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
19
from trace import mutter
20
from errors import BzrError, BzrCheckError
23
"""File stat cache to speed up tree comparisons.
25
This module basically gives a quick way to find the SHA-1 and related
26
information of a file in the working directory, without actually
27
reading and hashing the whole file.
34
Users of this module should not need to know about how this is
35
implemented, and in particular should not depend on the particular
36
data which is stored or its format.
38
This is done by maintaining a cache indexed by a file fingerprint of
39
(path, size, mtime, ctime, ino, dev) pointing to the SHA-1. If the
40
fingerprint has changed, we assume the file content has not changed
41
either and the SHA-1 is therefore the same.
43
If any of the fingerprint fields have changed then the file content
44
*may* have changed, or it may not have. We need to reread the file
45
contents to make sure, but this is not visible to the user or
46
higher-level code (except as a delay of course).
48
The mtime and ctime are stored with nanosecond fields, but not all
49
filesystems give this level of precision. There is therefore a
50
possible race: the file might be modified twice within a second
51
without changing the size or mtime, and a SHA-1 cached from the first
52
version would be wrong. We handle this by not recording a cached hash
53
for any files which were modified in the current second and that
54
therefore have the chance to change again before the second is up.
56
The only known hole in this design is if the system clock jumps
57
backwards crossing invocations of bzr. Please don't do that; use ntp
58
to gradually adjust your clock or don't use bzr over the step.
60
At the moment this is stored in a simple textfile; it might be nice
63
The cache is represented as a map from file_id to a tuple of (file_id,
64
sha1, path, size, mtime, ctime, ino, dev).
66
The SHA-1 is stored in memory as a hexdigest.
68
File names and file-ids are written out with non-ascii or whitespace
69
characters given as python-style unicode escapes. (file-ids shouldn't
70
contain wierd characters, but it might happen.)
73
# order of fields returned by fingerprint()
80
# order of fields in the statcache file and in the in-memory map
92
CACHE_HEADER = "### bzr statcache v3"
95
def fingerprint(abspath):
97
fs = os.lstat(abspath)
99
# might be missing, etc
102
if stat.S_ISDIR(fs.st_mode):
105
return (fs.st_size, fs.st_mtime,
106
fs.st_ctime, fs.st_ino, fs.st_dev)
111
return s.encode('unicode_escape') \
112
.replace('\n', '\\u000a') \
113
.replace(' ', '\\u0020') \
114
.replace('\r', '\\u000d')
117
def _write_cache(basedir, entries):
118
from atomicfile import AtomicFile
120
cachefn = os.path.join(basedir, '.bzr', 'stat-cache')
121
outf = AtomicFile(cachefn, 'wb')
122
outf.write(CACHE_HEADER + '\n')
124
for entry in entries:
126
raise ValueError("invalid statcache entry tuple %r" % entry)
127
outf.write(safe_quote(entry[0])) # file id
129
outf.write(entry[1]) # hex sha1
131
outf.write(safe_quote(entry[2])) # name
133
outf.write(' %d' % nf)
142
def _try_write_cache(basedir, entries):
144
return _write_cache(basedir, entries)
146
mutter("cannot update statcache in %s: %s" % (basedir, e))
148
mutter("cannot update statcache in %s: %s" % (basedir, e))
152
def load_cache(basedir):
157
sha_re = re.compile(r'[a-f0-9]{40}')
160
cachefn = os.path.join(basedir, '.bzr', 'stat-cache')
161
cachefile = open(cachefn, 'rb')
165
line1 = cachefile.readline().rstrip('\r\n')
166
if line1 != CACHE_HEADER:
167
mutter('cache header marker not found at top of %s' % cachefn)
173
file_id = f[0].decode('unicode_escape')
175
raise BzrCheckError("duplicated file_id in cache: {%s}" % file_id)
178
if len(text_sha) != 40 or not sha_re.match(text_sha):
179
raise BzrCheckError("invalid file SHA-1 in cache: %r" % text_sha)
181
path = f[2].decode('unicode_escape')
182
if path in seen_paths:
183
raise BzrCheckError("duplicated path in cache: %r" % path)
184
seen_paths[path] = True
186
entry = (file_id, text_sha, path) + tuple([long(x) for x in f[3:]])
188
raise ValueError("invalid statcache entry tuple %r" % entry)
190
cache[file_id] = entry
196
def _files_from_inventory(inv):
197
for path, ie in inv.iter_entries():
198
if ie.kind != 'file':
200
yield ie.file_id, path
204
def update_cache(basedir, inv, flush=False):
205
"""Update and return the cache for the branch.
207
The returned cache may contain entries that have not been written
208
to disk for files recently touched.
210
flush -- discard any previous cache and recalculate from scratch.
213
# load the existing cache; use information there to find a list of
214
# files ordered by inode, which is alleged to be the fastest order
217
to_update = _files_from_inventory(inv)
219
assert isinstance(flush, bool)
223
cache = load_cache(basedir)
227
for file_id, path in to_update:
229
by_inode.append((cache[file_id][SC_INO], file_id, path))
231
without_inode.append((file_id, path))
234
to_update = [a[1:] for a in by_inode] + without_inode
236
stat_cnt = missing_cnt = new_cnt = hardcheck = change_cnt = 0
238
# dangerfiles have been recently touched and can't be committed to
239
# a persistent cache yet, but they are returned to the caller.
242
now = int(time.time())
244
## mutter('update statcache under %r' % basedir)
245
for file_id, path in to_update:
246
abspath = os.path.join(basedir, path)
247
fp = fingerprint(abspath)
250
cacheentry = cache.get(file_id)
252
if fp == None: # not here
261
if (fp[FP_MTIME] >= now) or (fp[FP_CTIME] >= now):
262
dangerfiles.append(file_id)
264
if cacheentry and (cacheentry[3:] == fp):
265
continue # all stat fields unchanged
269
dig = sha.new(file(abspath, 'rb').read()).hexdigest()
271
# We update the cache even if the digest has not changed from
272
# last time we looked, so that the fingerprint fields will
274
cacheentry = (file_id, dig, path) + fp
275
cache[file_id] = cacheentry
278
mutter('statcache: statted %d files, read %d files, %d changed, %d dangerous, '
279
'%d deleted, %d new, '
281
% (stat_cnt, hardcheck, change_cnt, len(dangerfiles),
282
missing_cnt, new_cnt, len(cache)))
285
mutter('updating on-disk statcache')
288
safe_cache = cache.copy()
289
for file_id in dangerfiles:
290
del safe_cache[file_id]
294
_try_write_cache(basedir, safe_cache.itervalues())
added
removed
bzrlib/statcache.py
