1
# Copyright (C) 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""DirState objects record the state of a directory and its bzr metadata.
19
Pseudo EBNF grammar for the state file. Fields are separated by NULLs, and
20
lines by NL. The field delimiters are ommitted in the grammar, line delimiters
21
are not - this is done for clarity of reading. All string data is in utf8.
23
MINIKIND = "f" | "d" | "l" | "a" | "r";
26
WHOLE_NUMBER = {digit}, digit;
28
REVISION_ID = a non-empty utf8 string;
30
dirstate format = header line, full checksum, row count, parent details,
31
ghost_details, entries;
32
header line = "#bazaar dirstate flat format 2", NL;
33
full checksum = "adler32: ", ["-"], WHOLE_NUMBER, NL;
34
row count = "num_entries: ", digit, NL;
35
parent_details = WHOLE NUMBER, {REVISION_ID}* NL;
36
ghost_details = WHOLE NUMBER, {REVISION_ID}*, NL;
38
entry = entry_key, current_entry_details, {parent_entry_details};
39
entry_key = dirname, basename, fileid;
40
current_entry_details = common_entry_details, working_entry_details;
41
parent_entry_details = common_entry_details, history_entry_details;
42
common_entry_details = MINIKIND, fingerprint, size, executable
43
working_entry_details = packed_stat
44
history_entry_details = REVISION_ID;
47
fingerprint = a nonempty utf8 sequence with meaning defined by minikind.
49
Given this definition, the following is useful to know:
50
entry (aka row) - all the data for a given key.
51
entry[0]: The key (dirname, basename, fileid)
55
entry[1]: The tree(s) data for this path and id combination.
56
entry[1][0]: The current tree
57
entry[1][1]: The second tree
59
For an entry for a tree, we have (using tree 0 - current tree) to demonstrate:
60
entry[1][0][0]: minikind
61
entry[1][0][1]: fingerprint
63
entry[1][0][3]: executable
64
entry[1][0][4]: packed_stat
66
entry[1][1][4]: revision_id
68
There may be multiple rows at the root, one per id present in the root, so the
69
in memory root row is now:
70
self._dirblocks[0] -> ('', [entry ...]),
71
and the entries in there are
74
entries[0][2]: file_id
75
entries[1][0]: The tree data for the current tree for this fileid at /
79
'r' is a relocated entry: This path is not present in this tree with this id,
80
but the id can be found at another location. The fingerprint is used to
81
point to the target location.
82
'a' is an absent entry: In that tree the id is not present at this path.
83
'd' is a directory entry: This path in this tree is a directory with the
84
current file id. There is no fingerprint for directories.
85
'f' is a file entry: As for directory, but its a file. The fingerprint is a
87
'l' is a symlink entry: As for directory, but a symlink. The fingerprint is the
91
--- Format 1 had the following different definition: ---
92
rows = dirname, NULL, basename, NULL, MINIKIND, NULL, fileid_utf8, NULL,
93
WHOLE NUMBER (* size *), NULL, packed stat, NULL, sha1|symlink target,
95
PARENT ROW = NULL, revision_utf8, NULL, MINIKIND, NULL, dirname, NULL,
96
basename, NULL, WHOLE NUMBER (* size *), NULL, "y" | "n", NULL,
99
PARENT ROW's are emitted for every parent that is not in the ghosts details
100
line. That is, if the parents are foo, bar, baz, and the ghosts are bar, then
101
each row will have a PARENT ROW for foo and baz, but not for bar.
104
In any tree, a kind of 'moved' indicates that the fingerprint field
105
(which we treat as opaque data specific to the 'kind' anyway) has the
106
details for the id of this row in that tree.
108
I'm strongly tempted to add a id->path index as well, but I think that
109
where we need id->path mapping; we also usually read the whole file, so
110
I'm going to skip that for the moment, as we have the ability to locate
111
via bisect any path in any tree, and if we lookup things by path, we can
112
accumulate a id->path mapping as we go, which will tend to match what we
115
I plan to implement this asap, so please speak up now to alter/tweak the
116
design - and once we stabilise on this, I'll update the wiki page for
119
The rationale for all this is that we want fast operations for the
120
common case (diff/status/commit/merge on all files) and extremely fast
121
operations for the less common but still occurs a lot status/diff/commit
122
on specific files). Operations on specific files involve a scan for all
123
the children of a path, *in every involved tree*, which the current
124
format did not accommodate.
128
1) Fast end to end use for bzr's top 5 uses cases. (commmit/diff/status/merge/???)
129
2) fall back current object model as needed.
130
3) scale usably to the largest trees known today - say 50K entries. (mozilla
131
is an example of this)
135
Eventually reuse dirstate objects across locks IFF the dirstate file has not
136
been modified, but will require that we flush/ignore cached stat-hit data
137
because we wont want to restat all files on disk just because a lock was
138
acquired, yet we cannot trust the data after the previous lock was released.
140
Memory representation:
141
vector of all directories, and vector of the childen ?
143
root_entrie = (direntry for root, [parent_direntries_for_root]),
145
('', ['data for achild', 'data for bchild', 'data for cchild'])
146
('dir', ['achild', 'cchild', 'echild'])
148
- single bisect to find N subtrees from a path spec
149
- in-order for serialisation - this is 'dirblock' grouping.
150
- insertion of a file '/a' affects only the '/' child-vector, that is, to
151
insert 10K elements from scratch does not generates O(N^2) memoves of a
152
single vector, rather each individual, which tends to be limited to a
153
manageable number. Will scale badly on trees with 10K entries in a
154
single directory. compare with Inventory.InventoryDirectory which has
155
a dictionary for the children. No bisect capability, can only probe for
156
exact matches, or grab all elements and sorta.
157
- Whats the risk of error here? Once we have the base format being processed
158
we should have a net win regardless of optimality. So we are going to
159
go with what seems reasonably.
162
maybe we should do a test profile of these core structure - 10K simulated searches/lookups/etc?
164
Objects for each row?
165
The lifetime of Dirstate objects is current per lock, but see above for
166
possible extensions. The lifetime of a row from a dirstate is expected to be
167
very short in the optimistic case: which we are optimising for. For instance,
168
subtree status will determine from analysis of the disk data what rows need to
169
be examined at all, and will be able to determine from a single row whether
170
that file has altered or not, so we are aiming to process tens of thousands of
171
entries each second within the dirstate context, before exposing anything to
172
the larger codebase. This suggests we want the time for a single file
173
comparison to be < 0.1 milliseconds. That would give us 10000 paths per second
174
processed, and to scale to 100 thousand we'll another order of magnitude to do
175
that. Now, as the lifetime for all unchanged entries is the time to parse, stat
176
the file on disk, and then immediately discard, the overhead of object creation
177
becomes a significant cost.
179
Figures: Creating a tuple from from 3 elements was profiled at 0.0625
180
microseconds, whereas creating a object which is subclassed from tuple was
181
0.500 microseconds, and creating an object with 3 elements and slots was 3
182
microseconds long. 0.1 milliseconds is 100 microseconds, and ideally we'll get
183
down to 10 microseconds for the total processing - having 33% of that be object
184
creation is a huge overhead. There is a potential cost in using tuples within
185
each row which is that the conditional code to do comparisons may be slower
186
than method invocation, but method invocation is known to be slow due to stack
187
frame creation, so avoiding methods in these tight inner loops in unfortunately
188
desirable. We can consider a pyrex version of this with objects in future if
211
from bzrlib.osutils import (
219
class _Bisector(object):
220
"""This just keeps track of information as we are bisecting."""
223
class DirState(object):
224
"""Record directory and metadata state for fast access.
226
A dirstate is a specialised data structure for managing local working
227
tree state information. Its not yet well defined whether it is platform
228
specific, and if it is how we detect/parameterise that.
231
_kind_to_minikind = {'absent':'a', 'file':'f', 'directory':'d', 'relocated':'r', 'symlink':'l'}
232
_minikind_to_kind = {'a':'absent', 'f':'file', 'd':'directory', 'l':'symlink', 'r':'relocated'}
233
_to_yesno = {True:'y', False: 'n'} # TODO profile the performance gain
234
# of using int conversion rather than a dict here. AND BLAME ANDREW IF
237
# TODO: jam 20070221 Make sure we handle when there are duplicated records
238
# (like when we remove + add the same path, or we have a rename)
239
# TODO: jam 20070221 Figure out what to do if we have a record that exceeds
240
# the BISECT_PAGE_SIZE. For now, we just have to make it large enough
241
# that we are sure a single record will always fit.
242
BISECT_PAGE_SIZE = 4096
245
IN_MEMORY_UNMODIFIED = 1
246
IN_MEMORY_MODIFIED = 2
248
# A pack_stat (the x's) that is just noise and will never match the output
251
NULL_PARENT_DETAILS = ('a', '', 0, False, '')
253
HEADER_FORMAT_2 = '#bazaar dirstate flat format 2\n'
254
HEADER_FORMAT_3 = '#bazaar dirstate flat format 3\n'
256
def __init__(self, path):
257
"""Create a DirState object.
261
:attr _root_entrie: The root row of the directory/file information,
262
- contains the path to / - '', ''
263
- kind of 'directory',
264
- the file id of the root in utf8
267
- and no sha information.
268
:param path: The path at which the dirstate file on disk should live.
270
# _header_state and _dirblock_state represent the current state
271
# of the dirstate metadata and the per-row data respectiely.
272
# NOT_IN_MEMORY indicates that no data is in memory
273
# IN_MEMORY_UNMODIFIED indicates that what we have in memory
274
# is the same as is on disk
275
# IN_MEMORY_MODIFIED indicates that we have a modified version
276
# of what is on disk.
277
# In future we will add more granularity, for instance _dirblock_state
278
# will probably support partially-in-memory as a separate variable,
279
# allowing for partially-in-memory unmodified and partially-in-memory
281
self._header_state = DirState.NOT_IN_MEMORY
282
self._dirblock_state = DirState.NOT_IN_MEMORY
286
self._state_file = None
287
self._filename = path
288
self._lock_token = None
289
self._lock_state = None
290
self._id_index = None
291
self._end_of_header = None
292
self._split_path_cache = {}
293
self._bisect_page_size = DirState.BISECT_PAGE_SIZE
295
def add(self, path, file_id, kind, stat, link_or_sha1):
296
"""Add a path to be tracked.
298
:param path: The path within the dirstate - '' is the root, 'foo' is the
299
path foo within the root, 'foo/bar' is the path bar within foo
301
:param file_id: The file id of the path being added.
302
:param kind: The kind of the path.
303
:param stat: The output of os.lstat for the path.
304
:param link_or_sha1: The sha value of the file, or the target of a
305
symlink. '' for directories.
308
# find the block its in.
309
# find the location in the block.
310
# check its not there
312
#------- copied from inventory.make_entry
313
# --- normalized_filename wants a unicode basename only, so get one.
314
dirname, basename = osutils.split(path)
315
# we dont import normalized_filename directly because we want to be
316
# able to change the implementation at runtime for tests.
317
norm_name, can_access = osutils.normalized_filename(basename)
318
if norm_name != basename:
322
raise errors.InvalidNormalization(path)
323
# now that we've normalised, we need the correct utf8 path and
324
# dirname and basename elements. This single encode and split should be
325
# faster than three separate encodes.
326
utf8path = (dirname + '/' + basename).strip('/').encode('utf8')
327
dirname, basename = osutils.split(utf8path)
328
assert file_id.__class__ == str, \
329
"must be a utf8 file_id not %s" % (type(file_id))
330
# Make sure the file_id does not exist in this tree
331
file_id_entry = self._get_entry(0, fileid_utf8=file_id)
332
if file_id_entry != (None, None):
333
path = osutils.pathjoin(file_id_entry[0][0], file_id_entry[0][1])
334
kind = DirState._minikind_to_kind[file_id_entry[1][0][0]]
335
info = '%s:%s' % (kind, path)
336
raise errors.DuplicateFileId(file_id, info)
337
first_key = (dirname, basename, '')
338
block_index, present = self._find_block_index_from_key(first_key)
340
# check the path is not in the tree
341
block = self._dirblocks[block_index][1]
342
entry_index, _ = self._find_entry_index(first_key, block)
343
while (entry_index < len(block) and
344
block[entry_index][0][0:2] == first_key[0:2]):
345
if block[entry_index][1][0][0] not in 'ar':
346
# this path is in the dirstate in the current tree.
347
raise Exception, "adding already added path!"
350
# The block where we want to put the file is not present. But it
351
# might be because the directory was empty, or not loaded yet. Look
352
# for a parent entry, if not found, raise NotVersionedError
353
parent_dir, parent_base = osutils.split(dirname)
354
parent_block_idx, parent_entry_idx, _, parent_present = \
355
self._get_block_entry_index(parent_dir, parent_base, 0)
356
if not parent_present:
357
raise errors.NotVersionedError(path, str(self))
358
self._ensure_block(parent_block_idx, parent_entry_idx, dirname)
359
block = self._dirblocks[block_index][1]
360
entry_key = (dirname, basename, file_id)
363
packed_stat = DirState.NULLSTAT
366
packed_stat = pack_stat(stat)
367
parent_info = self._empty_parent_info()
368
minikind = DirState._kind_to_minikind[kind]
370
entry_data = entry_key, [
371
(minikind, link_or_sha1, size, False, packed_stat),
373
elif kind == 'directory':
374
entry_data = entry_key, [
375
(minikind, '', 0, False, packed_stat),
377
elif kind == 'symlink':
378
entry_data = entry_key, [
379
(minikind, link_or_sha1, size, False, packed_stat),
382
raise errors.BzrError('unknown kind %r' % kind)
383
entry_index, present = self._find_entry_index(entry_key, block)
384
assert not present, "basename %r already added" % basename
385
block.insert(entry_index, entry_data)
387
if kind == 'directory':
388
# insert a new dirblock
389
self._ensure_block(block_index, entry_index, utf8path)
390
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
392
self._id_index.setdefault(entry_key[2], set()).add(entry_key)
394
def _bisect(self, dir_name_list):
395
"""Bisect through the disk structure for specific rows.
397
:param dir_name_list: A list of (dir, name) pairs.
398
:return: A dict mapping (dir, name) => entry for found entries. Missing
399
entries will not be in the map.
401
self._requires_lock()
402
# We need the file pointer to be right after the initial header block
403
self._read_header_if_needed()
404
# If _dirblock_state was in memory, we should just return info from
405
# there, this function is only meant to handle when we want to read
407
assert self._dirblock_state == DirState.NOT_IN_MEMORY
409
# The disk representation is generally info + '\0\n\0' at the end. But
410
# for bisecting, it is easier to treat this as '\0' + info + '\0\n'
411
# Because it means we can sync on the '\n'
412
state_file = self._state_file
413
file_size = os.fstat(state_file.fileno()).st_size
414
# We end up with 2 extra fields, we should have a trailing '\n' to
415
# ensure that we read the whole record, and we should have a precursur
416
# '' which ensures that we start after the previous '\n'
417
entry_field_count = self._fields_per_entry() + 1
419
low = self._end_of_header
420
high = file_size - 1 # Ignore the final '\0'
421
# Map from (dir, name) => entry
424
# Avoid infinite seeking
425
max_count = 30*len(dir_name_list)
427
# pending is a list of places to look.
428
# each entry is a tuple of low, high, dir_names
429
# low -> the first byte offset to read (inclusive)
430
# high -> the last byte offset (inclusive)
431
# dir_names -> The list of (dir, name) pairs that should be found in
432
# the [low, high] range
433
pending = [(low, high, dir_name_list)]
435
page_size = self._bisect_page_size
437
fields_to_entry = self._get_fields_to_entry()
440
low, high, cur_files = pending.pop()
442
if not cur_files or low >= high:
447
if count > max_count:
448
raise errors.BzrError('Too many seeks, most likely a bug.')
450
mid = max(low, (low+high-page_size)/2)
453
# limit the read size, so we don't end up reading data that we have
455
read_size = min(page_size, (high-mid)+1)
456
block = state_file.read(read_size)
459
entries = block.split('\n')
462
# We didn't find a '\n', so we cannot have found any records.
463
# So put this range back and try again. But we know we have to
464
# increase the page size, because a single read did not contain
465
# a record break (so records must be larger than page_size)
467
pending.append((low, high, cur_files))
470
# Check the first and last entries, in case they are partial, or if
471
# we don't care about the rest of this page
473
first_fields = entries[0].split('\0')
474
if len(first_fields) < entry_field_count:
475
# We didn't get the complete first entry
476
# so move start, and grab the next, which
477
# should be a full entry
478
start += len(entries[0])+1
479
first_fields = entries[1].split('\0')
482
if len(first_fields) <= 2:
483
# We didn't even get a filename here... what do we do?
484
# Try a large page size and repeat this query
486
pending.append((low, high, cur_files))
489
# Find what entries we are looking for, which occur before and
490
# after this first record.
492
first_dir_name = (first_fields[1], first_fields[2])
493
first_loc = bisect.bisect_left(cur_files, first_dir_name)
495
# These exist before the current location
496
pre = cur_files[:first_loc]
497
# These occur after the current location, which may be in the
498
# data we read, or might be after the last entry
499
post = cur_files[first_loc:]
501
if post and len(first_fields) >= entry_field_count:
502
# We have files after the first entry
504
# Parse the last entry
505
last_entry_num = len(entries)-1
506
last_fields = entries[last_entry_num].split('\0')
507
if len(last_fields) < entry_field_count:
508
# The very last hunk was not complete,
509
# read the previous hunk
510
after = mid + len(block) - len(entries[-1])
512
last_fields = entries[last_entry_num].split('\0')
514
after = mid + len(block)
516
last_dir_name = (last_fields[1], last_fields[2])
517
last_loc = bisect.bisect_right(post, last_dir_name)
519
middle_files = post[:last_loc]
520
post = post[last_loc:]
523
# We have files that should occur in this block
524
# (>= first, <= last)
525
# Either we will find them here, or we can mark them as
528
if middle_files[0] == first_dir_name:
529
# We might need to go before this location
530
pre.append(first_dir_name)
531
if middle_files[-1] == last_dir_name:
532
post.insert(0, last_dir_name)
534
# Find out what paths we have
535
paths = {first_dir_name:[first_fields]}
536
# last_dir_name might == first_dir_name so we need to be
537
# careful if we should append rather than overwrite
538
if last_entry_num != first_entry_num:
539
paths.setdefault(last_dir_name, []).append(last_fields)
540
for num in xrange(first_entry_num+1, last_entry_num):
541
# TODO: jam 20070223 We are already splitting here, so
542
# shouldn't we just split the whole thing rather
543
# than doing the split again in add_one_record?
544
fields = entries[num].split('\0')
545
dir_name = (fields[1], fields[2])
546
paths.setdefault(dir_name, []).append(fields)
548
for dir_name in middle_files:
549
for fields in paths.get(dir_name, []):
550
# offset by 1 because of the opening '\0'
551
# consider changing fields_to_entry to avoid the
553
entry = fields_to_entry(fields[1:])
554
found.setdefault(dir_name, []).append(entry)
556
# Now we have split up everything into pre, middle, and post, and
557
# we have handled everything that fell in 'middle'.
558
# We add 'post' first, so that we prefer to seek towards the
559
# beginning, so that we will tend to go as early as we need, and
560
# then only seek forward after that.
562
pending.append((after, high, post))
564
pending.append((low, start-1, pre))
566
# Consider that we may want to return the directory entries in sorted
567
# order. For now, we just return them in whatever order we found them,
568
# and leave it up to the caller if they care if it is ordered or not.
571
def _bisect_dirblocks(self, dir_list):
572
"""Bisect through the disk structure to find entries in given dirs.
574
_bisect_dirblocks is meant to find the contents of directories, which
575
differs from _bisect, which only finds individual entries.
577
:param dir_list: An sorted list of directory names ['', 'dir', 'foo'].
578
:return: A map from dir => entries_for_dir
580
# TODO: jam 20070223 A lot of the bisecting logic could be shared
581
# between this and _bisect. It would require parameterizing the
582
# inner loop with a function, though. We should evaluate the
583
# performance difference.
584
self._requires_lock()
585
# We need the file pointer to be right after the initial header block
586
self._read_header_if_needed()
587
# If _dirblock_state was in memory, we should just return info from
588
# there, this function is only meant to handle when we want to read
590
assert self._dirblock_state == DirState.NOT_IN_MEMORY
592
# The disk representation is generally info + '\0\n\0' at the end. But
593
# for bisecting, it is easier to treat this as '\0' + info + '\0\n'
594
# Because it means we can sync on the '\n'
595
state_file = self._state_file
596
file_size = os.fstat(state_file.fileno()).st_size
597
# We end up with 2 extra fields, we should have a trailing '\n' to
598
# ensure that we read the whole record, and we should have a precursur
599
# '' which ensures that we start after the previous '\n'
600
entry_field_count = self._fields_per_entry() + 1
602
low = self._end_of_header
603
high = file_size - 1 # Ignore the final '\0'
604
# Map from dir => entry
607
# Avoid infinite seeking
608
max_count = 30*len(dir_list)
610
# pending is a list of places to look.
611
# each entry is a tuple of low, high, dir_names
612
# low -> the first byte offset to read (inclusive)
613
# high -> the last byte offset (inclusive)
614
# dirs -> The list of directories that should be found in
615
# the [low, high] range
616
pending = [(low, high, dir_list)]
618
page_size = self._bisect_page_size
620
fields_to_entry = self._get_fields_to_entry()
623
low, high, cur_dirs = pending.pop()
625
if not cur_dirs or low >= high:
630
if count > max_count:
631
raise errors.BzrError('Too many seeks, most likely a bug.')
633
mid = max(low, (low+high-page_size)/2)
636
# limit the read size, so we don't end up reading data that we have
638
read_size = min(page_size, (high-mid)+1)
639
block = state_file.read(read_size)
642
entries = block.split('\n')
645
# We didn't find a '\n', so we cannot have found any records.
646
# So put this range back and try again. But we know we have to
647
# increase the page size, because a single read did not contain
648
# a record break (so records must be larger than page_size)
650
pending.append((low, high, cur_dirs))
653
# Check the first and last entries, in case they are partial, or if
654
# we don't care about the rest of this page
656
first_fields = entries[0].split('\0')
657
if len(first_fields) < entry_field_count:
658
# We didn't get the complete first entry
659
# so move start, and grab the next, which
660
# should be a full entry
661
start += len(entries[0])+1
662
first_fields = entries[1].split('\0')
665
if len(first_fields) <= 1:
666
# We didn't even get a dirname here... what do we do?
667
# Try a large page size and repeat this query
669
pending.append((low, high, cur_dirs))
672
# Find what entries we are looking for, which occur before and
673
# after this first record.
675
first_dir = first_fields[1]
676
first_loc = bisect.bisect_left(cur_dirs, first_dir)
678
# These exist before the current location
679
pre = cur_dirs[:first_loc]
680
# These occur after the current location, which may be in the
681
# data we read, or might be after the last entry
682
post = cur_dirs[first_loc:]
684
if post and len(first_fields) >= entry_field_count:
685
# We have records to look at after the first entry
687
# Parse the last entry
688
last_entry_num = len(entries)-1
689
last_fields = entries[last_entry_num].split('\0')
690
if len(last_fields) < entry_field_count:
691
# The very last hunk was not complete,
692
# read the previous hunk
693
after = mid + len(block) - len(entries[-1])
695
last_fields = entries[last_entry_num].split('\0')
697
after = mid + len(block)
699
last_dir = last_fields[1]
700
last_loc = bisect.bisect_right(post, last_dir)
702
middle_files = post[:last_loc]
703
post = post[last_loc:]
706
# We have files that should occur in this block
707
# (>= first, <= last)
708
# Either we will find them here, or we can mark them as
711
if middle_files[0] == first_dir:
712
# We might need to go before this location
713
pre.append(first_dir)
714
if middle_files[-1] == last_dir:
715
post.insert(0, last_dir)
717
# Find out what paths we have
718
paths = {first_dir:[first_fields]}
719
# last_dir might == first_dir so we need to be
720
# careful if we should append rather than overwrite
721
if last_entry_num != first_entry_num:
722
paths.setdefault(last_dir, []).append(last_fields)
723
for num in xrange(first_entry_num+1, last_entry_num):
724
# TODO: jam 20070223 We are already splitting here, so
725
# shouldn't we just split the whole thing rather
726
# than doing the split again in add_one_record?
727
fields = entries[num].split('\0')
728
paths.setdefault(fields[1], []).append(fields)
730
for cur_dir in middle_files:
731
for fields in paths.get(cur_dir, []):
732
# offset by 1 because of the opening '\0'
733
# consider changing fields_to_entry to avoid the
735
entry = fields_to_entry(fields[1:])
736
found.setdefault(cur_dir, []).append(entry)
738
# Now we have split up everything into pre, middle, and post, and
739
# we have handled everything that fell in 'middle'.
740
# We add 'post' first, so that we prefer to seek towards the
741
# beginning, so that we will tend to go as early as we need, and
742
# then only seek forward after that.
744
pending.append((after, high, post))
746
pending.append((low, start-1, pre))
750
def _bisect_recursive(self, dir_name_list):
751
"""Bisect for entries for all paths and their children.
753
This will use bisect to find all records for the supplied paths. It
754
will then continue to bisect for any records which are marked as
755
directories. (and renames?)
757
:param paths: A sorted list of (dir, name) pairs
758
eg: [('', 'a'), ('', 'f'), ('a/b', 'c')]
759
:return: A dictionary mapping (dir, name, file_id) => [tree_info]
761
# Map from (dir, name, file_id) => [tree_info]
764
found_dir_names = set()
766
# Directories that have been read
767
processed_dirs = set()
768
# Get the ball rolling with the first bisect for all entries.
769
newly_found = self._bisect(dir_name_list)
772
# Directories that need to be read
774
paths_to_search = set()
775
for entry_list in newly_found.itervalues():
776
for dir_name_id, trees_info in entry_list:
777
found[dir_name_id] = trees_info
778
found_dir_names.add(dir_name_id[:2])
780
for tree_info in trees_info:
781
minikind = tree_info[0]
784
# We already processed this one as a directory,
785
# we don't need to do the extra work again.
787
subdir, name, file_id = dir_name_id
788
path = osutils.pathjoin(subdir, name)
790
if path not in processed_dirs:
791
pending_dirs.add(path)
792
elif minikind == 'r':
793
# Rename, we need to directly search the target
794
# which is contained in the fingerprint column
795
dir_name = osutils.split(tree_info[1])
796
if dir_name[0] in pending_dirs:
797
# This entry will be found in the dir search
799
# TODO: We need to check if this entry has
800
# already been found. Otherwise we might be
801
# hitting infinite recursion.
802
if dir_name not in found_dir_names:
803
paths_to_search.add(dir_name)
804
# Now we have a list of paths to look for directly, and
805
# directory blocks that need to be read.
806
# newly_found is mixing the keys between (dir, name) and path
807
# entries, but that is okay, because we only really care about the
809
newly_found = self._bisect(sorted(paths_to_search))
810
newly_found.update(self._bisect_dirblocks(sorted(pending_dirs)))
811
processed_dirs.update(pending_dirs)
814
def _empty_parent_info(self):
815
return [DirState.NULL_PARENT_DETAILS] * (len(self._parents) -
818
def _ensure_block(self, parent_block_index, parent_row_index, dirname):
819
"""Ensure a block for dirname exists.
821
This function exists to let callers which know that there is a
822
directory dirname ensure that the block for it exists. This block can
823
fail to exist because of demand loading, or because a directory had no
824
children. In either case it is not an error. It is however an error to
825
call this if there is no parent entry for the directory, and thus the
826
function requires the coordinates of such an entry to be provided.
828
The root row is special cased and can be indicated with a parent block
831
:param parent_block_index: The index of the block in which dirname's row
833
:param parent_row_index: The index in the parent block where the row
835
:param dirname: The utf8 dirname to ensure there is a block for.
836
:return: The index for the block.
838
if dirname == '' and parent_row_index == 0 and parent_block_index == 0:
839
# This is the signature of the root row, and the
840
# contents-of-root row is always index 1
842
# the basename of the directory must be the end of its full name.
843
if not (parent_block_index == -1 and
844
parent_block_index == -1 and dirname == ''):
845
assert dirname.endswith(
846
self._dirblocks[parent_block_index][1][parent_row_index][0][1])
847
block_index, present = self._find_block_index_from_key((dirname, '', ''))
849
## In future, when doing partial parsing, this should load and
850
# populate the entire block.
851
self._dirblocks.insert(block_index, (dirname, []))
854
def _entries_to_current_state(self, new_entries):
855
"""Load new_entries into self.dirblocks.
857
Process new_entries into the current state object, making them the active
860
:param new_entries: A sorted list of entries. This function does not sort
861
to prevent unneeded overhead when callers have a sorted list already.
864
assert new_entries[0][0][0:2] == ('', ''), \
865
"Missing root row %r" % (new_entries[0][0],)
866
# The two blocks here are deliberate: the root block and the
867
# contents-of-root block.
868
self._dirblocks = [('', []), ('', [])]
869
current_block = self._dirblocks[0][1]
872
append_entry = current_block.append
873
for entry in new_entries:
874
if entry[0][0] != current_dirname:
875
# new block - different dirname
877
current_dirname = entry[0][0]
878
self._dirblocks.append((current_dirname, current_block))
879
append_entry = current_block.append
880
# append the entry to the current block
882
self._split_root_dirblock_into_contents()
884
def _split_root_dirblock_into_contents(self):
885
"""Split the root dirblocks into root and contents-of-root.
887
After parsing by path, we end up with root entries and contents-of-root
888
entries in the same block. This loop splits them out again.
890
# The above loop leaves the "root block" entries mixed with the
891
# "contents-of-root block". But we don't want an if check on
892
# all entries, so instead we just fix it up here.
893
assert self._dirblocks[1] == ('', [])
895
contents_of_root_block = []
896
for entry in self._dirblocks[0][1]:
897
if not entry[0][1]: # This is a root entry
898
root_block.append(entry)
900
contents_of_root_block.append(entry)
901
self._dirblocks[0] = ('', root_block)
902
self._dirblocks[1] = ('', contents_of_root_block)
904
def _entry_to_line(self, entry):
905
"""Serialize entry to a NULL delimited line ready for _get_output_lines.
907
:param entry: An entry_tuple as defined in the module docstring.
909
entire_entry = list(entry[0])
910
for tree_number, tree_data in enumerate(entry[1]):
911
# (minikind, fingerprint, size, executable, tree_specific_string)
912
entire_entry.extend(tree_data)
913
# 3 for the key, 5 for the fields per tree.
914
tree_offset = 3 + tree_number * 5
916
entire_entry[tree_offset + 0] = tree_data[0]
918
entire_entry[tree_offset + 2] = str(tree_data[2])
920
entire_entry[tree_offset + 3] = DirState._to_yesno[tree_data[3]]
921
return '\0'.join(entire_entry)
923
def _fields_per_entry(self):
924
"""How many null separated fields should be in each entry row.
926
Each line now has an extra '\n' field which is not used
927
so we just skip over it
930
+ number of fields per tree_data (5) * tree count
933
tree_count = 1 + self._num_present_parents()
934
return 3 + 5 * tree_count + 1
936
def _find_block(self, key, add_if_missing=False):
937
"""Return the block that key should be present in.
939
:param key: A dirstate entry key.
940
:return: The block tuple.
942
block_index, present = self._find_block_index_from_key(key)
944
if not add_if_missing:
945
# check to see if key is versioned itself - we might want to
946
# add it anyway, because dirs with no entries dont get a
947
# dirblock at parse time.
948
# This is an uncommon branch to take: most dirs have children,
949
# and most code works with versioned paths.
950
parent_base, parent_name = osutils.split(key[0])
951
if not self._get_block_entry_index(parent_base, parent_name, 0)[3]:
952
# some parent path has not been added - its an error to add
954
raise errors.NotVersionedError(key[0:2], str(self))
955
self._dirblocks.insert(block_index, (key[0], []))
956
return self._dirblocks[block_index]
958
def _find_block_index_from_key(self, key):
959
"""Find the dirblock index for a key.
961
:return: The block index, True if the block for the key is present.
963
if key[0:2] == ('', ''):
965
block_index = bisect_dirblock(self._dirblocks, key[0], 1,
966
cache=self._split_path_cache)
967
# _right returns one-past-where-key is so we have to subtract
968
# one to use it. we use _right here because there are two
969
# '' blocks - the root, and the contents of root
970
# we always have a minimum of 2 in self._dirblocks: root and
971
# root-contents, and for '', we get 2 back, so this is
972
# simple and correct:
973
present = (block_index < len(self._dirblocks) and
974
self._dirblocks[block_index][0] == key[0])
975
return block_index, present
977
def _find_entry_index(self, key, block):
978
"""Find the entry index for a key in a block.
980
:return: The entry index, True if the entry for the key is present.
982
entry_index = bisect.bisect_left(block, (key, []))
983
present = (entry_index < len(block) and
984
block[entry_index][0] == key)
985
return entry_index, present
988
def from_tree(tree, dir_state_filename):
989
"""Create a dirstate from a bzr Tree.
991
:param tree: The tree which should provide parent information and
993
:return: a DirState object which is currently locked for writing.
994
(it was locked by DirState.initialize)
996
result = DirState.initialize(dir_state_filename)
1000
parent_ids = tree.get_parent_ids()
1001
num_parents = len(parent_ids)
1003
for parent_id in parent_ids:
1004
parent_tree = tree.branch.repository.revision_tree(parent_id)
1005
parent_trees.append((parent_id, parent_tree))
1006
parent_tree.lock_read()
1007
result.set_parent_trees(parent_trees, [])
1008
result.set_state_from_inventory(tree.inventory)
1010
for revid, parent_tree in parent_trees:
1011
parent_tree.unlock()
1014
# The caller won't have a chance to unlock this, so make sure we
1020
def get_sha1_for_entry(self, entry, abspath, stat_value=None):
1021
"""Return the sha1 sum for a given file.
1023
:param entry: This is the dirblock entry for the file in question.
1024
:param abspath: The path on disk for this file.
1025
:param stat_value: (optional) if we already have done a stat on the
1027
:return: The sha1 hexdigest of the file (40 bytes)
1029
# This code assumes that the entry passed in is directly held in one of
1030
# the internal _dirblocks. So the dirblock state must have already been
1032
assert self._dirblock_state != DirState.NOT_IN_MEMORY
1033
# TODO: jam 20070301 Because we now allow kind changes (files => dirs)
1034
# we should actually base this check on the stat value, since
1035
# that is the absolute measurement of whether we have a file or
1036
# directory or link. That means that this function might actually
1037
# change an entry from being a file => dir or dir => file, etc.
1038
if entry[1][0][0] != 'f':
1040
if stat_value is None:
1041
stat_value = os.lstat(abspath)
1042
packed_stat = pack_stat(stat_value)
1043
saved_sha1_digest = entry[1][0][1]
1044
saved_file_size = entry[1][0][2]
1045
saved_packed_stat = entry[1][0][4]
1046
if (packed_stat == saved_packed_stat
1047
and saved_sha1_digest != ''
1048
# size should also be in packed_stat
1049
and saved_file_size == stat_value.st_size):
1050
# The stat hasn't changed since we saved, so we can potentially
1051
# re-use the saved sha hash.
1052
cutoff = self._sha_cutoff_time()
1053
if (stat_value.st_mtime < cutoff
1054
and stat_value.st_ctime < cutoff):
1055
# Return the existing fingerprint
1056
return saved_sha1_digest
1057
# If we have gotten this far, that means that we need to actually
1059
new_sha1_digest = self._sha1_file(abspath)
1060
# TODO: jam 20070301 Is it worth checking to see if the new sha is
1061
# actually different? I'm guessing not, because we wouldn't have
1062
# gotten this far otherwise.
1063
entry[1][0] = ('f', new_sha1_digest, stat_value.st_size,
1064
entry[1][0][3], # Executable?
1067
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1068
return new_sha1_digest
1070
def _sha_cutoff_time(self):
1071
"""Return cutoff time.
1073
Files modified more recently than this time are at risk of being
1074
undetectably modified and so can't be cached.
1076
return int(time.time()) - 3
1078
def _sha1_file(self, abspath):
1079
"""Calculate the SHA1 of a file by reading the full text"""
1080
f = file(abspath, 'rb', buffering=65000)
1082
return osutils.sha_file(f)
1086
def get_ghosts(self):
1087
"""Return a list of the parent tree revision ids that are ghosts."""
1088
self._read_header_if_needed()
1091
def get_lines(self):
1092
"""Serialise the entire dirstate to a sequence of lines."""
1093
if (self._header_state == DirState.IN_MEMORY_UNMODIFIED and
1094
self._dirblock_state == DirState.IN_MEMORY_UNMODIFIED):
1095
# read whats on disk.
1096
self._state_file.seek(0)
1097
return self._state_file.readlines()
1099
lines.append(self._get_parents_line(self.get_parent_ids()))
1100
lines.append(self._get_ghosts_line(self._ghosts))
1101
# append the root line which is special cased
1102
lines.extend(map(self._entry_to_line, self._iter_entries()))
1103
return self._get_output_lines(lines)
1105
def _get_ghosts_line(self, ghost_ids):
1106
"""Create a line for the state file for ghost information."""
1107
return '\0'.join([str(len(ghost_ids))] + ghost_ids)
1109
def _get_parents_line(self, parent_ids):
1110
"""Create a line for the state file for parents information."""
1111
return '\0'.join([str(len(parent_ids))] + parent_ids)
1113
def _get_fields_to_entry(self):
1114
"""Get a function which converts entry fields into a entry record.
1116
This handles size and executable, as well as parent records.
1118
:return: A function which takes a list of fields, and returns an
1119
appropriate record for storing in memory.
1121
# This is intentionally unrolled for performance
1122
num_present_parents = self._num_present_parents()
1123
if num_present_parents == 0:
1124
def fields_to_entry_0_parents(fields, _int=int):
1125
path_name_file_id_key = (fields[0], fields[1], fields[2])
1126
return (path_name_file_id_key, [
1128
fields[3], # minikind
1129
fields[4], # fingerprint
1130
_int(fields[5]), # size
1131
fields[6] == 'y', # executable
1132
fields[7], # packed_stat or revision_id
1134
return fields_to_entry_0_parents
1135
elif num_present_parents == 1:
1136
def fields_to_entry_1_parent(fields, _int=int):
1137
path_name_file_id_key = (fields[0], fields[1], fields[2])
1138
return (path_name_file_id_key, [
1140
fields[3], # minikind
1141
fields[4], # fingerprint
1142
_int(fields[5]), # size
1143
fields[6] == 'y', # executable
1144
fields[7], # packed_stat or revision_id
1147
fields[8], # minikind
1148
fields[9], # fingerprint
1149
_int(fields[10]), # size
1150
fields[11] == 'y', # executable
1151
fields[12], # packed_stat or revision_id
1154
return fields_to_entry_1_parent
1155
elif num_present_parents == 2:
1156
def fields_to_entry_2_parents(fields, _int=int):
1157
path_name_file_id_key = (fields[0], fields[1], fields[2])
1158
return (path_name_file_id_key, [
1160
fields[3], # minikind
1161
fields[4], # fingerprint
1162
_int(fields[5]), # size
1163
fields[6] == 'y', # executable
1164
fields[7], # packed_stat or revision_id
1167
fields[8], # minikind
1168
fields[9], # fingerprint
1169
_int(fields[10]), # size
1170
fields[11] == 'y', # executable
1171
fields[12], # packed_stat or revision_id
1174
fields[13], # minikind
1175
fields[14], # fingerprint
1176
_int(fields[15]), # size
1177
fields[16] == 'y', # executable
1178
fields[17], # packed_stat or revision_id
1181
return fields_to_entry_2_parents
1183
def fields_to_entry_n_parents(fields, _int=int):
1184
path_name_file_id_key = (fields[0], fields[1], fields[2])
1185
trees = [(fields[cur], # minikind
1186
fields[cur+1], # fingerprint
1187
_int(fields[cur+2]), # size
1188
fields[cur+3] == 'y', # executable
1189
fields[cur+4], # stat or revision_id
1190
) for cur in xrange(3, len(fields)-1, 5)]
1191
return path_name_file_id_key, trees
1192
return fields_to_entry_n_parents
1194
def get_parent_ids(self):
1195
"""Return a list of the parent tree ids for the directory state."""
1196
self._read_header_if_needed()
1197
return list(self._parents)
1199
def _get_block_entry_index(self, dirname, basename, tree_index):
1200
"""Get the coordinates for a path in the state structure.
1202
:param dirname: The utf8 dirname to lookup.
1203
:param basename: The utf8 basename to lookup.
1204
:param tree_index: The index of the tree for which this lookup should
1206
:return: A tuple describing where the path is located, or should be
1207
inserted. The tuple contains four fields: the block index, the row
1208
index, anda two booleans are True when the directory is present, and
1209
when the entire path is present. There is no guarantee that either
1210
coordinate is currently reachable unless the found field for it is
1211
True. For instance, a directory not present in the searched tree
1212
may be returned with a value one greater than the current highest
1213
block offset. The directory present field will always be True when
1214
the path present field is True. The directory present field does
1215
NOT indicate that the directory is present in the searched tree,
1216
rather it indicates that there are at least some files in some
1219
self._read_dirblocks_if_needed()
1220
key = dirname, basename, ''
1221
block_index, present = self._find_block_index_from_key(key)
1223
# no such directory - return the dir index and 0 for the row.
1224
return block_index, 0, False, False
1225
block = self._dirblocks[block_index][1] # access the entries only
1226
entry_index, present = self._find_entry_index(key, block)
1227
# linear search through present entries at this path to find the one
1229
while entry_index < len(block) and block[entry_index][0][1] == basename:
1230
if block[entry_index][1][tree_index][0] not in \
1231
('a', 'r'): # absent, relocated
1232
return block_index, entry_index, True, True
1234
return block_index, entry_index, True, False
1236
def _get_entry(self, tree_index, fileid_utf8=None, path_utf8=None):
1237
"""Get the dirstate entry for path in tree tree_index
1239
If either file_id or path is supplied, it is used as the key to lookup.
1240
If both are supplied, the fastest lookup is used, and an error is
1241
raised if they do not both point at the same row.
1243
:param tree_index: The index of the tree we wish to locate this path
1244
in. If the path is present in that tree, the entry containing its
1245
details is returned, otherwise (None, None) is returned
1246
:param fileid_utf8: A utf8 file_id to look up.
1247
:param path_utf8: An utf8 path to be looked up.
1248
:return: The dirstate entry tuple for path, or (None, None)
1250
self._read_dirblocks_if_needed()
1251
if path_utf8 is not None:
1252
assert path_utf8.__class__ == str, 'path_utf8 is not a str: %s %s' % (type(path_utf8), path_utf8)
1253
# path lookups are faster
1254
dirname, basename = osutils.split(path_utf8)
1255
block_index, entry_index, dir_present, file_present = \
1256
self._get_block_entry_index(dirname, basename, tree_index)
1257
if not file_present:
1259
entry = self._dirblocks[block_index][1][entry_index]
1260
assert entry[0][2] and entry[1][tree_index][0] not in ('a', 'r'), 'unversioned entry?!?!'
1262
if entry[0][2] != fileid_utf8:
1263
raise errors.BzrError('integrity error ? : mismatching'
1264
' tree_index, file_id and path')
1267
assert fileid_utf8 is not None
1268
possible_keys = self._get_id_index().get(fileid_utf8, None)
1269
if not possible_keys:
1271
for key in possible_keys:
1272
block_index, present = \
1273
self._find_block_index_from_key(key)
1274
# strange, probably indicates an out of date
1275
# id index - for now, allow this.
1278
# WARNING: DO not change this code to use _get_block_entry_index
1279
# as that function is not suitable: it does not use the key
1280
# to lookup, and thus the wront coordinates are returned.
1281
block = self._dirblocks[block_index][1]
1282
entry_index, present = self._find_entry_index(key, block)
1284
entry = self._dirblocks[block_index][1][entry_index]
1285
if entry[1][tree_index][0] in 'fdl':
1286
# this is the result we are looking for: the
1287
# real home of this file_id in this tree.
1289
if entry[1][tree_index][0] == 'a':
1290
# there is no home for this entry in this tree
1292
assert entry[1][tree_index][0] == 'r'
1293
real_path = entry[1][tree_index][1]
1294
return self._get_entry(tree_index, fileid_utf8=fileid_utf8,
1295
path_utf8=real_path)
1299
def initialize(cls, path):
1300
"""Create a new dirstate on path.
1302
The new dirstate will be an empty tree - that is it has no parents,
1303
and only a root node - which has id ROOT_ID.
1305
The object will be write locked when returned to the caller,
1306
unless there was an exception in the writing, in which case it
1309
:param path: The name of the file for the dirstate.
1310
:return: A DirState object.
1312
# This constructs a new DirState object on a path, sets the _state_file
1313
# to a new empty file for that path. It then calls _set_data() with our
1314
# stock empty dirstate information - a root with ROOT_ID, no children,
1315
# and no parents. Finally it calls save() to ensure that this data will
1318
# root dir and root dir contents with no children.
1319
empty_tree_dirblocks = [('', []), ('', [])]
1320
# a new root directory, with a NULLSTAT.
1321
empty_tree_dirblocks[0][1].append(
1322
(('', '', inventory.ROOT_ID), [
1323
('d', '', 0, False, DirState.NULLSTAT),
1327
result._set_data([], empty_tree_dirblocks)
1334
def _inv_entry_to_details(self, inv_entry):
1335
"""Convert an inventory entry (from a revision tree) to state details.
1337
:param inv_entry: An inventory entry whose sha1 and link targets can be
1338
relied upon, and which has a revision set.
1339
:return: A details tuple - the details for a single tree at a path +
1342
kind = inv_entry.kind
1343
minikind = DirState._kind_to_minikind[kind]
1344
tree_data = inv_entry.revision
1345
assert len(tree_data) > 0, 'empty revision for the inv_entry.'
1346
if kind == 'directory':
1350
elif kind == 'symlink':
1351
fingerprint = inv_entry.symlink_target or ''
1354
elif kind == 'file':
1355
fingerprint = inv_entry.text_sha1 or ''
1356
size = inv_entry.text_size or 0
1357
executable = inv_entry.executable
1360
return (minikind, fingerprint, size, executable, tree_data)
1362
def _iter_entries(self):
1363
"""Iterate over all the entries in the dirstate.
1365
Each yelt item is an entry in the standard format described in the
1366
docstring of bzrlib.dirstate.
1368
self._read_dirblocks_if_needed()
1369
for directory in self._dirblocks:
1370
for entry in directory[1]:
1373
def _get_id_index(self):
1374
"""Get an id index of self._dirblocks."""
1375
if self._id_index is None:
1377
for key, tree_details in self._iter_entries():
1378
id_index.setdefault(key[2], set()).add(key)
1379
self._id_index = id_index
1380
return self._id_index
1382
def _get_output_lines(self, lines):
1383
"""format lines for final output.
1385
:param lines: A sequece of lines containing the parents list and the
1388
output_lines = [DirState.HEADER_FORMAT_3]
1389
lines.append('') # a final newline
1390
inventory_text = '\0\n\0'.join(lines)
1391
output_lines.append('adler32: %s\n' % (zlib.adler32(inventory_text),))
1392
# -3, 1 for num parents, 1 for ghosts, 1 for final newline
1393
num_entries = len(lines)-3
1394
output_lines.append('num_entries: %s\n' % (num_entries,))
1395
output_lines.append(inventory_text)
1398
def _make_deleted_row(self, fileid_utf8, parents):
1399
"""Return a deleted for for fileid_utf8."""
1400
return ('/', 'RECYCLED.BIN', 'file', fileid_utf8, 0, DirState.NULLSTAT,
1403
def _num_present_parents(self):
1404
"""The number of parent entries in each record row."""
1405
return len(self._parents) - len(self._ghosts)
1409
"""Construct a DirState on the file at path path.
1411
:return: An unlocked DirState object, associated with the given path.
1413
result = DirState(path)
1416
def _read_dirblocks_if_needed(self):
1417
"""Read in all the dirblocks from the file if they are not in memory.
1419
This populates self._dirblocks, and sets self._dirblock_state to
1420
IN_MEMORY_UNMODIFIED. It is not currently ready for incremental block
1423
self._read_header_if_needed()
1424
if self._dirblock_state == DirState.NOT_IN_MEMORY:
1425
# move the _state_file pointer to after the header (in case bisect
1426
# has been called in the mean time)
1427
self._state_file.seek(self._end_of_header)
1428
text = self._state_file.read()
1429
# TODO: check the adler checksums. adler_measured = zlib.adler32(text)
1431
fields = text.split('\0')
1432
# Remove the last blank entry
1433
trailing = fields.pop()
1434
assert trailing == ''
1435
# consider turning fields into a tuple.
1437
# skip the first field which is the trailing null from the header.
1439
# Each line now has an extra '\n' field which is not used
1440
# so we just skip over it
1442
# 3 fields for the key
1443
# + number of fields per tree_data (5) * tree count
1445
num_present_parents = self._num_present_parents()
1446
tree_count = 1 + num_present_parents
1447
entry_size = self._fields_per_entry()
1448
expected_field_count = entry_size * self._num_entries
1449
if len(fields) - cur > expected_field_count:
1450
fields = fields[:expected_field_count + cur]
1451
trace.mutter('Unexpectedly long dirstate field count!')
1452
print "XXX: incorrectly truncated dirstate file bug triggered."
1453
field_count = len(fields)
1454
# this checks our adjustment, and also catches file too short.
1455
assert field_count - cur == expected_field_count, \
1456
'field count incorrect %s != %s, entry_size=%s, '\
1457
'num_entries=%s fields=%r' % (
1458
field_count - cur, expected_field_count, entry_size,
1459
self._num_entries, fields)
1461
if num_present_parents == 1:
1462
# Bind external functions to local names
1464
# We access all fields in order, so we can just iterate over
1465
# them. Grab an straight iterator over the fields. (We use an
1466
# iterator because we don't want to do a lot of additions, nor
1467
# do we want to do a lot of slicing)
1468
next = iter(fields).next
1469
# Move the iterator to the current position
1470
for x in xrange(cur):
1472
# The two blocks here are deliberate: the root block and the
1473
# contents-of-root block.
1474
self._dirblocks = [('', []), ('', [])]
1475
current_block = self._dirblocks[0][1]
1476
current_dirname = ''
1477
append_entry = current_block.append
1478
for count in xrange(self._num_entries):
1482
if dirname != current_dirname:
1483
# new block - different dirname
1485
current_dirname = dirname
1486
self._dirblocks.append((current_dirname, current_block))
1487
append_entry = current_block.append
1488
# we know current_dirname == dirname, so re-use it to avoid
1489
# creating new strings
1490
entry = ((current_dirname, name, file_id),
1493
next(), # fingerprint
1494
_int(next()), # size
1495
next() == 'y', # executable
1496
next(), # packed_stat or revision_id
1500
next(), # fingerprint
1501
_int(next()), # size
1502
next() == 'y', # executable
1503
next(), # packed_stat or revision_id
1507
assert trailing == '\n'
1508
# append the entry to the current block
1510
self._split_root_dirblock_into_contents()
1512
fields_to_entry = self._get_fields_to_entry()
1513
entries = [fields_to_entry(fields[pos:pos+entry_size])
1514
for pos in xrange(cur, field_count, entry_size)]
1515
self._entries_to_current_state(entries)
1516
# To convert from format 2 => format 3
1517
# self._dirblocks = sorted(self._dirblocks,
1518
# key=lambda blk:blk[0].split('/'))
1519
# To convert from format 3 => format 2
1520
# self._dirblocks = sorted(self._dirblocks)
1521
self._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
1523
def _read_header(self):
1524
"""This reads in the metadata header, and the parent ids.
1526
After reading in, the file should be positioned at the null
1527
just before the start of the first record in the file.
1529
:return: (expected adler checksum, number of entries, parent list)
1531
self._read_prelude()
1532
parent_line = self._state_file.readline()
1533
info = parent_line.split('\0')
1534
num_parents = int(info[0])
1535
assert num_parents == len(info)-2, 'incorrect parent info line'
1536
self._parents = info[1:-1]
1538
ghost_line = self._state_file.readline()
1539
info = ghost_line.split('\0')
1540
num_ghosts = int(info[1])
1541
assert num_ghosts == len(info)-3, 'incorrect ghost info line'
1542
self._ghosts = info[2:-1]
1543
self._header_state = DirState.IN_MEMORY_UNMODIFIED
1544
self._end_of_header = self._state_file.tell()
1546
def _read_header_if_needed(self):
1547
"""Read the header of the dirstate file if needed."""
1548
# inline this as it will be called a lot
1549
if not self._lock_token:
1550
raise errors.ObjectNotLocked(self)
1551
if self._header_state == DirState.NOT_IN_MEMORY:
1554
def _read_prelude(self):
1555
"""Read in the prelude header of the dirstate file
1557
This only reads in the stuff that is not connected to the adler
1558
checksum. The position will be correct to read in the rest of
1559
the file and check the checksum after this point.
1560
The next entry in the file should be the number of parents,
1561
and their ids. Followed by a newline.
1563
header = self._state_file.readline()
1564
assert header == DirState.HEADER_FORMAT_3, \
1565
'invalid header line: %r' % (header,)
1566
adler_line = self._state_file.readline()
1567
assert adler_line.startswith('adler32: '), 'missing adler32 checksum'
1568
self.adler_expected = int(adler_line[len('adler32: '):-1])
1569
num_entries_line = self._state_file.readline()
1570
assert num_entries_line.startswith('num_entries: '), 'missing num_entries line'
1571
self._num_entries = int(num_entries_line[len('num_entries: '):-1])
1574
"""Save any pending changes created during this session.
1576
We reuse the existing file, because that prevents race conditions with
1577
file creation, and use oslocks on it to prevent concurrent modification
1578
and reads - because dirstates incremental data aggretation is not
1579
compatible with reading a modified file, and replacing a file in use by
1580
another process is impossible on windows.
1582
A dirstate in read only mode should be smart enough though to validate
1583
that the file has not changed, and otherwise discard its cache and
1584
start over, to allow for fine grained read lock duration, so 'status'
1585
wont block 'commit' - for example.
1587
if (self._header_state == DirState.IN_MEMORY_MODIFIED or
1588
self._dirblock_state == DirState.IN_MEMORY_MODIFIED):
1590
if self._lock_state == 'w':
1591
out_file = self._state_file
1594
# Try to grab a write lock so that we can update the file.
1596
wlock = lock.WriteLock(self._filename)
1597
except (errors.LockError, errors.LockContention), e:
1598
# We couldn't grab the lock, so just leave things dirty in
1602
# This may be a read-only tree, or someone else may have a
1603
# ReadLock. so handle the case when we cannot grab a write
1605
if e.errno in (errno.ENOENT, errno.EPERM, errno.EACCES,
1607
# Ignore these errors and just don't save anything
1613
out_file.writelines(self.get_lines())
1616
self._header_state = DirState.IN_MEMORY_UNMODIFIED
1617
self._dirblock_state = DirState.IN_MEMORY_UNMODIFIED
1619
if wlock is not None:
1622
def _set_data(self, parent_ids, dirblocks):
1623
"""Set the full dirstate data in memory.
1625
This is an internal function used to completely replace the objects
1626
in memory state. It puts the dirstate into state 'full-dirty'.
1628
:param parent_ids: A list of parent tree revision ids.
1629
:param dirblocks: A list containing one tuple for each directory in the
1630
tree. Each tuple contains the directory path and a list of entries
1631
found in that directory.
1633
# our memory copy is now authoritative.
1634
self._dirblocks = dirblocks
1635
self._header_state = DirState.IN_MEMORY_MODIFIED
1636
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1637
self._parents = list(parent_ids)
1638
self._id_index = None
1640
def set_path_id(self, path, new_id):
1641
"""Change the id of path to new_id in the current working tree.
1643
:param path: The path inside the tree to set - '' is the root, 'foo'
1644
is the path foo in the root.
1645
:param new_id: The new id to assign to the path. This must be a utf8
1646
file id (not unicode, and not None).
1648
# TODO: start warning here.
1649
assert new_id.__class__ == str
1650
self._read_dirblocks_if_needed()
1652
import pdb;pdb.set_trace()
1654
raise NotImplementedError(self.set_path_id)
1655
# TODO: check new id is unique
1656
entry = self._get_entry(0, path_utf8=path)
1657
if entry[0][2] == new_id:
1658
# Nothing to change.
1660
# mark the old path absent, and insert a new root path
1661
self._make_absent(entry)
1662
self.update_minimal(('', '', new_id), 'd',
1663
path_utf8='', packed_stat=entry[1][0][4])
1664
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1665
if self._id_index is not None:
1666
self._id_index.setdefault(new_id, set()).add(entry[0])
1668
def set_parent_trees(self, trees, ghosts):
1669
"""Set the parent trees for the dirstate.
1671
:param trees: A list of revision_id, tree tuples. tree must be provided
1672
even if the revision_id refers to a ghost: supply an empty tree in
1674
:param ghosts: A list of the revision_ids that are ghosts at the time
1677
# TODO: generate a list of parent indexes to preserve to save
1678
# processing specific parent trees. In the common case one tree will
1679
# be preserved - the left most parent.
1680
# TODO: if the parent tree is a dirstate, we might want to walk them
1681
# all by path in parallel for 'optimal' common-case performance.
1682
# generate new root row.
1683
self._read_dirblocks_if_needed()
1684
# TODO future sketch: Examine the existing parents to generate a change
1685
# map and then walk the new parent trees only, mapping them into the
1686
# dirstate. Walk the dirstate at the same time to remove unreferenced
1689
# sketch: loop over all entries in the dirstate, cherry picking
1690
# entries from the parent trees, if they are not ghost trees.
1691
# after we finish walking the dirstate, all entries not in the dirstate
1692
# are deletes, so we want to append them to the end as per the design
1693
# discussions. So do a set difference on ids with the parents to
1694
# get deletes, and add them to the end.
1695
# During the update process we need to answer the following questions:
1696
# - find other keys containing a fileid in order to create cross-path
1697
# links. We dont't trivially use the inventory from other trees
1698
# because this leads to either double touching, or to accessing
1700
# - find other keys containing a path
1701
# We accumulate each entry via this dictionary, including the root
1704
# we could do parallel iterators, but because file id data may be
1705
# scattered throughout, we dont save on index overhead: we have to look
1706
# at everything anyway. We can probably save cycles by reusing parent
1707
# data and doing an incremental update when adding an additional
1708
# parent, but for now the common cases are adding a new parent (merge),
1709
# and replacing completely (commit), and commit is more common: so
1710
# optimise merge later.
1712
# ---- start generation of full tree mapping data
1713
# what trees should we use?
1714
parent_trees = [tree for rev_id, tree in trees if rev_id not in ghosts]
1715
# how many trees do we end up with
1716
parent_count = len(parent_trees)
1718
# one: the current tree
1719
for entry in self._iter_entries():
1720
# skip entries not in the current tree
1721
if entry[1][0][0] in ('a', 'r'): # absent, relocated
1723
by_path[entry[0]] = [entry[1][0]] + \
1724
[DirState.NULL_PARENT_DETAILS] * parent_count
1725
id_index[entry[0][2]] = set([entry[0]])
1727
# now the parent trees:
1728
for tree_index, tree in enumerate(parent_trees):
1729
# the index is off by one, adjust it.
1730
tree_index = tree_index + 1
1731
# when we add new locations for a fileid we need these ranges for
1732
# any fileid in this tree as we set the by_path[id] to:
1733
# already_processed_tree_details + new_details + new_location_suffix
1734
# the suffix is from tree_index+1:parent_count+1.
1735
new_location_suffix = [DirState.NULL_PARENT_DETAILS] * (parent_count - tree_index)
1736
# now stitch in all the entries from this tree
1737
for path, entry in tree.inventory.iter_entries_by_dir():
1738
# here we process each trees details for each item in the tree.
1739
# we first update any existing entries for the id at other paths,
1740
# then we either create or update the entry for the id at the
1741
# right path, and finally we add (if needed) a mapping from
1742
# file_id to this path. We do it in this order to allow us to
1743
# avoid checking all known paths for the id when generating a
1744
# new entry at this path: by adding the id->path mapping last,
1745
# all the mappings are valid and have correct relocation
1746
# records where needed.
1747
file_id = entry.file_id
1748
path_utf8 = path.encode('utf8')
1749
dirname, basename = osutils.split(path_utf8)
1750
new_entry_key = (dirname, basename, file_id)
1751
# tree index consistency: All other paths for this id in this tree
1752
# index must point to the correct path.
1753
for entry_key in id_index.setdefault(file_id, set()):
1754
# TODO:PROFILING: It might be faster to just update
1755
# rather than checking if we need to, and then overwrite
1756
# the one we are located at.
1757
if entry_key != new_entry_key:
1758
# this file id is at a different path in one of the
1759
# other trees, so put absent pointers there
1760
# This is the vertical axis in the matrix, all pointing
1762
by_path[entry_key][tree_index] = ('r', path_utf8, 0, False, '')
1763
# by path consistency: Insert into an existing path record (trivial), or
1764
# add a new one with relocation pointers for the other tree indexes.
1765
if new_entry_key in id_index[file_id]:
1766
# there is already an entry where this data belongs, just insert it.
1767
by_path[new_entry_key][tree_index] = \
1768
self._inv_entry_to_details(entry)
1770
# add relocated entries to the horizontal axis - this row
1771
# mapping from path,id. We need to look up the correct path
1772
# for the indexes from 0 to tree_index -1
1774
for lookup_index in xrange(tree_index):
1775
# boundary case: this is the first occurence of file_id
1776
# so there are no id_indexs, possibly take this out of
1778
if not len(id_index[file_id]):
1779
new_details.append(DirState.NULL_PARENT_DETAILS)
1781
# grab any one entry, use it to find the right path.
1782
# TODO: optimise this to reduce memory use in highly
1783
# fragmented situations by reusing the relocation
1785
a_key = iter(id_index[file_id]).next()
1786
if by_path[a_key][lookup_index][0] in ('r', 'a'):
1787
# its a pointer or missing statement, use it as is.
1788
new_details.append(by_path[a_key][lookup_index])
1790
# we have the right key, make a pointer to it.
1791
real_path = ('/'.join(a_key[0:2])).strip('/')
1792
new_details.append(('r', real_path, 0, False, ''))
1793
new_details.append(self._inv_entry_to_details(entry))
1794
new_details.extend(new_location_suffix)
1795
by_path[new_entry_key] = new_details
1796
id_index[file_id].add(new_entry_key)
1797
# --- end generation of full tree mappings
1799
# sort and output all the entries
1800
new_entries = sorted(by_path.items(),
1801
key=lambda entry:(entry[0][0].split('/'), entry[0][1]))
1802
self._entries_to_current_state(new_entries)
1803
self._parents = [rev_id for rev_id, tree in trees]
1804
self._ghosts = list(ghosts)
1805
self._header_state = DirState.IN_MEMORY_MODIFIED
1806
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1807
self._id_index = id_index
1809
def set_state_from_inventory(self, new_inv):
1810
"""Set new_inv as the current state.
1812
This API is called by tree transform, and will usually occur with
1813
existing parent trees.
1815
:param new_inv: The inventory object to set current state from.
1817
self._read_dirblocks_if_needed()
1819
# incremental algorithm:
1820
# two iterators: current data and new data, both in dirblock order.
1821
new_iterator = new_inv.iter_entries_by_dir()
1822
# we will be modifying the dirstate, so we need a stable iterator. In
1823
# future we might write one, for now we just clone the state into a
1824
# list - which is a shallow copy, so each
1825
old_iterator = iter(list(self._iter_entries()))
1826
# both must have roots so this is safe:
1827
current_new = new_iterator.next()
1828
current_old = old_iterator.next()
1829
def advance(iterator):
1831
return iterator.next()
1832
except StopIteration:
1834
while current_new or current_old:
1835
# skip entries in old that are not really there
1836
if current_old and current_old[1][0][0] in ('r', 'a'):
1837
# relocated or absent
1838
current_old = advance(old_iterator)
1841
# convert new into dirblock style
1842
new_path_utf8 = current_new[0].encode('utf8')
1843
new_dirname, new_basename = osutils.split(new_path_utf8)
1844
new_id = current_new[1].file_id
1845
new_entry_key = (new_dirname, new_basename, new_id)
1846
current_new_minikind = \
1847
DirState._kind_to_minikind[current_new[1].kind]
1849
# for safety disable variables
1850
new_path_utf8 = new_dirname = new_basename = new_id = new_entry_key = None
1851
# 5 cases, we dont have a value that is strictly greater than everything, so
1852
# we make both end conditions explicit
1854
# old is finished: insert current_new into the state.
1855
self.update_minimal(new_entry_key, current_new_minikind,
1856
executable=current_new[1].executable,
1857
path_utf8=new_path_utf8)
1858
current_new = advance(new_iterator)
1859
elif not current_new:
1861
self._make_absent(current_old)
1862
current_old = advance(old_iterator)
1863
elif new_entry_key == current_old[0]:
1864
# same - common case
1865
# TODO: update the record if anything significant has changed.
1866
# the minimal required trigger is if the execute bit or cached
1868
if (current_old[1][0][3] != current_new[1].executable or
1869
current_old[1][0][0] != current_new_minikind):
1870
self.update_minimal(current_old[0], current_new_minikind,
1871
executable=current_new[1].executable,
1872
path_utf8=new_path_utf8)
1873
# both sides are dealt with, move on
1874
current_old = advance(old_iterator)
1875
current_new = advance(new_iterator)
1876
elif new_entry_key < current_old[0]:
1878
# add a entry for this and advance new
1879
self.update_minimal(new_entry_key, current_new_minikind,
1880
executable=current_new[1].executable,
1881
path_utf8=new_path_utf8)
1882
current_new = advance(new_iterator)
1885
self._make_absent(current_old)
1886
current_old = advance(old_iterator)
1887
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1888
self._id_index = None
1890
def _make_absent(self, current_old):
1891
"""Mark current_old - an entry - as absent for tree 0.
1893
:return: True if this was the last details entry for they entry key:
1894
that is, if the underlying block has had the entry removed, thus
1895
shrinking in length.
1897
# build up paths that this id will be left at after the change is made,
1898
# so we can update their cross references in tree 0
1899
all_remaining_keys = set()
1900
# Dont check the working tree, because its going.
1901
for details in current_old[1][1:]:
1902
if details[0] not in ('a', 'r'): # absent, relocated
1903
all_remaining_keys.add(current_old[0])
1904
elif details[0] == 'r': # relocated
1905
# record the key for the real path.
1906
all_remaining_keys.add(tuple(osutils.split(details[1])) + (current_old[0][2],))
1907
# absent rows are not present at any path.
1908
last_reference = current_old[0] not in all_remaining_keys
1910
# the current row consists entire of the current item (being marked
1911
# absent), and relocated or absent entries for the other trees:
1912
# Remove it, its meaningless.
1913
block = self._find_block(current_old[0])
1914
entry_index, present = self._find_entry_index(current_old[0], block[1])
1915
assert present, 'could not find entry for %s' % (current_old,)
1916
block[1].pop(entry_index)
1917
# if we have an id_index in use, remove this key from it for this id.
1918
if self._id_index is not None:
1919
self._id_index[current_old[0][2]].remove(current_old[0])
1920
# update all remaining keys for this id to record it as absent. The
1921
# existing details may either be the record we are making as deleted
1922
# (if there were other trees with the id present at this path), or may
1924
for update_key in all_remaining_keys:
1925
update_block_index, present = \
1926
self._find_block_index_from_key(update_key)
1927
assert present, 'could not find block for %s' % (update_key,)
1928
update_entry_index, present = \
1929
self._find_entry_index(update_key, self._dirblocks[update_block_index][1])
1930
assert present, 'could not find entry for %s' % (update_key,)
1931
update_tree_details = self._dirblocks[update_block_index][1][update_entry_index][1]
1932
# it must not be absent at the moment
1933
assert update_tree_details[0][0] != 'a' # absent
1934
update_tree_details[0] = DirState.NULL_PARENT_DETAILS
1935
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
1936
return last_reference
1938
def update_minimal(self, key, minikind, executable=False, fingerprint='',
1939
packed_stat=None, size=0, path_utf8=None):
1940
"""Update an entry to the state in tree 0.
1942
This will either create a new entry at 'key' or update an existing one.
1943
It also makes sure that any other records which might mention this are
1946
:param key: (dir, name, file_id) for the new entry
1947
:param minikind: The type for the entry ('f' == 'file', 'd' ==
1949
:param executable: Should the executable bit be set?
1950
:param fingerprint: Simple fingerprint for new entry.
1951
:param packed_stat: packed stat value for new entry.
1952
:param size: Size information for new entry
1953
:param path_utf8: key[0] + '/' + key[1], just passed in to avoid doing
1956
block = self._find_block(key)[1]
1957
if packed_stat is None:
1958
packed_stat = DirState.NULLSTAT
1959
entry_index, present = self._find_entry_index(key, block)
1960
new_details = (minikind, fingerprint, size, executable, packed_stat)
1961
id_index = self._get_id_index()
1963
# new entry, synthesis cross reference here,
1964
existing_keys = id_index.setdefault(key[2], set())
1965
if not existing_keys:
1966
# not currently in the state, simplest case
1967
new_entry = key, [new_details] + self._empty_parent_info()
1969
# present at one or more existing other paths.
1970
# grab one of them and use it to generate parent
1971
# relocation/absent entries.
1972
new_entry = key, [new_details]
1973
for other_key in existing_keys:
1974
# change the record at other to be a pointer to this new
1975
# record. The loop looks similar to the change to
1976
# relocations when updating an existing record but its not:
1977
# the test for existing kinds is different: this can be
1978
# factored out to a helper though.
1979
other_block_index, present = self._find_block_index_from_key(other_key)
1981
import pdb; pdb.set_trace()
1982
assert present, 'could not find block for %s' % (other_key,)
1983
other_entry_index, present = self._find_entry_index(other_key,
1984
self._dirblocks[other_block_index][1])
1986
import pdb; pdb.set_trace()
1987
assert present, 'could not find entry for %s' % (other_key,)
1988
assert path_utf8 is not None
1989
self._dirblocks[other_block_index][1][other_entry_index][1][0] = \
1990
('r', path_utf8, 0, False, '')
1992
num_present_parents = self._num_present_parents()
1993
for lookup_index in xrange(1, num_present_parents + 1):
1994
# grab any one entry, use it to find the right path.
1995
# TODO: optimise this to reduce memory use in highly
1996
# fragmented situations by reusing the relocation
1998
update_block_index, present = \
1999
self._find_block_index_from_key(other_key)
2000
assert present, 'could not find block for %s' % (other_key,)
2001
update_entry_index, present = \
2002
self._find_entry_index(other_key, self._dirblocks[update_block_index][1])
2003
assert present, 'could not find entry for %s' % (other_key,)
2004
update_details = self._dirblocks[update_block_index][1][update_entry_index][1][lookup_index]
2005
if update_details[0] in ('r', 'a'): # relocated, absent
2006
# its a pointer or absent in lookup_index's tree, use
2008
new_entry[1].append(update_details)
2010
# we have the right key, make a pointer to it.
2011
pointer_path = osutils.pathjoin(*other_key[0:2])
2012
new_entry[1].append(('r', pointer_path, 0, False, ''))
2013
block.insert(entry_index, new_entry)
2014
existing_keys.add(key)
2016
# Does the new state matter?
2017
block[entry_index][1][0] = new_details
2018
# parents cannot be affected by what we do.
2019
# other occurences of this id can be found
2020
# from the id index.
2022
# tree index consistency: All other paths for this id in this tree
2023
# index must point to the correct path. We have to loop here because
2024
# we may have passed entries in the state with this file id already
2025
# that were absent - where parent entries are - and they need to be
2026
# converted to relocated.
2027
assert path_utf8 is not None
2028
for entry_key in id_index.setdefault(key[2], set()):
2029
# TODO:PROFILING: It might be faster to just update
2030
# rather than checking if we need to, and then overwrite
2031
# the one we are located at.
2032
if entry_key != key:
2033
# this file id is at a different path in one of the
2034
# other trees, so put absent pointers there
2035
# This is the vertical axis in the matrix, all pointing
2037
block_index, present = self._find_block_index_from_key(entry_key)
2039
entry_index, present = self._find_entry_index(entry_key, self._dirblocks[block_index][1])
2041
self._dirblocks[block_index][1][entry_index][1][0] = \
2042
('r', path_utf8, 0, False, '')
2043
# add a containing dirblock if needed.
2044
if new_details[0] == 'd':
2045
subdir_key = (osutils.pathjoin(*key[0:2]), '', '')
2046
block_index, present = self._find_block_index_from_key(subdir_key)
2048
self._dirblocks.insert(block_index, (subdir_key[0], []))
2050
self._dirblock_state = DirState.IN_MEMORY_MODIFIED
2052
def _validate(self):
2053
"""Check that invariants on the dirblock are correct.
2055
This can be useful in debugging; it shouldn't be necessary in
2058
from pprint import pformat
2059
if len(self._dirblocks) > 0:
2060
assert self._dirblocks[0][0] == '', \
2061
"dirblocks don't start with root block:\n" + \
2063
if len(self._dirblocks) > 1:
2064
assert self._dirblocks[1][0] == '', \
2065
"dirblocks missing root directory:\n" + \
2067
assert self._dirblocks[1:] == sorted(self._dirblocks[1:]), \
2068
"dirblocks are not in sorted order:\n" + \
2069
pformat(self._dirblocks)
2070
for dirblock in self._dirblocks:
2071
assert dirblock[1] == sorted(dirblock[1]), \
2072
"dirblock for %r is not sorted:\n%s" % \
2073
(dirblock[0], pformat(dirblock))
2075
def _wipe_state(self):
2076
"""Forget all state information about the dirstate."""
2077
self._header_state = DirState.NOT_IN_MEMORY
2078
self._dirblock_state = DirState.NOT_IN_MEMORY
2081
self._dirblocks = []
2083
def lock_read(self):
2084
"""Acquire a read lock on the dirstate"""
2085
if self._lock_token is not None:
2086
raise errors.LockContention(self._lock_token)
2087
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2088
# already in memory, we could read just the header and check for
2089
# any modification. If not modified, we can just leave things
2091
self._lock_token = lock.ReadLock(self._filename)
2092
self._lock_state = 'r'
2093
self._state_file = self._lock_token.f
2096
def lock_write(self):
2097
"""Acquire a write lock on the dirstate"""
2098
if self._lock_token is not None:
2099
raise errors.LockContention(self._lock_token)
2100
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2101
# already in memory, we could read just the header and check for
2102
# any modification. If not modified, we can just leave things
2104
self._lock_token = lock.WriteLock(self._filename)
2105
self._lock_state = 'w'
2106
self._state_file = self._lock_token.f
2110
"""Drop any locks held on the dirstate"""
2111
if self._lock_token is None:
2112
raise errors.LockNotHeld(self)
2113
# TODO: jam 20070301 Rather than wiping completely, if the blocks are
2114
# already in memory, we could read just the header and check for
2115
# any modification. If not modified, we can just leave things
2117
self._state_file = None
2118
self._lock_state = None
2119
self._lock_token.unlock()
2120
self._lock_token = None
2121
self._split_path_cache = {}
2123
def _requires_lock(self):
2124
"""Checks that a lock is currently held by someone on the dirstate"""
2125
if not self._lock_token:
2126
raise errors.ObjectNotLocked(self)
2129
def bisect_dirblock(dirblocks, dirname, lo=0, hi=None, cache={}):
2130
"""Return the index where to insert dirname into the dirblocks.
2132
The return value idx is such that all directories blocks in dirblock[:idx]
2133
have names < dirname, and all blocks in dirblock[idx:] have names >=
2136
Optional args lo (default 0) and hi (default len(dirblocks)) bound the
2137
slice of a to be searched.
2142
dirname_split = cache[dirname]
2144
dirname_split = dirname.split('/')
2145
cache[dirname] = dirname_split
2148
# Grab the dirname for the current dirblock
2149
cur = dirblocks[mid][0]
2151
cur_split = cache[cur]
2153
cur_split = cur.split('/')
2154
cache[cur] = cur_split
2155
if cur_split < dirname_split: lo = mid+1
2160
_base64_encoder = codecs.getencoder('base64')
2163
def pack_stat(st, _encode=_base64_encoder, _pack=struct.pack):
2164
"""Convert stat values into a packed representation."""
2165
# jam 20060614 it isn't really worth removing more entries if we
2166
# are going to leave it in packed form.
2167
# With only st_mtime and st_mode filesize is 5.5M and read time is 275ms
2168
# With all entries filesize is 5.9M and read time is mabye 280ms
2169
# well within the noise margin
2171
# base64.encode always adds a final newline, so strip it off
2172
return _encode(_pack('>llllll'
2173
, st.st_size, int(st.st_mtime), int(st.st_ctime)
2174
, st.st_dev, st.st_ino, st.st_mode))[0][:-1]
added
removed
bzrlib/dirstate.py
