def md5_file_checksum_failure(filename):
try:
with open("/dev/null", "a") as devnull:
with util.archive_created(filename, "/bin", "--file-cksum", "md5", stderr=devnull) as path:
raise AssertionError("xar succeeded when it should have failed")
except subprocess.CalledProcessError:
pass
def _test_xattr_on_file_with_contents(filename, file_contents, xattrs=[], xar_create_flags=[], xar_extract_flags=[]):
try:
# Write file out
with open(filename, "w") as f:
f.write(file_contents)
for (key, value) in xattrs:
xattr.setxattr(f, key, value)
# Store it into a xarchive
archive_name = "{f}.xar".format(f=filename)
with util.archive_created(archive_name, filename, *xar_create_flags) as path:
# Extract xarchive
with util.directory_created("extracted") as directory:
# Validate resulting xattrs
subprocess.check_call(["xar", "-x", "-C", directory, "-f", path] + xar_extract_flags)
for (key, value) in xattrs:
try:
assert xattr.getxattr(os.path.join(directory, filename), key) == value, "extended attribute \"{n}\" has incorrect contents after extraction".format(n=key)
except KeyError:
raise MissingExtendedAttributeError("extended attribute \"{n}\" missing after extraction".format(n=key))
# Validate file contents
with open(os.path.join(directory, filename), "r") as f:
if f.read() != file_contents:
raise MissingExtendedAttributeError("archived file \"{f}\" has has incorrect contents after extraction".format(f=filename))
finally:
os.unlink(filename)
def broken_toc_default_checksum(filename): with util.archive_created(filename, "/bin") as path: # Mess up the archive toc_start = _get_header_size(path) _clobber_bytes_at(range(toc_start + 4, toc_start + 4 + 100), path) # Why did the original test specify 4? No idea. # Try to extract it _verify_extraction_failed(filename)
def hard_link_in_cwd(filename):
with util.directory_created("hardlink_scratch") as directory:
with util.chdir(directory):
util.touch("a")
os.link("a", "b")
with util.archive_created(os.path.join("..", "hardlink_archive.xar"), ".") as path:
with util.directory_created(os.path.join("..", "hardlink_extracted")) as extracted:
subprocess.check_call(["xar", "-x", "-C", extracted, "-f", path])
_assert_same_inodes(os.path.join(extracted, "a"), os.path.join(extracted, "b"))
def _data_test(filename, *args, **kwargs):
with util.directory_created("data_scratch") as directory:
yield directory
# Files are now created
with util.archive_created("data.xar", "./data_scratch", *args, **kwargs) as path:
_process_toc(path)
with util.directory_created("data_extracted") as extracted:
subprocess.check_call(["xar", "-x", "-f", path, "-C", extracted])
util.assert_identical_directories(directory, os.path.join(extracted, "data_scratch"))
def broken_heap_default_checksum(filename): with util.archive_created(filename, "/bin") as path: # Mess up the archive toc_start = _get_header_size(path) toc_size = _get_toc_size(path) # Why 32? That's the size of the default sha256 checksum, which is stored before the heap. _clobber_bytes_at(range(toc_start + toc_size + 32, toc_start + toc_size + 32 + 100), path) # Try to extract it _verify_extraction_failed(filename)
def broken_toc_sha1_checksum(filename):
with util.archive_created(filename, "/bin", "--toc-cksum", "sha1") as path:
# Mess up the archive
toc_start = _get_header_size(path)
_clobber_bytes_at(
range(toc_start + 4, toc_start + 4 + 100),
path) # Why did the original test specify 4? No idea.
# Try to extract it
_verify_extraction_failed(filename)
def hard_link_identical_files(filename):
with util.directory_created("hardlink_scratch") as directory:
with open(os.path.join(directory, "a"), "w") as f:
f.write("1234samecontent")
with open(os.path.join(directory, "b"), "w") as f:
f.write("1234samecontent")
with util.archive_created("hardlink_archive.xar", "hardlink_scratch", "--link-same") as path:
with util.directory_created("hardlink_extracted") as extracted:
subprocess.check_call(["xar", "-x", "-C", extracted, "-f", path])
_assert_same_inodes(os.path.join(extracted, "hardlink_scratch", "a"), os.path.join(extracted, "hardlink_scratch", "b"))
def md5_file_checksum_failure(filename):
try:
with open("/dev/null", "a") as devnull:
with util.archive_created(filename,
"/bin",
"--file-cksum",
"md5",
stderr=devnull) as path:
raise AssertionError(
"xar succeeded when it should have failed")
except subprocess.CalledProcessError:
pass
def broken_heap_default_checksum(filename):
with util.archive_created(filename, "/bin") as path:
# Mess up the archive
toc_start = _get_header_size(path)
toc_size = _get_toc_size(path)
# Why 32? That's the size of the default sha256 checksum, which is stored before the heap.
_clobber_bytes_at(
range(toc_start + toc_size + 32, toc_start + toc_size + 32 + 100),
path)
# Try to extract it
_verify_extraction_failed(filename)
def default_checksum_algorithm(filename):
with util.archive_created(filename, "/bin") as path:
header = subprocess.check_output(["xar", "--dump-header", "-f", path])
found = False
for line in header.splitlines():
matchdata = re.match("^Checksum algorithm:\s+(\d+)\s+\\((\w+)\\)$", line)
if not matchdata:
continue
found = True
algorithm = matchdata.groups()[1]
assert algorithm == "SHA1", "unexpected checksum algorithm default: received {a}, expected SHA1".format(a=algorithm)
assert found, "unexpected output from `xar --dump-header`:\n{h}".format(h=header)
def _test_truncation(filename, path_to_be_archived, bytes_to_chop, *args):
with util.archive_created(filename, path_to_be_archived) as path:
with open("/dev/null", "w") as bitbucket:
size = os.stat(path).st_size
while size > 0:
last_size = size
size = max(0, size - bytes_to_chop)
with open(path, "w+") as f:
f.truncate(size)
with util.directory_created("scratch") as directory:
returncode = subprocess.call(["xar", "-x", "-f", path, "-C", directory], stderr=bitbucket)
assert returncode != 0, "xar claimed to succeed when extracting a truncated archive"
def _test_truncation(filename, path_to_be_archived, bytes_to_chop, *args):
with util.archive_created(filename, path_to_be_archived) as path:
with open("/dev/null", "w") as bitbucket:
size = os.stat(path).st_size
while size > 0:
last_size = size
size = max(0, size - bytes_to_chop)
with open(path, "w+") as f:
f.truncate(size)
with util.directory_created("scratch") as directory:
returncode = subprocess.call(["xar", "-x", "-f", path, "-C", directory], stderr=bitbucket)
assert returncode != 0, "xar claimed to succeed when extracting a truncated archive"
def hard_link_in_cwd(filename):
with util.directory_created("hardlink_scratch") as directory:
with util.chdir(directory):
util.touch("a")
os.link("a", "b")
with util.archive_created(
os.path.join("..", "hardlink_archive.xar"), ".") as path:
with util.directory_created(
os.path.join("..", "hardlink_extracted")) as extracted:
subprocess.check_call(
["xar", "-x", "-C", extracted, "-f", path])
_assert_same_inodes(os.path.join(extracted, "a"),
os.path.join(extracted, "b"))
def coalesce_heap(filename):
with util.directory_created("scratch") as directory:
shutil.copy("/bin/ls", os.path.join(directory, "ls"))
shutil.copy(os.path.join(directory, "ls"), os.path.join(directory, "foo"))
with util.chdir(directory):
with util.archive_created(os.path.join("..", "heap.xar"), ".", "--coalesce-heap") as path:
# Verify file offsets are as we expect
offsets = _file_offsets_for_archive(path, os.path.join("..", "heap1.xsl"))
(f1, o1) = offsets[0]
(f2, o2) = offsets[1]
# Make sure extraction goes all right
with util.directory_created("extracted") as extracted:
subprocess.check_call(["xar", "-x", "-f", path, "-C", extracted])
def hard_link_identical_files(filename):
with util.directory_created("hardlink_scratch") as directory:
with open(os.path.join(directory, "a"), "w") as f:
f.write("1234samecontent")
with open(os.path.join(directory, "b"), "w") as f:
f.write("1234samecontent")
with util.archive_created("hardlink_archive.xar", "hardlink_scratch",
"--link-same") as path:
with util.directory_created("hardlink_extracted") as extracted:
subprocess.check_call(
["xar", "-x", "-C", extracted, "-f", path])
_assert_same_inodes(
os.path.join(extracted, "hardlink_scratch", "a"),
os.path.join(extracted, "hardlink_scratch", "b"))
def default_checksum_algorithm(filename):
with util.archive_created(filename, "/bin") as path:
header = subprocess.check_output(["xar", "--dump-header", "-f", path])
found = False
for line in header.splitlines():
matchdata = re.match("^Checksum algorithm:\s+(\d+)\s+\\((\w+)\\)$",
line)
if not matchdata:
continue
found = True
algorithm = matchdata.groups()[1]
assert algorithm == "SHA1", "unexpected checksum algorithm default: received {a}, expected SHA1".format(
a=algorithm)
assert found, "unexpected output from `xar --dump-header`:\n{h}".format(
h=header)
def _verify_checksum_algorithm(filename, algorithm):
additional_args = []
if algorithm:
additional_args = ["--file-cksum", algorithm]
else:
algorithm = "sha1"
with util.archive_created(filename, "/bin", *additional_args) as path:
toc = subprocess.check_output(["xar", "--dump-toc=-", "-f", path])
found = False
for line in toc.splitlines():
if '<unarchived-checksum style="{a}">'.format(a=algorithm) in line or '<archived-checksum style="{a}">'.format(a=algorithm) in line:
break
else:
raise AssertionError("unexpected output from `xar --dump-toc=-`:\n{t}".format(t=toc))
def normal_heap(filename):
with util.directory_created("scratch") as directory:
shutil.copy("/bin/ls", os.path.join(directory, "ls"))
shutil.copy(os.path.join(directory, "ls"), os.path.join(directory, "foo"))
with util.chdir(directory):
with util.archive_created(os.path.join("..", "heap.xar"), ".") as path:
# Verify file offsets are as we expect
offsets = _file_offsets_for_archive(path, os.path.join("..", "heap1.xsl"))
(f1, o1) = offsets[0]
(f2, o2) = offsets[1]
assert o1 < o2, "offset for first file \"{f1}\" ({o1}) greater than or equal to offset for last file \"{f2}\" ({o2})".format(f1=f1, o1=o1, f2=f2, o2=o2)
# Make sure extraction goes all right
with util.directory_created("extracted") as extracted:
subprocess.check_call(["xar", "-x", "-f", path, "-C", extracted])
def _verify_checksum_algorithm(filename, algorithm):
additional_args = []
if algorithm:
additional_args = ["--file-cksum", algorithm]
else:
algorithm = "sha1"
with util.archive_created(filename, "/bin", *additional_args) as path:
toc = subprocess.check_output(["xar", "--dump-toc=-", "-f", path])
found = False
for line in toc.splitlines():
if '<unarchived-checksum style="{a}">'.format(
a=algorithm
) in line or '<archived-checksum style="{a}">'.format(
a=algorithm) in line:
break
else:
raise AssertionError(
"unexpected output from `xar --dump-toc=-`:\n{t}".format(
t=toc))
def _check_compression(filename, *args, **kwargs):
with util.archive_created(filename, "/bin", *args, **kwargs) as path:
with util.directory_created("extracted") as directory:
subprocess.check_call(["xar", "-x", "-f", path, "-C", directory])
util.assert_identical_directories("/bin", os.path.join(directory, "bin"))
def default_toc_checksum_validity(filename):
with util.archive_created(filename, "/bin") as path:
_verify_header_checksum(path, "sha1")
def sha512_toc_checksum_validity(filename):
with util.archive_created(filename, "/bin", "--toc-cksum",
"sha512") as path:
_verify_header_checksum(path, "sha512")
def sha512_toc_checksum_validity(filename): with util.archive_created(filename, "/bin", "--toc-cksum", "sha512") as path: _verify_header_checksum(path, "sha512")
def default_toc_checksum_validity(filename): with util.archive_created(filename, "/bin") as path: _verify_header_checksum(path, "sha1")