+
+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, e:
+ print e
+
projects / dak.git / blobdiff