« back to all changes in this revision
Viewing changes to bzrlib/hashcache.py
- browse files at revision 950
- compare with another revision
- download diff
- download tarball
- view history from revision 950
- Committer: Martin Pool
- Date: 2005-07-21 11:57:44 UTC
- Revision ID: mbp@sourcefrog.net-20050721115744-e450602f7f11efa5
- fix up self reference in msvc lock object
- files removed:
- bzrlib/selftest/testinv.py
- files modified:
- README
added
removed
bzrlib/hashcache.py
14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
16
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# out-of-date.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.
17
18
19
# TODO: Perhaps have a way to stat all the files in inode order, and
20
# then remember that they're all fresh for the lifetime of the object?
21
22
# TODO: Keep track of whether there are in-memory updates that need to
23
# be flushed.
22
24
23
25
# TODO: Perhaps return more details on the file to avoid statting it
24
26
# again: nonexistent, file type, size, etc
25
27
26
28
27
29
30
28
31
CACHE_HEADER = "### bzr hashcache v5\n"
29
32
30
import os, stat, time
31
32
from bzrlib.osutils import sha_file
33
from bzrlib.trace import mutter, warning
34
35
36
33
37
34
def _fingerprint(abspath):
35
import os, stat
36
38
37
try:
39
38
fs = os.lstat(abspath)
40
39
except OSError:
92
91
self.miss_count = 0
93
92
self.stat_count = 0
94
93
self.danger_count = 0
95
self.removed_count = 0
96
self.update_count = 0
97
94
self._cache = {}
98
95
99
96
97
100
98
def cache_file_name(self):
101
return os.sep.join([self.basedir, '.bzr', 'stat-cache'])
99
import os.path
100
return os.path.join(self.basedir, '.bzr', 'stat-cache')
102
101
103
102
104
103
112
111
self._cache = {}
113
112
114
113
115
def scan(self):
116
"""Scan all files and remove entries where the cache entry is obsolete.
117
118
Obsolete entries are those where the file has been modified or deleted
119
since the entry was inserted.
120
"""
121
prep = [(ce[1][3], path, ce) for (path, ce) in self._cache.iteritems()]
122
prep.sort()
123
124
for inum, path, cache_entry in prep:
125
abspath = os.sep.join([self.basedir, path])
126
fp = _fingerprint(abspath)
127
self.stat_count += 1
128
129
cache_fp = cache_entry[1]
130
131
if (not fp) or (cache_fp != fp):
132
# not here or not a regular file anymore
133
self.removed_count += 1
134
self.needs_write = True
135
del self._cache[path]
136
137
138
139
114
def get_sha1(self, path):
140
"""Return the sha1 of a file.
115
"""Return the hex SHA-1 of the contents of the file at path.
116
117
XXX: If the file does not exist or is not a plain file???
141
118
"""
142
abspath = os.sep.join([self.basedir, path])
119
120
import os, time
121
from bzrlib.osutils import sha_file
122
from bzrlib.trace import mutter
123
124
abspath = os.path.join(self.basedir, path)
125
fp = _fingerprint(abspath)
126
c = self._cache.get(path)
127
if c:
128
cache_sha1, cache_fp = c
129
else:
130
cache_sha1, cache_fp = None, None
131
143
132
self.stat_count += 1
144
file_fp = _fingerprint(abspath)
145
146
if not file_fp:
147
# not a regular file or not existing
148
if path in self._cache:
149
self.removed_count += 1
150
self.needs_write = True
151
del self._cache[path]
152
return None
153
154
if path in self._cache:
155
cache_sha1, cache_fp = self._cache[path]
156
else:
157
cache_sha1, cache_fp = None, None
158
159
if cache_fp == file_fp:
133
134
if not fp:
135
# not a regular file
136
return None
137
elif cache_fp and (cache_fp == fp):
160
138
self.hit_count += 1
161
139
return cache_sha1
162
163
self.miss_count += 1
164
digest = sha_file(file(abspath, 'rb', buffering=65000))
140
else:
141
self.miss_count += 1
142
digest = sha_file(file(abspath, 'rb', buffering=65000))
165
143
166
now = int(time.time())
167
if file_fp[1] >= now or file_fp[2] >= now:
168
# changed too recently; can't be cached. we can
169
# return the result and it could possibly be cached
170
# next time.
171
self.danger_count += 1
172
if cache_fp:
173
self.removed_count += 1
144
now = int(time.time())
145
if fp[1] >= now or fp[2] >= now:
146
# changed too recently; can't be cached. we can
147
# return the result and it could possibly be cached
148
# next time.
149
self.danger_count += 1
150
if cache_fp:
151
mutter("remove outdated entry for %s" % path)
152
self.needs_write = True
153
del self._cache[path]
154
elif (fp != cache_fp) or (digest != cache_sha1):
155
mutter("update entry for %s" % path)
156
mutter(" %r" % (fp,))
157
mutter(" %r" % (cache_fp,))
174
158
self.needs_write = True
175
del self._cache[path]
176
else:
177
self.update_count += 1
178
self.needs_write = True
179
self._cache[path] = (digest, file_fp)
180
181
return digest
182
159
self._cache[path] = (digest, fp)
160
161
return digest
183
162
184
163
185
164
215
194
216
195
If the cache file has the wrong version marker, this just clears
217
196
the cache."""
197
from bzrlib.trace import mutter, warning
198
218
199
self._cache = {}
219
200
220
201
fn = self.cache_file_name()
Loggerhead is a web-based interface for Breezy
Version: 2.0.1