18
18
from binascii import b2a_qp, a2b_qp
20
20
from trace import mutter
21
from errors import BzrError
24
"""File stat cache to speed up tree comparisons.
26
This module basically gives a quick way to find the SHA-1 and related
27
information of a file in the working directory, without actually
28
reading and hashing the whole file.
30
This is done by maintaining a cache indexed by a file fingerprint of
31
(path, size, mtime, ctime, ino, dev) pointing to the SHA-1. If the
32
fingerprint has changed, we assume the file content has not changed
33
either and the SHA-1 is therefore the same.
35
If any of the fingerprint fields have changed then the file content
36
*may* have changed, or it may not have. We need to reread the file
37
contents to make sure, but this is not visible to the user or
38
higher-level code (except as a delay of course).
40
The mtime and ctime are stored with nanosecond fields, but not all
41
filesystems give this level of precision. There is therefore a
42
possible race: the file might be modified twice within a second
43
without changing the size or mtime, and a SHA-1 cached from the first
44
version would be wrong. We handle this by not recording a cached hash
45
for any files which were modified in the current second and that
46
therefore have the chance to change again before the second is up.
48
The only known hole in this design is if the system clock jumps
49
backwards crossing invocations of bzr. Please don't do that; use ntp
50
to gradually adjust your clock or don't use bzr over the step.
52
At the moment this is stored in a simple textfile; it might be nice
55
The cache is represented as a map from file_id to a tuple of (file_id,
56
sha1, path, size, mtime, ctime, ino, dev).
23
# file fingerprints are: (path, size, mtime, ctime, ino, dev).
25
# if this is the same for this file as in the previous revision, we
26
# assume the content is the same and the SHA-1 is the same.
28
# This is stored in a fingerprint file that also contains the file-id
29
# and the content SHA-1.
31
# Thus for any given file we can quickly get the SHA-1, either from
32
# the cache or if the cache is out of date.
34
# At the moment this is stored in a simple textfile; it might be nice
35
# to use a tdb instead.
40
# build a new cache from scratch
41
# load cache, incrementally update it
43
# TODO: Have a paranoid mode where we always compare the texts and
44
# always recalculate the digest, to trap modification without stat
45
# change and SHA collisions.
71
49
def fingerprint(path, abspath):
82
60
fs.st_ctime, fs.st_ino, fs.st_dev)
85
def _write_cache(branch, entry_iter, dangerfiles):
86
from atomicfile import AtomicFile
88
outf = AtomicFile(branch.controlfilename('stat-cache'), 'wb', 'utf-8')
90
for entry in entry_iter:
91
if entry[0] in dangerfiles:
93
outf.write(entry[0] + ' ' + entry[1] + ' ')
94
outf.write(b2a_qp(entry[2], True))
95
outf.write(' %d %d %d %d %d\n' % entry[3:])
63
def write_cache(branch, entry_iter):
64
outf = branch.controlfile('work-cache.tmp', 'wt')
65
for entry in entry_iter:
66
outf.write(entry[0] + ' ' + entry[1] + ' ')
67
outf.write(b2a_qp(entry[2], True))
68
outf.write(' %d %d %d %d %d\n' % entry[3:])
71
os.rename(branch.controlfilename('work-cache.tmp'),
72
branch.controlfilename('work-cache'))
103
76
def load_cache(branch):
107
cachefile = branch.controlfile('stat-cache', 'r')
80
cachefile = branch.controlfile('work-cache', 'rt')
126
99
yield ie.file_id, path
130
def update_cache(branch, inv=None, flush=False):
131
"""Update and return the cache for the branch.
133
The returned cache may contain entries that have not been written
134
to disk for files recently touched.
136
flush -- discard any previous cache and recalculate from scratch.
102
def build_cache(branch):
103
inv = branch.read_working_inventory()
106
_update_cache_from_list(branch, cache, _files_from_inventory(inv))
110
def update_cache(branch, inv):
140
111
# TODO: It's supposed to be faster to stat the files in order by inum.
141
112
# We don't directly know the inum of the files of course but we do
142
113
# know where they were last sighted, so we can sort by that.
144
assert isinstance(flush, bool)
148
cache = load_cache(branch)
150
inv = branch.read_working_inventory()
115
cache = load_cache(branch)
151
116
return _update_cache_from_list(branch, cache, _files_from_inventory(inv))
155
120
def _update_cache_from_list(branch, cache, to_update):
156
"""Update and return the cache for given files.
158
cache -- Previously cached values to be validated.
160
to_update -- Sequence of (file_id, path) pairs to check.
121
"""Update the cache to have info on the named files.
123
to_update is a sequence of (file_id, path) pairs.
165
125
hardcheck = dirty = 0
167
# files that have been recently touched and can't be
168
# committed to a persistent cache yet.
171
now = int(time.time())
173
126
for file_id, path in to_update:
174
127
fap = branch.abspath(path)
175
128
fp = fingerprint(fap, path)
198
148
cache[file_id] = cacheentry
201
mutter('statcache: read %d files, %d changed, %d dangerous, '
203
% (hardcheck, dirty, len(dangerfiles), len(cache)))
151
mutter('work cache: read %d files, %d changed' % (hardcheck, dirty))
206
mutter('updating on-disk statcache')
207
_write_cache(branch, cache.itervalues(), dangerfiles)
154
write_cache(branch, cache.itervalues())