingTree if they are in the working inventory
404
and the working file exists.
407
for key, tree_details in self.current_dirstate()._iter_entries():
408
if tree_details[0][0] in ('a', 'r'): # absent, relocated
409
# not relevant to the working tree
411
path = pathjoin(self.basedir, key[0].decode('utf8'), key[1].decode('utf8'))
412
if osutils.lexists(path):
413
result.append(key[2])
417
def _last_revision(self):
418
"""See Mutable.last_revision."""
419
parent_ids = self.current_dirstate().get_parent_ids()
426
super(WorkingTree4, self).lock_read()
427
if self._dirstate is None:
428
self.current_dirstate()
429
self._dirstate.lock_read()
431
def lock_tree_write(self):
432
super(WorkingTree4, self).lock_tree_write()
433
if self._dirstate is None:
434
self.current_dirstate()
435
self._dirstate.lock_write()
437
def lock_write(self):
438
super(WorkingTree4, self).lock_write()
439
if self._dirstate is None:
440
self.current_dirstate()
441
self._dirstate.lock_write()
443
@needs_tree_write_lock
444
def move(self, from_paths, to_dir, after=False):
445
"""See WorkingTree.move()."""
450
state = self.current_dirstate()
452
assert not isinstance(from_paths, basestring)
453
to_dir_utf8 = to_dir.encode('utf8')
454
to_entry_dirname, to_basename = os.path.split(to_dir_utf8)
455
id_index = state._get_id_index()
456
# check destination directory
457
# get the details for it
458
to_entry_block_index, to_entry_entry_index, dir_present, entry_present = \
459
state._get_block_entry_index(to_entry_dirname, to_basename, 0)
460
if not entry_present:
461
raise errors.BzrMoveFailedError('', to_dir,
462
errors.NotInWorkingDirectory(to_dir))
463
to_entry = state._dirblocks[to_entry_block_index][1][to_entry_entry_index]
464
# get a handle on the block itself.
465
to_block_index = state._ensure_block(
466
to_entry_block_index, to_entry_entry_index, to_dir_utf8)
467
to_block = state._dirblocks[to_block_index]
468
to_abs = self.abspath(to_dir)
469
if not isdir(to_abs):
470
raise errors.BzrMoveFailedError('',to_dir,
471
errors.NotADirectory(to_abs))
473
if to_entry[1][0][0] != 'd':
474
raise errors.BzrMoveFailedError('',to_dir,
475
errors.NotADirectory(to_abs))
477
if self._inventory is not None:
478
update_inventory = True
480
to_dir_ie = inv[to_dir_id]
481
to_dir_id = to_entry[0][2]
483
update_inventory = False
486
def move_one(old_entry, from_path_utf8, minikind, executable,
487
fingerprint, packed_stat, size,
488
to_block, to_key, to_path_utf8):
489
state._make_absent(old_entry)
490
from_key = old_entry[0]
492
lambda:state.update_minimal(from_key,
494
executable=executable,
495
fingerprint=fingerprint,
496
packed_stat=packed_stat,
498
path_utf8=from_path_utf8))
499
state.update_minimal(to_key,
501
executable=executable,
502
fingerprint=fingerprint,
503
packed_stat=packed_stat,
505
path_utf8=to_path_utf8)
506
added_entry_index, _ = state._find_entry_index(to_key, to_block[1])
507
new_entry = to_block[1][added_entry_index]
508
rollbacks.append(lambda:state._make_absent(new_entry))
510
# create rename entries and tuples
511
for from_rel in from_paths:
512
# from_rel is 'pathinroot/foo/bar'
513
from_rel_utf8 = from_rel.encode('utf8')
514
from_dirname, from_tail = osutils.split(from_rel)
515
from_dirname, from_tail_utf8 = osutils.split(from_rel_utf8)
516
from_entry = self._get_entry(path=from_rel)
517
if from_entry == (None, None):
518
raise errors.BzrMoveFailedError(from_rel,to_dir,
519
errors.NotVersionedError(path=str(from_rel)))
521
from_id = from_entry[0][2]
522
to_rel = pathjoin(to_dir, from_tail)
523
to_rel_utf8 = pathjoin(to_dir_utf8, from_tail_utf8)
524
item_to_entry = self._get_entry(path=to_rel)
525
if item_to_entry != (None, None):
526
raise errors.BzrMoveFailedError(from_rel, to_rel,
527
"Target is already versioned.")
529
if from_rel == to_rel:
530
raise errors.BzrMoveFailedError(from_rel, to_rel,
531
"Source and target are identical.")
533
from_missing = not self.has_filename(from_rel)
534
to_missing = not self.has_filename(to_rel)
541
raise errors.BzrMoveFailedError(from_rel, to_rel,
542
errors.NoSuchFile(path=to_rel,
543
extra="New file has not been created yet"))
545
# neither path exists
546
raise errors.BzrRenameFailedError(from_rel, to_rel,
547
errors.PathsDoNotExist(paths=(from_rel, to_rel)))
549
if from_missing: # implicitly just update our path mapping
552
raise errors.RenameFailedFilesExist(from_rel, to_rel,
553
extra="(Use --after to update the Bazaar id)")
556
def rollback_rename():
557
"""A single rename has failed, roll it back."""
559
for rollback in reversed(rollbacks):
563
import pdb;pdb.set_trace()
564
exc_info = sys.exc_info()
566
raise exc_info[0], exc_info[1], exc_info[2]
568
# perform the disk move first - its the most likely failure point.
570
from_rel_abs = self.abspath(from_rel)
571
to_rel_abs = self.abspath(to_rel)
573
osutils.rename(from_rel_abs, to_rel_abs)
575
raise errors.BzrMoveFailedError(from_rel, to_rel, e[1])
576
rollbacks.append(lambda: osutils.rename(to_rel_abs, from_rel_abs))
578
# perform the rename in the inventory next if needed: its easy
582
from_entry = inv[from_id]
583
current_parent = from_entry.parent_id
584
inv.rename(from_id, to_dir_id, from_tail)
586
lambda: inv.rename(from_id, current_parent, from_tail))
587
# finally do the rename in the dirstate, which is a little
588
# tricky to rollback, but least likely to need it.
589
old_block_index, old_entry_index, dir_present, file_present = \
590
state._get_block_entry_index(from_dirname, from_tail_utf8, 0)
591
old_block = state._dirblocks[old_block_index][1]
592
old_entry = old_block[old_entry_index]
593
from_key, old_entry_details = old_entry
594
cur_details = old_entry_details[0]
596
to_key = ((to_block[0],) + from_key[1:3])
597
minikind = cur_details[0]
598
move_one(old_entry, from_path_utf8=from_rel_utf8,
600
executable=cur_details[3],
601
fingerprint=cur_details[1],
602
packed_stat=cur_details[4],
606
to_path_utf8=to_rel_utf8)
609
def update_dirblock(from_dir, to_key, to_dir_utf8):
610
"""all entries in this block need updating.
612
TODO: This is pretty ugly, and doesn't support
613
reverting, but it works.
615
assert from_dir != '', "renaming root not supported"
616
from_key = (from_dir, '')
617
from_block_idx, present = \
618
state._find_block_index_from_key(from_key)
620
# This is the old record, if it isn't present, then
621
# there is theoretically nothing to update.
622
# (Unless it isn't present because of lazy loading,
623
# but we don't do that yet)
625
from_block = state._dirblocks[from_block_idx]
626
to_block_index, to_entry_index, _, _ = \
627
state._get_block_entry_index(to_key[0], to_key[1], 0)
628
to_block_index = state._ensure_block(
629
to_block_index, to_entry_index, to_dir_utf8)
630
to_block = state._dirblocks[to_block_index]
631
for entry in from_block[1]:
632
assert entry[0][0] == from_dir
633
cur_details = entry[1][0]
634
to_key = (to_dir_utf8, entry[0][1], entry[0][2])
635
from_path_utf8 = osutils.pathjoin(entry[0][0], entry[0][1])
636
to_path_utf8 = osutils.pathjoin(to_dir_utf8, entry[0][1])
637
minikind = cur_details[0]
638
move_one(entry, from_path_utf8=from_path_utf8,
640
executable=cur_details[3],
641
fingerprint=cur_details[1],
642
packed_stat=cur_details[4],
646
to_path_utf8=to_rel_utf8)
648
# We need to move all the children of this
650
update_dirblock(from_path_utf8, to_key,
652
update_dirblock(from_rel_utf8, to_key, to_rel_utf8)
656
result.append((from_rel, to_rel))
657
state._dirblock_state = dirstate.DirState.IN_MEMORY_MODIFIED
663
"""Initialize the state in this tree to be a new tree."""
667
def path2id(self, path):
668
"""Return the id for path in this tree."""
669
entry = self._get_entry(path=path)
670
if entry == (None, None):
674
def paths2ids(self, paths, trees=[], require_versioned=True):
675
"""See Tree.paths2ids().
677
This specialisation fast-paths the case where all the trees are in the
682
parents = self.get_parent_ids()
684
if not (isinstance(tree, DirStateRevisionTree) and tree._revision_id in
686
return super(WorkingTree4, self).paths2ids(paths, trees, require_versioned)
687
search_indexes = [0] + [1 + parents.index(tree._revision_id) for tree in trees]
688
# -- make all paths utf8 --
691
paths_utf8.add(path.encode('utf8'))
693
# -- paths is now a utf8 path set --
694
# -- get the state object and prepare it.
695
state = self.current_dirstate()
696
if False and (state._dirblock_state == dirstate.DirState.NOT_IN_MEMORY
697
and '' not in paths):
698
paths2ids = self._paths2ids_using_bisect
700
paths2ids = self._paths2ids_in_memory
701
return paths2ids(paths, search_indexes,
702
require_versioned=require_versioned)
704
def _paths2ids_in_memory(self, paths, search_indexes,
705
require_versioned=True):
706
state = self.current_dirstate()
707
state._read_dirblocks_if_needed()
708
def _entries_for_path(path):
709
"""Return a list with all the entries that match path for all ids.
711
dirname, basename = os.path.split(path)
712
key = (dirname, basename, '')
713
block_index, present = state._find_block_index_from_key(key)
715
# the block which should contain path is absent.
718
block = state._dirblocks[block_index][1]
719
entry_index, _ = state._find_entry_index(key, block)
720
# we may need to look at multiple entries at this path: walk while the paths match.
721
while (entry_index < len(block) and
722
block[entry_index][0][0:2] == key[0:2]):
723
result.append(block[entry_index])
726
if require_versioned:
727
# -- check all supplied paths are versioned in a search tree. --
730
path_entries = _entries_for_path(path)
732
# this specified path is not present at all: error
733
all_versioned = False
735
found_versioned = False
736
# for each id at this path
737
for entry in path_entries:
739
for index in search_indexes:
740
if entry[1][index][0] != 'a': # absent
741
found_versioned = True
742
# all good: found a versioned cell
744
if not found_versioned:
745
# none of the indexes was not 'absent' at all ids for this
747
all_versioned = False
749
if not all_versioned:
750
raise errors.PathsNotVersionedError(paths)
751
# -- remove redundancy in supplied paths to prevent over-scanning --
754
other_paths = paths.difference(set([path]))
755
if not osutils.is_inside_any(other_paths, path):
756
# this is a top level path, we must check it.
757
search_paths.add(path)
759
# for all search_indexs in each path at or under each element of
760
# search_paths, if the detail is relocated: add the id, and add the
761
# relocated path as one to search if its not searched already. If the
762
# detail is not relocated, add the id.
763
searched_paths = set()
765
def _process_entry(entry):
766
"""Look at search_indexes within entry.
768
If a specific tree's details are relocated, add the relocation
769
target to search_paths if not searched already. If it is absent, do
770
nothing. Otherwise add the id to found_ids.
772
for index in search_indexes:
773
if entry[1][index][0] == 'r': # relocated
774
if not osutils.is_inside_any(searched_paths, entry[1][index][1]):
775
search_paths.add(entry[1][index][1])
776
elif entry[1][index][0] != 'a': # absent
777
found_ids.add(entry[0][2])
779
current_root = search_paths.pop()
780
searched_paths.add(current_root)
781
# process the entries for this containing directory: the rest will be
782
# found by their parents recursively.
783
root_entries = _entries_for_path(current_root)
785
# this specified path is not present at all, skip it.
787
for entry in root_entries:
788
_process_entry(entry)
789
initial_key = (current_root, '', '')
790
block_index, _ = state._find_block_index_from_key(initial_key)
791
while (block_index < len(state._dirblocks) and
792
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
793
for entry in state._dirblocks[block_index][1]:
794
_process_entry(entry)
798
def _paths2ids_using_bisect(self, paths, search_indexes,
799
require_versioned=True):
800
state = self.current_dirstate()
803
split_paths = sorted(osutils.split(p) for p in paths)
804
found = state._bisect_recursive(split_paths)
806
if require_versioned:
807
found_dir_names = set(dir_name_id[:2] for dir_name_id in found)
808
for dir_name in split_paths:
809
if dir_name not in found_dir_names:
810
raise errors.PathsNotVersionedError(paths)
812
for dir_name_id, trees_info in found.iteritems():
813
for index in search_indexes:
814
if trees_info[index][0] not in ('r', 'a'):
815
found_ids.add(dir_name_id[2])
818
def read_working_inventory(self):
819
"""Read the working inventory.
821
This is a meaningless operation for dirstate, but we obey it anyhow.
823
return self.inventory
826
def revision_tree(self, revision_id):
827
"""See Tree.revision_tree.
829
WorkingTree4 supplies revision_trees for any basis tree.
831
revision_id = osutils.safe_revision_id(revision_id)
832
dirstate = self.current_dirstate()
833
parent_ids = dirstate.get_parent_ids()
834
if revision_id not in parent_ids:
835
raise errors.NoSuchRevisionInTree(self, revision_id)
836
if revision_id in dirstate.get_ghosts():
837
raise errors.NoSuchRevisionInTree(self, revision_id)
838
return DirStateRevisionTree(dirstate, revision_id,
839
self.branch.repository)
841
@needs_tree_write_lock
842
def set_last_revision(self, new_revision):
843
"""Change the last revision in the working tree."""
844
new_revision = osutils.safe_revision_id(new_revision)
845
parents = self.get_parent_ids()
846
if new_revision in (NULL_REVISION, None):
847
assert len(parents) < 2, (
848
"setting the last parent to none with a pending merge is "
850
self.set_parent_ids([])
852
self.set_parent_ids([new_revision] + parents[1:],
853
allow_leftmost_as_ghost=True)
855
@needs_tree_write_lock
856
def set_parent_ids(self, revision_ids, allow_leftmost_as_ghost=False):
857
"""Set the parent ids to revision_ids.
859
See also set_parent_trees. This api will try to retrieve the tree data
860
for each element of revision_ids from the trees repository. If you have
861
tree data already available, it is more efficient to use
862
set_parent_trees rather than set_parent_ids. set_parent_ids is however
863
an easier API to use.
865
:param revision_ids: The revision_ids to set as the parent ids of this
866
working tree. Any of these may be ghosts.
868
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
870
for revision_id in revision_ids:
872
revtree = self.branch.repository.revision_tree(revision_id)
873
# TODO: jam 20070213 KnitVersionedFile raises
874
# RevisionNotPresent rather than NoSuchRevision if a
875
# given revision_id is not present. Should Repository be
876
# catching it and re-raising NoSuchRevision?
877
except (errors.NoSuchRevision, errors.RevisionNotPresent):
879
trees.append((revision_id, revtree))
880
self.set_parent_trees(trees,
881
allow_leftmost_as_ghost=allow_leftmost_as_ghost)
883
@needs_tree_write_lock
884
def set_parent_trees(self, parents_list, allow_leftmost_as_ghost=False):
885
"""Set the parents of the working tree.
887
:param parents_list: A list of (revision_id, tree) tuples.
888
If tree is None, then that element is treated as an unreachable
889
parent tree - i.e. a ghost.
891
dirstate = self.current_dirstate()
892
if len(parents_list) > 0:
893
if not allow_leftmost_as_ghost and parents_list[0][1] is None:
894
raise errors.GhostRevisionUnusableHere(parents_list[0][0])
897
# convert absent trees to the null tree, which we convert back to
899
for rev_id, tree in parents_list:
900
rev_id = osutils.safe_revision_id(rev_id)
902
real_trees.append((rev_id, tree))
904
real_trees.append((rev_id,
905
self.branch.repository.revision_tree(None)))
906
ghosts.append(rev_id)
907
dirstate.set_parent_trees(real_trees, ghosts=ghosts)
910
def _set_root_id(self, file_id):
911
"""See WorkingTree.set_root_id."""
912
state = self.current_dirstate()
913
state.set_path_id('', file_id)
914
self._dirty = state._dirblock_state == dirstate.DirState.IN_MEMORY_MODIFIED
917
"""Unlock in format 4 trees needs to write the entire dirstate."""
918
if self._control_files._lock_count == 1:
919
self._write_hashcache_if_dirty()
920
# eventually we should do signature checking during read locks for
922
if self._control_files._lock_mode == 'w':
925
if self._dirstate is not None:
926
self._dirstate.unlock()
927
self._dirstate = None
928
self._inventory = None
929
# reverse order of locking.
931
return self._control_files.unlock()
935
@needs_tree_write_lock
936
def unversion(self, file_ids):
937
"""Remove the file ids in file_ids from the current versioned set.
939
When a file_id is unversioned, all of its children are automatically
942
:param file_ids: The file ids to stop versioning.
943
:raises: NoSuchId if any fileid is not currently versioned.
947
state = self.current_dirstate()
948
state._read_dirblocks_if_needed()
949
ids_to_unversion = set()
950
for file_id in file_ids:
951
ids_to_unversion.add(osutils.safe_file_id(file_id))
952
paths_to_unversion = set()
954
# check if the root is to be unversioned, if so, assert for now.
955
# walk the state marking unversioned things as absent.
956
# if there are any un-unversioned ids at the end, raise
957
for key, details in state._dirblocks[0][1]:
958
if (details[0][0] not in ('a', 'r') and # absent or relocated
959
key[2] in ids_to_unversion):
960
# I haven't written the code to unversion / yet - it should be
962
raise errors.BzrError('Unversioning the / is not currently supported')
964
while block_index < len(state._dirblocks):
965
# process one directory at a time.
966
block = state._dirblocks[block_index]
967
# first check: is the path one to remove - it or its children
969
for path in paths_to_unversion:
970
if (block[0].startswith(path) and
971
(len(block[0]) == len(path) or
972
block[0][len(path)] == '/')):
973
# this entire block should be deleted - its the block for a
974
# path to unversion; or the child of one
977
# TODO: trim paths_to_unversion as we pass by paths
979
# this block is to be deleted: process it.
980
# TODO: we can special case the no-parents case and
981
# just forget the whole block.
983
while entry_index < len(block[1]):
984
# Mark this file id as having been removed
985
ids_to_unversion.discard(block[1][entry_index][0][2])
986
if not state._make_absent(block[1][entry_index]):
988
# go to the next block. (At the moment we dont delete empty
993
while entry_index < len(block[1]):
994
entry = block[1][entry_index]
995
if (entry[1][0][0] in ('a', 'r') or # absent, relocated
997
entry[0][2] not in ids_to_unversion):
998
# ^ not an id to unversion
1001
if entry[1][0][0] == 'd':
1002
paths_to_unversion.add(pathjoin(entry[0][0], entry[0][1]))
1003
if not state._make_absent(entry):
1005
# we have unversioned this id
1006
ids_to_unversion.remove(entry[0][2])
1008
if ids_to_unversion:
1009
raise errors.NoSuchId(self, iter(ids_to_unversion).next())
1011
# have to change the legacy inventory too.
1012
if self._inventory is not None:
1013
for file_id in file_ids:
1014
self._inventory.remove_recursive_id(file_id)
1016
@needs_tree_write_lock
1017
def _write_inventory(self, inv):
1018
"""Write inventory as the current inventory."""
1019
assert not self._dirty, "attempting to write an inventory when the dirstate is dirty will cause data loss"
1020
self.current_dirstate().set_state_from_inventory(inv)
1025
class WorkingTreeFormat4(WorkingTreeFormat3):
1026
"""The first consolidated dirstate working tree format.
1029
- exists within a metadir controlling .bzr
1030
- includes an explicit version marker for the workingtree control
1031
files, separate from the BzrDir format
1032
- modifies the hash cache format
1033
- is new in bzr TODO FIXME SETBEFOREMERGE
1034
- uses a LockDir to guard access to it.
1037
def get_format_string(self):
1038
"""See WorkingTreeFormat.get_format_string()."""
1039
return "Bazaar Working Tree format 4\n"
1041
def get_format_description(self):
1042
"""See WorkingTreeFormat.get_format_description()."""
1043
return "Working tree format 4"
1045
def initialize(self, a_bzrdir, revision_id=None):
1046
"""See WorkingTreeFormat.initialize().
1048
revision_id allows creating a working tree at a different
1049
revision than the branch is at.
1051
revision_id = osutils.safe_revision_id(revision_id)
1052
if not isinstance(a_bzrdir.transport, LocalTransport):
1053
raise errors.NotLocalUrl(a_bzrdir.transport.base)
1054
transport = a_bzrdir.get_workingtree_transport(self)
1055
control_files = self._open_control_files(a_bzrdir)
1056
control_files.create_lock()
1057
control_files.lock_write()
1058
control_files.put_utf8('format', self.get_format_string())
1059
branch = a_bzrdir.open_branch()
1060
if revision_id is None:
1061
revision_id = branch.last_revision()
1062
local_path = transport.local_abspath('dirstate')
1063
state = dirstate.DirState.initialize(local_path)
1065
wt = WorkingTree4(a_bzrdir.root_transport.local_abspath('.'),
1069
_control_files=control_files)
1071
wt.lock_tree_write()
1073
#wt.current_dirstate().set_path_id('', NEWROOT)
1074
wt.set_last_revision(revision_id)
1076
basis = wt.basis_tree()
1078
transform.build_tree(basis, wt)
1081
control_files.unlock()
1086
def _open(self, a_bzrdir, control_files):
1087
"""Open the tree itself.
1089
:param a_bzrdir: the dir for the tree.
1090
:param control_files: the control files for the tree.
1092
return WorkingTree4(a_bzrdir.root_transport.local_abspath('.'),
1093
branch=a_bzrdir.open_branch(),
1096
_control_files=control_files)
1099
class DirStateRevisionTree(Tree):
1100
"""A revision tree pulling the inventory from a dirstate."""
1102
def __init__(self, dirstate, revision_id, repository):
1103
self._dirstate = dirstate
1104
self._revision_id = osutils.safe_revision_id(revision_id)
1105
self._repository = repository
1106
self._inventory = None
1108
self._dirstate_locked = False
1110
def annotate_iter(self, file_id):
1111
"""See Tree.annotate_iter"""
1112
w = self._repository.weave_store.get_weave(file_id,
1113
self._repository.get_transaction())
1114
return w.annotate_iter(self.inventory[file_id].revision)
1116
def _comparison_data(self, entry, path):
1117
"""See Tree._comparison_data."""
1119
return None, False, None
1120
# trust the entry as RevisionTree does, but this may not be
1121
# sensible: the entry might not have come from us?
1122
return entry.kind, entry.executable, None
1124
def _file_size(self, entry, stat_value):
1125
return entry.text_size
1127
def filter_unversioned_files(self, paths):
1128
"""Filter out paths that are not versioned.
1130
:return: set of paths.
1132
pred = self.has_filename
1133
return set((p for p in paths if not pred(p)))
1135
def _get_parent_index(self):
1136
"""Return the index in the dirstate referenced by this tree."""
1137
return self._dirstate.get_parent_ids().index(self._revision_id) + 1
1139
def _get_entry(self, file_id=None, path=None):
1140
"""Get the dirstate row for file_id or path.
1142
If either file_id or path is supplied, it is used as the key to lookup.
1143
If both are supplied, the fastest lookup is used, and an error is
1144
raised if they do not both point at the same row.
1146
:param file_id: An optional unicode file_id to be looked up.
1147
:param path: An optional unicode path to be looked up.
1148
:return: The dirstate row tuple for path/file_id, or (None, None)
1150
if file_id is None and path is None:
1151
raise errors.BzrError('must supply file_id or path')
1152
file_id = osutils.safe_file_id(file_id)
1153
if path is not None:
1154
path = path.encode('utf8')
1155
parent_index = self._get_parent_index()
1156
return self._dirstate._get_entry(parent_index, fileid_utf8=file_id, path_utf8=path)
1158
def _generate_inventory(self):
1159
"""Create and set self.inventory from the dirstate object.
1161
This is relatively expensive: we have to walk the entire dirstate.
1162
Ideally we would not, and instead would """
1163
assert self._locked, 'cannot generate inventory of an unlocked '\
1164
'dirstate revision tree'
1165
# separate call for profiling - makes it clear where the costs are.
1166
self._dirstate._read_dirblocks_if_needed()
1167
assert self._revision_id in self._dirstate.get_parent_ids(), \
1168
'parent %s has disappeared from %s' % (
1169
self._revision_id, self._dirstate.get_parent_ids())
1170
parent_index = self._dirstate.get_parent_ids().index(self._revision_id) + 1
1171
# This is identical now to the WorkingTree _generate_inventory except
1172
# for the tree index use.
1173
root_key, current_entry = self._dirstate._get_entry(parent_index, path_utf8='')
1174
current_id = root_key[2]
1175
assert current_entry[parent_index][0] == 'd'
1176
inv = Inventory(root_id=current_id, revision_id=self._revision_id)
1177
inv.root.revision = current_entry[parent_index][4]
1178
# Turn some things into local variables
1179
minikind_to_kind = dirstate.DirState._minikind_to_kind
1180
factory = entry_factory
1181
utf8_decode = cache_utf8._utf8_decode
1182
inv_byid = inv._byid
1183
# we could do this straight out of the dirstate; it might be fast
1184
# and should be profiled - RBC 20070216
1185
parent_ies = {'' : inv.root}
1186
for block in self._dirstate._dirblocks[1:]: #skip root
1189
parent_ie = parent_ies[dirname]
1191
# all the paths in this block are not versioned in this tree
1193
for key, entry in block[1]:
1194
minikind, link_or_sha1, size, executable, revid = entry[parent_index]
1195
if minikind in ('a', 'r'): # absent, relocated
1199
name_unicode = utf8_decode(name)[0]
1201
kind = minikind_to_kind[minikind]
1202
inv_entry = factory[kind](file_id, name_unicode,
1204
inv_entry.revision = revid
1206
inv_entry.executable = executable
1207
inv_entry.text_size = size
1208
inv_entry.text_sha1 = link_or_sha1
1209
elif kind == 'directory':
1210
parent_ies[(dirname + '/' + name).strip('/')] = inv_entry
1211
elif kind == 'symlink':
1212
inv_entry.executable = False
1213
inv_entry.text_size = size
1214
inv_entry.symlink_target = utf8_decode(link_or_sha1)[0]
1216
raise Exception, kind
1217
# These checks cost us around 40ms on a 55k entry tree
1218
assert file_id not in inv_byid
1219
assert name_unicode not in parent_ie.children
1220
inv_byid[file_id] = inv_entry
1221
parent_ie.children[name_unicode] = inv_entry
1222
self._inventory = inv
1224
def get_file_mtime(self, file_id, path=None):
1225
"""Return the modification time for this record.
1227
We return the timestamp of the last-changed revision.
1229
# Make sure the file exists
1230
entry = self._get_entry(file_id, path=path)
1231
if entry == (None, None): # do we raise?
1233
parent_index = self._get_parent_index()
1234
last_changed_revision = entry[1][parent_index][4]
1235
return self._repository.get_revision(last_changed_revision).timestamp
1237
def get_file_sha1(self, file_id, path=None, stat_value=None):
1238
# TODO: if path is present, fast-path on that, as inventory
1239
# might not be present
1240
ie = self.inventory[file_id]
1241
if ie.kind == "file":
1245
def get_file(self, file_id):
1246
return StringIO(self.get_file_text(file_id))
1248
def get_file_lines(self, file_id):
1249
ie = self.inventory[file_id]
1250
return self._repository.weave_store.get_weave(file_id,
1251
self._repository.get_transaction()).get_lines(ie.revision)
1253
def get_file_size(self, file_id):
1254
return self.inventory[file_id].text_size
1256
def get_file_text(self, file_id):
1257
return ''.join(self.get_file_lines(file_id))
1259
def get_symlink_target(self, file_id):
1260
entry = self._get_entry(file_id=file_id)
1261
parent_index = self._get_parent_index()
1262
if entry[1][parent_index][0] != 'l':
1265
# At present, none of the tree implementations supports non-ascii
1266
# symlink targets. So we will just assume that the dirstate path is
1268
return entry[1][parent_index][1]
1270
def get_revision_id(self):
1271
"""Return the revision id for this tree."""
1272
return self._revision_id
1274
def _get_inventory(self):
1275
if self._inventory is not None:
1276
return self._inventory
1277
self._generate_inventory()
1278
return self._inventory
1280
inventory = property(_get_inventory,
1281
doc="Inventory of this Tree")
1283
def get_parent_ids(self):
1284
"""The parents of a tree in the dirstate are not cached."""
1285
return self._repository.get_revision(self._revision_id).parent_ids
1287
def has_filename(self, filename):
1288
return bool(self.path2id(filename))
1290
def kind(self, file_id):
1291
return self.inventory[file_id].kind
1293
def is_executable(self, file_id, path=None):
1294
ie = self.inventory[file_id]
1295
if ie.kind != "file":
1297
return ie.executable
1299
def list_files(self, include_root=False):
1300
# We use a standard implementation, because DirStateRevisionTree is
1301
# dealing with one of the parents of the current state
1302
inv = self._get_inventory()
1303
entries = inv.iter_entries()
1304
if self.inventory.root is not None and not include_root:
1306
for path, entry in entries:
1307
yield path, 'V', entry.kind, entry.file_id, entry
1309
def lock_read(self):
1310
"""Lock the tree for a set of operations."""
1311
if not self._locked:
1312
self._repository.lock_read()
1313
if self._dirstate._lock_token is None:
1314
self._dirstate.lock_read()
1315
self._dirstate_locked = True
1319
def path2id(self, path):
1320
"""Return the id for path in this tree."""
1321
# lookup by path: faster than splitting and walking the ivnentory.
1322
entry = self._get_entry(path=path)
1323
if entry == (None, None):
1328
"""Unlock, freeing any cache memory used during the lock."""
1329
# outside of a lock, the inventory is suspect: release it.
1331
if not self._locked:
1332
self._inventory = None
1334
if self._dirstate_locked:
1335
self._dirstate.unlock()
1336
self._dirstate_locked = False
1337
self._repository.unlock()
1339
def walkdirs(self, prefix=""):
1340
# TODO: jam 20070215 This is the cheap way by cheating and using the
1341
# RevisionTree implementation.
1342
# This should be cleaned up to use the much faster Dirstate code
1343
# This is a little tricky, though, because the dirstate is
1344
# indexed by current path, not by parent path.
1345
# So for now, we just build up the parent inventory, and extract
1346
# it the same way RevisionTree does.
1347
_directory = 'directory'
1348
inv = self._get_inventory()
1349
top_id = inv.path2id(prefix)
1353
pending = [(prefix, top_id)]
1356
relpath, file_id = pending.pop()
1357
# 0 - relpath, 1- file-id
1359
relroot = relpath + '/'
1362
# FIXME: stash the node in pending
1363
entry = inv[file_id]
1364
for name, child in entry.sorted_children():
1365
toppath = relroot + name
1366
dirblock.append((toppath, name, child.kind, None,
1367
child.file_id, child.kind
1369
yield (relpath, entry.file_id), dirblock
1370
# push the user specified dirs from dirblock
1371
for dir in reversed(dirblock):
1372
if dir[2] == _directory:
1373
pending.append((dir[0], dir[4]))
1376
class InterDirStateTree(InterTree):
1377
"""Fast path optimiser for changes_from with dirstate trees."""
1379
def __init__(self, source, target):
1380
super(InterDirStateTree, self).__init__(source, target)
1381
if not InterDirStateTree.is_compatible(source, target):
1382
raise Exception, "invalid source %r and target %r" % (source, target)
1385
def make_source_parent_tree(source, target):
1386
"""Change the source tree into a parent of the target."""
1387
revid = source.commit('record tree')
1388
target.branch.repository.fetch(source.branch.repository, revid)
1389
target.set_parent_ids([revid])
1390
return target.basis_tree(), target
1392
_matching_from_tree_format = WorkingTreeFormat4()
1393
_matching_to_tree_format = WorkingTreeFormat4()
1394
_test_mutable_trees_to_test_trees = make_source_parent_tree
1396
def _iter_changes(self, include_unchanged=False,
1397
specific_files=None, pb=None, extra_trees=[],
1398
require_versioned=True):
1399
"""Return the changes from source to target.
1401
:return: An iterator that yields tuples. See InterTree._iter_changes
1403
:param specific_files: An optional list of file paths to restrict the
1404
comparison to. When mapping filenames to ids, all matches in all
1405
trees (including optional extra_trees) are used, and all children of
1406
matched directories are included.
1407
:param include_unchanged: An optional boolean requesting the inclusion of
1408
unchanged entries in the result.
1409
:param extra_trees: An optional list of additional trees to use when
1410
mapping the contents of specific_files (paths) to file_ids.
1411
:param require_versioned: If True, all files in specific_files must be
1412
versioned in one of source, target, extra_trees or
1413
PathsNotVersionedError is raised.
1415
utf8_decode = cache_utf8._utf8_decode
1416
_minikind_to_kind = dirstate.DirState._minikind_to_kind
1417
# NB: show_status depends on being able to pass in non-versioned files
1418
# and report them as unknown
1419
# TODO: handle extra trees in the dirstate.
1421
for f in super(InterDirStateTree, self)._iter_changes(
1422
include_unchanged, specific_files, pb, extra_trees,
1426
parent_ids = self.target.get_parent_ids()
1428
if self.source._revision_id == NULL_REVISION:
1430
indices = (target_index,)
1432
assert (self.source._revision_id in parent_ids), \
1433
"Failure: source._revision_id: %s not in target.parent_ids(%s)" % (
1434
self.source._revision_id, parent_ids)
1435
source_index = 1 + parent_ids.index(self.source._revision_id)
1436
indices = (source_index,target_index)
1437
# -- make all specific_files utf8 --
1439
specific_files_utf8 = set()
1440
for path in specific_files:
1441
specific_files_utf8.add(path.encode('utf8'))
1442
specific_files = specific_files_utf8
1444
specific_files = set([''])
1445
# -- specific_files is now a utf8 path set --
1446
# -- get the state object and prepare it.
1447
state = self.target.current_dirstate()
1448
state._read_dirblocks_if_needed()
1449
def _entries_for_path(path):
1450
"""Return a list with all the entries that match path for all ids.
1452
dirname, basename = os.path.split(path)
1453
key = (dirname, basename, '')
1454
block_index, present = state._find_block_index_from_key(key)
1456
# the block which should contain path is absent.
1459
block = state._dirblocks[block_index][1]
1460
entry_index, _ = state._find_entry_index(key, block)
1461
# we may need to look at multiple entries at this path: walk while the specific_files match.
1462
while (entry_index < len(block) and
1463
block[entry_index][0][0:2] == key[0:2]):
1464
result.append(block[entry_index])
1467
if require_versioned:
1468
# -- check all supplied paths are versioned in a search tree. --
1469
all_versioned = True
1470
for path in specific_files:
1471
path_entries = _entries_for_path(path)
1472
if not path_entries:
1473
# this specified path is not present at all: error
1474
all_versioned = False
1476
found_versioned = False
1477
# for each id at this path
1478
for entry in path_entries:
1480
for index in indices:
1481
if entry[1][index][0] != 'a': # absent
1482
found_versioned = True
1483
# all good: found a versioned cell
1485
if not found_versioned:
1486
# none of the indexes was not 'absent' at all ids for this
1488
all_versioned = False
1490
if not all_versioned:
1491
raise errors.PathsNotVersionedError(paths)
1492
# -- remove redundancy in supplied specific_files to prevent over-scanning --
1493
search_specific_files = set()
1494
for path in specific_files:
1495
other_specific_files = specific_files.difference(set([path]))
1496
if not osutils.is_inside_any(other_specific_files, path):
1497
# this is a top level path, we must check it.
1498
search_specific_files.add(path)
1500
# compare source_index and target_index at or under each element of search_specific_files.
1501
# follow the following comparison table. Note that we only want to do diff operations when
1502
# the target is fdl because thats when the walkdirs logic will have exposed the pathinfo
1506
# Source | Target | disk | action
1507
# r | fdl | | add source to search, add id path move and perform
1508
# | | | diff check on source-target
1509
# r | fdl | a | dangling file that was present in the basis.
1511
# r | a | | add source to search
1513
# r | r | | this path is present in a non-examined tree, skip.
1514
# r | r | a | this path is present in a non-examined tree, skip.
1515
# a | fdl | | add new id
1516
# a | fdl | a | dangling locally added file, skip
1517
# a | a | | not present in either tree, skip
1518
# a | a | a | not present in any tree, skip
1519
# a | r | | not present in either tree at this path, skip as it
1520
# | | | may not be selected by the users list of paths.
1521
# a | r | a | not present in either tree at this path, skip as it
1522
# | | | may not be selected by the users list of paths.
1523
# fdl | fdl | | content in both: diff them
1524
# fdl | fdl | a | deleted locally, but not unversioned - show as deleted ?
1525
# fdl | a | | unversioned: output deleted id for now
1526
# fdl | a | a | unversioned and deleted: output deleted id
1527
# fdl | r | | relocated in this tree, so add target to search.
1528
# | | | Dont diff, we will see an r,fd; pair when we reach
1529
# | | | this id at the other path.
1530
# fdl | r | a | relocated in this tree, so add target to search.
1531
# | | | Dont diff, we will see an r,fd; pair when we reach
1532
# | | | this id at the other path.
1534
# for all search_indexs in each path at or under each element of
1535
# search_specific_files, if the detail is relocated: add the id, and add the
1536
# relocated path as one to search if its not searched already. If the
1537
# detail is not relocated, add the id.
1538
searched_specific_files = set()
1539
NULL_PARENT_DETAILS = dirstate.DirState.NULL_PARENT_DETAILS
1540
# Using a list so that we can access the values and change them in
1541
# nested scope. Each one is [path, file_id, entry]
1542
last_source_parent = [None, None, None]
1543
last_target_parent = [None, None, None]
1545
def _process_entry(entry, path_info):
1546
"""Compare an entry and real disk to generate delta information.
1548
:param path_info: top_relpath, basename, kind, lstat, abspath for
1549
the path of entry. If None, then the path is considered absent.
1550
(Perhaps we should pass in a concrete entry for this ?)
1552
# TODO: when a parent has been renamed, dont emit path renames for children,
1553
if source_index is None:
1554
source_details = NULL_PARENT_DETAILS
1556
source_details = entry[1][source_index]
1557
target_details = entry[1][target_index]
1558
source_minikind = source_details[0]
1559
target_minikind = target_details[0]
1560
if source_minikind in 'fdlr' and target_minikind in 'fdl':
1561
# claimed content in both: diff
1562
# r | fdl | | add source to search, add id path move and perform
1563
# | | | diff check on source-target
1564
# r | fdl | a | dangling file that was present in the basis.
1566
if source_minikind in 'r':
1567
# add the source to the search path to find any children it
1568
# has. TODO ? : only add if it is a container ?
1569
if not osutils.is_inside_any(searched_specific_files,
1571
search_specific_files.add(source_details[1])
1572
# generate the old path; this is needed for stating later
1574
old_path = source_details[1]
1575
old_dirname, old_basename = os.path.split(old_path)
1576
path = pathjoin(entry[0][0], entry[0][1])
1577
old_entry = state._get_entry(source_index,
1579
# update the source details variable to be the real
1581
source_details = old_entry[1][source_index]
1582
source_minikind = source_details[0]
1584
old_dirname = entry[0][0]
1585
old_basename = entry[0][1]
1586
old_path = path = pathjoin(old_dirname, old_basename)
1587
if path_info is None:
1588
# the file is missing on disk, show as removed.
1589
old_path = pathjoin(entry[0][0], entry[0][1])
1590
content_change = True
1594
# source and target are both versioned and disk file is present.
1595
target_kind = path_info[2]
1596
if target_kind == 'directory':
1597
if source_minikind != 'd':
1598
content_change = True
1600
# directories have no fingerprint
1601
content_change = False
1603
elif target_kind == 'file':
1604
if source_minikind != 'f':
1605
content_change = True
1607
# has it changed? fast path: size, slow path: sha1.
1608
if source_details[2] != path_info[3].st_size:
1609
content_change = True
1611
# maybe the same. Get the hash
1612
new_hash = self.target._hashcache.get_sha1(
1614
content_change = (new_hash != source_details[1])
1616
stat.S_ISREG(path_info[3].st_mode)
1617
and stat.S_IEXEC & path_info[3].st_mode)
1618
elif target_kind == 'symlink':
1619
if source_minikind != 'l':
1620
content_change = True
1622
# TODO: check symlink supported for windows users
1623
# and grab from target state here.
1624
link_target = os.readlink(path_info[4])
1625
content_change = (link_target != source_details[1])
1628
raise Exception, "unknown kind %s" % path_info[2]
1629
# parent id is the entry for the path in the target tree
1630
if old_dirname == last_source_parent[0]:
1631
source_parent_id = last_source_parent[1]
1633
source_parent_entry = state._get_entry(source_index,
1634
path_utf8=old_dirname)
1635
source_parent_id = source_parent_entry[0][2]
1636
if source_parent_id == entry[0][2]:
1637
# This is the root, so the parent is None
1638
source_parent_id = None
1640
last_source_parent[0] = old_dirname
1641
last_source_parent[1] = source_parent_id
1642
last_source_parent[2] = source_parent_entry
1644
new_dirname = entry[0][0]
1645
if new_dirname == last_target_parent[0]:
1646
target_parent_id = last_target_parent[1]
1648
# TODO: We don't always need to do the lookup, because the
1649
# parent entry will be the same as the source entry.
1650
target_parent_entry = state._get_entry(target_index,
1651
path_utf8=new_dirname)
1652
target_parent_id = target_parent_entry[0][2]
1653
if target_parent_id == entry[0][2]:
1654
# This is the root, so the parent is None
1655
target_parent_id = None
1657
last_target_parent[0] = new_dirname
1658
last_target_parent[1] = target_parent_id
1659
last_target_parent[2] = target_parent_entry
1661
source_exec = source_details[3]
1662
path_unicode = utf8_decode(path)[0]
1663
return ((entry[0][2], path_unicode, content_change,
1665
(source_parent_id, target_parent_id),
1666
(old_basename, entry[0][1]),
1667
(_minikind_to_kind[source_minikind], target_kind),
1668
(source_exec, target_exec)),)
1669
elif source_minikind in 'a' and target_minikind in 'fdl':
1670
# looks like a new file
1671
if path_info is not None:
1672
path = pathjoin(entry[0][0], entry[0][1])
1673
# parent id is the entry for the path in the target tree
1674
# TODO: these are the same for an entire directory: cache em.
1675
parent_id = state._get_entry(target_index, path_utf8=entry[0][0])[0][2]
1676
if parent_id == entry[0][2]:
1679
new_executable = bool(
1680
stat.S_ISREG(path_info[3].st_mode)
1681
and stat.S_IEXEC & path_info[3].st_mode)
1682
path_unicode = utf8_decode(path)[0]
1683
return ((entry[0][2], path_unicode, True,
1686
(None, entry[0][1]),
1687
(None, path_info[2]),
1688
(None, new_executable)),)
1690
# but its not on disk: we deliberately treat this as just
1691
# never-present. (Why ?! - RBC 20070224)
1693
elif source_minikind in 'fdl' and target_minikind in 'a':
1694
# unversioned, possibly, or possibly not deleted: we dont care.
1695
# if its still on disk, *and* theres no other entry at this
1696
# path [we dont know this in this routine at the moment -
1697
# perhaps we should change this - then it would be an unknown.
1698
old_path = pathjoin(entry[0][0], entry[0][1])
1699
# parent id is the entry for the path in the target tree
1700
parent_id = state._get_entry(source_index, path_utf8=entry[0][0])[0][2]
1701
if parent_id == entry[0][2]:
1703
old_path_unicode = utf8_decode(old_path)[0]
1704
return ((entry[0][2], old_path_unicode, True,
1707
(entry[0][1], None),
1708
(_minikind_to_kind[source_minikind], None),
1709
(source_details[3], None)),)
1710
elif source_minikind in 'fdl' and target_minikind in 'r':
1711
# a rename; could be a true rename, or a rename inherited from
1712
# a renamed parent. TODO: handle this efficiently. Its not
1713
# common case to rename dirs though, so a correct but slow
1714
# implementation will do.
1715
if not osutils.is_inside_any(searched_specific_files, target_details[1]):
1716
search_specific_files.add(target_details[1])
1717
elif source_minikind in 'r' and target_minikind in 'r':
1718
# neither of the selected trees contain this file,
1719
# so skip over it. This is not currently directly tested, but
1720
# is indirectly via test_too_much.TestCommands.test_conflicts.
1723
print "*******", source_minikind, target_minikind
1724
import pdb;pdb.set_trace()
1726
while search_specific_files:
1727
# TODO: the pending list should be lexically sorted?
1728
current_root = search_specific_files.pop()
1729
searched_specific_files.add(current_root)
1730
# process the entries for this containing directory: the rest will be
1731
# found by their parents recursively.
1732
root_entries = _entries_for_path(current_root)
1733
root_abspath = self.target.abspath(current_root)
1735
root_stat = os.lstat(root_abspath)
1737
if e.errno == errno.ENOENT:
1738
# the path does not exist: let _process_entry know that.
1739
root_dir_info = None
1741
# some other random error: hand it up.
1744
root_dir_info = ('', current_root,
1745
osutils.file_kind_from_stat_mode(root_stat.st_mode), root_stat,
1747
if not root_entries and not root_dir_info:
1748
# this specified path is not present at all, skip it.
1750
for entry in root_entries:
1751
for result in _process_entry(entry, root_dir_info):
1752
# this check should probably be outside the loop: one
1753
# 'iterate two trees' api, and then _iter_changes filters
1754
# unchanged pairs. - RBC 20070226
1755
if (include_unchanged
1756
or result[2] # content change
1757
or result[3][0] != result[3][1] # versioned status
1758
or result[4][0] != result[4][1] # parent id
1759
or result[5][0] != result[5][1] # name
1760
or result[6][0] != result[6][1] # kind
1761
or result[7][0] != result[7][1] # executable
1764
dir_iterator = osutils._walkdirs_utf8(root_abspath, prefix=current_root)
1765
initial_key = (current_root, '', '')
1766
block_index, _ = state._find_block_index_from_key(initial_key)
1767
if block_index == 0:
1768
# we have processed the total root already, but because the
1769
# initial key matched it we should skip it here.
1772
current_dir_info = dir_iterator.next()
1774
if e.errno in (errno.ENOENT, errno.ENOTDIR):
1775
# there may be directories in the inventory even though
1776
# this path is not a file on disk: so mark it as end of
1778
current_dir_info = None
1782
if current_dir_info[0][0] == '':
1783
# remove .bzr from iteration
1784
bzr_index = bisect_left(current_dir_info[1], ('.bzr',))
1785
assert current_dir_info[1][bzr_index][0] == '.bzr'
1786
del current_dir_info[1][bzr_index]
1787
# walk until both the directory listing and the versioned metadata
1788
# are exhausted. TODO: reevaluate this, perhaps we should stop when
1789
# the versioned data runs out.
1790
if (block_index < len(state._dirblocks) and
1791
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
1792
current_block = state._dirblocks[block_index]
1794
current_block = None
1795
while (current_dir_info is not None or
1796
current_block is not None):
1797
if (current_dir_info and current_block
1798
and current_dir_info[0][0] != current_block[0]):
1799
if current_dir_info[0][0] < current_block[0] :
1800
# import pdb; pdb.set_trace()
1801
# print 'unversioned dir'
1802
# filesystem data refers to paths not covered by the dirblock.
1803
# this has two possibilities:
1804
# A) it is versioned but empty, so there is no block for it
1805
# B) it is not versioned.
1806
# in either case it was processed by the containing directories walk:
1807
# if it is root/foo, when we walked root we emitted it,
1808
# or if we ere given root/foo to walk specifically, we
1809
# emitted it when checking the walk-root entries
1810
# advance the iterator and loop - we dont need to emit it.
1812
current_dir_info = dir_iterator.next()
1813
except StopIteration:
1814
current_dir_info = None
1816
# We have a dirblock entry for this location, but there
1817
# is no filesystem path for this. This is most likely
1818
# because a directory was removed from the disk.
1819
# We don't have to report the missing directory,
1820
# because that should have already been handled, but we
1821
# need to handle all of the files that are contained
1823
for current_entry in current_block[1]:
1824
# entry referring to file not present on disk.
1825
# advance the entry only, after processing.
1826
for result in _process_entry(current_entry, None):
1827
# this check should probably be outside the loop: one
1828
# 'iterate two trees' api, and then _iter_changes filters
1829
# unchanged pairs. - RBC 20070226
1830
if (include_unchanged
1831
or result[2] # content change
1832
or result[3][0] != result[3][1] # versioned status
1833
or result[4][0] != result[4][1] # parent id
1834
or result[5][0] != result[5][1] # name
1835
or result[6][0] != result[6][1] # kind
1836
or result[7][0] != result[7][1] # executable
1840
if (block_index < len(state._dirblocks) and
1841
osutils.is_inside(current_root,
1842
state._dirblocks[block_index][0])):
1843
current_block = state._dirblocks[block_index]
1845
current_block = None
1848
if current_block and entry_index < len(current_block[1]):
1849
current_entry = current_block[1][entry_index]
1851
current_entry = None
1852
advance_entry = True
1854
if current_dir_info and path_index < len(current_dir_info[1]):
1855
current_path_info = current_dir_info[1][path_index]
1857
current_path_info = None
1859
while (current_entry is not None or
1860
current_path_info is not None):
1861
if current_entry is None:
1862
# no more entries: yield current_pathinfo as an
1863
# unversioned file: its not the same as a path in any
1864
# tree in the dirstate.
1865
new_executable = bool(
1866
stat.S_ISREG(current_path_info[3].st_mode)
1867
and stat.S_IEXEC & current_path_info[3].st_mode)
1868
yield (None, current_path_info[0], True,
1871
(None, current_path_info[1]),
1872
(None, current_path_info[2]),
1873
(None, new_executable))
1874
elif current_path_info is None:
1875
# no path is fine: the per entry code will handle it.
1876
for result in _process_entry(current_entry, current_path_info):
1877
# this check should probably be outside the loop: one
1878
# 'iterate two trees' api, and then _iter_changes filters
1879
# unchanged pairs. - RBC 20070226
1880
if (include_unchanged
1881
or result[2] # content change
1882
or result[3][0] != result[3][1] # versioned status
1883
or result[4][0] != result[4][1] # parent id
1884
or result[5][0] != result[5][1] # name
1885
or result[6][0] != result[6][1] # kind
1886
or result[7][0] != result[7][1] # executable
1889
elif current_entry[0][1] != current_path_info[1]:
1890
if current_path_info[1] < current_entry[0][1]:
1891
# extra file on disk: pass for now, but only
1892
# increment the path, not the entry
1893
# import pdb; pdb.set_trace()
1894
# print 'unversioned file'
1895
advance_entry = False
1897
# entry referring to file not present on disk.
1898
# advance the entry only, after processing.
1899
for result in _process_entry(current_entry, None):
1900
# this check should probably be outside the loop: one
1901
# 'iterate two trees' api, and then _iter_changes filters
1902
# unchanged pairs. - RBC 20070226
1903
if (include_unchanged
1904
or result[2] # content change
1905
or result[3][0] != result[3][1] # versioned status
1906
or result[4][0] != result[4][1] # parent id
1907
or result[5][0] != result[5][1] # name
1908
or result[6][0] != result[6][1] # kind
1909
or result[7][0] != result[7][1] # executable
1912
advance_path = False
1914
for result in _process_entry(current_entry, current_path_info):
1915
# this check should probably be outside the loop: one
1916
# 'iterate two trees' api, and then _iter_changes filters
1917
# unchanged pairs. - RBC 20070226
1918
if (include_unchanged
1919
or result[2] # content change
1920
or result[3][0] != result[3][1] # versioned status
1921
or result[4][0] != result[4][1] # parent id
1922
or result[5][0] != result[5][1] # name
1923
or result[6][0] != result[6][1] # kind
1924
or result[7][0] != result[7][1] # executable
1927
if advance_entry and current_entry is not None:
1929
if entry_index < len(current_block[1]):
1930
current_entry = current_block[1][entry_index]
1932
current_entry = None
1934
advance_entry = True # reset the advance flaga
1935
if advance_path and current_path_info is not None:
1937
if path_index < len(current_dir_info[1]):
1938
current_path_info = current_dir_info[1][path_index]
1940
current_path_info = None
1942
advance_path = True # reset the advance flagg.
1943
if current_block is not None:
1945
if (block_index < len(state._dirblocks) and
1946
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
1947
current_block = state._dirblocks[block_index]
1949
current_block = None
1950
if current_dir_info is not None:
1952
current_dir_info = dir_iterator.next()
1953
except StopIteration:
1954
current_dir_info = None
1958
def is_compatible(source, target):
1959
# the target must be a dirstate working tree
1960
if not isinstance(target, WorkingTree4):
1962
# the source must be a revtreee or dirstate rev tree.
1963
if not isinstance(source,
1964
(revisiontree.RevisionTree, DirStateRevisionTree)):
1966
# the source revid must be in the target dirstate
1967
if not (source._revision_id == NULL_REVISION or
1968
source._revision_id in target.get_parent_ids()):
1969
# TODO: what about ghosts? it may well need to
1970
# check for them explicitly.
1974
InterTree.register_optimiser(InterDirStateTree)
added
removed
bzrlib/workingtree_4.py
