1
# Copyright (C) 2006, 2007, 2008 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" | "t";
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 3", NL;
33
full checksum = "crc32: ", ["-"], WHOLE_NUMBER, NL;
34
row count = "num_entries: ", WHOLE_NUMBER, 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
89
't' is a reference to a nested subtree; the fingerprint is the referenced
94
The entries on disk and in memory are ordered according to the following keys:
96
directory, as a list of components
100
--- Format 1 had the following different definition: ---
101
rows = dirname, NULL, basename, NULL, MINIKIND, NULL, fileid_utf8, NULL,
102
WHOLE NUMBER (* size *), NULL, packed stat, NULL, sha1|symlink target,
104
PARENT ROW = NULL, revision_utf8, NULL, MINIKIND, NULL, dirname, NULL,
105
basename, NULL, WHOLE NUMBER (* size *), NULL, "y" | "n", NULL,
108
PARENT ROW's are emitted for every parent that is not in the ghosts details
109
line. That is, if the parents are foo, bar, baz, and the ghosts are bar, then
110
each row will have a PARENT ROW for foo and baz, but not for bar.
113
In any tree, a kind of 'moved' indicates that the fingerprint field
114
(which we treat as opaque data specific to the 'kind' anyway) has the
115
details for the id of this row in that tree.
117
I'm strongly tempted to add a id->path index as well, but I think that
118
where we need id->path mapping; we also usually read the whole file, so
119
I'm going to skip that for the moment, as we have the ability to locate
120
via bisect any path in any tree, and if we lookup things by path, we can
121
accumulate an id->path mapping as we go, which will tend to match what we
124
I plan to implement this asap, so please speak up now to alter/tweak the
125
design - and once we stabilise on this, I'll update the wiki page for
128
The rationale for all this is that we want fast operations for the
129
common case (diff/status/commit/merge on all files) and extremely fast
130
operations for the less common but still occurs a lot status/diff/commit
131
on specific files). Operations on specific files involve a scan for all
132
the children of a path, *in every involved tree*, which the current
133
format did not accommodate.
137
1) Fast end to end use for bzr's top 5 uses cases. (commmit/diff/status/merge/???)
138
2) fall back current object model as needed.
139
3) scale usably to the largest trees known today - say 50K entries. (mozilla
140
is an example of this)
144
Eventually reuse dirstate objects across locks IFF the dirstate file has not
145
been modified, but will require that we flush/ignore cached stat-hit data
146
because we won't want to restat all files on disk just because a lock was
147
acquired, yet we cannot trust the data after the previous lock was released.
149
Memory representation:
150
vector of all directories, and vector of the childen ?
152
root_entrie = (direntry for root, [parent_direntries_for_root]),
154
('', ['data for achild', 'data for bchild', 'data for cchild'])
155
('dir', ['achild', 'cchild', 'echild'])
157
- single bisect to find N subtrees from a path spec
158
- in-order for serialisation - this is 'dirblock' grouping.
159
- insertion of a file '/a' affects only the '/' child-vector, that is, to
160
insert 10K elements from scratch does not generates O(N^2) memoves of a
161
single vector, rather each individual, which tends to be limited to a
162
manageable number. Will scale badly on trees with 10K entries in a
163
single directory. compare with Inventory.InventoryDirectory which has
164
a dictionary for the children. No bisect capability, can only probe for
165
exact matches, or grab all elements and sort.
166
- What's the risk of error here? Once we have the base format being processed
167
we should have a net win regardless of optimality. So we are going to
168
go with what seems reasonable.
171
Maybe we should do a test profile of the core structure - 10K simulated
172
searches/lookups/etc?
174
Objects for each row?
175
The lifetime of Dirstate objects is current per lock, but see above for
176
possible extensions. The lifetime of a row from a dirstate is expected to be
177
very short in the optimistic case: which we are optimising for. For instance,
178
subtree status will determine from analysis of the disk data what rows need to
179
be examined at all, and will be able to determine from a single row whether
180
that file has altered or not, so we are aiming to process tens of thousands of
181
entries each second within the dirstate context, before exposing anything to
182
the larger codebase. This suggests we want the time for a single file
183
comparison to be < 0.1 milliseconds. That would give us 10000 paths per second
184
processed, and to scale to 100 thousand we'll another order of magnitude to do
185
that. Now, as the lifetime for all unchanged entries is the time to parse, stat
186
the file on disk, and then immediately discard, the overhead of object creation
187
becomes a significant cost.
189
Figures: Creating a tuple from from 3 elements was profiled at 0.0625
190
microseconds, whereas creating a object which is subclassed from tuple was
191
0.500 microseconds, and creating an object with 3 elements and slots was 3
192
microseconds long. 0.1 milliseconds is 100 microseconds, and ideally we'll get
193
down to 10 microseconds for the total processing - having 33% of that be object
194
creation is a huge overhead. There is a potential cost in using tuples within
195
each row which is that the conditional code to do comparisons may be slower
196
than method invocation, but method invocation is known to be slow due to stack
197
frame creation, so avoiding methods in these tight inner loops in unfortunately
198
desirable. We can consider a pyrex version of this with objects in future if
207
from stat import S_IEXEC
225
def pack_stat(st, _encode=binascii.b2a_base64, _pack=struct.pack):
226
"""Convert stat values into a packed representation."""
227
# jam 20060614 it isn't really worth removing more entries if we
228
# are going to leave it in packed form.
229
# With only st_mtime and st_mode filesize is 5.5M and read time is 275ms
230
# With all entries, filesize is 5.9M and read time is maybe 280ms
231
# well within the noise margin
233
# base64 encoding always adds a final newline, so strip it off
234
# The current version
235
return _encode(_pack('>LLLLLL'
236
, st.st_size, int(st.st_mtime), int(st.st_ctime)
237
, st.st_dev, st.st_ino & 0xFFFFFFFF, st.st_mode))[:-1]
238
# This is 0.060s / 1.520s faster by not encoding as much information
239
# return _encode(_pack('>LL', int(st.st_mtime), st.st_mode))[:-1]
240
# This is not strictly faster than _encode(_pack())[:-1]
241
# return '%X.%X.%X.%X.%X.%X' % (
242
# st.st_size, int(st.st_mtime), int(st.st_ctime),
243
# st.st_dev, st.st_ino, st.st_mode)
244
# Similar to the _encode(_pack('>LL'))
245
# return '%X.%X' % (int(st.st_mtime), st.st_mode)
248
class DirState(object):
249
"""Record directory and metadata state for fast access.
251
A dirstate is a specialised data structure for managing local working
252
tree state information. Its not yet well defined whether it is platform
253
specific, and if it is how we detect/parameterize that.
255
Dirstates use the usual lock_write, lock_read and unlock mechanisms.
256
Unlike most bzr disk formats, DirStates must be locked for reading, using
257
lock_read. (This is an os file lock internally.) This is necessary
258
because the file can be rewritten in place.
260
DirStates must be explicitly written with save() to commit changes; just
261
unlocking them does not write the changes to disk.
264
_kind_to_minikind = {
270
'tree-reference': 't',
272
_minikind_to_kind = {
278
't': 'tree-reference',
280
_stat_to_minikind = {
285
_to_yesno = {True:'y', False: 'n'} # TODO profile the performance gain
286
# of using int conversion rather than a dict here. AND BLAME ANDREW IF
289
# TODO: jam 20070221 Figure out what to do if we have a record that exceeds
290
# the BISECT_PAGE_SIZE. For now, we just have to make it large enough
291
# that we are sure a single record will always fit.
292
BISECT_PAGE_SIZE = 4096
295
IN_MEMORY_UNMODIFIED = 1
296
IN_MEMORY_MODIFIED = 2
298
# A pack_stat (the x's) that is just noise and will never match the output
301
NULL_PARENT_DETAILS = ('a', '', 0, False, '')
303
HEADER_FORMAT_2 = '#bazaar dirstate flat format 2\n'
304
HEADER_FORMAT_3 = '#bazaar dirstate flat format 3\n'
306
def __init__(self, path):
307
"""Create a DirState object.
309
:param path: The path at which the dirstate file on disk should live.
311
# _header_state and _dirblock_state represent the current state
312
# of the dirstate metadata and the per-row data respectiely.
313
# NOT_IN_MEMORY indicates that no data is in memory
314
# IN_MEMORY_UNMODIFIED indicates that what we have in memory
315
# is the same as is on disk
316
# IN_MEMORY_MODIFIED indicates that we have a modified version
317
# of what is on disk.
318
# In future we will add more granularity, for instance _dirblock_state
319
# will probably support partially-in-memory as a separate variable,
320
# allowing for partially-in-memory unmodified and partially-in-memory
322
self._header_state = DirState.NOT_IN_MEMORY
323
self._dirblock_state = DirState.NOT_IN_MEMORY
324
# If true, an error has been detected while updating the dirstate, and
325
# for safety we're not going to commit to disk.
326
self._changes_aborted = False
330
self._state_file = None
331
self._filename = path
332
self._lock_token = None
333
self._lock_state = None
334
self._id_index = None
335
# a map from packed_stat to sha's.
336
self._packed_stat_index = None
337
self._end_of_header = None
338
self._cutoff_time = None
339
self._split_path_cache = {}
340
self._bisect_page_size = DirState.BISECT_PAGE_SIZE
341
if 'hashcache' in debug.debug_flags:
342
self._sha1_file = self._sha1_file_and_mutter
344
self._sha1_file = osutils.sha_file_by_name
345
# These two attributes provide a simple cache for lookups into the
346
# dirstate in-memory vectors. By probing respectively for the last
347
# block, and for the next entry, we save nearly 2 bisections per path
349
self._last_block_index = None
350
self._last_entry_index = None
354
(self.__class__.__name__, self._filename)
356
def add(self, path, file_id, kind, stat, fingerprint):
357
"""Add a path to be tracked.
359
:param path: The path within the dirstate - '' is the root, 'foo' is the
360
path foo within the root, 'foo/bar' is the path bar within foo
362
:param file_id: The file id of the path being added.
363
:param kind: The kind of the path, as a string like 'file',
365
:param stat: The output of os.lstat for the path.
366
:param fingerprint: The sha value of the file,
367
or the target of a symlink,
368
or the referenced revision id for tree-references,
369
or '' for directories.
372
# find the block its in.
373
# find the location in the block.
374
# check its not there
376
#------- copied from inventory.ensure_normalized_name - keep synced.
377
# --- normalized_filename wants a unicode basename only, so get one.
378
dirname, basename = osutils.split(path)
379
# we dont import normalized_filename directly because we want to be
380
# able to change the implementation at runtime for tests.
381
norm_name, can_access = osutils.normalized_filename(basename)
382
if norm_name != basename:
386
raise errors.InvalidNormalization(path)
387
# you should never have files called . or ..; just add the directory
388
# in the parent, or according to the special treatment for the root
389
if basename == '.' or basename == '..':
390
raise errors.InvalidEntryName(path)
391
# now that we've normalised, we need the correct utf8 path and
392
# dirname and basename elements. This single encode and split should be
393
# faster than three separate encodes.
394
utf8path = (dirname + '/' + basename).strip('/').encode('utf8')
395
dirname, basename = osutils.split(utf8path)
396
# uses __class__ for speed; the check is needed for safety
397
if file_id.__class__ is not str:
398
raise AssertionError(
399
"must be a utf8 file_id not %s" % (type(file_id), ))
400
# Make sure the file_id does not exist in this tree
401
file_id_entry = self._get_entry(0, fileid_utf8=file_id)
402
if file_id_entry != (None, None):
403
path = osutils.pathjoin(file_id_entry[0][0], file_id_entry[0][1])
404
kind = DirState._minikind_to_kind[file_id_entry[1][0][0]]
405
info = '%s:%s' % (kind, path)
406
raise errors.DuplicateFileId(file_id, info)
407
first_key = (dirname, basename, '')
408
block_index, present = self._find_block_index_from_key(first_key)
410
# check the path is not in the tree
411
block = self._dirblocks[block_index][1]
412
entry_index, _ = self._find_entry_index(first_key, block)
413
while (entry_index < len(block) and
414
block[entry_index][0][0:2] == first_key[0:2]):
415
if block[entry_index][1][0][0] not in 'ar':
416
# this path is in the dirstate in the current tree.
417
raise Exception, "adding already added path!"
420
# The block where we want to put the file is not present. But it
421
# might be because the directory was empty, or not loaded yet. Look
422
# for a parent entry, if not found, raise NotVersionedError
423
parent_dir, parent_base = osutils.split(dirname)
424
parent_block_idx, parent_entry_idx, _, parent_present = \
425
self._get_block_entry_index(parent_dir, parent_base, 0)
426
if not parent_present:
427
raise errors.NotVersionedError(path, str(self))
428
self._ensure_block(parent_block_idx, parent_entry_idx, dirname)
429
block = self._dirblocks[block_index][1]
430
entry_key = (dirname, basename, file_id)
433
packed_stat = DirState.NULLSTAT
436
packed_stat = pack_stat(stat)
437
parent_info = self._empty_parent_info()
438
minikind = DirState._kind_to_minikind[kind]
440
entry_data = entry_key, [
441
(minikind, fingerprint, size, False, packed_stat),
443
elif kind == 'directory':
444
entry_data = entry_key, [
445
(minikind, '', 0, False, packed_stat),
447
elif kind == 'symlink':
448
entry_data = entry_key, [
449
(minikind, fingerprint, size, False, packed_stat),
451
elif kind == 'tree-reference':
452
entry_data = entry_key, [
453
(minikind, fingerprint, 0, False, packed_stat),
456
raise errors.BzrError('unknown kind %r' % kind)
457
entry_index, present = self._find_entry_index(entry_key, block)
459
block.insert(entry_index, entry_data)
461
if block[entry_index][1][0][0] != 'a':
462
raise AssertionError(" %r(%r) already added" % (basename, file_id))
463
block[entry_index][1][0] = entry_data[1][0]
465
if kind == 'directory':
466
# insert a new dirblock
467
self._ensure_block(block_index, entry_index, utf8path)
468
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
470
self._id_index.setdefault(entry_key[2], set()).add(entry_key)
472
def _bisect(self, paths):
473
"""Bisect through the disk structure for specific rows.
475
:param paths: A list of paths to find
476
:return: A dict mapping path => entries for found entries. Missing
477
entries will not be in the map.
478
The list is not sorted, and entries will be populated
479
based on when they were read.
481
self._requires_lock()
482
# We need the file pointer to be right after the initial header block
483
self._read_header_if_needed()
484
# If _dirblock_state was in memory, we should just return info from
485
# there, this function is only meant to handle when we want to read
487
if self._dirblock_state != DirState.NOT_IN_MEMORY:
488
raise AssertionError("bad dirblock state %r" % self._dirblock_state)
490
# The disk representation is generally info + '\0\n\0' at the end. But
491
# for bisecting, it is easier to treat this as '\0' + info + '\0\n'
492
# Because it means we can sync on the '\n'
493
state_file = self._state_file
494
file_size = os.fstat(state_file.fileno()).st_size
495
# We end up with 2 extra fields, we should have a trailing '\n' to
496
# ensure that we read the whole record, and we should have a precursur
497
# '' which ensures that we start after the previous '\n'
498
entry_field_count = self._fields_per_entry() + 1
500
low = self._end_of_header
501
high = file_size - 1 # Ignore the final '\0'
502
# Map from (dir, name) => entry
505
# Avoid infinite seeking
506
max_count = 30*len(paths)
508
# pending is a list of places to look.
509
# each entry is a tuple of low, high, dir_names
510
# low -> the first byte offset to read (inclusive)
511
# high -> the last byte offset (inclusive)
512
# dir_names -> The list of (dir, name) pairs that should be found in
513
# the [low, high] range
514
pending = [(low, high, paths)]
516
page_size = self._bisect_page_size
518
fields_to_entry = self._get_fields_to_entry()
521
low, high, cur_files = pending.pop()
523
if not cur_files or low >= high:
528
if count > max_count:
529
raise errors.BzrError('Too many seeks, most likely a bug.')
531
mid = max(low, (low+high-page_size)/2)
534
# limit the read size, so we don't end up reading data that we have
536
read_size = min(page_size, (high-mid)+1)
537
block = state_file.read(read_size)
540
entries = block.split('\n')
543
# We didn't find a '\n', so we cannot have found any records.
544
# So put this range back and try again. But we know we have to
545
# increase the page size, because a single read did not contain
546
# a record break (so records must be larger than page_size)
548
pending.append((low, high, cur_files))
551
# Check the first and last entries, in case they are partial, or if
552
# we don't care about the rest of this page
554
first_fields = entries[0].split('\0')
555
if len(first_fields) < entry_field_count:
556
# We didn't get the complete first entry
557
# so move start, and grab the next, which
558
# should be a full entry
559
start += len(entries[0])+1
560
first_fields = entries[1].split('\0')
563
if len(first_fields) <= 2:
564
# We didn't even get a filename here... what do we do?
565
# Try a large page size and repeat this query
567
pending.append((low, high, cur_files))
570
# Find what entries we are looking for, which occur before and
571
# after this first record.
574
first_path = first_fields[1] + '/' + first_fields[2]
576
first_path = first_fields[2]
577
first_loc = _bisect_path_left(cur_files, first_path)
579
# These exist before the current location
580
pre = cur_files[:first_loc]
581
# These occur after the current location, which may be in the
582
# data we read, or might be after the last entry
583
post = cur_files[first_loc:]
585
if post and len(first_fields) >= entry_field_count:
586
# We have files after the first entry
588
# Parse the last entry
589
last_entry_num = len(entries)-1
590
last_fields = entries[last_entry_num].split('\0')
591
if len(last_fields) < entry_field_count:
592
# The very last hunk was not complete,
593
# read the previous hunk
594
after = mid + len(block) - len(entries[-1])
596
last_fields = entries[last_entry_num].split('\0')
598
after = mid + len(block)
601
last_path = last_fields[1] + '/' + last_fields[2]
603
last_path = last_fields[2]
604
last_loc = _bisect_path_right(post, last_path)
606
middle_files = post[:last_loc]
607
post = post[last_loc:]
610
# We have files that should occur in this block
611
# (>= first, <= last)
612
# Either we will find them here, or we can mark them as
615
if middle_files[0] == first_path:
616
# We might need to go before this location
617
pre.append(first_path)
618
if middle_files[-1] == last_path:
619
post.insert(0, last_path)
621
# Find out what paths we have
622
paths = {first_path:[first_fields]}
623
# last_path might == first_path so we need to be
624
# careful if we should append rather than overwrite
625
if last_entry_num != first_entry_num:
626
paths.setdefault(last_path, []).append(last_fields)
627
for num in xrange(first_entry_num+1, last_entry_num):
628
# TODO: jam 20070223 We are already splitting here, so
629
# shouldn't we just split the whole thing rather
630
# than doing the split again in add_one_record?
631
fields = entries[num].split('\0')
633
path = fields[1] + '/' + fields[2]
636
paths.setdefault(path, []).append(fields)
638
for path in middle_files:
639
for fields in paths.get(path, []):
640
# offset by 1 because of the opening '\0'
641
# consider changing fields_to_entry to avoid the
643
entry = fields_to_entry(fields[1:])
644
found.setdefault(path, []).append(entry)
646
# Now we have split up everything into pre, middle, and post, and
647
# we have handled everything that fell in 'middle'.
648
# We add 'post' first, so that we prefer to seek towards the
649
# beginning, so that we will tend to go as early as we need, and
650
# then only seek forward after that.
652
pending.append((after, high, post))
654
pending.append((low, start-1, pre))
656
# Consider that we may want to return the directory entries in sorted
657
# order. For now, we just return them in whatever order we found them,
658
# and leave it up to the caller if they care if it is ordered or not.
661
def _bisect_dirblocks(self, dir_list):
662
"""Bisect through the disk structure to find entries in given dirs.
664
_bisect_dirblocks is meant to find the contents of directories, which
665
differs from _bisect, which only finds individual entries.
667
:param dir_list: A sorted list of directory names ['', 'dir', 'foo'].
668
:return: A map from dir => entries_for_dir
670
# TODO: jam 20070223 A lot of the bisecting logic could be shared
671
# between this and _bisect. It would require parameterizing the
672
# inner loop with a function, though. We should evaluate the
673
# performance difference.
674
self._requires_lock()
675
# We need the file pointer to be right after the initial header block
676
self._read_header_if_needed()
677
# If _dirblock_state was in memory, we should just return info from
678
# there, this function is only meant to handle when we want to read
680
if self._dirblock_state != DirState.NOT_IN_MEMORY:
681
raise AssertionError("bad dirblock state %r" % self._dirblock_state)
682
# The disk representation is generally info + '\0\n\0' at the end. But
683
# for bisecting, it is easier to treat this as '\0' + info + '\0\n'
684
# Because it means we can sync on the '\n'
685
state_file = self._state_file
686
file_size = os.fstat(state_file.fileno()).st_size
687
# We end up with 2 extra fields, we should have a trailing '\n' to
688
# ensure that we read the whole record, and we should have a precursur
689
# '' which ensures that we start after the previous '\n'
690
entry_field_count = self._fields_per_entry() + 1
692
low = self._end_of_header
693
high = file_size - 1 # Ignore the final '\0'
694
# Map from dir => entry
697
# Avoid infinite seeking
698
max_count = 30*len(dir_list)
700
# pending is a list of places to look.
701
# each entry is a tuple of low, high, dir_names
702
# low -> the first byte offset to read (inclusive)
703
# high -> the last byte offset (inclusive)
704
# dirs -> The list of directories that should be found in
705
# the [low, high] range
706
pending = [(low, high, dir_list)]
708
page_size = self._bisect_page_size
710
fields_to_entry = self._get_fields_to_entry()
713
low, high, cur_dirs = pending.pop()
715
if not cur_dirs or low >= high:
720
if count > max_count:
721
raise errors.BzrError('Too many seeks, most likely a bug.')
723
mid = max(low, (low+high-page_size)/2)
726
# limit the read size, so we don't end up reading data that we have
728
read_size = min(page_size, (high-mid)+1)
729
block = state_file.read(read_size)
732
entries = block.split('\n')
735
# We didn't find a '\n', so we cannot have found any records.
736
# So put this range back and try again. But we know we have to
737
# increase the page size, because a single read did not contain
738
# a record break (so records must be larger than page_size)
740
pending.append((low, high, cur_dirs))
743
# Check the first and last entries, in case they are partial, or if
744
# we don't care about the rest of this page
746
first_fields = entries[0].split('\0')
747
if len(first_fields) < entry_field_count:
748
# We didn't get the complete first entry
749
# so move start, and grab the next, which
750
# should be a full entry
751
start += len(entries[0])+1
752
first_fields = entries[1].split('\0')
755
if len(first_fields) <= 1:
756
# We didn't even get a dirname here... what do we do?
757
# Try a large page size and repeat this query
759
pending.append((low, high, cur_dirs))
762
# Find what entries we are looking for, which occur before and
763
# after this first record.
765
first_dir = first_fields[1]
766
first_loc = bisect.bisect_left(cur_dirs, first_dir)
768
# These exist before the current location
769
pre = cur_dirs[:first_loc]
770
# These occur after the current location, which may be in the
771
# data we read, or might be after the last entry
772
post = cur_dirs[first_loc:]
774
if post and len(first_fields) >= entry_field_count:
775
# We have records to look at after the first entry
777
# Parse the last entry
778
last_entry_num = len(entries)-1
779
last_fields = entries[last_entry_num].split('\0')
780
if len(last_fields) < entry_field_count:
781
# The very last hunk was not complete,
782
# read the previous hunk
783
after = mid + len(block) - len(entries[-1])
785
last_fields = entries[last_entry_num].split('\0')
787
after = mid + len(block)
789
last_dir = last_fields[1]
790
last_loc = bisect.bisect_right(post, last_dir)
792
middle_files = post[:last_loc]
793
post = post[last_loc:]
796
# We have files that should occur in this block
797
# (>= first, <= last)
798
# Either we will find them here, or we can mark them as
801
if middle_files[0] == first_dir:
802
# We might need to go before this location
803
pre.append(first_dir)
804
if middle_files[-1] == last_dir:
805
post.insert(0, last_dir)
807
# Find out what paths we have
808
paths = {first_dir:[first_fields]}
809
# last_dir might == first_dir so we need to be
810
# careful if we should append rather than overwrite
811
if last_entry_num != first_entry_num:
812
paths.setdefault(last_dir, []).append(last_fields)
813
for num in xrange(first_entry_num+1, last_entry_num):
814
# TODO: jam 20070223 We are already splitting here, so
815
# shouldn't we just split the whole thing rather
816
# than doing the split again in add_one_record?
817
fields = entries[num].split('\0')
818
paths.setdefault(fields[1], []).append(fields)
820
for cur_dir in middle_files:
821
for fields in paths.get(cur_dir, []):
822
# offset by 1 because of the opening '\0'
823
# consider changing fields_to_entry to avoid the
825
entry = fields_to_entry(fields[1:])
826
found.setdefault(cur_dir, []).append(entry)
828
# Now we have split up everything into pre, middle, and post, and
829
# we have handled everything that fell in 'middle'.
830
# We add 'post' first, so that we prefer to seek towards the
831
# beginning, so that we will tend to go as early as we need, and
832
# then only seek forward after that.
834
pending.append((after, high, post))
836
pending.append((low, start-1, pre))
840
def _bisect_recursive(self, paths):
841
"""Bisect for entries for all paths and their children.
843
This will use bisect to find all records for the supplied paths. It
844
will then continue to bisect for any records which are marked as
845
directories. (and renames?)
847
:param paths: A sorted list of (dir, name) pairs
848
eg: [('', 'a'), ('', 'f'), ('a/b', 'c')]
849
:return: A dictionary mapping (dir, name, file_id) => [tree_info]
851
# Map from (dir, name, file_id) => [tree_info]
854
found_dir_names = set()
856
# Directories that have been read
857
processed_dirs = set()
858
# Get the ball rolling with the first bisect for all entries.
859
newly_found = self._bisect(paths)
862
# Directories that need to be read
864
paths_to_search = set()
865
for entry_list in newly_found.itervalues():
866
for dir_name_id, trees_info in entry_list:
867
found[dir_name_id] = trees_info
868
found_dir_names.add(dir_name_id[:2])
870
for tree_info in trees_info:
871
minikind = tree_info[0]
874
# We already processed this one as a directory,
875
# we don't need to do the extra work again.
877
subdir, name, file_id = dir_name_id
878
path = osutils.pathjoin(subdir, name)
880
if path not in processed_dirs:
881
pending_dirs.add(path)
882
elif minikind == 'r':
883
# Rename, we need to directly search the target
884
# which is contained in the fingerprint column
885
dir_name = osutils.split(tree_info[1])
886
if dir_name[0] in pending_dirs:
887
# This entry will be found in the dir search
889
if dir_name not in found_dir_names:
890
paths_to_search.add(tree_info[1])
891
# Now we have a list of paths to look for directly, and
892
# directory blocks that need to be read.
893
# newly_found is mixing the keys between (dir, name) and path
894
# entries, but that is okay, because we only really care about the
896
newly_found = self._bisect(sorted(paths_to_search))
897
newly_found.update(self._bisect_dirblocks(sorted(pending_dirs)))
898
processed_dirs.update(pending_dirs)
901
def _discard_merge_parents(self):
902
"""Discard any parents trees beyond the first.
904
Note that if this fails the dirstate is corrupted.
906
After this function returns the dirstate contains 2 trees, neither of
909
self._read_header_if_needed()
910
parents = self.get_parent_ids()
913
# only require all dirblocks if we are doing a full-pass removal.
914
self._read_dirblocks_if_needed()
915
dead_patterns = set([('a', 'r'), ('a', 'a'), ('r', 'r'), ('r', 'a')])
916
def iter_entries_removable():
917
for block in self._dirblocks:
918
deleted_positions = []
919
for pos, entry in enumerate(block[1]):
921
if (entry[1][0][0], entry[1][1][0]) in dead_patterns:
922
deleted_positions.append(pos)
923
if deleted_positions:
924
if len(deleted_positions) == len(block[1]):
927
for pos in reversed(deleted_positions):
929
# if the first parent is a ghost:
930
if parents[0] in self.get_ghosts():
931
empty_parent = [DirState.NULL_PARENT_DETAILS]
932
for entry in iter_entries_removable():
933
entry[1][1:] = empty_parent
935
for entry in iter_entries_removable():
939
self._parents = [parents[0]]
940
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
941
self._header_state = DirState.IN_MEMORY_MODIFIED
943
def _empty_parent_info(self):
944
return [DirState.NULL_PARENT_DETAILS] * (len(self._parents) -
947
def _ensure_block(self, parent_block_index, parent_row_index, dirname):
948
"""Ensure a block for dirname exists.
950
This function exists to let callers which know that there is a
951
directory dirname ensure that the block for it exists. This block can
952
fail to exist because of demand loading, or because a directory had no
953
children. In either case it is not an error. It is however an error to
954
call this if there is no parent entry for the directory, and thus the
955
function requires the coordinates of such an entry to be provided.
957
The root row is special cased and can be indicated with a parent block
960
:param parent_block_index: The index of the block in which dirname's row
962
:param parent_row_index: The index in the parent block where the row
964
:param dirname: The utf8 dirname to ensure there is a block for.
965
:return: The index for the block.
967
if dirname == '' and parent_row_index == 0 and parent_block_index == 0:
968
# This is the signature of the root row, and the
969
# contents-of-root row is always index 1
971
# the basename of the directory must be the end of its full name.
972
if not (parent_block_index == -1 and
973
parent_block_index == -1 and dirname == ''):
974
if not dirname.endswith(
975
self._dirblocks[parent_block_index][1][parent_row_index][0][1]):
976
raise AssertionError("bad dirname %r" % dirname)
977
block_index, present = self._find_block_index_from_key((dirname, '', ''))
979
## In future, when doing partial parsing, this should load and
980
# populate the entire block.
981
self._dirblocks.insert(block_index, (dirname, []))
984
def _entries_to_current_state(self, new_entries):
985
"""Load new_entries into self.dirblocks.
987
Process new_entries into the current state object, making them the active
988
state. The entries are grouped together by directory to form dirblocks.
990
:param new_entries: A sorted list of entries. This function does not sort
991
to prevent unneeded overhead when callers have a sorted list already.
994
if new_entries[0][0][0:2] != ('', ''):
995
raise AssertionError(
996
"Missing root row %r" % (new_entries[0][0],))
997
# The two blocks here are deliberate: the root block and the
998
# contents-of-root block.
999
self._dirblocks = [('', []), ('', [])]
1000
current_block = self._dirblocks[0][1]
1001
current_dirname = ''
1003
append_entry = current_block.append
1004
for entry in new_entries:
1005
if entry[0][0] != current_dirname:
1006
# new block - different dirname
1008
current_dirname = entry[0][0]
1009
self._dirblocks.append((current_dirname, current_block))
1010
append_entry = current_block.append
1011
# append the entry to the current block
1013
self._split_root_dirblock_into_contents()
1015
def _split_root_dirblock_into_contents(self):
1016
"""Split the root dirblocks into root and contents-of-root.
1018
After parsing by path, we end up with root entries and contents-of-root
1019
entries in the same block. This loop splits them out again.
1021
# The above loop leaves the "root block" entries mixed with the
1022
# "contents-of-root block". But we don't want an if check on
1023
# all entries, so instead we just fix it up here.
1024
if self._dirblocks[1] != ('', []):
1025
raise ValueError("bad dirblock start %r" % (self._dirblocks[1],))
1027
contents_of_root_block = []
1028
for entry in self._dirblocks[0][1]:
1029
if not entry[0][1]: # This is a root entry
1030
root_block.append(entry)
1032
contents_of_root_block.append(entry)
1033
self._dirblocks[0] = ('', root_block)
1034
self._dirblocks[1] = ('', contents_of_root_block)
1036
def _entry_to_line(self, entry):
1037
"""Serialize entry to a NULL delimited line ready for _get_output_lines.
1039
:param entry: An entry_tuple as defined in the module docstring.
1041
entire_entry = list(entry[0])
1042
for tree_number, tree_data in enumerate(entry[1]):
1043
# (minikind, fingerprint, size, executable, tree_specific_string)
1044
entire_entry.extend(tree_data)
1045
# 3 for the key, 5 for the fields per tree.
1046
tree_offset = 3 + tree_number * 5
1048
entire_entry[tree_offset + 0] = tree_data[0]
1050
entire_entry[tree_offset + 2] = str(tree_data[2])
1052
entire_entry[tree_offset + 3] = DirState._to_yesno[tree_data[3]]
1053
return '\0'.join(entire_entry)
1055
def _fields_per_entry(self):
1056
"""How many null separated fields should be in each entry row.
1058
Each line now has an extra '\n' field which is not used
1059
so we just skip over it
1061
3 fields for the key
1062
+ number of fields per tree_data (5) * tree count
1065
tree_count = 1 + self._num_present_parents()
1066
return 3 + 5 * tree_count + 1
1068
def _find_block(self, key, add_if_missing=False):
1069
"""Return the block that key should be present in.
1071
:param key: A dirstate entry key.
1072
:return: The block tuple.
1074
block_index, present = self._find_block_index_from_key(key)
1076
if not add_if_missing:
1077
# check to see if key is versioned itself - we might want to
1078
# add it anyway, because dirs with no entries dont get a
1079
# dirblock at parse time.
1080
# This is an uncommon branch to take: most dirs have children,
1081
# and most code works with versioned paths.
1082
parent_base, parent_name = osutils.split(key[0])
1083
if not self._get_block_entry_index(parent_base, parent_name, 0)[3]:
1084
# some parent path has not been added - its an error to add
1086
raise errors.NotVersionedError(key[0:2], str(self))
1087
self._dirblocks.insert(block_index, (key[0], []))
1088
return self._dirblocks[block_index]
1090
def _find_block_index_from_key(self, key):
1091
"""Find the dirblock index for a key.
1093
:return: The block index, True if the block for the key is present.
1095
if key[0:2] == ('', ''):
1098
if (self._last_block_index is not None and
1099
self._dirblocks[self._last_block_index][0] == key[0]):
1100
return self._last_block_index, True
1103
block_index = bisect_dirblock(self._dirblocks, key[0], 1,
1104
cache=self._split_path_cache)
1105
# _right returns one-past-where-key is so we have to subtract
1106
# one to use it. we use _right here because there are two
1107
# '' blocks - the root, and the contents of root
1108
# we always have a minimum of 2 in self._dirblocks: root and
1109
# root-contents, and for '', we get 2 back, so this is
1110
# simple and correct:
1111
present = (block_index < len(self._dirblocks) and
1112
self._dirblocks[block_index][0] == key[0])
1113
self._last_block_index = block_index
1114
# Reset the entry index cache to the beginning of the block.
1115
self._last_entry_index = -1
1116
return block_index, present
1118
def _find_entry_index(self, key, block):
1119
"""Find the entry index for a key in a block.
1121
:return: The entry index, True if the entry for the key is present.
1123
len_block = len(block)
1125
if self._last_entry_index is not None:
1127
entry_index = self._last_entry_index + 1
1128
# A hit is when the key is after the last slot, and before or
1129
# equal to the next slot.
1130
if ((entry_index > 0 and block[entry_index - 1][0] < key) and
1131
key <= block[entry_index][0]):
1132
self._last_entry_index = entry_index
1133
present = (block[entry_index][0] == key)
1134
return entry_index, present
1137
entry_index = bisect.bisect_left(block, (key, []))
1138
present = (entry_index < len_block and
1139
block[entry_index][0] == key)
1140
self._last_entry_index = entry_index
1141
return entry_index, present
1144
def from_tree(tree, dir_state_filename):
1145
"""Create a dirstate from a bzr Tree.
1147
:param tree: The tree which should provide parent information and
1149
:return: a DirState object which is currently locked for writing.
1150
(it was locked by DirState.initialize)
1152
result = DirState.initialize(dir_state_filename)
1156
parent_ids = tree.get_parent_ids()
1157
num_parents = len(parent_ids)
1159
for parent_id in parent_ids:
1160
parent_tree = tree.branch.repository.revision_tree(parent_id)
1161
parent_trees.append((parent_id, parent_tree))
1162
parent_tree.lock_read()
1163
result.set_parent_trees(parent_trees, [])
1164
result.set_state_from_inventory(tree.inventory)
1166
for revid, parent_tree in parent_trees:
1167
parent_tree.unlock()
1170
# The caller won't have a chance to unlock this, so make sure we
1176
def update_by_delta(self, delta):
1177
"""Apply an inventory delta to the dirstate for tree 0
1179
:param delta: An inventory delta. See Inventory.apply_delta for
1182
self._read_dirblocks_if_needed()
1185
for old_path, new_path, file_id, inv_entry in sorted(delta, reverse=True):
1186
if (file_id in insertions) or (file_id in removals):
1187
raise AssertionError("repeated file id in delta %r" % (file_id,))
1188
if old_path is not None:
1189
old_path = old_path.encode('utf-8')
1190
removals[file_id] = old_path
1191
if new_path is not None:
1192
new_path = new_path.encode('utf-8')
1193
dirname, basename = osutils.split(new_path)
1194
key = (dirname, basename, file_id)
1195
minikind = DirState._kind_to_minikind[inv_entry.kind]
1197
fingerprint = inv_entry.reference_revision
1200
insertions[file_id] = (key, minikind, inv_entry.executable,
1201
fingerprint, new_path)
1202
# Transform moves into delete+add pairs
1203
if None not in (old_path, new_path):
1204
for child in self._iter_child_entries(0, old_path):
1205
if child[0][2] in insertions or child[0][2] in removals:
1207
child_dirname = child[0][0]
1208
child_basename = child[0][1]
1209
minikind = child[1][0][0]
1210
fingerprint = child[1][0][4]
1211
executable = child[1][0][3]
1212
old_child_path = osutils.pathjoin(child[0][0],
1214
removals[child[0][2]] = old_child_path
1215
child_suffix = child_dirname[len(old_path):]
1216
new_child_dirname = (new_path + child_suffix)
1217
key = (new_child_dirname, child_basename, child[0][2])
1218
new_child_path = os.path.join(new_child_dirname,
1220
insertions[child[0][2]] = (key, minikind, executable,
1221
fingerprint, new_child_path)
1222
self._apply_removals(removals.values())
1223
self._apply_insertions(insertions.values())
1225
def _apply_removals(self, removals):
1226
for path in sorted(removals, reverse=True):
1227
dirname, basename = osutils.split(path)
1228
block_i, entry_i, d_present, f_present = \
1229
self._get_block_entry_index(dirname, basename, 0)
1230
entry = self._dirblocks[block_i][1][entry_i]
1231
self._make_absent(entry)
1232
# See if we have a malformed delta: deleting a directory must not
1233
# leave crud behind. This increases the number of bisects needed
1234
# substantially, but deletion or renames of large numbers of paths
1235
# is rare enough it shouldn't be an issue (famous last words?) RBC
1237
block_i, entry_i, d_present, f_present = \
1238
self._get_block_entry_index(path, '', 0)
1240
# The dir block is still present in the dirstate; this could
1241
# be due to it being in a parent tree, or a corrupt delta.
1242
for child_entry in self._dirblocks[block_i][1]:
1243
if child_entry[1][0][0] not in ('r', 'a'):
1244
raise errors.InconsistentDelta(path, entry[0][2],
1245
"The file id was deleted but its children were "
1248
def _apply_insertions(self, adds):
1249
for key, minikind, executable, fingerprint, path_utf8 in sorted(adds):
1250
self.update_minimal(key, minikind, executable, fingerprint,
1251
path_utf8=path_utf8)
1253
def update_basis_by_delta(self, delta, new_revid):
1254
"""Update the parents of this tree after a commit.
1256
This gives the tree one parent, with revision id new_revid. The
1257
inventory delta is applied to the current basis tree to generate the
1258
inventory for the parent new_revid, and all other parent trees are
1261
Note that an exception during the operation of this method will leave
1262
the dirstate in a corrupt state where it should not be saved.
1264
Finally, we expect all changes to be synchronising the basis tree with
1267
:param new_revid: The new revision id for the trees parent.
1268
:param delta: An inventory delta (see apply_inventory_delta) describing
1269
the changes from the current left most parent revision to new_revid.
1271
self._read_dirblocks_if_needed()
1272
self._discard_merge_parents()
1273
if self._ghosts != []:
1274
raise NotImplementedError(self.update_basis_by_delta)
1275
if len(self._parents) == 0:
1276
# setup a blank tree, the most simple way.
1277
empty_parent = DirState.NULL_PARENT_DETAILS
1278
for entry in self._iter_entries():
1279
entry[1].append(empty_parent)
1280
self._parents.append(new_revid)
1282
self._parents[0] = new_revid
1284
delta = sorted(delta, reverse=True)
1288
# The paths this function accepts are unicode and must be encoded as we
1290
encode = cache_utf8.encode
1291
inv_to_entry = self._inv_entry_to_details
1292
# delta is now (deletes, changes), (adds) in reverse lexographical
1294
# deletes in reverse lexographic order are safe to process in situ.
1295
# renames are not, as a rename from any path could go to a path
1296
# lexographically lower, so we transform renames into delete, add pairs,
1297
# expanding them recursively as needed.
1298
# At the same time, to reduce interface friction we convert the input
1299
# inventory entries to dirstate.
1300
root_only = ('', '')
1301
for old_path, new_path, file_id, inv_entry in delta:
1302
if old_path is None:
1303
adds.append((None, encode(new_path), file_id,
1304
inv_to_entry(inv_entry), True))
1305
elif new_path is None:
1306
deletes.append((encode(old_path), None, file_id, None, True))
1307
elif (old_path, new_path) != root_only:
1309
# Because renames must preserve their children we must have
1310
# processed all relocations and removes before hand. The sort
1311
# order ensures we've examined the child paths, but we also
1312
# have to execute the removals, or the split to an add/delete
1313
# pair will result in the deleted item being reinserted, or
1314
# renamed items being reinserted twice - and possibly at the
1315
# wrong place. Splitting into a delete/add pair also simplifies
1316
# the handling of entries with ('f', ...), ('r' ...) because
1317
# the target of the 'r' is old_path here, and we add that to
1318
# deletes, meaning that the add handler does not need to check
1319
# for 'r' items on every pass.
1320
self._update_basis_apply_deletes(deletes)
1322
new_path_utf8 = encode(new_path)
1323
# Split into an add/delete pair recursively.
1324
adds.append((None, new_path_utf8, file_id,
1325
inv_to_entry(inv_entry), False))
1326
# Expunge deletes that we've seen so that deleted/renamed
1327
# children of a rename directory are handled correctly.
1328
new_deletes = reversed(list(self._iter_child_entries(1,
1330
# Remove the current contents of the tree at orig_path, and
1331
# reinsert at the correct new path.
1332
for entry in new_deletes:
1334
source_path = entry[0][0] + '/' + entry[0][1]
1336
source_path = entry[0][1]
1338
target_path = new_path_utf8 + source_path[len(old_path):]
1341
raise AssertionError("cannot rename directory to"
1343
target_path = source_path[len(old_path) + 1:]
1344
adds.append((None, target_path, entry[0][2], entry[1][1], False))
1346
(source_path, target_path, entry[0][2], None, False))
1348
(encode(old_path), new_path, file_id, None, False))
1350
# changes to just the root should not require remove/insertion
1352
changes.append((encode(old_path), encode(new_path), file_id,
1353
inv_to_entry(inv_entry)))
1355
# Finish expunging deletes/first half of renames.
1356
self._update_basis_apply_deletes(deletes)
1357
# Reinstate second half of renames and new paths.
1358
self._update_basis_apply_adds(adds)
1359
# Apply in-situ changes.
1360
self._update_basis_apply_changes(changes)
1362
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1363
self._header_state = DirState.IN_MEMORY_MODIFIED
1364
self._id_index = None
1367
def _update_basis_apply_adds(self, adds):
1368
"""Apply a sequence of adds to tree 1 during update_basis_by_delta.
1370
They may be adds, or renames that have been split into add/delete
1373
:param adds: A sequence of adds. Each add is a tuple:
1374
(None, new_path_utf8, file_id, (entry_details), real_add). real_add
1375
is False when the add is the second half of a remove-and-reinsert
1376
pair created to handle renames and deletes.
1378
# Adds are accumulated partly from renames, so can be in any input
1381
# adds is now in lexographic order, which places all parents before
1382
# their children, so we can process it linearly.
1384
for old_path, new_path, file_id, new_details, real_add in adds:
1385
# the entry for this file_id must be in tree 0.
1386
entry = self._get_entry(0, file_id, new_path)
1387
if entry[0] is None or entry[0][2] != file_id:
1388
self._changes_aborted = True
1389
raise errors.InconsistentDelta(new_path, file_id,
1390
'working tree does not contain new entry')
1391
if real_add and entry[1][1][0] not in absent:
1392
self._changes_aborted = True
1393
raise errors.InconsistentDelta(new_path, file_id,
1394
'The entry was considered to be a genuinely new record,'
1395
' but there was already an old record for it.')
1396
# We don't need to update the target of an 'r' because the handling
1397
# of renames turns all 'r' situations into a delete at the original
1399
entry[1][1] = new_details
1401
def _update_basis_apply_changes(self, changes):
1402
"""Apply a sequence of changes to tree 1 during update_basis_by_delta.
1404
:param adds: A sequence of changes. Each change is a tuple:
1405
(path_utf8, path_utf8, file_id, (entry_details))
1408
for old_path, new_path, file_id, new_details in changes:
1409
# the entry for this file_id must be in tree 0.
1410
entry = self._get_entry(0, file_id, new_path)
1411
if entry[0] is None or entry[0][2] != file_id:
1412
self._changes_aborted = True
1413
raise errors.InconsistentDelta(new_path, file_id,
1414
'working tree does not contain new entry')
1415
if (entry[1][0][0] in absent or
1416
entry[1][1][0] in absent):
1417
self._changes_aborted = True
1418
raise errors.InconsistentDelta(new_path, file_id,
1419
'changed considered absent')
1420
entry[1][1] = new_details
1422
def _update_basis_apply_deletes(self, deletes):
1423
"""Apply a sequence of deletes to tree 1 during update_basis_by_delta.
1425
They may be deletes, or renames that have been split into add/delete
1428
:param deletes: A sequence of deletes. Each delete is a tuple:
1429
(old_path_utf8, new_path_utf8, file_id, None, real_delete).
1430
real_delete is True when the desired outcome is an actual deletion
1431
rather than the rename handling logic temporarily deleting a path
1432
during the replacement of a parent.
1434
null = DirState.NULL_PARENT_DETAILS
1435
for old_path, new_path, file_id, _, real_delete in deletes:
1436
if real_delete != (new_path is None):
1437
raise AssertionError("bad delete delta")
1438
# the entry for this file_id must be in tree 1.
1439
dirname, basename = osutils.split(old_path)
1440
block_index, entry_index, dir_present, file_present = \
1441
self._get_block_entry_index(dirname, basename, 1)
1442
if not file_present:
1443
self._changes_aborted = True
1444
raise errors.InconsistentDelta(old_path, file_id,
1445
'basis tree does not contain removed entry')
1446
entry = self._dirblocks[block_index][1][entry_index]
1447
if entry[0][2] != file_id:
1448
self._changes_aborted = True
1449
raise errors.InconsistentDelta(old_path, file_id,
1450
'mismatched file_id in tree 1')
1452
if entry[1][0][0] != 'a':
1453
self._changes_aborted = True
1454
raise errors.InconsistentDelta(old_path, file_id,
1455
'This was marked as a real delete, but the WT state'
1456
' claims that it still exists and is versioned.')
1457
del self._dirblocks[block_index][1][entry_index]
1459
if entry[1][0][0] == 'a':
1460
self._changes_aborted = True
1461
raise errors.InconsistentDelta(old_path, file_id,
1462
'The entry was considered a rename, but the source path'
1463
' is marked as absent.')
1464
# For whatever reason, we were asked to rename an entry
1465
# that was originally marked as deleted. This could be
1466
# because we are renaming the parent directory, and the WT
1467
# current state has the file marked as deleted.
1468
elif entry[1][0][0] == 'r':
1469
# implement the rename
1470
del self._dirblocks[block_index][1][entry_index]
1472
# it is being resurrected here, so blank it out temporarily.
1473
self._dirblocks[block_index][1][entry_index][1][1] = null
1475
def _observed_sha1(self, entry, sha1, stat_value,
1476
_stat_to_minikind=_stat_to_minikind, _pack_stat=pack_stat):
1477
"""Note the sha1 of a file.
1479
:param entry: The entry the sha1 is for.
1480
:param sha1: The observed sha1.
1481
:param stat_value: The os.lstat for the file.
1484
minikind = _stat_to_minikind[stat_value.st_mode & 0170000]
1488
packed_stat = _pack_stat(stat_value)
1490
if self._cutoff_time is None:
1491
self._sha_cutoff_time()
1492
if (stat_value.st_mtime < self._cutoff_time
1493
and stat_value.st_ctime < self._cutoff_time):
1494
entry[1][0] = ('f', sha1, entry[1][0][2], entry[1][0][3],
1496
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1498
def update_entry(self, entry, abspath, stat_value,
1499
_stat_to_minikind=_stat_to_minikind,
1500
_pack_stat=pack_stat):
1501
"""Update the entry based on what is actually on disk.
1503
:param entry: This is the dirblock entry for the file in question.
1504
:param abspath: The path on disk for this file.
1505
:param stat_value: (optional) if we already have done a stat on the
1507
:return: The sha1 hexdigest of the file (40 bytes) or link target of a
1511
minikind = _stat_to_minikind[stat_value.st_mode & 0170000]
1515
packed_stat = _pack_stat(stat_value)
1516
(saved_minikind, saved_link_or_sha1, saved_file_size,
1517
saved_executable, saved_packed_stat) = entry[1][0]
1519
if (minikind == saved_minikind
1520
and packed_stat == saved_packed_stat):
1521
# The stat hasn't changed since we saved, so we can re-use the
1526
# size should also be in packed_stat
1527
if saved_file_size == stat_value.st_size:
1528
return saved_link_or_sha1
1530
# If we have gotten this far, that means that we need to actually
1531
# process this entry.
1534
link_or_sha1 = self._sha1_file(abspath)
1535
executable = self._is_executable(stat_value.st_mode,
1537
if self._cutoff_time is None:
1538
self._sha_cutoff_time()
1539
if (stat_value.st_mtime < self._cutoff_time
1540
and stat_value.st_ctime < self._cutoff_time):
1541
entry[1][0] = ('f', link_or_sha1, stat_value.st_size,
1542
executable, packed_stat)
1544
entry[1][0] = ('f', '', stat_value.st_size,
1545
executable, DirState.NULLSTAT)
1546
elif minikind == 'd':
1548
entry[1][0] = ('d', '', 0, False, packed_stat)
1549
if saved_minikind != 'd':
1550
# This changed from something into a directory. Make sure we
1551
# have a directory block for it. This doesn't happen very
1552
# often, so this doesn't have to be super fast.
1553
block_index, entry_index, dir_present, file_present = \
1554
self._get_block_entry_index(entry[0][0], entry[0][1], 0)
1555
self._ensure_block(block_index, entry_index,
1556
osutils.pathjoin(entry[0][0], entry[0][1]))
1557
elif minikind == 'l':
1558
link_or_sha1 = self._read_link(abspath, saved_link_or_sha1)
1559
if self._cutoff_time is None:
1560
self._sha_cutoff_time()
1561
if (stat_value.st_mtime < self._cutoff_time
1562
and stat_value.st_ctime < self._cutoff_time):
1563
entry[1][0] = ('l', link_or_sha1, stat_value.st_size,
1566
entry[1][0] = ('l', '', stat_value.st_size,
1567
False, DirState.NULLSTAT)
1568
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1571
def _sha_cutoff_time(self):
1572
"""Return cutoff time.
1574
Files modified more recently than this time are at risk of being
1575
undetectably modified and so can't be cached.
1577
# Cache the cutoff time as long as we hold a lock.
1578
# time.time() isn't super expensive (approx 3.38us), but
1579
# when you call it 50,000 times it adds up.
1580
# For comparison, os.lstat() costs 7.2us if it is hot.
1581
self._cutoff_time = int(time.time()) - 3
1582
return self._cutoff_time
1584
def _lstat(self, abspath, entry):
1585
"""Return the os.lstat value for this path."""
1586
return os.lstat(abspath)
1588
def _sha1_file_and_mutter(self, abspath):
1589
# when -Dhashcache is turned on, this is monkey-patched in to log
1591
trace.mutter("dirstate sha1 " + abspath)
1592
return osutils.sha_file_by_name(abspath)
1594
def _is_executable(self, mode, old_executable):
1595
"""Is this file executable?"""
1596
return bool(S_IEXEC & mode)
1598
def _is_executable_win32(self, mode, old_executable):
1599
"""On win32 the executable bit is stored in the dirstate."""
1600
return old_executable
1602
if sys.platform == 'win32':
1603
_is_executable = _is_executable_win32
1605
def _read_link(self, abspath, old_link):
1606
"""Read the target of a symlink"""
1607
# TODO: jam 200700301 On Win32, this could just return the value
1608
# already in memory. However, this really needs to be done at a
1609
# higher level, because there either won't be anything on disk,
1610
# or the thing on disk will be a file.
1611
return os.readlink(abspath)
1613
def get_ghosts(self):
1614
"""Return a list of the parent tree revision ids that are ghosts."""
1615
self._read_header_if_needed()
1618
def get_lines(self):
1619
"""Serialise the entire dirstate to a sequence of lines."""
1620
if (self._header_state == DirState.IN_MEMORY_UNMODIFIED and
1621
self._dirblock_state == DirState.IN_MEMORY_UNMODIFIED):
1622
# read whats on disk.
1623
self._state_file.seek(0)
1624
return self._state_file.readlines()
1626
lines.append(self._get_parents_line(self.get_parent_ids()))
1627
lines.append(self._get_ghosts_line(self._ghosts))
1628
# append the root line which is special cased
1629
lines.extend(map(self._entry_to_line, self._iter_entries()))
1630
return self._get_output_lines(lines)
1632
def _get_ghosts_line(self, ghost_ids):
1633
"""Create a line for the state file for ghost information."""
1634
return '\0'.join([str(len(ghost_ids))] + ghost_ids)
1636
def _get_parents_line(self, parent_ids):
1637
"""Create a line for the state file for parents information."""
1638
return '\0'.join([str(len(parent_ids))] + parent_ids)
1640
def _get_fields_to_entry(self):
1641
"""Get a function which converts entry fields into a entry record.
1643
This handles size and executable, as well as parent records.
1645
:return: A function which takes a list of fields, and returns an
1646
appropriate record for storing in memory.
1648
# This is intentionally unrolled for performance
1649
num_present_parents = self._num_present_parents()
1650
if num_present_parents == 0:
1651
def fields_to_entry_0_parents(fields, _int=int):
1652
path_name_file_id_key = (fields[0], fields[1], fields[2])
1653
return (path_name_file_id_key, [
1655
fields[3], # minikind
1656
fields[4], # fingerprint
1657
_int(fields[5]), # size
1658
fields[6] == 'y', # executable
1659
fields[7], # packed_stat or revision_id
1661
return fields_to_entry_0_parents
1662
elif num_present_parents == 1:
1663
def fields_to_entry_1_parent(fields, _int=int):
1664
path_name_file_id_key = (fields[0], fields[1], fields[2])
1665
return (path_name_file_id_key, [
1667
fields[3], # minikind
1668
fields[4], # fingerprint
1669
_int(fields[5]), # size
1670
fields[6] == 'y', # executable
1671
fields[7], # packed_stat or revision_id
1674
fields[8], # minikind
1675
fields[9], # fingerprint
1676
_int(fields[10]), # size
1677
fields[11] == 'y', # executable
1678
fields[12], # packed_stat or revision_id
1681
return fields_to_entry_1_parent
1682
elif num_present_parents == 2:
1683
def fields_to_entry_2_parents(fields, _int=int):
1684
path_name_file_id_key = (fields[0], fields[1], fields[2])
1685
return (path_name_file_id_key, [
1687
fields[3], # minikind
1688
fields[4], # fingerprint
1689
_int(fields[5]), # size
1690
fields[6] == 'y', # executable
1691
fields[7], # packed_stat or revision_id
1694
fields[8], # minikind
1695
fields[9], # fingerprint
1696
_int(fields[10]), # size
1697
fields[11] == 'y', # executable
1698
fields[12], # packed_stat or revision_id
1701
fields[13], # minikind
1702
fields[14], # fingerprint
1703
_int(fields[15]), # size
1704
fields[16] == 'y', # executable
1705
fields[17], # packed_stat or revision_id
1708
return fields_to_entry_2_parents
1710
def fields_to_entry_n_parents(fields, _int=int):
1711
path_name_file_id_key = (fields[0], fields[1], fields[2])
1712
trees = [(fields[cur], # minikind
1713
fields[cur+1], # fingerprint
1714
_int(fields[cur+2]), # size
1715
fields[cur+3] == 'y', # executable
1716
fields[cur+4], # stat or revision_id
1717
) for cur in xrange(3, len(fields)-1, 5)]
1718
return path_name_file_id_key, trees
1719
return fields_to_entry_n_parents
1721
def get_parent_ids(self):
1722
"""Return a list of the parent tree ids for the directory state."""
1723
self._read_header_if_needed()
1724
return list(self._parents)
1726
def _get_block_entry_index(self, dirname, basename, tree_index):
1727
"""Get the coordinates for a path in the state structure.
1729
:param dirname: The utf8 dirname to lookup.
1730
:param basename: The utf8 basename to lookup.
1731
:param tree_index: The index of the tree for which this lookup should
1733
:return: A tuple describing where the path is located, or should be
1734
inserted. The tuple contains four fields: the block index, the row
1735
index, the directory is present (boolean), the entire path is
1736
present (boolean). There is no guarantee that either
1737
coordinate is currently reachable unless the found field for it is
1738
True. For instance, a directory not present in the searched tree
1739
may be returned with a value one greater than the current highest
1740
block offset. The directory present field will always be True when
1741
the path present field is True. The directory present field does
1742
NOT indicate that the directory is present in the searched tree,
1743
rather it indicates that there are at least some files in some
1746
self._read_dirblocks_if_needed()
1747
key = dirname, basename, ''
1748
block_index, present = self._find_block_index_from_key(key)
1750
# no such directory - return the dir index and 0 for the row.
1751
return block_index, 0, False, False
1752
block = self._dirblocks[block_index][1] # access the entries only
1753
entry_index, present = self._find_entry_index(key, block)
1754
# linear search through entries at this path to find the one
1756
while entry_index < len(block) and block[entry_index][0][1] == basename:
1757
if block[entry_index][1][tree_index][0] not in 'ar':
1758
# neither absent or relocated
1759
return block_index, entry_index, True, True
1761
return block_index, entry_index, True, False
1763
def _get_entry(self, tree_index, fileid_utf8=None, path_utf8=None):
1764
"""Get the dirstate entry for path in tree tree_index.
1766
If either file_id or path is supplied, it is used as the key to lookup.
1767
If both are supplied, the fastest lookup is used, and an error is
1768
raised if they do not both point at the same row.
1770
:param tree_index: The index of the tree we wish to locate this path
1771
in. If the path is present in that tree, the entry containing its
1772
details is returned, otherwise (None, None) is returned
1773
0 is the working tree, higher indexes are successive parent
1775
:param fileid_utf8: A utf8 file_id to look up.
1776
:param path_utf8: An utf8 path to be looked up.
1777
:return: The dirstate entry tuple for path, or (None, None)
1779
self._read_dirblocks_if_needed()
1780
if path_utf8 is not None:
1781
if type(path_utf8) is not str:
1782
raise AssertionError('path_utf8 is not a str: %s %s'
1783
% (type(path_utf8), path_utf8))
1784
# path lookups are faster
1785
dirname, basename = osutils.split(path_utf8)
1786
block_index, entry_index, dir_present, file_present = \
1787
self._get_block_entry_index(dirname, basename, tree_index)
1788
if not file_present:
1790
entry = self._dirblocks[block_index][1][entry_index]
1791
if not (entry[0][2] and entry[1][tree_index][0] not in ('a', 'r')):
1792
raise AssertionError('unversioned entry?')
1794
if entry[0][2] != fileid_utf8:
1795
self._changes_aborted = True
1796
raise errors.BzrError('integrity error ? : mismatching'
1797
' tree_index, file_id and path')
1800
possible_keys = self._get_id_index().get(fileid_utf8, None)
1801
if not possible_keys:
1803
for key in possible_keys:
1804
block_index, present = \
1805
self._find_block_index_from_key(key)
1806
# strange, probably indicates an out of date
1807
# id index - for now, allow this.
1810
# WARNING: DO not change this code to use _get_block_entry_index
1811
# as that function is not suitable: it does not use the key
1812
# to lookup, and thus the wrong coordinates are returned.
1813
block = self._dirblocks[block_index][1]
1814
entry_index, present = self._find_entry_index(key, block)
1816
entry = self._dirblocks[block_index][1][entry_index]
1817
if entry[1][tree_index][0] in 'fdlt':
1818
# this is the result we are looking for: the
1819
# real home of this file_id in this tree.
1821
if entry[1][tree_index][0] == 'a':
1822
# there is no home for this entry in this tree
1824
if entry[1][tree_index][0] != 'r':
1825
raise AssertionError(
1826
"entry %r has invalid minikind %r for tree %r" \
1828
entry[1][tree_index][0],
1830
real_path = entry[1][tree_index][1]
1831
return self._get_entry(tree_index, fileid_utf8=fileid_utf8,
1832
path_utf8=real_path)
1836
def initialize(cls, path):
1837
"""Create a new dirstate on path.
1839
The new dirstate will be an empty tree - that is it has no parents,
1840
and only a root node - which has id ROOT_ID.
1842
:param path: The name of the file for the dirstate.
1843
:return: A write-locked DirState object.
1845
# This constructs a new DirState object on a path, sets the _state_file
1846
# to a new empty file for that path. It then calls _set_data() with our
1847
# stock empty dirstate information - a root with ROOT_ID, no children,
1848
# and no parents. Finally it calls save() to ensure that this data will
1851
# root dir and root dir contents with no children.
1852
empty_tree_dirblocks = [('', []), ('', [])]
1853
# a new root directory, with a NULLSTAT.
1854
empty_tree_dirblocks[0][1].append(
1855
(('', '', inventory.ROOT_ID), [
1856
('d', '', 0, False, DirState.NULLSTAT),
1860
result._set_data([], empty_tree_dirblocks)
1868
def _inv_entry_to_details(inv_entry):
1869
"""Convert an inventory entry (from a revision tree) to state details.
1871
:param inv_entry: An inventory entry whose sha1 and link targets can be
1872
relied upon, and which has a revision set.
1873
:return: A details tuple - the details for a single tree at a path +
1876
kind = inv_entry.kind
1877
minikind = DirState._kind_to_minikind[kind]
1878
tree_data = inv_entry.revision
1879
if kind == 'directory':
1883
elif kind == 'symlink':
1884
# We don't support non-ascii targets for symlinks yet.
1885
fingerprint = str(inv_entry.symlink_target or '')
1888
elif kind == 'file':
1889
fingerprint = inv_entry.text_sha1 or ''
1890
size = inv_entry.text_size or 0
1891
executable = inv_entry.executable
1892
elif kind == 'tree-reference':
1893
fingerprint = inv_entry.reference_revision or ''
1897
raise Exception("can't pack %s" % inv_entry)
1898
return (minikind, fingerprint, size, executable, tree_data)
1900
def _iter_child_entries(self, tree_index, path_utf8):
1901
"""Iterate over all the entries that are children of path_utf.
1903
This only returns entries that are present (not in 'a', 'r') in
1904
tree_index. tree_index data is not refreshed, so if tree 0 is used,
1905
results may differ from that obtained if paths were statted to
1906
determine what ones were directories.
1908
Asking for the children of a non-directory will return an empty
1912
next_pending_dirs = [path_utf8]
1914
while next_pending_dirs:
1915
pending_dirs = next_pending_dirs
1916
next_pending_dirs = []
1917
for path in pending_dirs:
1918
block_index, present = self._find_block_index_from_key(
1920
if block_index == 0:
1922
if len(self._dirblocks) == 1:
1923
# asked for the children of the root with no other
1927
# children of a non-directory asked for.
1929
block = self._dirblocks[block_index]
1930
for entry in block[1]:
1931
kind = entry[1][tree_index][0]
1932
if kind not in absent:
1936
path = entry[0][0] + '/' + entry[0][1]
1939
next_pending_dirs.append(path)
1941
def _iter_entries(self):
1942
"""Iterate over all the entries in the dirstate.
1944
Each yelt item is an entry in the standard format described in the
1945
docstring of bzrlib.dirstate.
1947
self._read_dirblocks_if_needed()
1948
for directory in self._dirblocks:
1949
for entry in directory[1]:
1952
def _get_id_index(self):
1953
"""Get an id index of self._dirblocks."""
1954
if self._id_index is None:
1956
for key, tree_details in self._iter_entries():
1957
id_index.setdefault(key[2], set()).add(key)
1958
self._id_index = id_index
1959
return self._id_index
1961
def _get_output_lines(self, lines):
1962
"""Format lines for final output.
1964
:param lines: A sequence of lines containing the parents list and the
1967
output_lines = [DirState.HEADER_FORMAT_3]
1968
lines.append('') # a final newline
1969
inventory_text = '\0\n\0'.join(lines)
1970
output_lines.append('crc32: %s\n' % (zlib.crc32(inventory_text),))
1971
# -3, 1 for num parents, 1 for ghosts, 1 for final newline
1972
num_entries = len(lines)-3
1973
output_lines.append('num_entries: %s\n' % (num_entries,))
1974
output_lines.append(inventory_text)
1977
def _make_deleted_row(self, fileid_utf8, parents):
1978
"""Return a deleted row for fileid_utf8."""
1979
return ('/', 'RECYCLED.BIN', 'file', fileid_utf8, 0, DirState.NULLSTAT,
1982
def _num_present_parents(self):
1983
"""The number of parent entries in each record row."""
1984
return len(self._parents) - len(self._ghosts)
1988
"""Construct a DirState on the file at path path.
1990
:return: An unlocked DirState object, associated with the given path.
1992
result = DirState(path)
1995
def _read_dirblocks_if_needed(self):
1996
"""Read in all the dirblocks from the file if they are not in memory.
1998
This populates self._dirblocks, and sets self._dirblock_state to
1999
IN_MEMORY_UNMODIFIED. It is not currently ready for incremental block
2002
self._read_header_if_needed()
2003
if self._dirblock_state == DirState.NOT_IN_MEMORY:
2004
_read_dirblocks(self)
2006
def _read_header(self):
2007
"""This reads in the metadata header, and the parent ids.
2009
After reading in, the file should be positioned at the null
2010
just before the start of the first record in the file.
2012
:return: (expected crc checksum, number of entries, parent list)
2014
self._read_prelude()
2015
parent_line = self._state_file.readline()
2016
info = parent_line.split('\0')
2017
num_parents = int(info[0])
2018
self._parents = info[1:-1]
2019
ghost_line = self._state_file.readline()
2020
info = ghost_line.split('\0')
2021
num_ghosts = int(info[1])
2022
self._ghosts = info[2:-1]
2023
self._header_state = DirState.IN_MEMORY_UNMODIFIED
2024
self._end_of_header = self._state_file.tell()
2026
def _read_header_if_needed(self):
2027
"""Read the header of the dirstate file if needed."""
2028
# inline this as it will be called a lot
2029
if not self._lock_token:
2030
raise errors.ObjectNotLocked(self)
2031
if self._header_state == DirState.NOT_IN_MEMORY:
2034
def _read_prelude(self):
2035
"""Read in the prelude header of the dirstate file.
2037
This only reads in the stuff that is not connected to the crc
2038
checksum. The position will be correct to read in the rest of
2039
the file and check the checksum after this point.
2040
The next entry in the file should be the number of parents,
2041
and their ids. Followed by a newline.
2043
header = self._state_file.readline()
2044
if header != DirState.HEADER_FORMAT_3:
2045
raise errors.BzrError(
2046
'invalid header line: %r' % (header,))
2047
crc_line = self._state_file.readline()
2048
if not crc_line.startswith('crc32: '):
2049
raise errors.BzrError('missing crc32 checksum: %r' % crc_line)
2050
self.crc_expected = int(crc_line[len('crc32: '):-1])
2051
num_entries_line = self._state_file.readline()
2052
if not num_entries_line.startswith('num_entries: '):
2053
raise errors.BzrError('missing num_entries line')
2054
self._num_entries = int(num_entries_line[len('num_entries: '):-1])
2056
def sha1_from_stat(self, path, stat_result, _pack_stat=pack_stat):
2057
"""Find a sha1 given a stat lookup."""
2058
return self._get_packed_stat_index().get(_pack_stat(stat_result), None)
2060
def _get_packed_stat_index(self):
2061
"""Get a packed_stat index of self._dirblocks."""
2062
if self._packed_stat_index is None:
2064
for key, tree_details in self._iter_entries():
2065
if tree_details[0][0] == 'f':
2066
index[tree_details[0][4]] = tree_details[0][1]
2067
self._packed_stat_index = index
2068
return self._packed_stat_index
2071
"""Save any pending changes created during this session.
2073
We reuse the existing file, because that prevents race conditions with
2074
file creation, and use oslocks on it to prevent concurrent modification
2075
and reads - because dirstate's incremental data aggregation is not
2076
compatible with reading a modified file, and replacing a file in use by
2077
another process is impossible on Windows.
2079
A dirstate in read only mode should be smart enough though to validate
2080
that the file has not changed, and otherwise discard its cache and
2081
start over, to allow for fine grained read lock duration, so 'status'
2082
wont block 'commit' - for example.
2084
if self._changes_aborted:
2085
# Should this be a warning? For now, I'm expecting that places that
2086
# mark it inconsistent will warn, making a warning here redundant.
2087
trace.mutter('Not saving DirState because '
2088
'_changes_aborted is set.')
2090
if (self._header_state == DirState.IN_MEMORY_MODIFIED or
2091
self._dirblock_state == DirState.IN_MEMORY_MODIFIED):
2093
grabbed_write_lock = False
2094
if self._lock_state != 'w':
2095
grabbed_write_lock, new_lock = self._lock_token.temporary_write_lock()
2096
# Switch over to the new lock, as the old one may be closed.
2097
# TODO: jam 20070315 We should validate the disk file has
2098
# not changed contents. Since temporary_write_lock may
2099
# not be an atomic operation.
2100
self._lock_token = new_lock
2101
self._state_file = new_lock.f
2102
if not grabbed_write_lock:
2103
# We couldn't grab a write lock, so we switch back to a read one
2106
self._state_file.seek(0)
2107
self._state_file.writelines(self.get_lines())
2108
self._state_file.truncate()
2109
self._state_file.flush()
2110
self._header_state = DirState.IN_MEMORY_UNMODIFIED
2111
self._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
2113
if grabbed_write_lock:
2114
self._lock_token = self._lock_token.restore_read_lock()
2115
self._state_file = self._lock_token.f
2116
# TODO: jam 20070315 We should validate the disk file has
2117
# not changed contents. Since restore_read_lock may
2118
# not be an atomic operation.
2120
def _set_data(self, parent_ids, dirblocks):
2121
"""Set the full dirstate data in memory.
2123
This is an internal function used to completely replace the objects
2124
in memory state. It puts the dirstate into state 'full-dirty'.
2126
:param parent_ids: A list of parent tree revision ids.
2127
:param dirblocks: A list containing one tuple for each directory in the
2128
tree. Each tuple contains the directory path and a list of entries
2129
found in that directory.
2131
# our memory copy is now authoritative.
2132
self._dirblocks = dirblocks
2133
self._header_state = DirState.IN_MEMORY_MODIFIED
2134
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2135
self._parents = list(parent_ids)
2136
self._id_index = None
2137
self._packed_stat_index = None
2139
def set_path_id(self, path, new_id):
2140
"""Change the id of path to new_id in the current working tree.
2142
:param path: The path inside the tree to set - '' is the root, 'foo'
2143
is the path foo in the root.
2144
:param new_id: The new id to assign to the path. This must be a utf8
2145
file id (not unicode, and not None).
2147
self._read_dirblocks_if_needed()
2149
# TODO: logic not written
2150
raise NotImplementedError(self.set_path_id)
2151
# TODO: check new id is unique
2152
entry = self._get_entry(0, path_utf8=path)
2153
if entry[0][2] == new_id:
2154
# Nothing to change.
2156
# mark the old path absent, and insert a new root path
2157
self._make_absent(entry)
2158
self.update_minimal(('', '', new_id), 'd',
2159
path_utf8='', packed_stat=entry[1][0][4])
2160
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2161
if self._id_index is not None:
2162
self._id_index.setdefault(new_id, set()).add(entry[0])
2164
def set_parent_trees(self, trees, ghosts):
2165
"""Set the parent trees for the dirstate.
2167
:param trees: A list of revision_id, tree tuples. tree must be provided
2168
even if the revision_id refers to a ghost: supply an empty tree in
2170
:param ghosts: A list of the revision_ids that are ghosts at the time
2173
# TODO: generate a list of parent indexes to preserve to save
2174
# processing specific parent trees. In the common case one tree will
2175
# be preserved - the left most parent.
2176
# TODO: if the parent tree is a dirstate, we might want to walk them
2177
# all by path in parallel for 'optimal' common-case performance.
2178
# generate new root row.
2179
self._read_dirblocks_if_needed()
2180
# TODO future sketch: Examine the existing parents to generate a change
2181
# map and then walk the new parent trees only, mapping them into the
2182
# dirstate. Walk the dirstate at the same time to remove unreferenced
2185
# sketch: loop over all entries in the dirstate, cherry picking
2186
# entries from the parent trees, if they are not ghost trees.
2187
# after we finish walking the dirstate, all entries not in the dirstate
2188
# are deletes, so we want to append them to the end as per the design
2189
# discussions. So do a set difference on ids with the parents to
2190
# get deletes, and add them to the end.
2191
# During the update process we need to answer the following questions:
2192
# - find other keys containing a fileid in order to create cross-path
2193
# links. We dont't trivially use the inventory from other trees
2194
# because this leads to either double touching, or to accessing
2196
# - find other keys containing a path
2197
# We accumulate each entry via this dictionary, including the root
2200
# we could do parallel iterators, but because file id data may be
2201
# scattered throughout, we dont save on index overhead: we have to look
2202
# at everything anyway. We can probably save cycles by reusing parent
2203
# data and doing an incremental update when adding an additional
2204
# parent, but for now the common cases are adding a new parent (merge),
2205
# and replacing completely (commit), and commit is more common: so
2206
# optimise merge later.
2208
# ---- start generation of full tree mapping data
2209
# what trees should we use?
2210
parent_trees = [tree for rev_id, tree in trees if rev_id not in ghosts]
2211
# how many trees do we end up with
2212
parent_count = len(parent_trees)
2214
# one: the current tree
2215
for entry in self._iter_entries():
2216
# skip entries not in the current tree
2217
if entry[1][0][0] in 'ar': # absent, relocated
2219
by_path[entry[0]] = [entry[1][0]] + \
2220
[DirState.NULL_PARENT_DETAILS] * parent_count
2221
id_index[entry[0][2]] = set([entry[0]])
2223
# now the parent trees:
2224
for tree_index, tree in enumerate(parent_trees):
2225
# the index is off by one, adjust it.
2226
tree_index = tree_index + 1
2227
# when we add new locations for a fileid we need these ranges for
2228
# any fileid in this tree as we set the by_path[id] to:
2229
# already_processed_tree_details + new_details + new_location_suffix
2230
# the suffix is from tree_index+1:parent_count+1.
2231
new_location_suffix = [DirState.NULL_PARENT_DETAILS] * (parent_count - tree_index)
2232
# now stitch in all the entries from this tree
2233
for path, entry in tree.inventory.iter_entries_by_dir():
2234
# here we process each trees details for each item in the tree.
2235
# we first update any existing entries for the id at other paths,
2236
# then we either create or update the entry for the id at the
2237
# right path, and finally we add (if needed) a mapping from
2238
# file_id to this path. We do it in this order to allow us to
2239
# avoid checking all known paths for the id when generating a
2240
# new entry at this path: by adding the id->path mapping last,
2241
# all the mappings are valid and have correct relocation
2242
# records where needed.
2243
file_id = entry.file_id
2244
path_utf8 = path.encode('utf8')
2245
dirname, basename = osutils.split(path_utf8)
2246
new_entry_key = (dirname, basename, file_id)
2247
# tree index consistency: All other paths for this id in this tree
2248
# index must point to the correct path.
2249
for entry_key in id_index.setdefault(file_id, set()):
2250
# TODO:PROFILING: It might be faster to just update
2251
# rather than checking if we need to, and then overwrite
2252
# the one we are located at.
2253
if entry_key != new_entry_key:
2254
# this file id is at a different path in one of the
2255
# other trees, so put absent pointers there
2256
# This is the vertical axis in the matrix, all pointing
2258
by_path[entry_key][tree_index] = ('r', path_utf8, 0, False, '')
2259
# by path consistency: Insert into an existing path record (trivial), or
2260
# add a new one with relocation pointers for the other tree indexes.
2261
if new_entry_key in id_index[file_id]:
2262
# there is already an entry where this data belongs, just insert it.
2263
by_path[new_entry_key][tree_index] = \
2264
self._inv_entry_to_details(entry)
2266
# add relocated entries to the horizontal axis - this row
2267
# mapping from path,id. We need to look up the correct path
2268
# for the indexes from 0 to tree_index -1
2270
for lookup_index in xrange(tree_index):
2271
# boundary case: this is the first occurence of file_id
2272
# so there are no id_indexs, possibly take this out of
2274
if not len(id_index[file_id]):
2275
new_details.append(DirState.NULL_PARENT_DETAILS)
2277
# grab any one entry, use it to find the right path.
2278
# TODO: optimise this to reduce memory use in highly
2279
# fragmented situations by reusing the relocation
2281
a_key = iter(id_index[file_id]).next()
2282
if by_path[a_key][lookup_index][0] in ('r', 'a'):
2283
# its a pointer or missing statement, use it as is.
2284
new_details.append(by_path[a_key][lookup_index])
2286
# we have the right key, make a pointer to it.
2287
real_path = ('/'.join(a_key[0:2])).strip('/')
2288
new_details.append(('r', real_path, 0, False, ''))
2289
new_details.append(self._inv_entry_to_details(entry))
2290
new_details.extend(new_location_suffix)
2291
by_path[new_entry_key] = new_details
2292
id_index[file_id].add(new_entry_key)
2293
# --- end generation of full tree mappings
2295
# sort and output all the entries
2296
new_entries = self._sort_entries(by_path.items())
2297
self._entries_to_current_state(new_entries)
2298
self._parents = [rev_id for rev_id, tree in trees]
2299
self._ghosts = list(ghosts)
2300
self._header_state = DirState.IN_MEMORY_MODIFIED
2301
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2302
self._id_index = id_index
2304
def _sort_entries(self, entry_list):
2305
"""Given a list of entries, sort them into the right order.
2307
This is done when constructing a new dirstate from trees - normally we
2308
try to keep everything in sorted blocks all the time, but sometimes
2309
it's easier to sort after the fact.
2312
# sort by: directory parts, file name, file id
2313
return entry[0][0].split('/'), entry[0][1], entry[0][2]
2314
return sorted(entry_list, key=_key)
2316
def set_state_from_inventory(self, new_inv):
2317
"""Set new_inv as the current state.
2319
This API is called by tree transform, and will usually occur with
2320
existing parent trees.
2322
:param new_inv: The inventory object to set current state from.
2324
if 'evil' in debug.debug_flags:
2325
trace.mutter_callsite(1,
2326
"set_state_from_inventory called; please mutate the tree instead")
2327
self._read_dirblocks_if_needed()
2329
# Two iterators: current data and new data, both in dirblock order.
2330
# We zip them together, which tells about entries that are new in the
2331
# inventory, or removed in the inventory, or present in both and
2334
# You might think we could just synthesize a new dirstate directly
2335
# since we're processing it in the right order. However, we need to
2336
# also consider there may be any number of parent trees and relocation
2337
# pointers, and we don't want to duplicate that here.
2338
new_iterator = new_inv.iter_entries_by_dir()
2339
# we will be modifying the dirstate, so we need a stable iterator. In
2340
# future we might write one, for now we just clone the state into a
2341
# list - which is a shallow copy.
2342
old_iterator = iter(list(self._iter_entries()))
2343
# both must have roots so this is safe:
2344
current_new = new_iterator.next()
2345
current_old = old_iterator.next()
2346
def advance(iterator):
2348
return iterator.next()
2349
except StopIteration:
2351
while current_new or current_old:
2352
# skip entries in old that are not really there
2353
if current_old and current_old[1][0][0] in 'ar':
2354
# relocated or absent
2355
current_old = advance(old_iterator)
2358
# convert new into dirblock style
2359
new_path_utf8 = current_new[0].encode('utf8')
2360
new_dirname, new_basename = osutils.split(new_path_utf8)
2361
new_id = current_new[1].file_id
2362
new_entry_key = (new_dirname, new_basename, new_id)
2363
current_new_minikind = \
2364
DirState._kind_to_minikind[current_new[1].kind]
2365
if current_new_minikind == 't':
2366
fingerprint = current_new[1].reference_revision or ''
2368
# We normally only insert or remove records, or update
2369
# them when it has significantly changed. Then we want to
2370
# erase its fingerprint. Unaffected records should
2371
# normally not be updated at all.
2374
# for safety disable variables
2375
new_path_utf8 = new_dirname = new_basename = new_id = \
2376
new_entry_key = None
2377
# 5 cases, we dont have a value that is strictly greater than everything, so
2378
# we make both end conditions explicit
2380
# old is finished: insert current_new into the state.
2381
self.update_minimal(new_entry_key, current_new_minikind,
2382
executable=current_new[1].executable,
2383
path_utf8=new_path_utf8, fingerprint=fingerprint)
2384
current_new = advance(new_iterator)
2385
elif not current_new:
2387
self._make_absent(current_old)
2388
current_old = advance(old_iterator)
2389
elif new_entry_key == current_old[0]:
2390
# same - common case
2391
# We're looking at the same path and id in both the dirstate
2392
# and inventory, so just need to update the fields in the
2393
# dirstate from the one in the inventory.
2394
# TODO: update the record if anything significant has changed.
2395
# the minimal required trigger is if the execute bit or cached
2397
if (current_old[1][0][3] != current_new[1].executable or
2398
current_old[1][0][0] != current_new_minikind):
2399
self.update_minimal(current_old[0], current_new_minikind,
2400
executable=current_new[1].executable,
2401
path_utf8=new_path_utf8, fingerprint=fingerprint)
2402
# both sides are dealt with, move on
2403
current_old = advance(old_iterator)
2404
current_new = advance(new_iterator)
2405
elif (cmp_by_dirs(new_dirname, current_old[0][0]) < 0
2406
or (new_dirname == current_old[0][0]
2407
and new_entry_key[1:] < current_old[0][1:])):
2409
# add a entry for this and advance new
2410
self.update_minimal(new_entry_key, current_new_minikind,
2411
executable=current_new[1].executable,
2412
path_utf8=new_path_utf8, fingerprint=fingerprint)
2413
current_new = advance(new_iterator)
2415
# we've advanced past the place where the old key would be,
2416
# without seeing it in the new list. so it must be gone.
2417
self._make_absent(current_old)
2418
current_old = advance(old_iterator)
2419
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2420
self._id_index = None
2421
self._packed_stat_index = None
2423
def _make_absent(self, current_old):
2424
"""Mark current_old - an entry - as absent for tree 0.
2426
:return: True if this was the last details entry for the entry key:
2427
that is, if the underlying block has had the entry removed, thus
2428
shrinking in length.
2430
# build up paths that this id will be left at after the change is made,
2431
# so we can update their cross references in tree 0
2432
all_remaining_keys = set()
2433
# Dont check the working tree, because it's going.
2434
for details in current_old[1][1:]:
2435
if details[0] not in 'ar': # absent, relocated
2436
all_remaining_keys.add(current_old[0])
2437
elif details[0] == 'r': # relocated
2438
# record the key for the real path.
2439
all_remaining_keys.add(tuple(osutils.split(details[1])) + (current_old[0][2],))
2440
# absent rows are not present at any path.
2441
last_reference = current_old[0] not in all_remaining_keys
2443
# the current row consists entire of the current item (being marked
2444
# absent), and relocated or absent entries for the other trees:
2445
# Remove it, its meaningless.
2446
block = self._find_block(current_old[0])
2447
entry_index, present = self._find_entry_index(current_old[0], block[1])
2449
raise AssertionError('could not find entry for %s' % (current_old,))
2450
block[1].pop(entry_index)
2451
# if we have an id_index in use, remove this key from it for this id.
2452
if self._id_index is not None:
2453
self._id_index[current_old[0][2]].remove(current_old[0])
2454
# update all remaining keys for this id to record it as absent. The
2455
# existing details may either be the record we are marking as deleted
2456
# (if there were other trees with the id present at this path), or may
2458
for update_key in all_remaining_keys:
2459
update_block_index, present = \
2460
self._find_block_index_from_key(update_key)
2462
raise AssertionError('could not find block for %s' % (update_key,))
2463
update_entry_index, present = \
2464
self._find_entry_index(update_key, self._dirblocks[update_block_index][1])
2466
raise AssertionError('could not find entry for %s' % (update_key,))
2467
update_tree_details = self._dirblocks[update_block_index][1][update_entry_index][1]
2468
# it must not be absent at the moment
2469
if update_tree_details[0][0] == 'a': # absent
2470
raise AssertionError('bad row %r' % (update_tree_details,))
2471
update_tree_details[0] = DirState.NULL_PARENT_DETAILS
2472
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2473
return last_reference
2475
def update_minimal(self, key, minikind, executable=False, fingerprint='',
2476
packed_stat=None, size=0, path_utf8=None):
2477
"""Update an entry to the state in tree 0.
2479
This will either create a new entry at 'key' or update an existing one.
2480
It also makes sure that any other records which might mention this are
2483
:param key: (dir, name, file_id) for the new entry
2484
:param minikind: The type for the entry ('f' == 'file', 'd' ==
2486
:param executable: Should the executable bit be set?
2487
:param fingerprint: Simple fingerprint for new entry: sha1 for files,
2488
referenced revision id for subtrees, etc.
2489
:param packed_stat: Packed stat value for new entry.
2490
:param size: Size information for new entry
2491
:param path_utf8: key[0] + '/' + key[1], just passed in to avoid doing
2494
If packed_stat and fingerprint are not given, they're invalidated in
2497
block = self._find_block(key)[1]
2498
if packed_stat is None:
2499
packed_stat = DirState.NULLSTAT
2500
# XXX: Some callers pass '' as the packed_stat, and it seems to be
2501
# sometimes present in the dirstate - this seems oddly inconsistent.
2503
entry_index, present = self._find_entry_index(key, block)
2504
new_details = (minikind, fingerprint, size, executable, packed_stat)
2505
id_index = self._get_id_index()
2507
# new entry, synthesis cross reference here,
2508
existing_keys = id_index.setdefault(key[2], set())
2509
if not existing_keys:
2510
# not currently in the state, simplest case
2511
new_entry = key, [new_details] + self._empty_parent_info()
2513
# present at one or more existing other paths.
2514
# grab one of them and use it to generate parent
2515
# relocation/absent entries.
2516
new_entry = key, [new_details]
2517
for other_key in existing_keys:
2518
# change the record at other to be a pointer to this new
2519
# record. The loop looks similar to the change to
2520
# relocations when updating an existing record but its not:
2521
# the test for existing kinds is different: this can be
2522
# factored out to a helper though.
2523
other_block_index, present = self._find_block_index_from_key(other_key)
2525
raise AssertionError('could not find block for %s' % (other_key,))
2526
other_entry_index, present = self._find_entry_index(other_key,
2527
self._dirblocks[other_block_index][1])
2529
raise AssertionError('could not find entry for %s' % (other_key,))
2530
if path_utf8 is None:
2531
raise AssertionError('no path')
2532
self._dirblocks[other_block_index][1][other_entry_index][1][0] = \
2533
('r', path_utf8, 0, False, '')
2535
num_present_parents = self._num_present_parents()
2536
for lookup_index in xrange(1, num_present_parents + 1):
2537
# grab any one entry, use it to find the right path.
2538
# TODO: optimise this to reduce memory use in highly
2539
# fragmented situations by reusing the relocation
2541
update_block_index, present = \
2542
self._find_block_index_from_key(other_key)
2544
raise AssertionError('could not find block for %s' % (other_key,))
2545
update_entry_index, present = \
2546
self._find_entry_index(other_key, self._dirblocks[update_block_index][1])
2548
raise AssertionError('could not find entry for %s' % (other_key,))
2549
update_details = self._dirblocks[update_block_index][1][update_entry_index][1][lookup_index]
2550
if update_details[0] in 'ar': # relocated, absent
2551
# its a pointer or absent in lookup_index's tree, use
2553
new_entry[1].append(update_details)
2555
# we have the right key, make a pointer to it.
2556
pointer_path = osutils.pathjoin(*other_key[0:2])
2557
new_entry[1].append(('r', pointer_path, 0, False, ''))
2558
block.insert(entry_index, new_entry)
2559
existing_keys.add(key)
2561
# Does the new state matter?
2562
block[entry_index][1][0] = new_details
2563
# parents cannot be affected by what we do.
2564
# other occurences of this id can be found
2565
# from the id index.
2567
# tree index consistency: All other paths for this id in this tree
2568
# index must point to the correct path. We have to loop here because
2569
# we may have passed entries in the state with this file id already
2570
# that were absent - where parent entries are - and they need to be
2571
# converted to relocated.
2572
if path_utf8 is None:
2573
raise AssertionError('no path')
2574
for entry_key in id_index.setdefault(key[2], set()):
2575
# TODO:PROFILING: It might be faster to just update
2576
# rather than checking if we need to, and then overwrite
2577
# the one we are located at.
2578
if entry_key != key:
2579
# this file id is at a different path in one of the
2580
# other trees, so put absent pointers there
2581
# This is the vertical axis in the matrix, all pointing
2583
block_index, present = self._find_block_index_from_key(entry_key)
2585
raise AssertionError('not present: %r', entry_key)
2586
entry_index, present = self._find_entry_index(entry_key, self._dirblocks[block_index][1])
2588
raise AssertionError('not present: %r', entry_key)
2589
self._dirblocks[block_index][1][entry_index][1][0] = \
2590
('r', path_utf8, 0, False, '')
2591
# add a containing dirblock if needed.
2592
if new_details[0] == 'd':
2593
subdir_key = (osutils.pathjoin(*key[0:2]), '', '')
2594
block_index, present = self._find_block_index_from_key(subdir_key)
2596
self._dirblocks.insert(block_index, (subdir_key[0], []))
2598
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2600
def _validate(self):
2601
"""Check that invariants on the dirblock are correct.
2603
This can be useful in debugging; it shouldn't be necessary in
2606
This must be called with a lock held.
2608
# NOTE: This must always raise AssertionError not just assert,
2609
# otherwise it may not behave properly under python -O
2611
# TODO: All entries must have some content that's not 'a' or 'r',
2612
# otherwise it could just be removed.
2614
# TODO: All relocations must point directly to a real entry.
2616
# TODO: No repeated keys.
2619
from pprint import pformat
2620
self._read_dirblocks_if_needed()
2621
if len(self._dirblocks) > 0:
2622
if not self._dirblocks[0][0] == '':
2623
raise AssertionError(
2624
"dirblocks don't start with root block:\n" + \
2625
pformat(self._dirblocks))
2626
if len(self._dirblocks) > 1:
2627
if not self._dirblocks[1][0] == '':
2628
raise AssertionError(
2629
"dirblocks missing root directory:\n" + \
2630
pformat(self._dirblocks))
2631
# the dirblocks are sorted by their path components, name, and dir id
2632
dir_names = [d[0].split('/')
2633
for d in self._dirblocks[1:]]
2634
if dir_names != sorted(dir_names):
2635
raise AssertionError(
2636
"dir names are not in sorted order:\n" + \
2637
pformat(self._dirblocks) + \
2640
for dirblock in self._dirblocks:
2641
# within each dirblock, the entries are sorted by filename and
2643
for entry in dirblock[1]:
2644
if dirblock[0] != entry[0][0]:
2645
raise AssertionError(
2647
"doesn't match directory name in\n%r" %
2648
(entry, pformat(dirblock)))
2649
if dirblock[1] != sorted(dirblock[1]):
2650
raise AssertionError(
2651
"dirblock for %r is not sorted:\n%s" % \
2652
(dirblock[0], pformat(dirblock)))
2654
def check_valid_parent():
2655
"""Check that the current entry has a valid parent.
2657
This makes sure that the parent has a record,
2658
and that the parent isn't marked as "absent" in the
2659
current tree. (It is invalid to have a non-absent file in an absent
2662
if entry[0][0:2] == ('', ''):
2663
# There should be no parent for the root row
2665
parent_entry = self._get_entry(tree_index, path_utf8=entry[0][0])
2666
if parent_entry == (None, None):
2667
raise AssertionError(
2668
"no parent entry for: %s in tree %s"
2669
% (this_path, tree_index))
2670
if parent_entry[1][tree_index][0] != 'd':
2671
raise AssertionError(
2672
"Parent entry for %s is not marked as a valid"
2673
" directory. %s" % (this_path, parent_entry,))
2675
# For each file id, for each tree: either
2676
# the file id is not present at all; all rows with that id in the
2677
# key have it marked as 'absent'
2678
# OR the file id is present under exactly one name; any other entries
2679
# that mention that id point to the correct name.
2681
# We check this with a dict per tree pointing either to the present
2682
# name, or None if absent.
2683
tree_count = self._num_present_parents() + 1
2684
id_path_maps = [dict() for i in range(tree_count)]
2685
# Make sure that all renamed entries point to the correct location.
2686
for entry in self._iter_entries():
2687
file_id = entry[0][2]
2688
this_path = osutils.pathjoin(entry[0][0], entry[0][1])
2689
if len(entry[1]) != tree_count:
2690
raise AssertionError(
2691
"wrong number of entry details for row\n%s" \
2692
",\nexpected %d" % \
2693
(pformat(entry), tree_count))
2694
absent_positions = 0
2695
for tree_index, tree_state in enumerate(entry[1]):
2696
this_tree_map = id_path_maps[tree_index]
2697
minikind = tree_state[0]
2698
if minikind in 'ar':
2699
absent_positions += 1
2700
# have we seen this id before in this column?
2701
if file_id in this_tree_map:
2702
previous_path, previous_loc = this_tree_map[file_id]
2703
# any later mention of this file must be consistent with
2704
# what was said before
2706
if previous_path is not None:
2707
raise AssertionError(
2708
"file %s is absent in row %r but also present " \
2710
(file_id, entry, previous_path))
2711
elif minikind == 'r':
2712
target_location = tree_state[1]
2713
if previous_path != target_location:
2714
raise AssertionError(
2715
"file %s relocation in row %r but also at %r" \
2716
% (file_id, entry, previous_path))
2718
# a file, directory, etc - may have been previously
2719
# pointed to by a relocation, which must point here
2720
if previous_path != this_path:
2721
raise AssertionError(
2722
"entry %r inconsistent with previous path %r "
2724
(entry, previous_path, previous_loc))
2725
check_valid_parent()
2728
# absent; should not occur anywhere else
2729
this_tree_map[file_id] = None, this_path
2730
elif minikind == 'r':
2731
# relocation, must occur at expected location
2732
this_tree_map[file_id] = tree_state[1], this_path
2734
this_tree_map[file_id] = this_path, this_path
2735
check_valid_parent()
2736
if absent_positions == tree_count:
2737
raise AssertionError(
2738
"entry %r has no data for any tree." % (entry,))
2740
def _wipe_state(self):
2741
"""Forget all state information about the dirstate."""
2742
self._header_state = DirState.NOT_IN_MEMORY
2743
self._dirblock_state = DirState.NOT_IN_MEMORY
2744
self._changes_aborted = False
2747
self._dirblocks = []
2748
self._id_index = None
2749
self._packed_stat_index = None
2750
self._end_of_header = None
2751
self._cutoff_time = None
2752
self._split_path_cache = {}
2754
def lock_read(self):
2755
"""Acquire a read lock on the dirstate."""
2756
if self._lock_token is not None:
2757
raise errors.LockContention(self._lock_token)
2758
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2759
# already in memory, we could read just the header and check for
2760
# any modification. If not modified, we can just leave things
2762
self._lock_token = lock.ReadLock(self._filename)
2763
self._lock_state = 'r'
2764
self._state_file = self._lock_token.f
2767
def lock_write(self):
2768
"""Acquire a write lock on the dirstate."""
2769
if self._lock_token is not None:
2770
raise errors.LockContention(self._lock_token)
2771
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2772
# already in memory, we could read just the header and check for
2773
# any modification. If not modified, we can just leave things
2775
self._lock_token = lock.WriteLock(self._filename)
2776
self._lock_state = 'w'
2777
self._state_file = self._lock_token.f
2781
"""Drop any locks held on the dirstate."""
2782
if self._lock_token is None:
2783
raise errors.LockNotHeld(self)
2784
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2785
# already in memory, we could read just the header and check for
2786
# any modification. If not modified, we can just leave things
2788
self._state_file = None
2789
self._lock_state = None
2790
self._lock_token.unlock()
2791
self._lock_token = None
2792
self._split_path_cache = {}
2794
def _requires_lock(self):
2795
"""Check that a lock is currently held by someone on the dirstate."""
2796
if not self._lock_token:
2797
raise errors.ObjectNotLocked(self)
2800
# Try to load the compiled form if possible
2802
from bzrlib._dirstate_helpers_c import (
2803
_read_dirblocks_c as _read_dirblocks,
2804
bisect_dirblock_c as bisect_dirblock,
2805
_bisect_path_left_c as _bisect_path_left,
2806
_bisect_path_right_c as _bisect_path_right,
2807
cmp_by_dirs_c as cmp_by_dirs,
2810
from bzrlib._dirstate_helpers_py import (
2811
_read_dirblocks_py as _read_dirblocks,
2812
bisect_dirblock_py as bisect_dirblock,
2813
_bisect_path_left_py as _bisect_path_left,
2814
_bisect_path_right_py as _bisect_path_right,
2815
cmp_by_dirs_py as cmp_by_dirs,
added
removed
bzrlib/dirstate.py
