def testOutOfOrderRead(self):
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
ent = src.next()
# Advance to the next entry, the first one should no longer be able to
# read
src.next()
self.assertRaises(cpiostream.OutOfOrderRead, ent.payload.read)
def testIterateAndReadAlternate(self):
# We only read every other file - this tests that we are properly
# rolling the cpio stream forward
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
for i, ent in enumerate(src):
if i % 2 == 0:
ent.payload.read()
def testOutOfOrderRead2(self):
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
ent = src.next()
# The cpio stream advances one byte. This should be enough to kill the
# reads for the entry
src.read(1)
self.assertRaises(cpiostream.OutOfOrderRead, ent.payload.read)
def testIterateAndPartialRead2(self):
# Read only a portion of the payload
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
for i, ent in enumerate(src):
if i % 2 == 0:
amt = min(10, ent.header.filesize)
else:
amt = min(155, ent.header.filesize)
ent.payload.read(amt)
def testIterate(self):
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
fileNames = [ x.filename for x in src ]
self.assertEqual(fileNames,
[
'./usr/bin/ptar',
'./usr/bin/ptardiff',
'./usr/lib/perl5/5.10.0/Archive/Tar',
'./usr/lib/perl5/5.10.0/Archive/Tar.pm',
'./usr/lib/perl5/5.10.0/Archive/Tar/Constant.pm',
'./usr/lib/perl5/5.10.0/Archive/Tar/File.pm',
'./usr/share/man/man1/ptar.1.gz',
'./usr/share/man/man1/ptardiff.1.gz',
'./usr/share/man/man3/Archive::Tar.3pm.gz',
'./usr/share/man/man3/Archive::Tar::File.3pm.gz',
])
def oldtest(self):
cpioPath = self._createCpio()
sha1sum = digestlib.sha1(file(cpioPath).read()).hexdigest()
resultFilePath = os.path.join(self.workDir, 'result.cpio')
# Use a variety of sizes, to try to come up with different chunking
# solutions
for bufferSize in [1001, 1003, 3001]:
f = file(resultFilePath, "w")
src = cpiostream.CpioStream(file(cpioPath))
while 1:
buf = src.read(bufferSize)
if not buf:
break
f.write(buf)
f.close()
nsha1sum = digestlib.sha1(file(resultFilePath).read()).hexdigest()
self.assertEqual(nsha1sum, sha1sum)
def extractFilesFromCpio(fileIn, fileList, tmpDir='/tmp'):
"""
Returns a list of open files parallel to fileList
Hardlinked files will share contents, so make sure you seek() back to the
beginning before you read.
"""
# Map device/inode to catch hardlinks
inodeMap = {}
# Map the path in fileList to header and device/inode
fileNameMap = dict((_normpath(x), x) for x in fileList)
fileNameInodeMap = {}
# Empty files will be shared to avoid consuming fd
EmptyFile = tempfile.TemporaryFile(dir=tmpDir, prefix='tmppayload-')
cpioObj = cpiostream.CpioStream(fileIn)
for entry in cpioObj:
if entry.header.mode & 0170000 != 0100000:
# Not a regular file
continue
fileName = _normpath(entry.filename)
devmajor = entry.header.devmajor
devminor = entry.header.devminor
inode = entry.header.inode
key = (devmajor, devminor, inode)
# This file may not be the one we're looking for, but it may be the
# one that provides the contents for hardlinked files we care about
if fileName not in fileNameMap and key not in inodeMap:
continue
if entry.header.filesize == 0:
fobj = EmptyFile
else:
fobj = tempfile.TemporaryFile(dir=tmpDir, prefix='tmppayload-')
util.copyfileobj(entry.payload, fobj)
fobj.seek(0)
inodeMap[key] = fobj
# in case we'll ever want to use the information from the cpio header
# entry to restore file permissions, we should also save the header
# here
fileNameInodeMap[fileName] = key
def testIterateAndRead(self):
cpioPath = self._createCpio()
src = cpiostream.CpioStream(file(cpioPath))
for ent in src:
ent.payload.read()