1
# (C) 2005 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
import stat, os, sha, time
18
from binascii import b2a_qp, a2b_qp
20
from trace import mutter
21
from errors import BzrError, BzrCheckError
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.
35
Users of this module should not need to know about how this is
36
implemented, and in particular should not depend on the particular
37
data which is stored or its format.
39
This is done by maintaining a cache indexed by a file fingerprint of
40
(path, size, mtime, ctime, ino, dev) pointing to the SHA-1. If the
41
fingerprint has changed, we assume the file content has not changed
42
either and the SHA-1 is therefore the same.
44
If any of the fingerprint fields have changed then the file content
45
*may* have changed, or it may not have. We need to reread the file
46
contents to make sure, but this is not visible to the user or
47
higher-level code (except as a delay of course).
49
The mtime and ctime are stored with nanosecond fields, but not all
50
filesystems give this level of precision. There is therefore a
51
possible race: the file might be modified twice within a second
52
without changing the size or mtime, and a SHA-1 cached from the first
53
version would be wrong. We handle this by not recording a cached hash
54
for any files which were modified in the current second and that
55
therefore have the chance to change again before the second is up.
57
The only known hole in this design is if the system clock jumps
58
backwards crossing invocations of bzr. Please don't do that; use ntp
59
to gradually adjust your clock or don't use bzr over the step.
61
At the moment this is stored in a simple textfile; it might be nice
64
The cache is represented as a map from file_id to a tuple of (file_id,
65
sha1, path, size, mtime, ctime, ino, dev).
67
The SHA-1 is stored in memory as a hexdigest.
69
File names and file-ids are written out as the quoted-printable
70
encoding of their UTF-8 representation. (file-ids shouldn't contain
71
wierd characters, but it might happen.)
74
# order of fields returned by fingerprint()
81
# order of fields in the statcache file and in the in-memory map
93
CACHE_HEADER = "### bzr statcache v2"
96
def fingerprint(abspath):
98
fs = os.lstat(abspath)
100
# might be missing, etc
103
if stat.S_ISDIR(fs.st_mode):
106
return (fs.st_size, fs.st_mtime,
107
fs.st_ctime, fs.st_ino, fs.st_dev)
110
def _write_cache(basedir, entry_iter, dangerfiles):
111
from atomicfile import AtomicFile
113
cachefn = os.path.join(basedir, '.bzr', 'stat-cache')
114
outf = AtomicFile(cachefn, 'wb')
115
outf.write(CACHE_HEADER + '\n')
117
for entry in entry_iter:
119
raise ValueError("invalid statcache entry tuple %r" % entry)
121
if entry[SC_FILE_ID] in dangerfiles:
122
continue # changed too recently
123
outf.write(b2a_qp(entry[0].encode('utf-8'))) # file id
125
outf.write(entry[1]) # hex sha1
127
outf.write(b2a_qp(entry[2].encode('utf-8'), True)) # name
129
outf.write(' %d' % nf)
138
def load_cache(basedir):
144
sha_re = re.compile(r'[a-f0-9]{40}')
147
cachefn = os.path.join(basedir, '.bzr', 'stat-cache')
148
cachefile = open(cachefn, 'rb')
152
line1 = cachefile.readline().rstrip('\r\n')
153
if line1 != CACHE_HEADER:
154
mutter('cache header marker not found at top of %s' % cachefn)
160
file_id = a2b_qp(f[0]).decode('utf-8')
162
raise BzrCheckError("duplicated file_id in cache: {%s}" % file_id)
165
if len(text_sha) != 40 or not sha_re.match(text_sha):
166
raise BzrCheckError("invalid file SHA-1 in cache: %r" % text_sha)
168
path = a2b_qp(f[2]).decode('utf-8')
169
if path in seen_paths:
170
raise BzrCheckError("duplicated path in cache: %r" % path)
173
entry = (file_id, text_sha, path) + tuple([long(x) for x in f[3:]])
175
raise ValueError("invalid statcache entry tuple %r" % entry)
177
cache[file_id] = entry
183
def _files_from_inventory(inv):
184
for path, ie in inv.iter_entries():
185
if ie.kind != 'file':
187
yield ie.file_id, path
191
def update_cache(basedir, inv, flush=False):
192
"""Update and return the cache for the branch.
194
The returned cache may contain entries that have not been written
195
to disk for files recently touched.
197
flush -- discard any previous cache and recalculate from scratch.
201
# TODO: It's supposed to be faster to stat the files in order by inum.
202
# We don't directly know the inum of the files of course but we do
203
# know where they were last sighted, so we can sort by that.
205
assert isinstance(flush, bool)
209
cache = load_cache(basedir)
210
return _update_cache_from_list(basedir, cache, _files_from_inventory(inv))
214
def _update_cache_from_list(basedir, cache, to_update):
215
"""Update and return the cache for given files.
217
cache -- Previously cached values to be validated.
219
to_update -- Sequence of (file_id, path) pairs to check.
224
stat_cnt = missing_cnt = hardcheck = change_cnt = 0
226
# files that have been recently touched and can't be
227
# committed to a persistent cache yet.
230
now = int(time.time())
232
## mutter('update statcache under %r' % basedir)
233
for file_id, path in to_update:
234
abspath = os.path.join(basedir, path)
235
fp = fingerprint(abspath)
238
cacheentry = cache.get(file_id)
240
if fp == None: # not here
247
if (fp[FP_MTIME] >= now) or (fp[FP_CTIME] >= now):
248
dangerfiles.add(file_id)
250
if cacheentry and (cacheentry[3:] == fp):
251
continue # all stat fields unchanged
255
dig = sha.new(file(abspath, 'rb').read()).hexdigest()
257
# We update the cache even if the digest has not changed from
258
# last time we looked, so that the fingerprint fields will
260
cacheentry = (file_id, dig, path) + fp
261
cache[file_id] = cacheentry
264
mutter('statcache: statted %d files, read %d files, %d changed, %d dangerous, '
266
% (stat_cnt, hardcheck, change_cnt, len(dangerfiles), len(cache)))
269
mutter('updating on-disk statcache')
270
_write_cache(basedir, cache.itervalues(), dangerfiles)
added
removed
bzrlib/statcache.py
