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 = os.path.join(*key[0:2]).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()."""
449
state = self.current_dirstate()
451
assert not isinstance(from_paths, basestring)
452
to_dir_utf8 = to_dir.encode('utf8')
453
to_entry_dirname, to_basename = os.path.split(to_dir_utf8)
454
id_index = state._get_id_index()
455
# check destination directory
456
# get the details for it
457
to_entry_block_index, to_entry_entry_index, dir_present, entry_present = \
458
state._get_block_entry_index(to_entry_dirname, to_basename, 0)
459
if not entry_present:
460
raise errors.BzrMoveFailedError('', to_dir,
461
errors.NotInWorkingDirectory(to_dir))
462
to_entry = state._dirblocks[to_entry_block_index][1][to_entry_entry_index]
463
# get a handle on the block itself.
464
to_block_index = state._ensure_block(
465
to_entry_block_index, to_entry_entry_index, to_dir_utf8)
466
to_block = state._dirblocks[to_block_index]
467
to_abs = self.abspath(to_dir)
468
if not isdir(to_abs):
469
raise errors.BzrMoveFailedError('',to_dir,
470
errors.NotADirectory(to_abs))
472
if to_entry[1][0][0] != 'd':
473
raise errors.BzrMoveFailedError('',to_dir,
474
errors.NotADirectory(to_abs))
476
if self._inventory is not None:
477
update_inventory = True
479
to_dir_ie = inv[to_dir_id]
480
to_dir_id = to_entry[0][2]
482
update_inventory = False
485
def move_one(old_entry, from_path_utf8, minikind, executable,
486
fingerprint, packed_stat, size,
487
to_block, to_key, to_path_utf8):
488
state._make_absent(old_entry)
489
from_key = old_entry[0]
491
lambda:state.update_minimal(from_key,
493
executable=executable,
494
fingerprint=fingerprint,
495
packed_stat=packed_stat,
497
path_utf8=from_path_utf8))
498
state.update_minimal(to_key,
500
executable=executable,
501
fingerprint=fingerprint,
502
packed_stat=packed_stat,
504
path_utf8=to_path_utf8)
505
added_entry_index, _ = state._find_entry_index(to_key, to_block[1])
506
new_entry = to_block[1][added_entry_index]
507
rollbacks.append(lambda:state._make_absent(new_entry))
509
# create rename entries and tuples
510
for from_rel in from_paths:
511
# from_rel is 'pathinroot/foo/bar'
512
from_rel_utf8 = from_rel.encode('utf8')
513
from_dirname, from_tail = osutils.split(from_rel)
514
from_dirname, from_tail_utf8 = osutils.split(from_rel_utf8)
515
from_entry = self._get_entry(path=from_rel)
516
if from_entry == (None, None):
517
raise errors.BzrMoveFailedError(from_rel,to_dir,
518
errors.NotVersionedError(path=str(from_rel)))
520
from_id = from_entry[0][2]
521
to_rel = pathjoin(to_dir, from_tail)
522
to_rel_utf8 = pathjoin(to_dir_utf8, from_tail_utf8)
523
item_to_entry = self._get_entry(path=to_rel)
524
if item_to_entry != (None, None):
525
raise errors.BzrMoveFailedError(from_rel, to_rel,
526
"Target is already versioned.")
528
if from_rel == to_rel:
529
raise errors.BzrMoveFailedError(from_rel, to_rel,
530
"Source and target are identical.")
532
from_missing = not self.has_filename(from_rel)
533
to_missing = not self.has_filename(to_rel)
540
raise errors.BzrMoveFailedError(from_rel, to_rel,
541
errors.NoSuchFile(path=to_rel,
542
extra="New file has not been created yet"))
544
# neither path exists
545
raise errors.BzrRenameFailedError(from_rel, to_rel,
546
errors.PathsDoNotExist(paths=(from_rel, to_rel)))
548
if from_missing: # implicitly just update our path mapping
551
raise errors.RenameFailedFilesExist(from_rel, to_rel,
552
extra="(Use --after to update the Bazaar id)")
555
def rollback_rename():
556
"""A single rename has failed, roll it back."""
558
for rollback in reversed(rollbacks):
562
import pdb;pdb.set_trace()
563
exc_info = sys.exc_info()
565
raise exc_info[0], exc_info[1], exc_info[2]
567
# perform the disk move first - its the most likely failure point.
569
from_rel_abs = self.abspath(from_rel)
570
to_rel_abs = self.abspath(to_rel)
572
osutils.rename(from_rel_abs, to_rel_abs)
574
raise errors.BzrMoveFailedError(from_rel, to_rel, e[1])
575
rollbacks.append(lambda: osutils.rename(to_rel_abs, from_rel_abs))
577
# perform the rename in the inventory next if needed: its easy
581
from_entry = inv[from_id]
582
current_parent = from_entry.parent_id
583
inv.rename(from_id, to_dir_id, from_tail)
585
lambda: inv.rename(from_id, current_parent, from_tail))
586
# finally do the rename in the dirstate, which is a little
587
# tricky to rollback, but least likely to need it.
588
old_block_index, old_entry_index, dir_present, file_present = \
589
state._get_block_entry_index(from_dirname, from_tail_utf8, 0)
590
old_block = state._dirblocks[old_block_index][1]
591
old_entry = old_block[old_entry_index]
592
from_key, old_entry_details = old_entry
593
cur_details = old_entry_details[0]
595
to_key = ((to_block[0],) + from_key[1:3])
596
minikind = cur_details[0]
597
move_one(old_entry, from_path_utf8=from_rel_utf8,
599
executable=cur_details[3],
600
fingerprint=cur_details[1],
601
packed_stat=cur_details[4],
605
to_path_utf8=to_rel_utf8)
608
def update_dirblock(from_dir, to_key, to_dir_utf8):
609
"""all entries in this block need updating.
611
TODO: This is pretty ugly, and doesn't support
612
reverting, but it works.
614
assert from_dir != '', "renaming root not supported"
615
from_key = (from_dir, '')
616
from_block_idx, present = \
617
state._find_block_index_from_key(from_key)
619
# This is the old record, if it isn't present, then
620
# there is theoretically nothing to update.
621
# (Unless it isn't present because of lazy loading,
622
# but we don't do that yet)
624
from_block = state._dirblocks[from_block_idx]
625
to_block_index, to_entry_index, _, _ = \
626
state._get_block_entry_index(to_key[0], to_key[1], 0)
627
to_block_index = state._ensure_block(
628
to_block_index, to_entry_index, to_dir_utf8)
629
to_block = state._dirblocks[to_block_index]
630
for entry in from_block[1]:
631
assert entry[0][0] == from_dir
632
cur_details = entry[1][0]
633
to_key = (to_dir_utf8, entry[0][1], entry[0][2])
634
from_path_utf8 = osutils.pathjoin(entry[0][0], entry[0][1])
635
to_path_utf8 = osutils.pathjoin(to_dir_utf8, entry[0][1])
636
minikind = cur_details[0]
637
move_one(entry, from_path_utf8=from_path_utf8,
639
executable=cur_details[3],
640
fingerprint=cur_details[1],
641
packed_stat=cur_details[4],
645
to_path_utf8=to_rel_utf8)
647
# We need to move all the children of this
649
update_dirblock(from_path_utf8, to_key,
651
update_dirblock(from_rel_utf8, to_key, to_rel_utf8)
655
state._dirblock_state = dirstate.DirState.IN_MEMORY_MODIFIED
658
return #rename_tuples
661
"""Initialize the state in this tree to be a new tree."""
665
def path2id(self, path):
666
"""Return the id for path in this tree."""
667
entry = self._get_entry(path=path)
668
if entry == (None, None):
672
def paths2ids(self, paths, trees=[], require_versioned=True):
673
"""See Tree.paths2ids().
675
This specialisation fast-paths the case where all the trees are in the
680
parents = self.get_parent_ids()
682
if not (isinstance(tree, DirStateRevisionTree) and tree._revision_id in
684
return super(WorkingTree4, self).paths2ids(paths, trees, require_versioned)
685
search_indexes = [0] + [1 + parents.index(tree._revision_id) for tree in trees]
686
# -- make all paths utf8 --
689
paths_utf8.add(path.encode('utf8'))
691
# -- paths is now a utf8 path set --
692
# -- get the state object and prepare it.
693
state = self.current_dirstate()
694
if False and (state._dirblock_state == dirstate.DirState.NOT_IN_MEMORY
695
and '' not in paths):
696
paths2ids = self._paths2ids_using_bisect
698
paths2ids = self._paths2ids_in_memory
699
return paths2ids(paths, search_indexes,
700
require_versioned=require_versioned)
702
def _paths2ids_in_memory(self, paths, search_indexes,
703
require_versioned=True):
704
state = self.current_dirstate()
705
state._read_dirblocks_if_needed()
706
def _entries_for_path(path):
707
"""Return a list with all the entries that match path for all ids.
709
dirname, basename = os.path.split(path)
710
key = (dirname, basename, '')
711
block_index, present = state._find_block_index_from_key(key)
713
# the block which should contain path is absent.
716
block = state._dirblocks[block_index][1]
717
entry_index, _ = state._find_entry_index(key, block)
718
# we may need to look at multiple entries at this path: walk while the paths match.
719
while (entry_index < len(block) and
720
block[entry_index][0][0:2] == key[0:2]):
721
result.append(block[entry_index])
724
if require_versioned:
725
# -- check all supplied paths are versioned in a search tree. --
728
path_entries = _entries_for_path(path)
730
# this specified path is not present at all: error
731
all_versioned = False
733
found_versioned = False
734
# for each id at this path
735
for entry in path_entries:
737
for index in search_indexes:
738
if entry[1][index][0] != 'a': # absent
739
found_versioned = True
740
# all good: found a versioned cell
742
if not found_versioned:
743
# none of the indexes was not 'absent' at all ids for this
745
all_versioned = False
747
if not all_versioned:
748
raise errors.PathsNotVersionedError(paths)
749
# -- remove redundancy in supplied paths to prevent over-scanning --
752
other_paths = paths.difference(set([path]))
753
if not osutils.is_inside_any(other_paths, path):
754
# this is a top level path, we must check it.
755
search_paths.add(path)
757
# for all search_indexs in each path at or under each element of
758
# search_paths, if the detail is relocated: add the id, and add the
759
# relocated path as one to search if its not searched already. If the
760
# detail is not relocated, add the id.
761
searched_paths = set()
763
def _process_entry(entry):
764
"""Look at search_indexes within entry.
766
If a specific tree's details are relocated, add the relocation
767
target to search_paths if not searched already. If it is absent, do
768
nothing. Otherwise add the id to found_ids.
770
for index in search_indexes:
771
if entry[1][index][0] == 'r': # relocated
772
if not osutils.is_inside_any(searched_paths, entry[1][index][1]):
773
search_paths.add(entry[1][index][1])
774
elif entry[1][index][0] != 'a': # absent
775
found_ids.add(entry[0][2])
777
current_root = search_paths.pop()
778
searched_paths.add(current_root)
779
# process the entries for this containing directory: the rest will be
780
# found by their parents recursively.
781
root_entries = _entries_for_path(current_root)
783
# this specified path is not present at all, skip it.
785
for entry in root_entries:
786
_process_entry(entry)
787
initial_key = (current_root, '', '')
788
block_index, _ = state._find_block_index_from_key(initial_key)
789
while (block_index < len(state._dirblocks) and
790
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
791
for entry in state._dirblocks[block_index][1]:
792
_process_entry(entry)
796
def _paths2ids_using_bisect(self, paths, search_indexes,
797
require_versioned=True):
798
state = self.current_dirstate()
801
split_paths = sorted(osutils.split(p) for p in paths)
802
found = state._bisect_recursive(split_paths)
804
if require_versioned:
805
found_dir_names = set(dir_name_id[:2] for dir_name_id in found)
806
for dir_name in split_paths:
807
if dir_name not in found_dir_names:
808
raise errors.PathsNotVersionedError(paths)
810
for dir_name_id, trees_info in found.iteritems():
811
for index in search_indexes:
812
if trees_info[index][0] not in ('r', 'a'):
813
found_ids.add(dir_name_id[2])
816
def read_working_inventory(self):
817
"""Read the working inventory.
819
This is a meaningless operation for dirstate, but we obey it anyhow.
821
return self.inventory
824
def revision_tree(self, revision_id):
825
"""See Tree.revision_tree.
827
WorkingTree4 supplies revision_trees for any basis tree.
829
revision_id = osutils.safe_revision_id(revision_id)
830
dirstate = self.current_dirstate()
831
parent_ids = dirstate.get_parent_ids()
832
if revision_id not in parent_ids:
833
raise errors.NoSuchRevisionInTree(self, revision_id)
834
if revision_id in dirstate.get_ghosts():
835
raise errors.NoSuchRevisionInTree(self, revision_id)
836
return DirStateRevisionTree(dirstate, revision_id,
837
self.branch.repository)
839
@needs_tree_write_lock
840
def set_last_revision(self, new_revision):
841
"""Change the last revision in the working tree."""
842
new_revision = osutils.safe_revision_id(new_revision)
843
parents = self.get_parent_ids()
844
if new_revision in (NULL_REVISION, None):
845
assert len(parents) < 2, (
846
"setting the last parent to none with a pending merge is "
848
self.set_parent_ids([])
850
self.set_parent_ids([new_revision] + parents[1:],
851
allow_leftmost_as_ghost=True)
853
@needs_tree_write_lock
854
def set_parent_ids(self, revision_ids, allow_leftmost_as_ghost=False):
855
"""Set the parent ids to revision_ids.
857
See also set_parent_trees. This api will try to retrieve the tree data
858
for each element of revision_ids from the trees repository. If you have
859
tree data already available, it is more efficient to use
860
set_parent_trees rather than set_parent_ids. set_parent_ids is however
861
an easier API to use.
863
:param revision_ids: The revision_ids to set as the parent ids of this
864
working tree. Any of these may be ghosts.
866
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
868
for revision_id in revision_ids:
870
revtree = self.branch.repository.revision_tree(revision_id)
871
# TODO: jam 20070213 KnitVersionedFile raises
872
# RevisionNotPresent rather than NoSuchRevision if a
873
# given revision_id is not present. Should Repository be
874
# catching it and re-raising NoSuchRevision?
875
except (errors.NoSuchRevision, errors.RevisionNotPresent):
877
trees.append((revision_id, revtree))
878
self.set_parent_trees(trees,
879
allow_leftmost_as_ghost=allow_leftmost_as_ghost)
881
@needs_tree_write_lock
882
def set_parent_trees(self, parents_list, allow_leftmost_as_ghost=False):
883
"""Set the parents of the working tree.
885
:param parents_list: A list of (revision_id, tree) tuples.
886
If tree is None, then that element is treated as an unreachable
887
parent tree - i.e. a ghost.
889
dirstate = self.current_dirstate()
890
if len(parents_list) > 0:
891
if not allow_leftmost_as_ghost and parents_list[0][1] is None:
892
raise errors.GhostRevisionUnusableHere(parents_list[0][0])
895
# convert absent trees to the null tree, which we convert back to
897
for rev_id, tree in parents_list:
898
rev_id = osutils.safe_revision_id(rev_id)
900
real_trees.append((rev_id, tree))
902
real_trees.append((rev_id,
903
self.branch.repository.revision_tree(None)))
904
ghosts.append(rev_id)
905
dirstate.set_parent_trees(real_trees, ghosts=ghosts)
908
def _set_root_id(self, file_id):
909
"""See WorkingTree.set_root_id."""
910
state = self.current_dirstate()
911
state.set_path_id('', file_id)
912
self._dirty = state._dirblock_state == dirstate.DirState.IN_MEMORY_MODIFIED
915
"""Unlock in format 4 trees needs to write the entire dirstate."""
916
if self._control_files._lock_count == 1:
917
self._write_hashcache_if_dirty()
918
# eventually we should do signature checking during read locks for
920
if self._control_files._lock_mode == 'w':
923
if self._dirstate is not None:
924
self._dirstate.unlock()
925
self._dirstate = None
926
self._inventory = None
927
# reverse order of locking.
929
return self._control_files.unlock()
933
@needs_tree_write_lock
934
def unversion(self, file_ids):
935
"""Remove the file ids in file_ids from the current versioned set.
937
When a file_id is unversioned, all of its children are automatically
940
:param file_ids: The file ids to stop versioning.
941
:raises: NoSuchId if any fileid is not currently versioned.
945
state = self.current_dirstate()
946
state._read_dirblocks_if_needed()
947
ids_to_unversion = set()
948
for file_id in file_ids:
949
ids_to_unversion.add(osutils.safe_file_id(file_id))
950
paths_to_unversion = set()
952
# check if the root is to be unversioned, if so, assert for now.
953
# walk the state marking unversioned things as absent.
954
# if there are any un-unversioned ids at the end, raise
955
for key, details in state._dirblocks[0][1]:
956
if (details[0][0] not in ('a', 'r') and # absent or relocated
957
key[2] in ids_to_unversion):
958
# I haven't written the code to unversion / yet - it should be
960
raise errors.BzrError('Unversioning the / is not currently supported')
961
details_length = len(state._dirblocks[0][1][0][1])
963
while block_index < len(state._dirblocks):
964
# process one directory at a time.
965
block = state._dirblocks[block_index]
966
# first check: is the path one to remove - it or its children
968
for path in paths_to_unversion:
969
if (block[0].startswith(path) and
970
(len(block[0]) == len(path) or
971
block[0][len(path)] == '/')):
972
# this entire block should be deleted - its the block for a
973
# path to unversion; or the child of one
976
# TODO: trim paths_to_unversion as we pass by paths
978
# this block is to be deleted: process it.
979
# TODO: we can special case the no-parents case and
980
# just forget the whole block.
982
while entry_index < len(block[1]):
983
if not state._make_absent(block[1][entry_index]):
985
# go to the next block. (At the moment we dont delete empty
990
while entry_index < len(block[1]):
991
entry = block[1][entry_index]
992
if (entry[1][0][0] in ('a', 'r') or # absent, relocated
994
entry[0][2] not in ids_to_unversion):
995
# ^ not an id to unversion
998
if entry[1][0][0] == 'd':
999
paths_to_unversion.add(os.path.join(*entry[0][0:2]))
1000
if not state._make_absent(entry):
1002
# we have unversioned this id
1003
ids_to_unversion.remove(entry[0][2])
1005
if ids_to_unversion:
1006
raise errors.NoSuchId(self, iter(ids_to_unversion).next())
1008
# have to change the legacy inventory too.
1009
if self._inventory is not None:
1010
for file_id in file_ids:
1011
self._inventory.remove_recursive_id(file_id)
1013
@needs_tree_write_lock
1014
def _write_inventory(self, inv):
1015
"""Write inventory as the current inventory."""
1016
assert not self._dirty, "attempting to write an inventory when the dirstate is dirty will cause data loss"
1017
self.current_dirstate().set_state_from_inventory(inv)
1022
class WorkingTreeFormat4(WorkingTreeFormat3):
1023
"""The first consolidated dirstate working tree format.
1026
- exists within a metadir controlling .bzr
1027
- includes an explicit version marker for the workingtree control
1028
files, separate from the BzrDir format
1029
- modifies the hash cache format
1030
- is new in bzr TODO FIXME SETBEFOREMERGE
1031
- uses a LockDir to guard access to it.
1034
def get_format_string(self):
1035
"""See WorkingTreeFormat.get_format_string()."""
1036
return "Bazaar Working Tree format 4\n"
1038
def get_format_description(self):
1039
"""See WorkingTreeFormat.get_format_description()."""
1040
return "Working tree format 4"
1042
def initialize(self, a_bzrdir, revision_id=None):
1043
"""See WorkingTreeFormat.initialize().
1045
revision_id allows creating a working tree at a different
1046
revision than the branch is at.
1048
revision_id = osutils.safe_revision_id(revision_id)
1049
if not isinstance(a_bzrdir.transport, LocalTransport):
1050
raise errors.NotLocalUrl(a_bzrdir.transport.base)
1051
transport = a_bzrdir.get_workingtree_transport(self)
1052
control_files = self._open_control_files(a_bzrdir)
1053
control_files.create_lock()
1054
control_files.lock_write()
1055
control_files.put_utf8('format', self.get_format_string())
1056
branch = a_bzrdir.open_branch()
1057
if revision_id is None:
1058
revision_id = branch.last_revision()
1059
local_path = transport.local_abspath('dirstate')
1060
state = dirstate.DirState.initialize(local_path)
1062
wt = WorkingTree4(a_bzrdir.root_transport.local_abspath('.'),
1066
_control_files=control_files)
1070
#wt.current_dirstate().set_path_id('', NEWROOT)
1071
wt.set_last_revision(revision_id)
1073
basis = wt.basis_tree()
1075
transform.build_tree(basis, wt)
1078
control_files.unlock()
1083
def _open(self, a_bzrdir, control_files):
1084
"""Open the tree itself.
1086
:param a_bzrdir: the dir for the tree.
1087
:param control_files: the control files for the tree.
1089
return WorkingTree4(a_bzrdir.root_transport.local_abspath('.'),
1090
branch=a_bzrdir.open_branch(),
1093
_control_files=control_files)
1096
class DirStateRevisionTree(Tree):
1097
"""A revision tree pulling the inventory from a dirstate."""
1099
def __init__(self, dirstate, revision_id, repository):
1100
self._dirstate = dirstate
1101
self._revision_id = osutils.safe_revision_id(revision_id)
1102
self._repository = repository
1103
self._inventory = None
1105
self._dirstate_locked = False
1107
def annotate_iter(self, file_id):
1108
"""See Tree.annotate_iter"""
1109
w = self._repository.weave_store.get_weave(file_id,
1110
self._repository.get_transaction())
1111
return w.annotate_iter(self.inventory[file_id].revision)
1113
def _comparison_data(self, entry, path):
1114
"""See Tree._comparison_data."""
1116
return None, False, None
1117
# trust the entry as RevisionTree does, but this may not be
1118
# sensible: the entry might not have come from us?
1119
return entry.kind, entry.executable, None
1121
def _file_size(self, entry, stat_value):
1122
return entry.text_size
1124
def filter_unversioned_files(self, paths):
1125
"""Filter out paths that are not versioned.
1127
:return: set of paths.
1129
pred = self.has_filename
1130
return set((p for p in paths if not pred(p)))
1132
def _get_parent_index(self):
1133
"""Return the index in the dirstate referenced by this tree."""
1134
return self._dirstate.get_parent_ids().index(self._revision_id) + 1
1136
def _get_entry(self, file_id=None, path=None):
1137
"""Get the dirstate row for file_id or path.
1139
If either file_id or path is supplied, it is used as the key to lookup.
1140
If both are supplied, the fastest lookup is used, and an error is
1141
raised if they do not both point at the same row.
1143
:param file_id: An optional unicode file_id to be looked up.
1144
:param path: An optional unicode path to be looked up.
1145
:return: The dirstate row tuple for path/file_id, or (None, None)
1147
if file_id is None and path is None:
1148
raise errors.BzrError('must supply file_id or path')
1149
file_id = osutils.safe_file_id(file_id)
1150
if path is not None:
1151
path = path.encode('utf8')
1152
parent_index = self._get_parent_index()
1153
return self._dirstate._get_entry(parent_index, fileid_utf8=file_id, path_utf8=path)
1155
def _generate_inventory(self):
1156
"""Create and set self.inventory from the dirstate object.
1158
This is relatively expensive: we have to walk the entire dirstate.
1159
Ideally we would not, and instead would """
1160
assert self._locked, 'cannot generate inventory of an unlocked '\
1161
'dirstate revision tree'
1162
# separate call for profiling - makes it clear where the costs are.
1163
self._dirstate._read_dirblocks_if_needed()
1164
assert self._revision_id in self._dirstate.get_parent_ids(), \
1165
'parent %s has disappeared from %s' % (
1166
self._revision_id, self._dirstate.get_parent_ids())
1167
parent_index = self._dirstate.get_parent_ids().index(self._revision_id) + 1
1168
# This is identical now to the WorkingTree _generate_inventory except
1169
# for the tree index use.
1170
root_key, current_entry = self._dirstate._get_entry(parent_index, path_utf8='')
1171
current_id = root_key[2]
1172
assert current_entry[parent_index][0] == 'd'
1173
inv = Inventory(root_id=current_id, revision_id=self._revision_id)
1174
inv.root.revision = current_entry[parent_index][4]
1175
# Turn some things into local variables
1176
minikind_to_kind = dirstate.DirState._minikind_to_kind
1177
factory = entry_factory
1178
utf8_decode = cache_utf8._utf8_decode
1179
inv_byid = inv._byid
1180
# we could do this straight out of the dirstate; it might be fast
1181
# and should be profiled - RBC 20070216
1182
parent_ies = {'' : inv.root}
1183
for block in self._dirstate._dirblocks[1:]: #skip root
1186
parent_ie = parent_ies[dirname]
1188
# all the paths in this block are not versioned in this tree
1190
for key, entry in block[1]:
1191
minikind, link_or_sha1, size, executable, revid = entry[parent_index]
1192
if minikind in ('a', 'r'): # absent, relocated
1196
name_unicode = utf8_decode(name)[0]
1198
kind = minikind_to_kind[minikind]
1199
inv_entry = factory[kind](file_id, name_unicode,
1201
inv_entry.revision = revid
1203
inv_entry.executable = executable
1204
inv_entry.text_size = size
1205
inv_entry.text_sha1 = link_or_sha1
1206
elif kind == 'directory':
1207
parent_ies[(dirname + '/' + name).strip('/')] = inv_entry
1208
elif kind == 'symlink':
1209
inv_entry.executable = False
1210
inv_entry.text_size = size
1211
inv_entry.symlink_target = utf8_decode(link_or_sha1)[0]
1213
raise Exception, kind
1214
# These checks cost us around 40ms on a 55k entry tree
1215
assert file_id not in inv_byid
1216
assert name_unicode not in parent_ie.children
1217
inv_byid[file_id] = inv_entry
1218
parent_ie.children[name_unicode] = inv_entry
1219
self._inventory = inv
1221
def get_file_mtime(self, file_id, path=None):
1222
"""Return the modification time for this record.
1224
We return the timestamp of the last-changed revision.
1226
# Make sure the file exists
1227
entry = self._get_entry(file_id, path=path)
1228
if entry == (None, None): # do we raise?
1230
parent_index = self._get_parent_index()
1231
last_changed_revision = entry[1][parent_index][4]
1232
return self._repository.get_revision(last_changed_revision).timestamp
1234
def get_file_sha1(self, file_id, path=None, stat_value=None):
1235
# TODO: if path is present, fast-path on that, as inventory
1236
# might not be present
1237
ie = self.inventory[file_id]
1238
if ie.kind == "file":
1242
def get_file(self, file_id):
1243
return StringIO(self.get_file_text(file_id))
1245
def get_file_lines(self, file_id):
1246
ie = self.inventory[file_id]
1247
return self._repository.weave_store.get_weave(file_id,
1248
self._repository.get_transaction()).get_lines(ie.revision)
1250
def get_file_size(self, file_id):
1251
return self.inventory[file_id].text_size
1253
def get_file_text(self, file_id):
1254
return ''.join(self.get_file_lines(file_id))
1256
def get_symlink_target(self, file_id):
1257
entry = self._get_entry(file_id=file_id)
1258
parent_index = self._get_parent_index()
1259
if entry[1][parent_index][0] != 'l':
1262
# At present, none of the tree implementations supports non-ascii
1263
# symlink targets. So we will just assume that the dirstate path is
1265
return entry[1][parent_index][1]
1267
def get_revision_id(self):
1268
"""Return the revision id for this tree."""
1269
return self._revision_id
1271
def _get_inventory(self):
1272
if self._inventory is not None:
1273
return self._inventory
1274
self._generate_inventory()
1275
return self._inventory
1277
inventory = property(_get_inventory,
1278
doc="Inventory of this Tree")
1280
def get_parent_ids(self):
1281
"""The parents of a tree in the dirstate are not cached."""
1282
return self._repository.get_revision(self._revision_id).parent_ids
1284
def has_filename(self, filename):
1285
return bool(self.path2id(filename))
1287
def kind(self, file_id):
1288
return self.inventory[file_id].kind
1290
def is_executable(self, file_id, path=None):
1291
ie = self.inventory[file_id]
1292
if ie.kind != "file":
1294
return ie.executable
1296
def list_files(self, include_root=False):
1297
# We use a standard implementation, because DirStateRevisionTree is
1298
# dealing with one of the parents of the current state
1299
inv = self._get_inventory()
1300
entries = inv.iter_entries()
1301
if self.inventory.root is not None and not include_root:
1303
for path, entry in entries:
1304
yield path, 'V', entry.kind, entry.file_id, entry
1306
def lock_read(self):
1307
"""Lock the tree for a set of operations."""
1308
if not self._locked:
1309
self._repository.lock_read()
1310
if self._dirstate._lock_token is None:
1311
self._dirstate.lock_read()
1312
self._dirstate_locked = True
1316
def path2id(self, path):
1317
"""Return the id for path in this tree."""
1318
# lookup by path: faster than splitting and walking the ivnentory.
1319
entry = self._get_entry(path=path)
1320
if entry == (None, None):
1325
"""Unlock, freeing any cache memory used during the lock."""
1326
# outside of a lock, the inventory is suspect: release it.
1328
if not self._locked:
1329
self._inventory = None
1331
if self._dirstate_locked:
1332
self._dirstate.unlock()
1333
self._dirstate_locked = False
1334
self._repository.unlock()
1336
def walkdirs(self, prefix=""):
1337
# TODO: jam 20070215 This is the cheap way by cheating and using the
1338
# RevisionTree implementation.
1339
# This should be cleaned up to use the much faster Dirstate code
1340
# This is a little tricky, though, because the dirstate is
1341
# indexed by current path, not by parent path.
1342
# So for now, we just build up the parent inventory, and extract
1343
# it the same way RevisionTree does.
1344
_directory = 'directory'
1345
inv = self._get_inventory()
1346
top_id = inv.path2id(prefix)
1350
pending = [(prefix, top_id)]
1353
relpath, file_id = pending.pop()
1354
# 0 - relpath, 1- file-id
1356
relroot = relpath + '/'
1359
# FIXME: stash the node in pending
1360
entry = inv[file_id]
1361
for name, child in entry.sorted_children():
1362
toppath = relroot + name
1363
dirblock.append((toppath, name, child.kind, None,
1364
child.file_id, child.kind
1366
yield (relpath, entry.file_id), dirblock
1367
# push the user specified dirs from dirblock
1368
for dir in reversed(dirblock):
1369
if dir[2] == _directory:
1370
pending.append((dir[0], dir[4]))
1373
class InterDirStateTree(InterTree):
1374
"""Fast path optimiser for changes_from with dirstate trees."""
1376
def __init__(self, source, target):
1377
super(InterDirStateTree, self).__init__(source, target)
1378
if not InterDirStateTree.is_compatible(source, target):
1379
raise Exception, "invalid source %r and target %r" % (source, target)
1382
def make_source_parent_tree(source, target):
1383
"""Change the source tree into a parent of the target."""
1384
revid = source.commit('record tree')
1385
target.branch.repository.fetch(source.branch.repository, revid)
1386
target.set_parent_ids([revid])
1387
return target.basis_tree(), target
1389
_matching_from_tree_format = WorkingTreeFormat4()
1390
_matching_to_tree_format = WorkingTreeFormat4()
1391
_test_mutable_trees_to_test_trees = make_source_parent_tree
1393
def _iter_changes(self, include_unchanged=False,
1394
specific_files=None, pb=None, extra_trees=[],
1395
require_versioned=True):
1396
"""Return the changes from source to target.
1398
:return: An iterator that yields tuples. See InterTree._iter_changes
1400
:param specific_files: An optional list of file paths to restrict the
1401
comparison to. When mapping filenames to ids, all matches in all
1402
trees (including optional extra_trees) are used, and all children of
1403
matched directories are included.
1404
:param include_unchanged: An optional boolean requesting the inclusion of
1405
unchanged entries in the result.
1406
:param extra_trees: An optional list of additional trees to use when
1407
mapping the contents of specific_files (paths) to file_ids.
1408
:param require_versioned: If True, all files in specific_files must be
1409
versioned in one of source, target, extra_trees or
1410
PathsNotVersionedError is raised.
1412
utf8_decode = cache_utf8._utf8_decode
1413
_minikind_to_kind = dirstate.DirState._minikind_to_kind
1414
# NB: show_status depends on being able to pass in non-versioned files
1415
# and report them as unknown
1416
# TODO: handle extra trees in the dirstate.
1418
for f in super(InterDirStateTree, self)._iter_changes(
1419
include_unchanged, specific_files, pb, extra_trees,
1423
parent_ids = self.target.get_parent_ids()
1425
if self.source._revision_id == NULL_REVISION:
1427
indices = (target_index,)
1429
assert (self.source._revision_id in parent_ids), \
1430
"Failure: source._revision_id: %s not in target.parent_ids(%s)" % (
1431
self.source._revision_id, parent_ids)
1432
source_index = 1 + parent_ids.index(self.source._revision_id)
1433
indices = (source_index,target_index)
1434
# -- make all specific_files utf8 --
1436
specific_files_utf8 = set()
1437
for path in specific_files:
1438
specific_files_utf8.add(path.encode('utf8'))
1439
specific_files = specific_files_utf8
1441
specific_files = set([''])
1442
# -- specific_files is now a utf8 path set --
1443
# -- get the state object and prepare it.
1444
state = self.target.current_dirstate()
1445
state._read_dirblocks_if_needed()
1446
def _entries_for_path(path):
1447
"""Return a list with all the entries that match path for all ids.
1449
dirname, basename = os.path.split(path)
1450
key = (dirname, basename, '')
1451
block_index, present = state._find_block_index_from_key(key)
1453
# the block which should contain path is absent.
1456
block = state._dirblocks[block_index][1]
1457
entry_index, _ = state._find_entry_index(key, block)
1458
# we may need to look at multiple entries at this path: walk while the specific_files match.
1459
while (entry_index < len(block) and
1460
block[entry_index][0][0:2] == key[0:2]):
1461
result.append(block[entry_index])
1464
if require_versioned:
1465
# -- check all supplied paths are versioned in a search tree. --
1466
all_versioned = True
1467
for path in specific_files:
1468
path_entries = _entries_for_path(path)
1469
if not path_entries:
1470
# this specified path is not present at all: error
1471
all_versioned = False
1473
found_versioned = False
1474
# for each id at this path
1475
for entry in path_entries:
1477
for index in indices:
1478
if entry[1][index][0] != 'a': # absent
1479
found_versioned = True
1480
# all good: found a versioned cell
1482
if not found_versioned:
1483
# none of the indexes was not 'absent' at all ids for this
1485
all_versioned = False
1487
if not all_versioned:
1488
raise errors.PathsNotVersionedError(paths)
1489
# -- remove redundancy in supplied specific_files to prevent over-scanning --
1490
search_specific_files = set()
1491
for path in specific_files:
1492
other_specific_files = specific_files.difference(set([path]))
1493
if not osutils.is_inside_any(other_specific_files, path):
1494
# this is a top level path, we must check it.
1495
search_specific_files.add(path)
1497
# compare source_index and target_index at or under each element of search_specific_files.
1498
# follow the following comparison table. Note that we only want to do diff operations when
1499
# the target is fdl because thats when the walkdirs logic will have exposed the pathinfo
1503
# Source | Target | disk | action
1504
# r | fdl | | add source to search, add id path move and perform
1505
# | | | diff check on source-target
1506
# r | fdl | a | dangling file that was present in the basis.
1508
# r | a | | add source to search
1510
# r | r | | this path is present in a non-examined tree, skip.
1511
# r | r | a | this path is present in a non-examined tree, skip.
1512
# a | fdl | | add new id
1513
# a | fdl | a | dangling locally added file, skip
1514
# a | a | | not present in either tree, skip
1515
# a | a | a | not present in any tree, skip
1516
# a | r | | not present in either tree at this path, skip as it
1517
# | | | may not be selected by the users list of paths.
1518
# a | r | a | not present in either tree at this path, skip as it
1519
# | | | may not be selected by the users list of paths.
1520
# fdl | fdl | | content in both: diff them
1521
# fdl | fdl | a | deleted locally, but not unversioned - show as deleted ?
1522
# fdl | a | | unversioned: output deleted id for now
1523
# fdl | a | a | unversioned and deleted: output deleted id
1524
# fdl | r | | relocated in this tree, so add target to search.
1525
# | | | Dont diff, we will see an r,fd; pair when we reach
1526
# | | | this id at the other path.
1527
# fdl | r | a | 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.
1531
# for all search_indexs in each path at or under each element of
1532
# search_specific_files, if the detail is relocated: add the id, and add the
1533
# relocated path as one to search if its not searched already. If the
1534
# detail is not relocated, add the id.
1535
searched_specific_files = set()
1536
NULL_PARENT_DETAILS = dirstate.DirState.NULL_PARENT_DETAILS
1537
# Using a list so that we can access the values and change them in
1538
# nested scope. Each one is [path, file_id, entry]
1539
last_source_parent = [None, None, None]
1540
last_target_parent = [None, None, None]
1542
def _process_entry(entry, path_info):
1543
"""Compare an entry and real disk to generate delta information.
1545
:param path_info: top_relpath, basename, kind, lstat, abspath for
1546
the path of entry. If None, then the path is considered absent.
1547
(Perhaps we should pass in a concrete entry for this ?)
1549
# TODO: when a parent has been renamed, dont emit path renames for children,
1550
if source_index is None:
1551
source_details = NULL_PARENT_DETAILS
1553
source_details = entry[1][source_index]
1554
target_details = entry[1][target_index]
1555
source_minikind = source_details[0]
1556
target_minikind = target_details[0]
1557
if source_minikind in 'fdlr' and target_minikind in 'fdl':
1558
# claimed content in both: diff
1559
# r | fdl | | add source to search, add id path move and perform
1560
# | | | diff check on source-target
1561
# r | fdl | a | dangling file that was present in the basis.
1563
if source_minikind in 'r':
1564
# add the source to the search path to find any children it
1565
# has. TODO ? : only add if it is a container ?
1566
if not osutils.is_inside_any(searched_specific_files,
1568
search_specific_files.add(source_details[1])
1569
# generate the old path; this is needed for stating later
1571
old_path = source_details[1]
1572
old_dirname, old_basename = os.path.split(old_path)
1573
path = os.path.join(entry[0][0], entry[0][1])
1574
old_entry = state._get_entry(source_index,
1576
# update the source details variable to be the real
1578
source_details = old_entry[1][source_index]
1579
source_minikind = source_details[0]
1581
old_dirname = entry[0][0]
1582
old_basename = entry[0][1]
1583
old_path = path = os.path.join(old_dirname, old_basename)
1584
if path_info is None:
1585
# the file is missing on disk, show as removed.
1586
old_path = os.path.join(entry[0][0], entry[0][1])
1587
content_change = True
1591
# source and target are both versioned and disk file is present.
1592
target_kind = path_info[2]
1593
if target_kind == 'directory':
1594
if source_minikind != 'd':
1595
content_change = True
1597
# directories have no fingerprint
1598
content_change = False
1600
elif target_kind == 'file':
1601
if source_minikind != 'f':
1602
content_change = True
1604
# has it changed? fast path: size, slow path: sha1.
1605
if source_details[2] != path_info[3].st_size:
1606
content_change = True
1608
# maybe the same. Get the hash
1609
new_hash = self.target._hashcache.get_sha1(
1611
content_change = (new_hash != source_details[1])
1613
stat.S_ISREG(path_info[3].st_mode)
1614
and stat.S_IEXEC & path_info[3].st_mode)
1615
elif target_kind == 'symlink':
1616
if source_minikind != 'l':
1617
content_change = True
1619
# TODO: check symlink supported for windows users
1620
# and grab from target state here.
1621
link_target = os.readlink(path_info[4])
1622
content_change = (link_target != source_details[1])
1625
raise Exception, "unknown kind %s" % path_info[2]
1626
# parent id is the entry for the path in the target tree
1627
if old_dirname == last_source_parent[0]:
1628
source_parent_id = last_source_parent[1]
1630
source_parent_entry = state._get_entry(source_index,
1631
path_utf8=old_dirname)
1632
source_parent_id = source_parent_entry[0][2]
1633
if source_parent_id == entry[0][2]:
1634
# This is the root, so the parent is None
1635
source_parent_id = None
1637
last_source_parent[0] = old_dirname
1638
last_source_parent[1] = source_parent_id
1639
last_source_parent[2] = source_parent_entry
1641
new_dirname = entry[0][0]
1642
if new_dirname == last_target_parent[0]:
1643
target_parent_id = last_target_parent[1]
1645
# TODO: We don't always need to do the lookup, because the
1646
# parent entry will be the same as the source entry.
1647
target_parent_entry = state._get_entry(target_index,
1648
path_utf8=new_dirname)
1649
target_parent_id = target_parent_entry[0][2]
1650
if target_parent_id == entry[0][2]:
1651
# This is the root, so the parent is None
1652
target_parent_id = None
1654
last_target_parent[0] = new_dirname
1655
last_target_parent[1] = target_parent_id
1656
last_target_parent[2] = target_parent_entry
1658
source_exec = source_details[3]
1659
path_unicode = utf8_decode(path)[0]
1660
return ((entry[0][2], path_unicode, content_change,
1662
(source_parent_id, target_parent_id),
1663
(old_basename, entry[0][1]),
1664
(_minikind_to_kind[source_minikind], target_kind),
1665
(source_exec, target_exec)),)
1666
elif source_minikind in 'a' and target_minikind in 'fdl':
1667
# looks like a new file
1668
if path_info is not None:
1669
path = os.path.join(*entry[0][0:2])
1670
# parent id is the entry for the path in the target tree
1671
# TODO: these are the same for an entire directory: cache em.
1672
parent_id = state._get_entry(target_index, path_utf8=entry[0][0])[0][2]
1673
if parent_id == entry[0][2]:
1676
new_executable = bool(
1677
stat.S_ISREG(path_info[3].st_mode)
1678
and stat.S_IEXEC & path_info[3].st_mode)
1679
path_unicode = utf8_decode(path)[0]
1680
return ((entry[0][2], path_unicode, True,
1683
(None, entry[0][1]),
1684
(None, path_info[2]),
1685
(None, new_executable)),)
1687
# but its not on disk: we deliberately treat this as just
1688
# never-present. (Why ?! - RBC 20070224)
1690
elif source_minikind in 'fdl' and target_minikind in 'a':
1691
# unversioned, possibly, or possibly not deleted: we dont care.
1692
# if its still on disk, *and* theres no other entry at this
1693
# path [we dont know this in this routine at the moment -
1694
# perhaps we should change this - then it would be an unknown.
1695
old_path = os.path.join(*entry[0][0:2])
1696
# parent id is the entry for the path in the target tree
1697
parent_id = state._get_entry(source_index, path_utf8=entry[0][0])[0][2]
1698
if parent_id == entry[0][2]:
1700
old_path_unicode = utf8_decode(old_path)[0]
1701
return ((entry[0][2], old_path_unicode, True,
1704
(entry[0][1], None),
1705
(_minikind_to_kind[source_minikind], None),
1706
(source_details[3], None)),)
1707
elif source_minikind in 'fdl' and target_minikind in 'r':
1708
# a rename; could be a true rename, or a rename inherited from
1709
# a renamed parent. TODO: handle this efficiently. Its not
1710
# common case to rename dirs though, so a correct but slow
1711
# implementation will do.
1712
if not osutils.is_inside_any(searched_specific_files, target_details[1]):
1713
search_specific_files.add(target_details[1])
1715
import pdb;pdb.set_trace()
1717
while search_specific_files:
1718
# TODO: the pending list should be lexically sorted?
1719
current_root = search_specific_files.pop()
1720
searched_specific_files.add(current_root)
1721
# process the entries for this containing directory: the rest will be
1722
# found by their parents recursively.
1723
root_entries = _entries_for_path(current_root)
1724
root_abspath = self.target.abspath(current_root)
1726
root_stat = os.lstat(root_abspath)
1728
if e.errno == errno.ENOENT:
1729
# the path does not exist: let _process_entry know that.
1730
root_dir_info = None
1732
# some other random error: hand it up.
1735
root_dir_info = ('', current_root,
1736
osutils.file_kind_from_stat_mode(root_stat.st_mode), root_stat,
1738
if not root_entries and not root_dir_info:
1739
# this specified path is not present at all, skip it.
1741
for entry in root_entries:
1742
for result in _process_entry(entry, root_dir_info):
1743
# this check should probably be outside the loop: one
1744
# 'iterate two trees' api, and then _iter_changes filters
1745
# unchanged pairs. - RBC 20070226
1746
if (include_unchanged
1747
or result[2] # content change
1748
or result[3][0] != result[3][1] # versioned status
1749
or result[4][0] != result[4][1] # parent id
1750
or result[5][0] != result[5][1] # name
1751
or result[6][0] != result[6][1] # kind
1752
or result[7][0] != result[7][1] # executable
1755
dir_iterator = osutils._walkdirs_utf8(root_abspath, prefix=current_root)
1756
initial_key = (current_root, '', '')
1757
block_index, _ = state._find_block_index_from_key(initial_key)
1758
if block_index == 0:
1759
# we have processed the total root already, but because the
1760
# initial key matched it we should skip it here.
1763
current_dir_info = dir_iterator.next()
1765
if e.errno in (errno.ENOENT, errno.ENOTDIR):
1766
# there may be directories in the inventory even though
1767
# this path is not a file on disk: so mark it as end of
1769
current_dir_info = None
1773
if current_dir_info[0][0] == '':
1774
# remove .bzr from iteration
1775
bzr_index = bisect_left(current_dir_info[1], ('.bzr',))
1776
assert current_dir_info[1][bzr_index][0] == '.bzr'
1777
del current_dir_info[1][bzr_index]
1778
# walk until both the directory listing and the versioned metadata
1779
# are exhausted. TODO: reevaluate this, perhaps we should stop when
1780
# the versioned data runs out.
1781
if (block_index < len(state._dirblocks) and
1782
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
1783
current_block = state._dirblocks[block_index]
1785
current_block = None
1786
while (current_dir_info is not None or
1787
current_block is not None):
1788
if (current_dir_info and current_block
1789
and current_dir_info[0][0] != current_block[0]):
1790
if current_dir_info[0][0] < current_block[0] :
1791
# import pdb; pdb.set_trace()
1792
# print 'unversioned dir'
1793
# filesystem data refers to paths not covered by the dirblock.
1794
# this has two possibilities:
1795
# A) it is versioned but empty, so there is no block for it
1796
# B) it is not versioned.
1797
# in either case it was processed by the containing directories walk:
1798
# if it is root/foo, when we walked root we emitted it,
1799
# or if we ere given root/foo to walk specifically, we
1800
# emitted it when checking the walk-root entries
1801
# advance the iterator and loop - we dont need to emit it.
1803
current_dir_info = dir_iterator.next()
1804
except StopIteration:
1805
current_dir_info = None
1807
# We have a dirblock entry for this location, but there
1808
# is no filesystem path for this. This is most likely
1809
# because a directory was removed from the disk.
1810
# We don't have to report the missing directory,
1811
# because that should have already been handled, but we
1812
# need to handle all of the files that are contained
1814
for current_entry in current_block[1]:
1815
# entry referring to file not present on disk.
1816
# advance the entry only, after processing.
1817
for result in _process_entry(current_entry, None):
1818
# this check should probably be outside the loop: one
1819
# 'iterate two trees' api, and then _iter_changes filters
1820
# unchanged pairs. - RBC 20070226
1821
if (include_unchanged
1822
or result[2] # content change
1823
or result[3][0] != result[3][1] # versioned status
1824
or result[4][0] != result[4][1] # parent id
1825
or result[5][0] != result[5][1] # name
1826
or result[6][0] != result[6][1] # kind
1827
or result[7][0] != result[7][1] # executable
1831
if (block_index < len(state._dirblocks) and
1832
osutils.is_inside(current_root,
1833
state._dirblocks[block_index][0])):
1834
current_block = state._dirblocks[block_index]
1836
current_block = None
1839
if current_block and entry_index < len(current_block[1]):
1840
current_entry = current_block[1][entry_index]
1842
current_entry = None
1843
advance_entry = True
1845
if current_dir_info and path_index < len(current_dir_info[1]):
1846
current_path_info = current_dir_info[1][path_index]
1848
current_path_info = None
1850
while (current_entry is not None or
1851
current_path_info is not None):
1852
if current_entry is None:
1853
# no more entries: yield current_pathinfo as an
1854
# unversioned file: its not the same as a path in any
1855
# tree in the dirstate.
1856
new_executable = bool(
1857
stat.S_ISREG(current_path_info[3].st_mode)
1858
and stat.S_IEXEC & current_path_info[3].st_mode)
1859
yield (None, current_path_info[0], True,
1862
(None, current_path_info[1]),
1863
(None, current_path_info[2]),
1864
(None, new_executable))
1865
elif current_path_info is None:
1866
# no path is fine: the per entry code will handle it.
1867
for result in _process_entry(current_entry, current_path_info):
1868
# this check should probably be outside the loop: one
1869
# 'iterate two trees' api, and then _iter_changes filters
1870
# unchanged pairs. - RBC 20070226
1871
if (include_unchanged
1872
or result[2] # content change
1873
or result[3][0] != result[3][1] # versioned status
1874
or result[4][0] != result[4][1] # parent id
1875
or result[5][0] != result[5][1] # name
1876
or result[6][0] != result[6][1] # kind
1877
or result[7][0] != result[7][1] # executable
1880
elif current_entry[0][1] != current_path_info[1]:
1881
if current_path_info[1] < current_entry[0][1]:
1882
# extra file on disk: pass for now, but only
1883
# increment the path, not the entry
1884
# import pdb; pdb.set_trace()
1885
# print 'unversioned file'
1886
advance_entry = False
1888
# entry referring to file not present on disk.
1889
# advance the entry only, after processing.
1890
for result in _process_entry(current_entry, None):
1891
# this check should probably be outside the loop: one
1892
# 'iterate two trees' api, and then _iter_changes filters
1893
# unchanged pairs. - RBC 20070226
1894
if (include_unchanged
1895
or result[2] # content change
1896
or result[3][0] != result[3][1] # versioned status
1897
or result[4][0] != result[4][1] # parent id
1898
or result[5][0] != result[5][1] # name
1899
or result[6][0] != result[6][1] # kind
1900
or result[7][0] != result[7][1] # executable
1903
advance_path = False
1905
for result in _process_entry(current_entry, current_path_info):
1906
# this check should probably be outside the loop: one
1907
# 'iterate two trees' api, and then _iter_changes filters
1908
# unchanged pairs. - RBC 20070226
1909
if (include_unchanged
1910
or result[2] # content change
1911
or result[3][0] != result[3][1] # versioned status
1912
or result[4][0] != result[4][1] # parent id
1913
or result[5][0] != result[5][1] # name
1914
or result[6][0] != result[6][1] # kind
1915
or result[7][0] != result[7][1] # executable
1918
if advance_entry and current_entry is not None:
1920
if entry_index < len(current_block[1]):
1921
current_entry = current_block[1][entry_index]
1923
current_entry = None
1925
advance_entry = True # reset the advance flaga
1926
if advance_path and current_path_info is not None:
1928
if path_index < len(current_dir_info[1]):
1929
current_path_info = current_dir_info[1][path_index]
1931
current_path_info = None
1933
advance_path = True # reset the advance flagg.
1934
if current_block is not None:
1936
if (block_index < len(state._dirblocks) and
1937
osutils.is_inside(current_root, state._dirblocks[block_index][0])):
1938
current_block = state._dirblocks[block_index]
1940
current_block = None
1941
if current_dir_info is not None:
1943
current_dir_info = dir_iterator.next()
1944
except StopIteration:
1945
current_dir_info = None
1949
def is_compatible(source, target):
1950
# the target must be a dirstate working tree
1951
if not isinstance(target, WorkingTree4):
1953
# the source must be a revtreee or dirstate rev tree.
1954
if not isinstance(source,
1955
(revisiontree.RevisionTree, DirStateRevisionTree)):
1957
# the source revid must be in the target dirstate
1958
if not (source._revision_id == NULL_REVISION or
1959
source._revision_id in target.get_parent_ids()):
1960
# TODO: what about ghosts? it may well need to
1961
# check for them explicitly.
1965
InterTree.register_optimiser(InterDirStateTree)
added
removed
bzrlib/workingtree_4.py
