"""Helper functions and classes for bup."""
+from __future__ import absolute_import, division
from collections import namedtuple
from contextlib import contextmanager
from ctypes import sizeof, c_void_p
+from math import floor
from os import environ
-from pipes import quote
from subprocess import PIPE, Popen
import sys, os, pwd, subprocess, errno, socket, select, mmap, stat, re, struct
import hashlib, heapq, math, operator, time, grp, tempfile
from bup import _helpers
from bup import compat
+from bup.compat import argv_bytes, byte_int
+from bup.io import byte_stream, path_msg
# This function should really be in helpers, not in bup.options. But we
# want options.py to be standalone so people can include it in other projects.
from bup.options import _tty_width as tty_width
if sc_arg_max == -1: # "no definite limit" - let's choose 2M
sc_arg_max = 2 * 1024 * 1024
-
def last(iterable):
result = None
for result in iterable:
def atoi(s):
- """Convert the string 's' to an integer. Return 0 if s is not a number."""
+ """Convert s (ascii bytes) to an integer. Return 0 if s is not a number."""
try:
- return int(s or '0')
+ return int(s or b'0')
except ValueError:
return 0
def atof(s):
- """Convert the string 's' to a float. Return 0 if s is not a number."""
+ """Convert s (ascii bytes) to a float. Return 0 if s is not a number."""
try:
- return float(s or '0')
+ return float(s or b'0')
except ValueError:
return 0
return (leading_matches(), rest())
+def merge_dict(*xs):
+ result = {}
+ for x in xs:
+ result.update(x)
+ return result
+
+
def lines_until_sentinel(f, sentinel, ex_type):
# sentinel must end with \n and must contain only one \n
while True:
line = f.readline()
- if not (line and line.endswith('\n')):
+ if not (line and line.endswith(b'\n')):
raise ex_type('Hit EOF while reading line')
if line == sentinel:
return
"""Print a log message to stderr."""
global _last_prog
sys.stdout.flush()
- _hard_write(sys.stderr.fileno(), s)
+ _hard_write(sys.stderr.fileno(), s if isinstance(s, bytes) else s.encode())
_last_prog = 0
raise
+class MergeIterItem:
+ def __init__(self, entry, read_it):
+ self.entry = entry
+ self.read_it = read_it
+ def __lt__(self, x):
+ return self.entry < x.entry
+
def merge_iter(iters, pfreq, pfunc, pfinal, key=None):
if key:
samekey = lambda e, pe: getattr(e, key) == getattr(pe, key, None)
total = sum(len(it) for it in iters)
iters = (iter(it) for it in iters)
heap = ((next(it, None),it) for it in iters)
- heap = [(e,it) for e,it in heap if e]
+ heap = [MergeIterItem(e, it) for e, it in heap if e]
heapq.heapify(heap)
pe = None
while heap:
if not count % pfreq:
pfunc(count, total)
- e, it = heap[0]
+ e, it = heap[0].entry, heap[0].read_it
if not samekey(e, pe):
pe = e
yield e
except StopIteration:
heapq.heappop(heap) # remove current
else:
- heapq.heapreplace(heap, (e, it)) # shift current to new location
+ # shift current to new location
+ heapq.heapreplace(heap, MergeIterItem(e, it))
pfinal(count, total)
raise
+_bq_simple_id_rx = re.compile(br'^[-_./a-zA-Z0-9]+$')
+_sq_simple_id_rx = re.compile(r'^[-_./a-zA-Z0-9]+$')
+
+def bquote(x):
+ if x == b'':
+ return b"''"
+ if _bq_simple_id_rx.match(x):
+ return x
+ return b"'%s'" % x.replace(b"'", b"'\"'\"'")
+
+def squote(x):
+ if x == '':
+ return "''"
+ if _sq_simple_id_rx.match(x):
+ return x
+ return "'%s'" % x.replace("'", "'\"'\"'")
+
+def quote(x):
+ if isinstance(x, bytes):
+ return bquote(x)
+ if isinstance(x, compat.str_type):
+ return squote(x)
+ assert False
+
def shstr(cmd):
- if isinstance(cmd, compat.str_type):
+ """Return a shell quoted string for cmd if it's a sequence, else cmd.
+
+ cmd must be a string, bytes, or a sequence of one or the other,
+ and the assumption is that if cmd is a string or bytes, then it's
+ already quoted (because it's what's actually being passed to
+ call() and friends. e.g. log(shstr(cmd)); call(cmd)
+
+ """
+ if isinstance(cmd, (bytes, compat.str_type)):
return cmd
- else:
- return ' '.join(map(quote, cmd))
+ elif all(isinstance(x, bytes) for x in cmd):
+ return b' '.join(map(bquote, cmd))
+ elif all(isinstance(x, compat.str_type) for x in cmd):
+ return ' '.join(map(squote, cmd))
+ raise TypeError('unsupported shstr argument: ' + repr(cmd))
+
exc = subprocess.check_call
stderr=None,
shell=False,
check=True,
- preexec_fn=None):
+ preexec_fn=None,
+ close_fds=True):
if input:
assert stdin in (None, PIPE)
stdin = PIPE
p = Popen(cmd,
stdin=stdin, stdout=PIPE, stderr=stderr,
shell=shell,
- preexec_fn=preexec_fn)
+ preexec_fn=preexec_fn,
+ close_fds=close_fds)
out, err = p.communicate(input)
if check and p.returncode != 0:
- raise Exception('subprocess %r failed with status %d, stderr: %r'
- % (' '.join(map(quote, cmd)), p.returncode, err))
+ raise Exception('subprocess %r failed with status %d%s'
+ % (b' '.join(map(quote, cmd)), p.returncode,
+ ', stderr: %r' % err if err else ''))
return out, err, p
def readpipe(argv, preexec_fn=None, shell=False):
"""Run a subprocess and return its output."""
- p = subprocess.Popen(argv, stdout=subprocess.PIPE, preexec_fn=preexec_fn,
- shell=shell)
- out, err = p.communicate()
- if p.returncode != 0:
- raise Exception('subprocess %r failed with status %d'
- % (' '.join(argv), p.returncode))
- return out
+ return exo(argv, preexec_fn=preexec_fn, shell=shell)[0]
def _argmax_base(command):
base_size = 2048
for c in command:
base_size += len(command) + 1
- for k, v in environ.iteritems():
+ for k, v in compat.items(environ):
base_size += len(k) + len(v) + 2 + sizeof(c_void_p)
return base_size
return os.getenv("FAKEROOTKEY") != None
-_warned_about_superuser_detection = None
-def is_superuser():
- if sys.platform.startswith('cygwin'):
- if sys.getwindowsversion()[0] > 5:
- # Sounds like situation is much more complicated here
- global _warned_about_superuser_detection
- if not _warned_about_superuser_detection:
- log("can't detect root status for OS version > 5; assuming not root")
- _warned_about_superuser_detection = True
- return False
- import ctypes
- return ctypes.cdll.shell32.IsUserAnAdmin()
- else:
+if sys.platform.startswith('cygwin'):
+ def is_superuser():
+ # https://cygwin.com/ml/cygwin/2015-02/msg00057.html
+ groups = os.getgroups()
+ return 544 in groups or 0 in groups
+else:
+ def is_superuser():
return os.geteuid() == 0
-def _cache_key_value(get_value, key, cache):
+def cache_key_value(get_value, key, cache):
"""Return (value, was_cached). If there is a value in the cache
for key, use that, otherwise, call get_value(key) which should
throw a KeyError if there is no value -- in which case the cached
return value, False
-_uid_to_pwd_cache = {}
-_name_to_pwd_cache = {}
-
-def pwd_from_uid(uid):
- """Return password database entry for uid (may be a cached value).
- Return None if no entry is found.
- """
- global _uid_to_pwd_cache, _name_to_pwd_cache
- entry, cached = _cache_key_value(pwd.getpwuid, uid, _uid_to_pwd_cache)
- if entry and not cached:
- _name_to_pwd_cache[entry.pw_name] = entry
- return entry
-
-
-def pwd_from_name(name):
- """Return password database entry for name (may be a cached value).
- Return None if no entry is found.
- """
- global _uid_to_pwd_cache, _name_to_pwd_cache
- entry, cached = _cache_key_value(pwd.getpwnam, name, _name_to_pwd_cache)
- if entry and not cached:
- _uid_to_pwd_cache[entry.pw_uid] = entry
- return entry
-
-
-_gid_to_grp_cache = {}
-_name_to_grp_cache = {}
-
-def grp_from_gid(gid):
- """Return password database entry for gid (may be a cached value).
- Return None if no entry is found.
- """
- global _gid_to_grp_cache, _name_to_grp_cache
- entry, cached = _cache_key_value(grp.getgrgid, gid, _gid_to_grp_cache)
- if entry and not cached:
- _name_to_grp_cache[entry.gr_name] = entry
- return entry
-
-
-def grp_from_name(name):
- """Return password database entry for name (may be a cached value).
- Return None if no entry is found.
- """
- global _gid_to_grp_cache, _name_to_grp_cache
- entry, cached = _cache_key_value(grp.getgrnam, name, _name_to_grp_cache)
- if entry and not cached:
- _gid_to_grp_cache[entry.gr_gid] = entry
- return entry
-
-
-_username = None
-def username():
- """Get the user's login name."""
- global _username
- if not _username:
- uid = os.getuid()
- _username = pwd_from_uid(uid)[0] or 'user%d' % uid
- return _username
-
-
-_userfullname = None
-def userfullname():
- """Get the user's full name."""
- global _userfullname
- if not _userfullname:
- uid = os.getuid()
- entry = pwd_from_uid(uid)
- if entry:
- _userfullname = entry[4].split(',')[0] or entry[0]
- if not _userfullname:
- _userfullname = 'user%d' % uid
- return _userfullname
-
-
_hostname = None
def hostname():
"""Get the FQDN of this machine."""
global _hostname
if not _hostname:
- _hostname = socket.getfqdn()
+ _hostname = socket.getfqdn().encode('iso-8859-1')
return _hostname
-_resource_path = None
-def resource_path(subdir=''):
- global _resource_path
- if not _resource_path:
- _resource_path = os.environ.get('BUP_RESOURCE_PATH') or '.'
- return os.path.join(_resource_path, subdir)
-
def format_filesize(size):
unit = 1024.0
size = float(size)
if size < unit:
return "%d" % (size)
- exponent = int(math.log(size) / math.log(unit))
+ exponent = int(math.log(size) // math.log(unit))
size_prefix = "KMGTPE"[exponent - 1]
return "%.1f%s" % (size / math.pow(unit, exponent), size_prefix)
def close(self):
while self._read(65536): pass
+ def _read(self, size):
+ raise NotImplementedError("Subclasses must implement _read")
+
def read(self, size):
"""Read 'size' bytes from input stream."""
self.outp.flush()
return self._read(size)
+ def _readline(self, size):
+ raise NotImplementedError("Subclasses must implement _readline")
+
def readline(self):
"""Read from input stream until a newline is found."""
self.outp.flush()
def has_input(self):
"""Return true if input stream is readable."""
- raise NotImplemented("Subclasses must implement has_input")
+ raise NotImplementedError("Subclasses must implement has_input")
def ok(self):
"""Indicate end of output from last sent command."""
- self.write('\nok\n')
+ self.write(b'\nok\n')
def error(self, s):
"""Indicate server error to the client."""
- s = re.sub(r'\s+', ' ', str(s))
- self.write('\nerror %s\n' % s)
+ s = re.sub(br'\s+', b' ', s)
+ self.write(b'\nerror %s\n' % s)
def _check_ok(self, onempty):
self.outp.flush()
- rl = ''
+ rl = b''
for rl in linereader(self):
#log('%d got line: %r\n' % (os.getpid(), rl))
if not rl: # empty line
continue
- elif rl == 'ok':
+ elif rl == b'ok':
return None
- elif rl.startswith('error '):
+ elif rl.startswith(b'error '):
#log('client: error: %s\n' % rl[6:])
return NotOk(rl[6:])
else:
BaseConn.__init__(self, outp)
# Anything that comes through before the sync string was not
# multiplexed and can be assumed to be debug/log before mux init.
- tail = ''
- while tail != 'BUPMUX':
+ tail = b''
+ while tail != b'BUPMUX':
b = os.read(infd, (len(tail) < 6) and (6-len(tail)) or 1)
if not b:
raise IOError('demux: unexpected EOF during initialization')
tail += b
- sys.stderr.write(tail[:-6]) # pre-mux log messages
+ byte_stream(sys.stderr).write(tail[:-6]) # pre-mux log messages
tail = tail[-6:]
self.infd = infd
self.reader = None
rl, wl, xl = select.select([self.infd], [], [], timeout)
if not rl: return False
assert(rl[0] == self.infd)
- ns = ''.join(checked_reader(self.infd, 5))
+ ns = b''.join(checked_reader(self.infd, 5))
n, fdw = struct.unpack('!IB', ns)
assert(n <= MAX_PACKET)
if fdw == 1:
self.reader = checked_reader(self.infd, n)
elif fdw == 2:
for buf in checked_reader(self.infd, n):
- sys.stderr.write(buf)
+ byte_stream(sys.stderr).write(buf)
elif fdw == 3:
self.closed = True
debug2("DemuxConn: marked closed\n")
def _readline(self):
def find_eol(buf):
try:
- return buf.index('\n')+1
+ return buf.index(b'\n')+1
except ValueError:
return None
- return ''.join(self._read_parts(find_eol))
+ return b''.join(self._read_parts(find_eol))
def _read(self, size):
csize = [size]
return None
else:
return csize[0]
- return ''.join(self._read_parts(until_size))
+ return b''.join(self._read_parts(until_size))
def has_input(self):
return self._load_buf(0)
def slashappend(s):
"""Append "/" to 's' if it doesn't aleady end in "/"."""
- if s and not s.endswith('/'):
- return s + '/'
+ assert isinstance(s, bytes)
+ if s and not s.endswith(b'/'):
+ return s + b'/'
else:
return s
pref_chunk_size = 64 * 1024 * 1024
chunk_size = sc_page_size
if (sc_page_size < pref_chunk_size):
- chunk_size = sc_page_size * (pref_chunk_size / sc_page_size)
+ chunk_size = sc_page_size * (pref_chunk_size // sc_page_size)
_fmincore_chunk_size = chunk_size
def fmincore(fd):
return bytearray(0)
if not _fmincore_chunk_size:
_set_fmincore_chunk_size()
- pages_per_chunk = _fmincore_chunk_size / sc_page_size;
- page_count = (st.st_size + sc_page_size - 1) / sc_page_size;
- chunk_count = page_count / _fmincore_chunk_size
- if chunk_count < 1:
- chunk_count = 1
+ pages_per_chunk = _fmincore_chunk_size // sc_page_size;
+ page_count = (st.st_size + sc_page_size - 1) // sc_page_size;
+ chunk_count = (st.st_size + _fmincore_chunk_size - 1) // _fmincore_chunk_size
result = bytearray(page_count)
- for ci in xrange(chunk_count):
+ for ci in compat.range(chunk_count):
pos = _fmincore_chunk_size * ci;
msize = min(_fmincore_chunk_size, st.st_size - pos)
try:
def parse_num(s):
- """Parse data size information into a float number.
+ """Parse string or bytes as a possibly unit suffixed number.
- Here are some examples of conversions:
+ For example:
199.2k means 203981 bytes
1GB means 1073741824 bytes
2.1 tb means 2199023255552 bytes
"""
+ if isinstance(s, bytes):
+ # FIXME: should this raise a ValueError for UnicodeDecodeError
+ # (perhaps with the latter as the context).
+ s = s.decode('ascii')
g = re.match(r'([-+\d.e]+)\s*(\w*)', str(s))
if not g:
raise ValueError("can't parse %r as a number" % s)
return int(num*mult)
-def count(l):
- """Count the number of elements in an iterator. (consumes the iterator)"""
- return reduce(lambda x,y: x+1, l)
-
-
saved_errors = []
def add_error(e):
"""Append an error message to the list of saved errors.
The number of columns is determined automatically based on the string
lengths.
"""
+ binary = isinstance(prefix, bytes)
+ nothing = b'' if binary else ''
+ nl = b'\n' if binary else '\n'
if not l:
- return ""
+ return nothing
l = l[:]
clen = max(len(s) for s in l)
ncols = (tty_width() - len(prefix)) // (clen + 2)
clen = 0
cols = []
while len(l) % ncols:
- l.append('')
+ l.append(nothing)
rows = len(l) // ncols
- for s in range(0, len(l), rows):
+ for s in compat.range(0, len(l), rows):
cols.append(l[s:s+rows])
- out = ''
+ out = nothing
+ fmt = b'%-*s' if binary else '%-*s'
for row in zip(*cols):
- out += prefix + ''.join(('%-*s' % (clen+2, s)) for s in row) + '\n'
+ out += prefix + nothing.join((fmt % (clen+2, s)) for s in row) + nl
return out
for flag in options:
(option, parameter) = flag
if option == '--exclude':
- excluded_paths.append(resolve_parent(parameter))
+ excluded_paths.append(resolve_parent(argv_bytes(parameter)))
elif option == '--exclude-from':
try:
- f = open(resolve_parent(parameter))
+ f = open(resolve_parent(argv_bytes(parameter)), 'rb')
except IOError as e:
- raise fatal("couldn't read %s" % parameter)
+ raise fatal("couldn't read %r" % parameter)
for exclude_path in f.readlines():
# FIXME: perhaps this should be rstrip('\n')
exclude_path = resolve_parent(exclude_path.strip())
(option, parameter) = flag
if option == '--exclude-rx':
try:
- excluded_patterns.append(re.compile(parameter))
+ excluded_patterns.append(re.compile(argv_bytes(parameter)))
except re.error as ex:
- fatal('invalid --exclude-rx pattern (%s): %s' % (parameter, ex))
+ fatal('invalid --exclude-rx pattern (%r): %s' % (parameter, ex))
elif option == '--exclude-rx-from':
try:
- f = open(resolve_parent(parameter))
+ f = open(resolve_parent(parameter), 'rb')
except IOError as e:
- raise fatal("couldn't read %s" % parameter)
+ raise fatal("couldn't read %r" % parameter)
for pattern in f.readlines():
- spattern = pattern.rstrip('\n')
+ spattern = pattern.rstrip(b'\n')
if not spattern:
continue
try:
excluded_patterns.append(re.compile(spattern))
except re.error as ex:
- fatal('invalid --exclude-rx pattern (%s): %s' % (spattern, ex))
+ fatal('invalid --exclude-rx pattern (%r): %s' % (spattern, ex))
return excluded_patterns
full_path_to_name). Path must start with '/'.
Example:
'/home/foo' -> [('', '/'), ('home', '/home'), ('foo', '/home/foo')]"""
- if not path.startswith('/'):
- raise Exception('path must start with "/": %s' % path)
+ if not path.startswith(b'/'):
+ raise Exception('path must start with "/": %s' % path_msg(path))
# Since we assume path startswith('/'), we can skip the first element.
- result = [('', '/')]
+ result = [(b'', b'/')]
norm_path = os.path.abspath(path)
- if norm_path == '/':
+ if norm_path == b'/':
return result
- full_path = ''
- for p in norm_path.split('/')[1:]:
- full_path += '/' + p
+ full_path = b''
+ for p in norm_path.split(b'/')[1:]:
+ full_path += b'/' + p
result.append((p, full_path))
return result
sorted_strip_prefixes = sorted(strip_prefixes, key=len, reverse=True)
for bp in sorted_strip_prefixes:
normalized_bp = os.path.abspath(bp)
- if normalized_bp == '/':
+ if normalized_bp == b'/':
continue
if normalized_path.startswith(normalized_bp):
prefix = normalized_path[:len(normalized_bp)]
result = []
- for p in normalized_path[len(normalized_bp):].split('/'):
+ for p in normalized_path[len(normalized_bp):].split(b'/'):
if p: # not root
- prefix += '/'
+ prefix += b'/'
prefix += p
result.append((p, prefix))
return result
new_prefix = os.path.normpath(new_prefix)
if clean_path.startswith(old_prefix):
escaped_prefix = re.escape(old_prefix)
- grafted_path = re.sub(r'^' + escaped_prefix, new_prefix, clean_path)
+ grafted_path = re.sub(br'^' + escaped_prefix, new_prefix, clean_path)
# Handle /foo=/ (at least) -- which produces //whatever.
- grafted_path = '/' + grafted_path.lstrip('/')
+ grafted_path = b'/' + grafted_path.lstrip(b'/')
clean_path_components = path_components(clean_path)
# Count the components that were stripped.
- strip_count = 0 if old_prefix == '/' else old_prefix.count('/')
- new_prefix_parts = new_prefix.split('/')
- result_prefix = grafted_path.split('/')[:new_prefix.count('/')]
+ strip_count = 0 if old_prefix == b'/' else old_prefix.count(b'/')
+ new_prefix_parts = new_prefix.split(b'/')
+ result_prefix = grafted_path.split(b'/')[:new_prefix.count(b'/')]
result = [(p, None) for p in result_prefix] \
+ clean_path_components[strip_count:]
# Now set the graft point name to match the end of new_prefix.
graft_point = len(result_prefix)
result[graft_point] = \
(new_prefix_parts[-1], clean_path_components[strip_count][1])
- if new_prefix == '/': # --graft ...=/ is a special case.
+ if new_prefix == b'/': # --graft ...=/ is a special case.
return result[1:]
return result
return path_components(clean_path)
# module, which doesn't appear willing to ignore the extra items.
if _localtime:
def localtime(time):
- return bup_time(*_helpers.localtime(time))
+ return bup_time(*_helpers.localtime(floor(time)))
def utc_offset_str(t):
"""Return the local offset from UTC as "+hhmm" or "-hhmm" for time t.
If the current UTC offset does not represent an integer number
offmin = abs(off) // 60
m = offmin % 60
h = (offmin - m) // 60
- return "%+03d%02d" % (-h if off < 0 else h, m)
+ return b'%+03d%02d' % (-h if off < 0 else h, m)
def to_py_time(x):
if isinstance(x, time.struct_time):
return x
else:
localtime = time.localtime
def utc_offset_str(t):
- return time.strftime('%z', localtime(t))
+ return time.strftime(b'%z', localtime(t))
def to_py_time(x):
return x
-_some_invalid_save_parts_rx = re.compile(r'[[ ~^:?*\\]|\.\.|//|@{')
+_some_invalid_save_parts_rx = re.compile(br'[\[ ~^:?*\\]|\.\.|//|@{')
def valid_save_name(name):
# Enforce a superset of the restrictions in git-check-ref-format(1)
- if name == '@' \
- or name.startswith('/') or name.endswith('/') \
- or name.endswith('.'):
+ if name == b'@' \
+ or name.startswith(b'/') or name.endswith(b'/') \
+ or name.endswith(b'.'):
return False
if _some_invalid_save_parts_rx.search(name):
return False
for c in name:
- if ord(c) < 0x20 or ord(c) == 0x7f:
+ if byte_int(c) < 0x20 or byte_int(c) == 0x7f:
return False
- for part in name.split('/'):
- if part.startswith('.') or part.endswith('.lock'):
+ for part in name.split(b'/'):
+ if part.startswith(b'.') or part.endswith(b'.lock'):
return False
return True
-_period_rx = re.compile(r'^([0-9]+)(s|min|h|d|w|m|y)$')
+_period_rx = re.compile(br'^([0-9]+)(s|min|h|d|w|m|y)$')
def period_as_secs(s):
- if s == 'forever':
+ if s == b'forever':
return float('inf')
match = _period_rx.match(s)
if not match:
return None
mag = int(match.group(1))
scale = match.group(2)
- return mag * {'s': 1,
- 'min': 60,
- 'h': 60 * 60,
- 'd': 60 * 60 * 24,
- 'w': 60 * 60 * 24 * 7,
- 'm': 60 * 60 * 24 * 31,
- 'y': 60 * 60 * 24 * 366}[scale]
+ return mag * {b's': 1,
+ b'min': 60,
+ b'h': 60 * 60,
+ b'd': 60 * 60 * 24,
+ b'w': 60 * 60 * 24 * 7,
+ b'm': 60 * 60 * 24 * 31,
+ b'y': 60 * 60 * 24 * 366}[scale]
projects / bup.git / blobdiff