+ file.write(item)
+ writer.close()
+
+
+def binary_helper(suite_id, arch_id, overridetype_id, component_id):
+ '''
+ This function is called in a new subprocess and multiprocessing wants a top
+ level function.
+ '''
+ session = DBConn().session(work_mem = 1000)
+ suite = Suite.get(suite_id, session)
+ architecture = Architecture.get(arch_id, session)
+ overridetype = OverrideType.get(overridetype_id, session)
+ component = Component.get(component_id, session)
+ log_message = [suite.suite_name, architecture.arch_string, \
+ overridetype.overridetype, component.component_name]
+ contents_writer = BinaryContentsWriter(suite, architecture, overridetype, component)
+ contents_writer.write_file()
+ session.close()
+ return log_message
+
+def source_helper(suite_id, component_id):
+ '''
+ This function is called in a new subprocess and multiprocessing wants a top
+ level function.
+ '''
+ session = DBConn().session(work_mem = 1000)
+ suite = Suite.get(suite_id, session)
+ component = Component.get(component_id, session)
+ log_message = [suite.suite_name, 'source', component.component_name]
+ contents_writer = SourceContentsWriter(suite, component)
+ contents_writer.write_file()
+ session.close()
+ return log_message
+
+class ContentsWriter(object):
+ '''
+ Loop over all suites, architectures, overridetypes, and components to write
+ all contents files.
+ '''
+ @classmethod
+ def log_result(class_, result):
+ '''
+ Writes a result message to the logfile.
+ '''
+ class_.logger.log(result)
+
+ @classmethod
+ def write_all(class_, logger, archive_names = [], suite_names = [], component_names = [], force = False):
+ '''
+ Writes all Contents files for suites in list suite_names which defaults
+ to all 'touchable' suites if not specified explicitely. Untouchable
+ suites will be included if the force argument is set to True.
+ '''
+ class_.logger = logger
+ session = DBConn().session()
+ suite_query = session.query(Suite)
+ if len(archive_names) > 0:
+ suite_query = suite_query.join(Suite.archive).filter(Archive.archive_name.in_(archive_names))
+ if len(suite_names) > 0:
+ suite_query = suite_query.filter(Suite.suite_name.in_(suite_names))
+ component_query = session.query(Component)
+ if len(component_names) > 0:
+ component_query = component_query.filter(Component.component_name.in_(component_names))
+ if not force:
+ suite_query = suite_query.filter(Suite.untouchable == False)
+ deb_id = get_override_type('deb', session).overridetype_id
+ udeb_id = get_override_type('udeb', session).overridetype_id
+ pool = Pool()
+ for suite in suite_query:
+ suite_id = suite.suite_id
+ for component in component_query:
+ component_id = component.component_id
+ # handle source packages
+ pool.apply_async(source_helper, (suite_id, component_id),
+ callback = class_.log_result)
+ for architecture in suite.get_architectures(skipsrc = True, skipall = True):
+ arch_id = architecture.arch_id
+ # handle 'deb' packages
+ pool.apply_async(binary_helper, (suite_id, arch_id, deb_id, component_id), \
+ callback = class_.log_result)
+ # handle 'udeb' packages
+ pool.apply_async(binary_helper, (suite_id, arch_id, udeb_id, component_id), \
+ callback = class_.log_result)
+ pool.close()
+ pool.join()
+ session.close()
+
+
+class BinaryContentsScanner(object):
+ '''
+ BinaryContentsScanner provides a threadsafe method scan() to scan the
+ contents of a DBBinary object.
+ '''
+ def __init__(self, binary_id):
+ '''
+ The argument binary_id is the id of the DBBinary object that
+ should be scanned.
+ '''
+ self.binary_id = binary_id
+
+ def scan(self, dummy_arg = None):
+ '''
+ This method does the actual scan and fills in the associated BinContents
+ property. It commits any changes to the database. The argument dummy_arg
+ is ignored but needed by our threadpool implementation.
+ '''
+ session = DBConn().session()
+ binary = session.query(DBBinary).get(self.binary_id)
+ fileset = set(binary.scan_contents())
+ if len(fileset) == 0:
+ fileset.add('EMPTY_PACKAGE')
+ for filename in fileset:
+ binary.contents.append(BinContents(file = filename))
+ session.commit()
+ session.close()
+
+ @classmethod
+ def scan_all(class_, limit = None):
+ '''
+ The class method scan_all() scans all binaries using multiple threads.
+ The number of binaries to be scanned can be limited with the limit
+ argument. Returns the number of processed and remaining packages as a
+ dict.
+ '''
+ session = DBConn().session()
+ query = session.query(DBBinary).filter(DBBinary.contents == None)
+ remaining = query.count
+ if limit is not None:
+ query = query.limit(limit)
+ processed = query.count()
+ pool = Pool()
+ for binary in query.yield_per(100):
+ pool.apply_async(binary_scan_helper, (binary.binary_id, ))
+ pool.close()
+ pool.join()
+ remaining = remaining()
+ session.close()
+ return { 'processed': processed, 'remaining': remaining }
+
+def binary_scan_helper(binary_id):
+ '''
+ This function runs in a subprocess.
+ '''
+ scanner = BinaryContentsScanner(binary_id)
+ scanner.scan()
+
+
+def subprocess_setup():
+ # Python installs a SIGPIPE handler by default. This is usually not what
+ # non-Python subprocesses expect.
+ signal.signal(signal.SIGPIPE, signal.SIG_DFL)
+
+class UnpackedSource(object):
+ '''
+ UnpackedSource extracts a source package into a temporary location and
+ gives you some convinient function for accessing it.
+ '''
+ def __init__(self, dscfilename):
+ '''
+ The dscfilename is a name of a DSC file that will be extracted.
+ '''
+ temp_directory = mkdtemp(dir = Config()['Dir::TempPath'])
+ self.root_directory = os.path.join(temp_directory, 'root')
+ command = ('dpkg-source', '--no-copy', '--no-check', '-q', '-x',
+ dscfilename, self.root_directory)
+ check_call(command, preexec_fn = subprocess_setup)
+
+ def get_root_directory(self):
+ '''
+ Returns the name of the package's root directory which is the directory
+ where the debian subdirectory is located.
+ '''
+ return self.root_directory
+
+ def get_changelog_file(self):
+ '''
+ Returns a file object for debian/changelog or None if no such file exists.
+ '''
+ changelog_name = os.path.join(self.root_directory, 'debian', 'changelog')
+ try:
+ return open(changelog_name)
+ except IOError:
+ return None
+
+ def get_all_filenames(self):
+ '''
+ Returns an iterator over all filenames. The filenames will be relative
+ to the root directory.
+ '''
+ skip = len(self.root_directory) + 1
+ for root, _, files in os.walk(self.root_directory):
+ for name in files:
+ yield os.path.join(root[skip:], name)
+
+ def cleanup(self):
+ '''
+ Removes all temporary files.
+ '''
+ if self.root_directory is None:
+ return
+ parent_directory = os.path.dirname(self.root_directory)
+ rmtree(parent_directory)
+ self.root_directory = None
+
+ def __del__(self):
+ '''
+ Enforce cleanup.
+ '''
+ self.cleanup()
+
+
+class SourceContentsScanner(object):
+ '''
+ SourceContentsScanner provides a method scan() to scan the contents of a
+ DBSource object.
+ '''
+ def __init__(self, source_id):
+ '''
+ The argument source_id is the id of the DBSource object that
+ should be scanned.
+ '''
+ self.source_id = source_id
+
+ def scan(self):
+ '''
+ This method does the actual scan and fills in the associated SrcContents
+ property. It commits any changes to the database.
+ '''
+ session = DBConn().session()
+ source = session.query(DBSource).get(self.source_id)
+ fileset = set(source.scan_contents())
+ for filename in fileset:
+ source.contents.append(SrcContents(file = filename))
+ session.commit()
+ session.close()
+
+ @classmethod
+ def scan_all(class_, limit = None):
+ '''
+ The class method scan_all() scans all source using multiple processes.
+ The number of sources to be scanned can be limited with the limit
+ argument. Returns the number of processed and remaining packages as a
+ dict.
+ '''
+ session = DBConn().session()
+ query = session.query(DBSource).filter(DBSource.contents == None)
+ remaining = query.count
+ if limit is not None:
+ query = query.limit(limit)
+ processed = query.count()
+ pool = Pool()
+ for source in query.yield_per(100):
+ pool.apply_async(source_scan_helper, (source.source_id, ))
+ pool.close()
+ pool.join()
+ remaining = remaining()
+ session.close()
+ return { 'processed': processed, 'remaining': remaining }
+
+def source_scan_helper(source_id):
+ '''
+ This function runs in a subprocess.
+ '''
+ try:
+ scanner = SourceContentsScanner(source_id)
+ scanner.scan()
+ except Exception as e:
+ print e
+