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
# TODO: Up-front, stat all files in order and remove those which are deleted or
18
# out-of-date. Don't actually re-read them until they're needed. That ought
19
# to bring all the inodes into core so that future stats to them are fast, and
20
# it preserves the nice property that any caller will always get up-to-date
21
# data except in unavoidable cases.
23
# TODO: Perhaps return more details on the file to avoid statting it
24
# again: nonexistent, file type, size, etc
26
# TODO: Perhaps use a Python pickle instead of a text file; might be faster.
30
CACHE_HEADER = "### bzr hashcache v5\n"
35
from bzrlib.osutils import sha_file
36
from bzrlib.trace import mutter, warning
37
from bzrlib.atomicfile import AtomicFile
42
def _fingerprint(abspath):
44
fs = os.lstat(abspath)
46
# might be missing, etc
49
if stat.S_ISDIR(fs.st_mode):
52
# we discard any high precision because it's not reliable; perhaps we
53
# could do better on some systems?
54
return (fs.st_size, long(fs.st_mtime),
55
long(fs.st_ctime), fs.st_ino, fs.st_dev, fs.st_mode)
58
class HashCache(object):
59
"""Cache for looking up file SHA-1.
61
Files are considered to match the cached value if the fingerprint
62
of the file has not changed. This includes its mtime, ctime,
63
device number, inode number, and size. This should catch
64
modifications or replacement of the file by a new one.
66
This may not catch modifications that do not change the file's
67
size and that occur within the resolution window of the
68
timestamps. To handle this we specifically do not cache files
69
which have changed since the start of the present second, since
70
they could undetectably change again.
72
This scheme may fail if the machine's clock steps backwards.
75
This does not canonicalize the paths passed in; that should be
79
Indexed by path, points to a two-tuple of the SHA-1 of the file.
83
number of times files have been statted
86
number of times files have been retrieved from the cache, avoiding a
90
number of misses (times files have been completely re-read)
94
def __init__(self, basedir):
95
self.basedir = basedir
100
self.removed_count = 0
101
self.update_count = 0
105
def cache_file_name(self):
106
return os.sep.join([self.basedir, '.bzr', 'stat-cache'])
112
"""Discard all cached information.
114
This does not reset the counters."""
116
self.needs_write = True
121
"""Scan all files and remove entries where the cache entry is obsolete.
123
Obsolete entries are those where the file has been modified or deleted
124
since the entry was inserted.
126
prep = [(ce[1][3], path, ce) for (path, ce) in self._cache.iteritems()]
129
for inum, path, cache_entry in prep:
130
abspath = os.sep.join([self.basedir, path])
131
fp = _fingerprint(abspath)
134
cache_fp = cache_entry[1]
136
if (not fp) or (cache_fp != fp):
137
# not here or not a regular file anymore
138
self.removed_count += 1
139
self.needs_write = True
140
del self._cache[path]
143
def get_sha1(self, path):
144
"""Return the sha1 of a file.
146
abspath = os.sep.join([self.basedir, path])
148
file_fp = _fingerprint(abspath)
151
# not a regular file or not existing
152
if path in self._cache:
153
self.removed_count += 1
154
self.needs_write = True
155
del self._cache[path]
158
if path in self._cache:
159
cache_sha1, cache_fp = self._cache[path]
161
cache_sha1, cache_fp = None, None
163
if cache_fp == file_fp:
170
mode = file_fp[FP_MODE_COLUMN]
171
if stat.S_ISREG(mode):
172
digest = sha_file(file(abspath, 'rb', buffering=65000))
173
elif stat.S_ISLNK(mode):
174
link_target = os.readlink(abspath)
175
digest = sha.new(os.readlink(abspath)).hexdigest()
177
raise BzrError("file %r: unknown file stat mode: %o"%(abspath,mode))
179
now = int(time.time())
180
if file_fp[1] >= now or file_fp[2] >= now:
181
# changed too recently; can't be cached. we can
182
# return the result and it could possibly be cached
184
self.danger_count += 1
186
self.removed_count += 1
187
self.needs_write = True
188
del self._cache[path]
190
self.update_count += 1
191
self.needs_write = True
192
self._cache[path] = (digest, file_fp)
196
"""Write contents of cache to file."""
197
outf = AtomicFile(self.cache_file_name(), 'wb')
199
print >>outf, CACHE_HEADER,
201
for path, c in self._cache.iteritems():
202
assert '//' not in path, path
203
outf.write(path.encode('utf-8'))
205
print >>outf, c[0], # hex sha1
207
print >>outf, "%d" % fld,
211
self.needs_write = False
219
"""Reinstate cache from file.
221
Overwrites existing cache.
223
If the cache file has the wrong version marker, this just clears
227
fn = self.cache_file_name()
229
inf = file(fn, 'rb', buffering=65000)
231
mutter("failed to open %s: %s" % (fn, e))
232
# better write it now so it is valid
233
self.needs_write = True
238
if hdr != CACHE_HEADER:
239
mutter('cache header marker not found at top of %s; discarding cache'
241
self.needs_write = True
246
path = l[:pos].decode('utf-8')
247
if path in self._cache:
248
warning('duplicated path %r in cache' % path)
252
fields = l[pos:].split(' ')
254
warning("bad line in hashcache: %r" % l)
259
warning("bad sha1 in hashcache: %r" % sha1)
262
fp = tuple(map(long, fields[1:]))
264
self._cache[path] = (sha1, fp)
266
self.needs_write = False
added
removed
bzrlib/hashcache.py
