def process_file(self, path, st, cache):
safe_path = make_path_safe(path)
# Is it a hard link?
if st.st_nlink > 1:
source = self.hard_links.get((st.st_ino, st.st_dev))
if (st.st_ino, st.st_dev) in self.hard_links:
item = self.stat_attrs(st, path)
item.update({b'path': safe_path, b'source': source})
self.add_item(item)
return
else:
self.hard_links[st.st_ino, st.st_dev] = safe_path
path_hash = self.key.id_hash(os.path.join(self.cwd, path).encode('utf-8', 'surrogateescape'))
ids = cache.file_known_and_unchanged(path_hash, st)
chunks = None
if ids is not None:
# Make sure all ids are available
for id_ in ids:
if not cache.seen_chunk(id_):
break
else:
chunks = [cache.chunk_incref(id_, self.stats) for id_ in ids]
# Only chunkify the file if needed
if chunks is None:
with open(path, 'rb') as fd:
chunks = []
for chunk in chunkify(fd, WINDOW_SIZE, CHUNK_MASK, CHUNK_MIN, self.key.chunk_seed):
chunks.append(cache.add_chunk(self.key.id_hash(chunk), chunk, self.stats))
cache.memorize_file(path_hash, st, [c[0] for c in chunks])
item = {b'path': safe_path, b'chunks': chunks}
item.update(self.stat_attrs(st, path))
self.stats.nfiles += 1
self.add_item(item)
def process_file(self, path, st, cache):
safe_path = make_path_safe(path)
# Is it a hard link?
if st.st_nlink > 1:
source = self.hard_links.get((st.st_ino, st.st_dev))
if (st.st_ino, st.st_dev) in self.hard_links:
item = self.stat_attrs(st, path)
item.update({b'path': safe_path, b'source': source})
self.add_item(item)
return
else:
self.hard_links[st.st_ino, st.st_dev] = safe_path
path_hash = self.key.id_hash(os.path.join(self.cwd, path).encode('utf-8', 'surrogateescape'))
ids = cache.file_known_and_unchanged(path_hash, st)
chunks = None
if ids is not None:
# Make sure all ids are available
for id_ in ids:
if not cache.seen_chunk(id_):
break
else:
chunks = [cache.chunk_incref(id_, self.stats) for id_ in ids]
# Only chunkify the file if needed
if chunks is None:
with open(path, 'rb') as fd:
chunks = []
for chunk in chunkify(fd, WINDOW_SIZE, CHUNK_MASK, CHUNK_MIN, self.key.chunk_seed):
chunks.append(cache.add_chunk(self.key.id_hash(chunk), chunk, self.stats))
cache.memorize_file(path_hash, st, [c[0] for c in chunks])
item = {b'path': safe_path, b'chunks': chunks}
item.update(self.stat_attrs(st, path))
self.stats.nfiles += 1
self.add_item(item)
def test(self):
self.assert_equal(make_path_safe('/foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/f/bar'), 'f/bar')
self.assert_equal(make_path_safe('fo/bar'), 'fo/bar')
self.assert_equal(make_path_safe('../foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('../../foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/'), '.')
self.assert_equal(make_path_safe('/'), '.')
def test(self):
self.assert_equal(make_path_safe('/foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/f/bar'), 'f/bar')
self.assert_equal(make_path_safe('fo/bar'), 'fo/bar')
self.assert_equal(make_path_safe('../foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('../../foo/bar'), 'foo/bar')
self.assert_equal(make_path_safe('/'), '.')
self.assert_equal(make_path_safe('/'), '.')
def process_dev(self, path, st):
item = {b'path': make_path_safe(path), b'rdev': st.st_rdev}
item.update(self.stat_attrs(st, path))
self.add_item(item)
if stat.S_ISCHR(st.st_mode):
return 'c' # char device
elif stat.S_ISBLK(st.st_mode):
return 'b' # block device
def process_file(self, path, st, cache):
status = None
safe_path = make_path_safe(path)
# Is it a hard link?
if st.st_nlink > 1:
source = self.hard_links.get((st.st_ino, st.st_dev))
if (st.st_ino, st.st_dev) in self.hard_links:
item = self.stat_attrs(st, path)
item.update({b'path': safe_path, b'source': source})
self.add_item(item)
status = 'h' # regular file, hardlink (to already seen inodes)
return status
else:
self.hard_links[st.st_ino, st.st_dev] = safe_path
path_hash = self.key.id_hash(os.path.join(self.cwd, path).encode('utf-8', 'surrogateescape'))
ids = cache.file_known_and_unchanged(path_hash, st)
chunks = None
if ids is not None:
# Make sure all ids are available
for id_ in ids:
if not cache.seen_chunk(id_):
break
else:
chunks = [cache.chunk_incref(id_, self.stats) for id_ in ids]
status = 'U' # regular file, unchanged
else:
status = 'A' # regular file, added
# Only chunkify the file if needed
if chunks is None:
fh = Archive._open_rb(path, st)
with os.fdopen(fh, 'rb') as fd:
chunks = []
for chunk in self.chunker.chunkify(fd, fh):
chunks.append(cache.add_chunk(self.key.id_hash(chunk), chunk, self.stats))
cache.memorize_file(path_hash, st, [c[0] for c in chunks])
status = status or 'M' # regular file, modified (if not 'A' already)
item = {b'path': safe_path, b'chunks': chunks}
item.update(self.stat_attrs(st, path))
self.stats.nfiles += 1
self.add_item(item)
return status
def process_symlink(self, path, st):
source = os.readlink(path)
item = {b'path': make_path_safe(path), b'source': source}
item.update(self.stat_attrs(st, path))
self.add_item(item)
def process_dev(self, path, st):
item = {b'path': make_path_safe(path), b'rdev': st.st_rdev}
item.update(self.stat_attrs(st, path))
self.add_item(item)
def process_item(self, path, st):
item = {b'path': make_path_safe(path)}
item.update(self.stat_attrs(st, path))
self.add_item(item)
def process_dir(self, path, st):
item = {b'path': make_path_safe(path)}
item.update(self.stat_attrs(st, path))
self.add_item(item)
return 'd' # directory
def process_symlink(self, path, st):
source = os.readlink(path)
item = {b'path': make_path_safe(path), b'source': source}
item.update(self.stat_attrs(st, path))
self.add_item(item)
def process_dev(self, path, st):
item = {b'path': make_path_safe(path), b'rdev': st.st_rdev}
item.update(self.stat_attrs(st, path))
self.add_item(item)
def process_item(self, path, st):
item = {b'path': make_path_safe(path)}
item.update(self.stat_attrs(st, path))
self.add_item(item)
