1
# Copyright (C) 2006, 2007 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
"""DirState objects record the state of a directory and its bzr metadata.
19
Pseudo EBNF grammar for the state file. Fields are separated by NULLs, and
20
lines by NL. The field delimiters are ommitted in the grammar, line delimiters
21
are not - this is done for clarity of reading. All string data is in utf8.
23
MINIKIND = "f" | "d" | "l" | "a" | "r";
26
WHOLE_NUMBER = {digit}, digit;
28
REVISION_ID = a non-empty utf8 string;
30
dirstate format = header line, full checksum, row count, parent details,
31
ghost_details, entries;
32
header line = "#bazaar dirstate flat format 2", NL;
33
full checksum = "adler32: ", ["-"], WHOLE_NUMBER, NL;
34
row count = "num_entries: ", digit, NL;
35
parent_details = WHOLE NUMBER, {REVISION_ID}* NL;
36
ghost_details = WHOLE NUMBER, {REVISION_ID}*, NL;
38
entry = entry_key, current_entry_details, {parent_entry_details};
39
entry_key = dirname, basename, fileid;
40
current_entry_details = common_entry_details, working_entry_details;
41
parent_entry_details = common_entry_details, history_entry_details;
42
common_entry_details = MINIKIND, fingerprint, size, executable
43
working_entry_details = packed_stat
44
history_entry_details = REVISION_ID;
47
fingerprint = a nonempty utf8 sequence with meaning defined by minikind.
49
Given this definition, the following is useful to know:
50
entry (aka row) - all the data for a given key.
51
entry[0]: The key (dirname, basename, fileid)
55
entry[1]: The tree(s) data for this path and id combination.
56
entry[1][0]: The current tree
57
entry[1][1]: The second tree
59
For an entry for a tree, we have (using tree 0 - current tree) to demonstrate:
60
entry[1][0][0]: minikind
61
entry[1][0][1]: fingerprint
63
entry[1][0][3]: executable
64
entry[1][0][4]: packed_stat
66
entry[1][1][4]: revision_id
68
There may be multiple rows at the root, one per id present in the root, so the
69
in memory root row is now:
70
self._dirblocks[0] -> ('', [entry ...]),
71
and the entries in there are
74
entries[0][2]: file_id
75
entries[1][0]: The tree data for the current tree for this fileid at /
79
'r' is a relocated entry: This path is not present in this tree with this id,
80
but the id can be found at another location. The fingerprint is used to
81
point to the target location.
82
'a' is an absent entry: In that tree the id is not present at this path.
83
'd' is a directory entry: This path in this tree is a directory with the
84
current file id. There is no fingerprint for directories.
85
'f' is a file entry: As for directory, but its a file. The fingerprint is a
87
'l' is a symlink entry: As for directory, but a symlink. The fingerprint is the
91
--- Format 1 had the following different definition: ---
92
rows = dirname, NULL, basename, NULL, MINIKIND, NULL, fileid_utf8, NULL,
93
WHOLE NUMBER (* size *), NULL, packed stat, NULL, sha1|symlink target,
95
PARENT ROW = NULL, revision_utf8, NULL, MINIKIND, NULL, dirname, NULL,
96
basename, NULL, WHOLE NUMBER (* size *), NULL, "y" | "n", NULL,
99
PARENT ROW's are emitted for every parent that is not in the ghosts details
100
line. That is, if the parents are foo, bar, baz, and the ghosts are bar, then
101
each row will have a PARENT ROW for foo and baz, but not for bar.
104
In any tree, a kind of 'moved' indicates that the fingerprint field
105
(which we treat as opaque data specific to the 'kind' anyway) has the
106
details for the id of this row in that tree.
108
I'm strongly tempted to add a id->path index as well, but I think that
109
where we need id->path mapping; we also usually read the whole file, so
110
I'm going to skip that for the moment, as we have the ability to locate
111
via bisect any path in any tree, and if we lookup things by path, we can
112
accumulate a id->path mapping as we go, which will tend to match what we
115
I plan to implement this asap, so please speak up now to alter/tweak the
116
design - and once we stabilise on this, I'll update the wiki page for
119
The rationale for all this is that we want fast operations for the
120
common case (diff/status/commit/merge on all files) and extremely fast
121
operations for the less common but still occurs a lot status/diff/commit
122
on specific files). Operations on specific files involve a scan for all
123
the children of a path, *in every involved tree*, which the current
124
format did not accommodate.
128
1) Fast end to end use for bzr's top 5 uses cases. (commmit/diff/status/merge/???)
129
2) fall back current object model as needed.
130
3) scale usably to the largest trees known today - say 50K entries. (mozilla
131
is an example of this)
135
Eventually reuse dirstate objects across locks IFF the dirstate file has not
136
been modified, but will require that we flush/ignore cached stat-hit data
137
because we wont want to restat all files on disk just because a lock was
138
acquired, yet we cannot trust the data after the previous lock was released.
140
Memory representation:
141
vector of all directories, and vector of the childen ?
143
root_entrie = (direntry for root, [parent_direntries_for_root]),
145
('', ['data for achild', 'data for bchild', 'data for cchild'])
146
('dir', ['achild', 'cchild', 'echild'])
148
- single bisect to find N subtrees from a path spec
149
- in-order for serialisation - this is 'dirblock' grouping.
150
- insertion of a file '/a' affects only the '/' child-vector, that is, to
151
insert 10K elements from scratch does not generates O(N^2) memoves of a
152
single vector, rather each individual, which tends to be limited to a
153
manageable number. Will scale badly on trees with 10K entries in a
154
single directory. compare with Inventory.InventoryDirectory which has
155
a dictionary for the children. No bisect capability, can only probe for
156
exact matches, or grab all elements and sorta.
157
- Whats the risk of error here? Once we have the base format being processed
158
we should have a net win regardless of optimality. So we are going to
159
go with what seems reasonably.
162
maybe we should do a test profile of these core structure - 10K simulated searches/lookups/etc?
164
Objects for each row?
165
The lifetime of Dirstate objects is current per lock, but see above for
166
possible extensions. The lifetime of a row from a dirstate is expected to be
167
very short in the optimistic case: which we are optimising for. For instance,
168
subtree status will determine from analysis of the disk data what rows need to
169
be examined at all, and will be able to determine from a single row whether
170
that file has altered or not, so we are aiming to process tens of thousands of
171
entries each second within the dirstate context, before exposing anything to
172
the larger codebase. This suggests we want the time for a single file
173
comparison to be < 0.1 milliseconds. That would give us 10000 paths per second
174
processed, and to scale to 100 thousand we'll another order of magnitude to do
175
that. Now, as the lifetime for all unchanged entries is the time to parse, stat
176
the file on disk, and then immediately discard, the overhead of object creation
177
becomes a significant cost.
179
Figures: Creating a tuple from from 3 elements was profiled at 0.0625
180
microseconds, whereas creating a object which is subclassed from tuple was
181
0.500 microseconds, and creating an object with 3 elements and slots was 3
182
microseconds long. 0.1 milliseconds is 100 microseconds, and ideally we'll get
183
down to 10 microseconds for the total processing - having 33% of that be object
184
creation is a huge overhead. There is a potential cost in using tuples within
185
each row which is that the conditional code to do comparisons may be slower
186
than method invocation, but method invocation is known to be slow due to stack
187
frame creation, so avoiding methods in these tight inner loops in unfortunately
188
desirable. We can consider a pyrex version of this with objects in future if
207
import bzrlib.inventory
208
from bzrlib import osutils
209
from bzrlib.osutils import (
217
class DirState(object):
218
"""Record directory and metadata state for fast access.
220
A dirstate is a specialised data structure for managing local working
221
tree state information. Its not yet well defined whether it is platform
222
specific, and if it is how we detect/parameterise that.
225
_kind_to_minikind = {'absent':'a', 'file':'f', 'directory':'d', 'relocated':'r', 'symlink':'l'}
226
_minikind_to_kind = {'a':'absent', 'f':'file', 'd':'directory', 'l':'symlink', 'r':'relocated'}
227
_to_yesno = {True:'y', False: 'n'} # TODO profile the performance gain
228
# of using int conversion rather than a dict here. AND BLAME ANDREW IF
231
# TODO: jam 20070221 Make sure we handle when there are duplicated records
232
# (like when we remove + add the same path, or we have a rename)
233
# TODO: jam 20070221 Figure out what to do if we have a record that exceeds
234
# the BISECT_PAGE_SIZE.
235
BISECT_PAGE_SIZE = 4096
238
IN_MEMORY_UNMODIFIED = 1
239
IN_MEMORY_MODIFIED = 2
241
# A pack_stat (the x's) that is just noise and will never match the output
244
NULL_PARENT_DETAILS = ('a', '', 0, False, '')
246
def __init__(self, path):
247
"""Create a DirState object.
251
:attr _root_entrie: The root row of the directory/file information,
252
- contains the path to / - '', ''
253
- kind of 'directory',
254
- the file id of the root in utf8
257
- and no sha information.
258
:param path: The path at which the dirstate file on disk should live.
260
# _header_state and _dirblock_state represent the current state
261
# of the dirstate metadata and the per-row data respectiely.
262
# NOT_IN_MEMORY indicates that no data is in memory
263
# IN_MEMORY_UNMODIFIED indicates that what we have in memory
264
# is the same as is on disk
265
# IN_MEMORY_MODIFIED indicates that we have a modified version
266
# of what is on disk.
267
# In future we will add more granularity, for instance _dirblock_state
268
# will probably support partially-in-memory as a separate variable,
269
# allowing for partially-in-memory unmodified and partially-in-memory
271
self._header_state = DirState.NOT_IN_MEMORY
272
self._dirblock_state = DirState.NOT_IN_MEMORY
276
self._state_file = None
277
self._filename = path
278
self._lock_token = None
280
def add(self, path, file_id, kind, stat, link_or_sha1):
281
"""Add a path to be tracked.
283
:param path: The path within the dirstate - '' is the root, 'foo' is the
284
path foo within the root, 'foo/bar' is the path bar within foo
286
:param file_id: The file id of the path being added.
287
:param kind: The kind of the path.
288
:param stat: The output of os.lstate for the path.
289
:param link_or_sha1: The sha value of the file, or the target of a
290
symlink. '' for directories.
293
# find the block its in.
294
# find the location in the block.
295
# check its not there
297
#------- copied from bzrlib.inventory.make_entry
298
# --- normalized_filename wants a unicode basename only, so get one.
299
dirname, basename = os.path.split(path)
300
# we dont import normalized_filename directly because we want to be
301
# able to change the implementation at runtime for tests.
302
norm_name, can_access = osutils.normalized_filename(basename)
303
if norm_name != basename:
307
raise errors.InvalidNormalization(path)
308
# now that we've normalised, we need the correct utf8 path and
309
# dirname and basename elements. This single encode and split should be
310
# faster than three separate encodes.
311
utf8path = (dirname + '/' + basename).strip('/').encode('utf8')
312
dirname, basename = os.path.split(utf8path)
313
assert file_id.__class__ == str, \
314
"must be a utf8 file_id not %s" % (type(file_id))
315
entry_key = (dirname, basename, file_id)
316
self._read_dirblocks_if_needed()
317
block_index, present = self._find_block_index_from_key(entry_key)
319
# TODO: This test is not complete - an empty directory, or a
320
# directory for a parent tree will fool it.
321
# some parent path has not been added - its an error to add this
323
raise errors.NotVersionedError(path, str(self))
324
block = self._dirblocks[block_index][1]
327
packed_stat = DirState.NULLSTAT
330
packed_stat = pack_stat(stat)
331
parent_info = self._empty_parent_info()
332
minikind = DirState._kind_to_minikind[kind]
334
entry_data = entry_key, [
335
(minikind, link_or_sha1, size, False, packed_stat),
337
elif kind == 'directory':
338
entry_data = entry_key, [
339
(minikind, '', 0, False, packed_stat),
341
elif kind == 'symlink':
342
entry_data = entry_key, [
343
(minikind, link_or_sha1, size, False, packed_stat),
346
raise errors.BzrError('unknown kind %r' % kind)
347
entry_index, present = self._find_entry_index(entry_key, block)
348
assert not present, "basename %r already added" % basename
349
block.insert(entry_index, entry_data)
351
if kind == 'directory':
352
# insert a new dirblock
353
self._ensure_block(block_index, entry_index, utf8path)
354
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
356
def _empty_parent_info(self):
357
return [DirState.NULL_PARENT_DETAILS] * (len(self._parents) -
360
def _ensure_block(self, parent_block_index, parent_row_index, dirname):
361
"""Enssure a block for dirname exists.
363
This function exists to let callers which know that there is a
364
directory dirname ensure that the block for it exists. This block can
365
fail to exist because of demand loading, or because a directory had no
366
children. In either case it is not an error. It is however an error to
367
call this if there is no parent entry for the directory, and thus the
368
function requires the coordinates of such an entry to be provided.
370
The root row is special cased and can be indicated with a parent block
373
:param parent_block_index: The index of the block in which dirname's row
375
:param parent_row_index: The index in the parent block where the row
377
:param dirname: The utf8 dirname to ensure there is a block for.
378
:return: The index for the block.
381
# the basename of the directory must be the end of its full name.
382
if not (parent_block_index == -1 and
383
parent_block_index == -1 and dirname == ''):
384
assert dirname.endswith(
385
self._dirblocks[parent_block_index][1][parent_row_index][0][1])
386
block_index, present = self._find_block_index_from_key((dirname, '', ''))
388
## In future, when doing partial parsing, this should load and
389
# populate the entire block.
390
self._dirblocks.insert(block_index, (dirname, []))
393
def _entries_to_current_state(self, new_entries):
394
"""Load new_entries into self.dirblocks.
396
Process new_entries into the current state object, making them the active
399
:param new_entries: A sorted list of entries. This function does not sort
400
to prevent unneeded overhead when callers have a sorted list already.
403
assert new_entries[0][0][0:2] == ('', ''), \
404
"Missing root row %r" % new_entries[0][0]
405
# The two blocks here are deliberate: the root block and the
406
# contents-of-root block.
407
self._dirblocks = [('', []), ('', [])]
408
current_block = self._dirblocks[0][1]
411
append_entry = current_block.append
412
for entry in new_entries:
413
if entry[0][0] != current_dirname:
414
# new block - different dirname
416
current_dirname = entry[0][0]
417
self._dirblocks.append((current_dirname, current_block))
418
append_entry = current_block.append
419
# append the entry to the current block
421
self._split_root_dirblock_into_contents()
423
def _split_root_dirblock_into_contents(self):
424
"""Split the root dirblocks into root and contents-of-root.
426
After parsing by path, we end up with root entries and contents-of-root
427
entries in the same block. This loop splits them out again.
429
# The above loop leaves the "root block" entries mixed with the
430
# "contents-of-root block". But we don't want an if check on
431
# all entries, so instead we just fix it up here.
432
assert self._dirblocks[1] == ('', [])
434
contents_of_root_block = []
435
for entry in self._dirblocks[0][1]:
436
if not entry[0][1]: # This is a root entry
437
root_block.append(entry)
439
contents_of_root_block.append(entry)
440
self._dirblocks[0] = ('', root_block)
441
self._dirblocks[1] = ('', contents_of_root_block)
443
def _entry_to_line(self, entry):
444
"""Serialize entry to a NULL delimited line ready for _get_output_lines.
446
:param entry: An entry_tuple as defined in the module docstring.
448
entire_entry = list(entry[0])
449
for tree_number, tree_data in enumerate(entry[1]):
450
# (minikind, fingerprint, size, executable, tree_specific_string)
451
entire_entry.extend(tree_data)
452
# 3 for the key, 5 for the fields per tree.
453
tree_offset = 3 + tree_number * 5
455
entire_entry[tree_offset + 0] = tree_data[0]
457
entire_entry[tree_offset + 2] = str(tree_data[2])
459
entire_entry[tree_offset + 3] = DirState._to_yesno[tree_data[3]]
460
return '\0'.join(entire_entry)
462
def _fields_per_row(self):
463
"""How many null separated fields should be in each entry row.
465
Each line now has an extra '\n' field which is not used
466
so we just skip over it
469
+ number of fields per tree_data (5) * tree count
472
tree_count = 1 + self._num_present_parents()
473
return 3 + 5 * tree_count + 1
475
def _find_block(self, key, add_if_missing=False):
476
"""Return the block that key should be present in.
478
:param key: A dirstate entry key.
479
:return: The block tuple.
481
block_index, present = self._find_block_index_from_key(key)
483
if not add_if_missing:
484
# check to see if key is versioned itself - we might want to
485
# add it anyway, because dirs with no entries dont get a
486
# dirblock at parse time.
487
# This is an uncommon branch to take: most dirs have children,
488
# and most code works with versioned paths.
489
parent_base, parent_name = os.path.split(key[0])
490
if not self._get_block_entry_index(parent_base, parent_name, 0)[3]:
491
# some parent path has not been added - its an error to add
493
raise errors.NotVersionedError(key[0:2], str(self))
494
self._dirblocks.insert(block_index, (key[0], []))
495
return self._dirblocks[block_index]
497
def _find_block_index_from_key(self, key):
498
"""Find the dirblock index for a key.
500
:return: The block index, True if the block for the key is present.
502
if key[0:2] == ('', ''):
504
block_index = bisect.bisect_left(self._dirblocks, (key[0], []), 1)
505
# _right returns one-past-where-key is so we have to subtract
506
# one to use it. we use _right here because there are two
507
# '' blocks - the root, and the contents of root
508
# we always have a minimum of 2 in self._dirblocks: root and
509
# root-contents, and for '', we get 2 back, so this is
510
# simple and correct:
511
present = (block_index < len(self._dirblocks) and
512
self._dirblocks[block_index][0] == key[0])
513
return block_index, present
515
def _find_entry_index(self, key, block):
516
"""Find the entry index for a key in a block.
518
:return: The entry index, True if the entry for the key is present.
520
entry_index = bisect.bisect_left(block, (key, []))
521
present = (entry_index < len(block) and
522
block[entry_index][0] == key)
523
return entry_index, present
526
def from_tree(tree, dir_state_filename):
527
"""Create a dirstate from a bzr Tree.
529
:param tree: The tree which should provide parent information and
531
:return: a DirState object which is currently locked for writing.
532
(it was locked by DirState.initialize)
534
result = DirState.initialize(dir_state_filename)
538
parent_ids = tree.get_parent_ids()
539
num_parents = len(parent_ids)
541
for parent_id in parent_ids:
542
parent_tree = tree.branch.repository.revision_tree(parent_id)
543
parent_trees.append((parent_id, parent_tree))
544
parent_tree.lock_read()
545
result.set_parent_trees(parent_trees, [])
546
result.set_state_from_inventory(tree.inventory)
548
for revid, parent_tree in parent_trees:
552
# The caller won't have a chance to unlock this, so make sure we
558
def get_ghosts(self):
559
"""Return a list of the parent tree revision ids that are ghosts."""
560
self._read_header_if_needed()
564
"""Serialise the entire dirstate to a sequence of lines."""
565
if (self._header_state == DirState.IN_MEMORY_UNMODIFIED and
566
self._dirblock_state == DirState.IN_MEMORY_UNMODIFIED):
567
# read whats on disk.
568
self._state_file.seek(0)
569
return self._state_file.readlines()
571
lines.append(self._get_parents_line(self.get_parent_ids()))
572
lines.append(self._get_ghosts_line(self._ghosts))
573
# append the root line which is special cased
574
lines.extend(map(self._entry_to_line, self._iter_entries()))
575
return self._get_output_lines(lines)
577
def _get_ghosts_line(self, ghost_ids):
578
"""Create a line for the state file for ghost information."""
579
return '\0'.join([str(len(ghost_ids))] + ghost_ids)
581
def _get_parents_line(self, parent_ids):
582
"""Create a line for the state file for parents information."""
583
return '\0'.join([str(len(parent_ids))] + parent_ids)
585
def _get_fields_to_entry(self):
586
"""Get a function which converts entry fields into a entry record.
588
This handles size and executable, as well as parent records.
590
:return: A function which takes a list of fields, and returns an
591
appropriate record for storing in memory.
593
# This is intentionally unrolled for performance
594
num_present_parents = self._num_present_parents()
595
if num_present_parents == 0:
596
def fields_to_entry_0_parents(fields, _int=int):
597
path_name_file_id_key = (fields[0], fields[1], fields[2])
598
return (path_name_file_id_key, [
600
fields[3], # minikind
601
fields[4], # fingerprint
602
_int(fields[5]), # size
603
fields[6] == 'y', # executable
604
fields[7], # packed_stat or revision_id
606
return fields_to_entry_0_parents
607
elif num_present_parents == 1:
608
def fields_to_entry_1_parent(fields, _int=int):
609
path_name_file_id_key = (fields[0], fields[1], fields[2])
610
return (path_name_file_id_key, [
612
fields[3], # minikind
613
fields[4], # fingerprint
614
_int(fields[5]), # size
615
fields[6] == 'y', # executable
616
fields[7], # packed_stat or revision_id
619
fields[8], # minikind
620
fields[9], # fingerprint
621
_int(fields[10]), # size
622
fields[11] == 'y', # executable
623
fields[12], # packed_stat or revision_id
626
return fields_to_entry_1_parent
627
elif num_present_parents == 2:
628
def fields_to_entry_2_parents(fields, _int=int):
629
path_name_file_id_key = (fields[0], fields[1], fields[2])
630
return (path_name_file_id_key, [
632
fields[3], # minikind
633
fields[4], # fingerprint
634
_int(fields[5]), # size
635
fields[6] == 'y', # executable
636
fields[7], # packed_stat or revision_id
639
fields[8], # minikind
640
fields[9], # fingerprint
641
_int(fields[10]), # size
642
fields[11] == 'y', # executable
643
fields[12], # packed_stat or revision_id
646
fields[13], # minikind
647
fields[14], # fingerprint
648
_int(fields[15]), # size
649
fields[16] == 'y', # executable
650
fields[17], # packed_stat or revision_id
653
return fields_to_entry_2_parents
655
def fields_to_entry_n_parents(fields, _int=int):
656
path_name_file_id_key = (fields[0], fields[1], fields[2])
657
trees = [(fields[cur], # minikind
658
fields[cur+1], # fingerprint
659
_int(fields[cur+2]), # size
660
fields[cur+3] == 'y', # executable
661
fields[cur+4], # stat or revision_id
662
) for cur in xrange(3, len(fields)-1, 5)]
663
return path_name_file_id_key, trees
664
return fields_to_entry_n_parents
666
def get_parent_ids(self):
667
"""Return a list of the parent tree ids for the directory state."""
668
self._read_header_if_needed()
669
return list(self._parents)
671
def _get_block_entry_index(self, dirname, basename, tree_index):
672
"""Get the coordinates for a path in the state structure.
674
:param dirname: The utf8 dirname to lookup.
675
:param basename: The utf8 basename to lookup.
676
:param tree_index: The index of the tree for which this lookup should
678
:return: A tuple describing where the path is located, or should be
679
inserted. The tuple contains four fields: the block index, the row
680
index, anda two booleans are True when the directory is present, and
681
when the entire path is present. There is no guarantee that either
682
coordinate is currently reachable unless the found field for it is
683
True. For instance, a directory not present in the searched tree
684
may be returned with a value one greater than the current highest
685
block offset. The directory present field will always be True when
686
the path present field is True. The directory present field does
687
NOT indicate that the directory is present in the searched tree,
688
rather it indicates that there are at least some files in some
691
self._read_dirblocks_if_needed()
692
key = dirname, basename, ''
693
block_index, present = self._find_block_index_from_key(key)
695
# no such directory - return the dir index and 0 for the row.
696
return block_index, 0, False, False
697
block = self._dirblocks[block_index][1] # access the entries only
698
entry_index, present = self._find_entry_index(key, block)
699
# linear search through present entries at this path to find the one
701
while entry_index < len(block) and block[entry_index][0][1] == basename:
702
if block[entry_index][1][tree_index][0] not in \
703
('a', 'r'): # absent, relocated
704
return block_index, entry_index, True, True
706
return block_index, entry_index, True, False
708
def _get_entry(self, tree_index, fileid_utf8=None, path_utf8=None):
709
"""Get the dirstate entry for path in tree tree_index
711
If either file_id or path is supplied, it is used as the key to lookup.
712
If both are supplied, the fastest lookup is used, and an error is
713
raised if they do not both point at the same row.
715
:param tree_index: The index of the tree we wish to locate this path
716
in. If the path is present in that tree, the entry containing its
717
details is returned, otherwise (None, None) is returned
718
:param fileid_utf8: A utf8 file_id to look up.
719
:param path_utf8: An utf8 path to be looked up.
720
:return: The dirstate entry tuple for path, or (None, None)
722
self._read_dirblocks_if_needed()
723
if path_utf8 is not None:
724
assert path_utf8.__class__ == str, 'path_utf8 is not a str: %s %s' % (type(path_utf8), path_utf8)
725
# path lookups are faster
726
dirname, basename = os.path.split(path_utf8)
727
block_index, entry_index, dir_present, file_present = \
728
self._get_block_entry_index(dirname, basename, tree_index)
731
entry = self._dirblocks[block_index][1][entry_index]
732
assert entry[0][2] and entry[1][tree_index][0] not in ('a', 'r'), 'unversioned entry?!?!'
734
if entry[0][2] != fileid_utf8:
735
raise BzrError('integrity error ? : mismatching tree_index, file_id and path')
738
for entry in self._iter_entries():
739
if entry[0][2] == fileid_utf8:
740
if entry[1][tree_index][0] == 'r': # relocated
741
# look up the real location directly by path
742
return self._get_entry(tree_index,
743
fileid_utf8=fileid_utf8,
744
path_utf8=entry[1][tree_index][1])
745
if entry[1][tree_index][0] == 'a': # absent
746
# not in the tree at all.
752
def initialize(path):
753
"""Create a new dirstate on path.
755
The new dirstate will be an empty tree - that is it has no parents,
756
and only a root node - which has id ROOT_ID.
758
The object will be write locked when returned to the caller,
759
unless there was an exception in the writing, in which case it
762
:param path: The name of the file for the dirstate.
763
:return: A DirState object.
765
# This constructs a new DirState object on a path, sets the _state_file
766
# to a new empty file for that path. It then calls _set_data() with our
767
# stock empty dirstate information - a root with ROOT_ID, no children,
768
# and no parents. Finally it calls save() to ensure that this data will
770
result = DirState(path)
771
# root dir and root dir contents with no children.
772
empty_tree_dirblocks = [('', []), ('', [])]
773
# a new root directory, with a NULLSTAT.
774
empty_tree_dirblocks[0][1].append(
775
(('', '', bzrlib.inventory.ROOT_ID), [
776
('d', '', 0, False, DirState.NULLSTAT),
780
result._set_data([], empty_tree_dirblocks)
787
def _inv_entry_to_details(self, inv_entry):
788
"""Convert an inventory entry (from a revision tree) to state details.
790
:param inv_entry: An inventory entry whose sha1 and link targets can be
791
relied upon, and which has a revision set.
792
:return: A details tuple - the details for a single tree at a path +
795
kind = inv_entry.kind
796
minikind = DirState._kind_to_minikind[kind]
797
tree_data = inv_entry.revision
798
assert len(tree_data) > 0, 'empty revision for the inv_entry.'
799
if kind == 'directory':
803
elif kind == 'symlink':
804
fingerprint = inv_entry.symlink_target or ''
808
fingerprint = inv_entry.text_sha1 or ''
809
size = inv_entry.text_size or 0
810
executable = inv_entry.executable
813
return (minikind, fingerprint, size, executable, tree_data)
815
def _iter_entries(self):
816
"""Iterate over all the entries in the dirstate.
818
Each yelt item is an entry in the standard format described in the
819
docstring of bzrlib.dirstate.
821
self._read_dirblocks_if_needed()
822
for directory in self._dirblocks:
823
for entry in directory[1]:
826
def _get_id_index(self):
827
"""Get an id index of self._dirblocks."""
829
for key, tree_details in self._iter_entries():
830
id_index.setdefault(key[2], set()).add(key)
833
def _get_output_lines(self, lines):
834
"""format lines for final output.
836
:param lines: A sequece of lines containing the parents list and the
839
output_lines = ['#bazaar dirstate flat format 2\n']
840
lines.append('') # a final newline
841
inventory_text = '\0\n\0'.join(lines)
842
output_lines.append('adler32: %s\n' % (zlib.adler32(inventory_text),))
843
# -3, 1 for num parents, 1 for ghosts, 1 for final newline
844
num_entries = len(lines)-3
845
output_lines.append('num_entries: %s\n' % (num_entries,))
846
output_lines.append(inventory_text)
849
def _make_deleted_row(self, fileid_utf8, parents):
850
"""Return a deleted for for fileid_utf8."""
851
return ('/', 'RECYCLED.BIN', 'file', fileid_utf8, 0, DirState.NULLSTAT,
854
def _num_present_parents(self):
855
"""The number of parent entries in each record row."""
856
return len(self._parents) - len(self._ghosts)
860
"""Construct a DirState on the file at path path.
862
:return: An unlocked DirState object, associated with the given path.
864
result = DirState(path)
867
def _read_dirblocks_if_needed(self):
868
"""Read in all the dirblocks from the file if they are not in memory.
870
This populates self._dirblocks, and sets self._dirblock_state to
871
IN_MEMORY_UNMODIFIED. It is not currently ready for incremental block
874
self._read_header_if_needed()
875
if self._dirblock_state == DirState.NOT_IN_MEMORY:
876
# the _state_file pointer will be positioned at the start of the
878
text = self._state_file.read()
879
# TODO: check the adler checksums. adler_measured = zlib.adler32(text)
881
fields = text.split('\0')
882
# Remove the last blank entry
883
trailing = fields.pop()
884
assert trailing == ''
885
# consider turning fields into a tuple.
887
# skip the first field which is the trailing null from the header.
889
# Each line now has an extra '\n' field which is not used
890
# so we just skip over it
892
# 3 fields for the key
893
# + number of fields per tree_data (5) * tree count
895
num_present_parents = self._num_present_parents()
896
tree_count = 1 + num_present_parents
897
entry_size = self._fields_per_row()
898
expected_field_count = entry_size * self._num_entries
899
if len(fields) - cur > expected_field_count:
900
fields = fields[:expected_field_count + cur]
901
trace.mutter('Unexpectedly long dirstate field count!')
902
print "XXX: incorrectly truncated dirstate file bug triggered."
903
field_count = len(fields)
904
# this checks our adjustment, and also catches file too short.
905
assert field_count - cur == expected_field_count, \
906
'field count incorrect %s != %s, entry_size=%s, '\
907
'num_entries=%s fields=%r' % (
908
field_count - cur, expected_field_count, entry_size,
909
self._num_entries, fields)
911
if num_present_parents == 1:
912
# Bind external functions to local names
914
# We access all fields in order, so we can just iterate over
915
# them. Grab an straight iterator over the fields. (We use an
916
# iterator because we don't want to do a lot of additions, nor
917
# do we want to do a lot of slicing)
918
next = iter(fields).next
919
# Move the iterator to the current position
920
for x in xrange(cur):
922
# The two blocks here are deliberate: the root block and the
923
# contents-of-root block.
924
self._dirblocks = [('', []), ('', [])]
925
current_block = self._dirblocks[0][1]
927
append_entry = current_block.append
928
for count in xrange(self._num_entries):
932
if dirname != current_dirname:
933
# new block - different dirname
935
current_dirname = dirname
936
self._dirblocks.append((current_dirname, current_block))
937
append_entry = current_block.append
938
# we know current_dirname == dirname, so re-use it to avoid
939
# creating new strings
940
entry = ((current_dirname, name, file_id),
943
next(), # fingerprint
945
next() == 'y', # executable
946
next(), # packed_stat or revision_id
950
next(), # fingerprint
952
next() == 'y', # executable
953
next(), # packed_stat or revision_id
957
assert trailing == '\n'
958
# append the entry to the current block
960
self._split_root_dirblock_into_contents()
962
fields_to_entry = self._get_fields_to_entry()
963
entries = [fields_to_entry(fields[pos:pos+entry_size])
964
for pos in xrange(cur, field_count, entry_size)]
965
self._entries_to_current_state(entries)
966
self._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
968
def _read_header(self):
969
"""This reads in the metadata header, and the parent ids.
971
After reading in, the file should be positioned at the null
972
just before the start of the first record in the file.
974
:return: (expected adler checksum, number of entries, parent list)
977
parent_line = self._state_file.readline()
978
info = parent_line.split('\0')
979
num_parents = int(info[0])
980
assert num_parents == len(info)-2, 'incorrect parent info line'
981
self._parents = info[1:-1]
983
ghost_line = self._state_file.readline()
984
info = ghost_line.split('\0')
985
num_ghosts = int(info[1])
986
assert num_ghosts == len(info)-3, 'incorrect ghost info line'
987
self._ghosts = info[2:-1]
988
self._header_state = DirState.IN_MEMORY_UNMODIFIED
990
def _read_header_if_needed(self):
991
"""Read the header of the dirstate file if needed."""
992
# inline this as it will be called a lot
993
if not self._lock_token:
994
raise errors.ObjectNotLocked(self)
995
if self._header_state == DirState.NOT_IN_MEMORY:
998
def _read_prelude(self):
999
"""Read in the prelude header of the dirstate file
1001
This only reads in the stuff that is not connected to the adler
1002
checksum. The position will be correct to read in the rest of
1003
the file and check the checksum after this point.
1004
The next entry in the file should be the number of parents,
1005
and their ids. Followed by a newline.
1007
header = self._state_file.readline()
1008
assert header == '#bazaar dirstate flat format 2\n', \
1009
'invalid header line: %r' % (header,)
1010
adler_line = self._state_file.readline()
1011
assert adler_line.startswith('adler32: '), 'missing adler32 checksum'
1012
self.adler_expected = int(adler_line[len('adler32: '):-1])
1013
num_entries_line = self._state_file.readline()
1014
assert num_entries_line.startswith('num_entries: '), 'missing num_entries line'
1015
self._num_entries = int(num_entries_line[len('num_entries: '):-1])
1018
"""Save any pending changes created during this session.
1020
We reuse the existing file, because that prevents race conditions with
1021
file creation, and we expect to be using oslocks on it in the near
1022
future to prevent concurrent modification and reads - because dirstates
1023
incremental data aggretation is not compatible with reading a modified
1024
file, and replacing a file in use by another process is impossible on
1027
A dirstate in read only mode should be smart enough though to validate
1028
that the file has not changed, and otherwise discard its cache and
1029
start over, to allow for fine grained read lock duration, so 'status'
1030
wont block 'commit' - for example.
1032
if (self._header_state == DirState.IN_MEMORY_MODIFIED or
1033
self._dirblock_state == DirState.IN_MEMORY_MODIFIED):
1034
self._state_file.seek(0)
1035
self._state_file.writelines(self.get_lines())
1036
self._state_file.truncate()
1037
self._state_file.flush()
1038
self._header_state = DirState.IN_MEMORY_UNMODIFIED
1039
self._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
1041
def _set_data(self, parent_ids, dirblocks):
1042
"""Set the full dirstate data in memory.
1044
This is an internal function used to completely replace the objects
1045
in memory state. It puts the dirstate into state 'full-dirty'.
1047
:param parent_ids: A list of parent tree revision ids.
1048
:param dirblocks: A list containing one tuple for each directory in the
1049
tree. Each tuple contains the directory path and a list of entries
1050
found in that directory.
1052
# our memory copy is now authoritative.
1053
self._dirblocks = dirblocks
1054
self._header_state = DirState.IN_MEMORY_MODIFIED
1055
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1056
self._parents = list(parent_ids)
1058
def set_path_id(self, path, new_id):
1059
"""Change the id of path to new_id in the current working tree.
1061
:param path: The path inside the tree to set - '' is the root, 'foo'
1062
is the path foo in the root.
1063
:param new_id: The new id to assign to the path. This must be a utf8
1064
file id (not unicode, and not None).
1066
# TODO: start warning here.
1067
assert new_id.__class__ == str
1068
self._read_dirblocks_if_needed()
1070
import pdb;pdb.set_trace()
1072
raise NotImplementedError(self.set_path_id)
1073
# TODO: check new id is unique
1074
entry = self._get_entry(0, path_utf8=path)
1075
# mark the old path absent, and insert a new root path
1076
present_parents = len(entry[1]) - 1
1077
self._make_absent(entry)
1078
id_index = self._get_id_index()
1079
self.update_minimal(('', '', new_id), 'directory', present_parents,
1080
path_utf8='', id_index=id_index, packed_stat=entry[1][0][4])
1081
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1083
def set_parent_trees(self, trees, ghosts):
1084
"""Set the parent trees for the dirstate.
1086
:param trees: A list of revision_id, tree tuples. tree must be provided
1087
even if the revision_id refers to a ghost: supply an empty tree in
1089
:param ghosts: A list of the revision_ids that are ghosts at the time
1092
# TODO: generate a list of parent indexes to preserve to save
1093
# processing specific parent trees. In the common case one tree will
1094
# be preserved - the left most parent.
1095
# TODO: if the parent tree is a dirstate, we might want to walk them
1096
# all by path in parallel for 'optimal' common-case performance.
1097
# generate new root row.
1098
self._read_dirblocks_if_needed()
1099
# TODO future sketch: Examine the existing parents to generate a change
1100
# map and then walk the new parent trees only, mapping them into the
1101
# dirstate. Walk the dirstate at the same time to remove unreferenced
1104
# sketch: loop over all entries in the dirstate, cherry picking
1105
# entries from the parent trees, if they are not ghost trees.
1106
# after we finish walking the dirstate, all entries not in the dirstate
1107
# are deletes, so we want to append them to the end as per the design
1108
# discussions. So do a set difference on ids with the parents to
1109
# get deletes, and add them to the end.
1110
# During the update process we need to answer the following questions:
1111
# - find other keys containing a fileid in order to create cross-path
1112
# links. We dont't trivially use the inventory from other trees
1113
# because this leads to either double touching, or to accessing
1115
# - find other keys containing a path
1116
# We accumulate each entry via this dictionary, including the root
1119
# we could do parallel iterators, but because file id data may be
1120
# scattered throughout, we dont save on index overhead: we have to look
1121
# at everything anyway. We can probably save cycles by reusing parent
1122
# data and doing an incremental update when adding an additional
1123
# parent, but for now the common cases are adding a new parent (merge),
1124
# and replacing completely (commit), and commit is more common: so
1125
# optimise merge later.
1127
# ---- start generation of full tree mapping data
1128
# what trees should we use?
1129
parent_trees = [tree for rev_id, tree in trees if rev_id not in ghosts]
1130
# how many trees do we end up with
1131
parent_count = len(parent_trees)
1133
# one: the current tree
1134
for entry in self._iter_entries():
1135
# skip entries not in the current tree
1136
if entry[1][0][0] in ('a', 'r'): # absent, relocated
1138
by_path[entry[0]] = [entry[1][0]] + \
1139
[DirState.NULL_PARENT_DETAILS] * parent_count
1140
id_index[entry[0][2]] = set([entry[0]])
1142
# now the parent trees:
1143
for tree_index, tree in enumerate(parent_trees):
1144
# the index is off by one, adjust it.
1145
tree_index = tree_index + 1
1146
# when we add new locations for a fileid we need these ranges for
1147
# any fileid in this tree as we set the by_path[id] to:
1148
# already_processed_tree_details + new_details + new_location_suffix
1149
# the suffix is from tree_index+1:parent_count+1.
1150
new_location_suffix = [DirState.NULL_PARENT_DETAILS] * (parent_count - tree_index)
1151
# now stitch in all the entries from this tree
1152
for path, entry in tree.inventory.iter_entries_by_dir():
1153
# here we process each trees details for each item in the tree.
1154
# we first update any existing entries for the id at other paths,
1155
# then we either create or update the entry for the id at the
1156
# right path, and finally we add (if needed) a mapping from
1157
# file_id to this path. We do it in this order to allow us to
1158
# avoid checking all known paths for the id when generating a
1159
# new entry at this path: by adding the id->path mapping last,
1160
# all the mappings are valid and have correct relocation
1161
# records where needed.
1162
file_id = entry.file_id
1163
path_utf8 = path.encode('utf8')
1164
dirname, basename = os.path.split(path_utf8)
1165
new_entry_key = (dirname, basename, file_id)
1166
# tree index consistency: All other paths for this id in this tree
1167
# index must point to the correct path.
1168
for entry_key in id_index.setdefault(file_id, set()):
1169
# TODO:PROFILING: It might be faster to just update
1170
# rather than checking if we need to, and then overwrite
1171
# the one we are located at.
1172
if entry_key != new_entry_key:
1173
# this file id is at a different path in one of the
1174
# other trees, so put absent pointers there
1175
# This is the vertical axis in the matrix, all pointing
1177
by_path[entry_key][tree_index] = ('r', path_utf8, 0, False, '')
1178
# by path consistency: Insert into an existing path record (trivial), or
1179
# add a new one with relocation pointers for the other tree indexes.
1180
if new_entry_key in id_index[file_id]:
1181
# there is already an entry where this data belongs, just insert it.
1182
by_path[new_entry_key][tree_index] = \
1183
self._inv_entry_to_details(entry)
1185
# add relocated entries to the horizontal axis - this row
1186
# mapping from path,id. We need to look up the correct path
1187
# for the indexes from 0 to tree_index -1
1189
for lookup_index in xrange(tree_index):
1190
# boundary case: this is the first occurence of file_id
1191
# so there are no id_indexs, possibly take this out of
1193
if not len(id_index[file_id]):
1194
new_details.append(DirState.NULL_PARENT_DETAILS)
1196
# grab any one entry, use it to find the right path.
1197
# TODO: optimise this to reduce memory use in highly
1198
# fragmented situations by reusing the relocation
1200
a_key = iter(id_index[file_id]).next()
1201
if by_path[a_key][lookup_index][0] in ('r', 'a'):
1202
# its a pointer or missing statement, use it as is.
1203
new_details.append(by_path[a_key][lookup_index])
1205
# we have the right key, make a pointer to it.
1206
real_path = ('/'.join(a_key[0:2])).strip('/')
1207
new_details.append(('r', real_path, 0, False, ''))
1208
new_details.append(self._inv_entry_to_details(entry))
1209
new_details.extend(new_location_suffix)
1210
by_path[new_entry_key] = new_details
1211
id_index[file_id].add(new_entry_key)
1212
# --- end generation of full tree mappings
1214
# sort and output all the entries
1215
new_entries = sorted(by_path.items())
1216
self._entries_to_current_state(new_entries)
1217
self._parents = [rev_id for rev_id, tree in trees]
1218
self._ghosts = list(ghosts)
1219
self._header_state = DirState.IN_MEMORY_MODIFIED
1220
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1222
def set_state_from_inventory(self, new_inv):
1223
"""Set new_inv as the current state.
1225
This API is called by tree transform, and will usually occur with
1226
existing parent trees.
1228
:param new_inv: The inventory object to set current state from.
1230
self._read_dirblocks_if_needed()
1232
# generate a byid index of the dirstate
1233
id_index = self._get_id_index()
1235
num_present_parents = len(self._parents) - len(self._ghosts)
1236
# incremental algorithm:
1237
# two iterators: current data and new data, both in dirblock order.
1238
new_iterator = new_inv.iter_entries_by_dir()
1239
# we will be modifying the dirstate, so we need a stable iterator. In
1240
# future we might write one, for now we just clone the state into a
1241
# list - which is a shallow copy, so each
1242
old_iterator = iter(list(self._iter_entries()))
1243
# both must have roots so this is safe:
1244
current_new = new_iterator.next()
1245
current_old = old_iterator.next()
1246
def advance(iterator):
1248
return iterator.next()
1249
except StopIteration:
1251
while current_new or current_old:
1252
# skip entries in old that are not really there
1253
if current_old and current_old[1][0][0] in ('r', 'a'):
1254
# relocated or absent
1255
current_old = advance(old_iterator)
1258
# convert new into dirblock style
1259
new_path_utf8 = current_new[0].encode('utf8')
1260
new_dirname, new_basename = os.path.split(new_path_utf8)
1261
new_id = current_new[1].file_id
1262
new_entry_key = (new_dirname, new_basename, new_id)
1264
# for safety disable variables
1265
new_path_utf8 = new_dirname = new_basename = new_id = new_entry_key = None
1266
# 5 cases, we dont have a value that is strictly greater than everything, so
1267
# we make both end conditions explicit
1269
# old is finished: insert current_new into the state.
1270
self.update_minimal(new_entry_key, current_new[1].kind,
1271
num_present_parents, executable=current_new[1].executable,
1272
id_index=id_index, path_utf8=new_path_utf8)
1273
current_new = advance(new_iterator)
1274
elif not current_new:
1276
self._make_absent(current_old, id_index)
1277
current_old = advance(old_iterator)
1278
elif new_entry_key == current_old[0]:
1279
# same - common case
1280
# TODO: update the record if anything significant has changed.
1281
# the minimal required trigger is if the execute bit or cached
1283
kind = DirState._minikind_to_kind[current_old[1][0][0]]
1284
if (current_old[1][0][3] != current_new[1].executable or
1285
kind != current_new[1].kind):
1286
self.update_minimal(current_old[0], current_new[1].kind,
1287
num_present_parents,
1288
executable=current_new[1].executable,
1289
id_index=id_index, path_utf8=new_path_utf8)
1290
# both sides are dealt with, move on
1291
current_old = advance(old_iterator)
1292
current_new = advance(new_iterator)
1293
elif new_entry_key < current_old[0]:
1295
# add a entry for this and advance new
1296
self.update_minimal(new_entry_key, current_new[1].kind,
1297
num_present_parents, executable=current_new[1].executable,
1298
id_index=id_index, path_utf8=new_path_utf8)
1299
current_new = advance(new_iterator)
1302
self._make_absent(current_old, id_index)
1303
current_old = advance(old_iterator)
1304
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1306
def _make_absent(self, current_old, id_index=None):
1307
"""Mark current_old - an entry - as absent for tree 0.
1309
:param id_index: An index from fileid_utf8 to sets of keys, used by
1310
some functions. If provided it will be updated if needed.
1311
:return: True if this was the last details entry for they entry key:
1312
that is, if the underlying block has had the entry removed, thus
1313
shrinking in legnth.
1315
# build up paths that this id will be left at after the change is made,
1316
# so we can update their cross references in tree 0
1317
all_remaining_keys = set()
1318
# Dont check the working tree, because its going.
1319
for details in current_old[1][1:]:
1320
if details[0] not in ('a', 'r'): # absent, relocated
1321
all_remaining_keys.add(current_old[0])
1322
elif details[0] == 'r': # relocated
1323
# record the key for the real path.
1324
all_remaining_keys.add(tuple(os.path.split(details[1])) + (current_old[0][2],))
1325
# absent rows are not present at any path.
1326
last_reference = current_old[0] not in all_remaining_keys
1328
# the current row consists entire of the current item (being marked
1329
# absent), and relocated or absent entries for the other trees:
1330
# Remove it, its meaningless.
1331
block = self._find_block(current_old[0])
1332
entry_index, present = self._find_entry_index(current_old[0], block[1])
1334
block[1].pop(entry_index)
1335
# if we have an id_index in use, remove this key from it for this id.
1336
if id_index is not None:
1337
id_index[current_old[0][2]].remove(current_old[0])
1338
# update all remaining keys for this id to record it as absent. The
1339
# existing details may either be the record we are making as deleted
1340
# (if there were other trees with the id present at this path), or may
1342
for update_key in all_remaining_keys:
1343
update_block_index, present = \
1344
self._find_block_index_from_key(update_key)
1346
update_entry_index, present = \
1347
self._find_entry_index(update_key, self._dirblocks[update_block_index][1])
1349
update_tree_details = self._dirblocks[update_block_index][1][update_entry_index][1]
1350
# it must not be absent at the moment
1351
assert update_tree_details[0][0] != 'a' # absent
1352
update_tree_details[0] = DirState.NULL_PARENT_DETAILS
1353
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1354
return last_reference
1356
def update_minimal(self, key, kind, num_present_parents, executable=False,
1357
fingerprint='', packed_stat=None, size=0, id_index=None,
1359
"""Update an entry to the state in tree 0."""
1360
block = self._find_block(key)[1]
1361
if packed_stat is None:
1362
packed_stat = DirState.NULLSTAT
1363
entry_index, present = self._find_entry_index(key, block)
1364
minikind = DirState._kind_to_minikind[kind]
1365
new_details = (minikind, fingerprint, size, executable, packed_stat)
1366
assert id_index is not None, 'need an id index to do updates for now !'
1368
# new entry, synthesis cross reference here,
1369
existing_keys = id_index.setdefault(key[2], set())
1370
if not existing_keys:
1371
# not currently in the state, simplest case
1372
new_entry = key, [new_details] + self._empty_parent_info()
1374
# present at one or more existing other paths.
1375
# grab one of them and use it to generate parent
1376
# relocation/absent entries.
1377
new_entry = key, [new_details]
1378
for other_key in existing_keys:
1379
# change the record at other to be a pointer to this new
1380
# record. The loop looks similar to the change to
1381
# relocations when updating an existing record but its not:
1382
# the test for existing kinds is different: this can be
1383
# factored out to a helper though.
1384
other_block_index, present = self._find_block_index_from_key(other_key)
1386
other_entry_index, present = self._find_entry_index(other_key, self._dirblocks[other_block_index][1])
1388
assert path_utf8 is not None
1389
self._dirblocks[other_block_index][1][other_entry_index][1][0] = \
1390
('r', path_utf8, 0, False, '')
1392
for lookup_index in xrange(1, num_present_parents + 1):
1393
# grab any one entry, use it to find the right path.
1394
# TODO: optimise this to reduce memory use in highly
1395
# fragmented situations by reusing the relocation
1397
update_block_index, present = \
1398
self._find_block_index_from_key(other_key)
1400
update_entry_index, present = \
1401
self._find_entry_index(other_key, self._dirblocks[update_block_index][1])
1403
update_details = self._dirblocks[update_block_index][1][update_entry_index][1][lookup_index]
1404
if update_details[0] in ('r', 'a'): # relocated, absent
1405
# its a pointer or absent in lookup_index's tree, use
1407
new_entry[1].append(update_details)
1409
# we have the right key, make a pointer to it.
1410
pointer_path = os.path.join(*other_key[0:2])
1411
new_entry[1].append(('r', pointer_path, 0, False, ''))
1412
block.insert(entry_index, new_entry)
1413
existing_keys.add(key)
1415
# Does the new state matter?
1416
block[entry_index][1][0] = new_details
1417
# parents cannot be affected by what we do.
1418
# other occurences of this id can be found
1419
# from the id index.
1421
# tree index consistency: All other paths for this id in this tree
1422
# index must point to the correct path. We have to loop here because
1423
# we may have passed entries in the state with this file id already
1424
# that were absent - where parent entries are - and they need to be
1425
# converted to relocated.
1426
assert path_utf8 is not None
1427
for entry_key in id_index.setdefault(key[2], set()):
1428
# TODO:PROFILING: It might be faster to just update
1429
# rather than checking if we need to, and then overwrite
1430
# the one we are located at.
1431
if entry_key != key:
1432
# this file id is at a different path in one of the
1433
# other trees, so put absent pointers there
1434
# This is the vertical axis in the matrix, all pointing
1436
block_index, present = self._find_block_index_from_key(entry_key)
1438
entry_index, present = self._find_entry_index(entry_key, self._dirblocks[block_index][1])
1440
self._dirblocks[block_index][1][entry_index][1][0] = \
1441
('r', path_utf8, 0, False, '')
1442
# add a containing dirblock if needed.
1443
if new_details[0] == 'd':
1444
subdir_key = (os.path.join(*key[0:2]), '', '')
1445
block_index, present = self._find_block_index_from_key(subdir_key)
1447
self._dirblocks.insert(block_index, (subdir_key[0], []))
1449
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1452
def _wipe_state(self):
1453
"""Forget all state information about the dirstate."""
1454
self._header_state = DirState.NOT_IN_MEMORY
1455
self._dirblock_state = DirState.NOT_IN_MEMORY
1458
self._dirblocks = []
1460
def lock_read(self):
1461
"""Acquire a read lock on the dirstate"""
1462
if self._lock_token is not None:
1463
raise errors.LockContention(self._lock_token)
1464
self._lock_token = lock.ReadLock(self._filename)
1465
self._state_file = self._lock_token.f
1468
def lock_write(self):
1469
"""Acquire a write lock on the dirstate"""
1470
if self._lock_token is not None:
1471
raise errors.LockContention(self._lock_token)
1472
self._lock_token = lock.WriteLock(self._filename)
1473
self._state_file = self._lock_token.f
1477
"""Drop any locks held on the dirstate"""
1478
if self._lock_token is None:
1479
raise errors.LockNotHeld(self)
1480
self._state_file = None
1481
self._lock_token.unlock()
1482
self._lock_token = None
1484
def _requires_lock(self):
1485
"""Checks that a lock is currently held by someone on the dirstate"""
1486
if not self._lock_token:
1487
raise errors.ObjectNotLocked(self)
1489
def pack_stat(st, _encode=base64.encodestring, _pack=struct.pack):
1490
"""Convert stat values into a packed representation."""
1491
# jam 20060614 it isn't really worth removing more entries if we
1492
# are going to leave it in packed form.
1493
# With only st_mtime and st_mode filesize is 5.5M and read time is 275ms
1494
# With all entries filesize is 5.9M and read time is mabye 280ms
1495
# well within the noise margin
1497
# base64.encode always adds a final newline, so strip it off
1498
return _encode(_pack('>llllll'
1499
, st.st_size, st.st_mtime, st.st_ctime
1500
, st.st_dev, st.st_ino, st.st_mode))[:-1]
added
removed
bzrlib/dirstate.py
