+def parse_changes(filename, signing_rules=0, dsc_file=0):
+ """
+ Parses a changes file and returns a dictionary where each field is a
+ key. The mandatory first argument is the filename of the .changes
+ file.
+
+ signing_rules is an optional argument:
+
+ - If signing_rules == -1, no signature is required.
+ - If signing_rules == 0 (the default), a signature is required.
+ - If signing_rules == 1, it turns on the same strict format checking
+ as dpkg-source.
+
+ The rules for (signing_rules == 1)-mode are:
+
+ - The PGP header consists of "-----BEGIN PGP SIGNED MESSAGE-----"
+ followed by any PGP header data and must end with a blank line.
+
+ - The data section must end with a blank line and must be followed by
+ "-----BEGIN PGP SIGNATURE-----".
+ """
+
+ changes_in = open_file(filename)
+ content = changes_in.read()
+ changes_in.close()
+ try:
+ unicode(content, 'utf-8')
+ except UnicodeError:
+ raise ChangesUnicodeError, "Changes file not proper utf-8"
+ changes = parse_deb822(content, signing_rules)
+
+
+ if not dsc_file:
+ # Finally ensure that everything needed for .changes is there
+ must_keywords = ('Format', 'Date', 'Source', 'Binary', 'Architecture', 'Version',
+ 'Distribution', 'Maintainer', 'Description', 'Changes', 'Files')
+
+ missingfields=[]
+ for keyword in must_keywords:
+ if not changes.has_key(keyword.lower()):
+ missingfields.append(keyword)
+
+ if len(missingfields):
+ raise ParseChangesError, "Missing mandantory field(s) in changes file (policy 5.5): %s" % (missingfields)
+
+ return changes
+
+################################################################################
+
+def hash_key(hashname):
+ return '%ssum' % hashname
+
+################################################################################
+
+def create_hash(where, files, hashname, hashfunc):
+ """
+ create_hash extends the passed files dict with the given hash by
+ iterating over all files on disk and passing them to the hashing
+ function given.
+ """
+
+ rejmsg = []
+ for f in files.keys():
+ try:
+ file_handle = open_file(f)
+ except CantOpenError:
+ rejmsg.append("Could not open file %s for checksumming" % (f))
+ continue
+
+ files[f][hash_key(hashname)] = hashfunc(file_handle)
+
+ file_handle.close()
+ return rejmsg
+
+################################################################################
+
+def check_hash(where, files, hashname, hashfunc):
+ """
+ check_hash checks the given hash in the files dict against the actual
+ files on disk. The hash values need to be present consistently in
+ all file entries. It does not modify its input in any way.
+ """
+
+ rejmsg = []
+ for f in files.keys():
+ file_handle = None
+ try:
+ try:
+ file_handle = open_file(f)
+
+ # Check for the hash entry, to not trigger a KeyError.
+ if not files[f].has_key(hash_key(hashname)):
+ rejmsg.append("%s: misses %s checksum in %s" % (f, hashname,
+ where))
+ continue
+
+ # Actually check the hash for correctness.
+ if hashfunc(file_handle) != files[f][hash_key(hashname)]:
+ rejmsg.append("%s: %s check failed in %s" % (f, hashname,
+ where))
+ except CantOpenError:
+ # TODO: This happens when the file is in the pool.
+ # warn("Cannot open file %s" % f)
+ continue
+ finally:
+ if file_handle:
+ file_handle.close()
+ return rejmsg
+
+################################################################################
+
+def check_size(where, files):
+ """
+ check_size checks the file sizes in the passed files dict against the
+ files on disk.
+ """
+
+ rejmsg = []
+ for f in files.keys():
+ try:
+ entry = os.stat(f)
+ except OSError, exc:
+ if exc.errno == 2:
+ # TODO: This happens when the file is in the pool.
+ continue
+ raise
+
+ actual_size = entry[stat.ST_SIZE]
+ size = int(files[f]["size"])
+ if size != actual_size:
+ rejmsg.append("%s: actual file size (%s) does not match size (%s) in %s"
+ % (f, actual_size, size, where))
+ return rejmsg
+
+################################################################################
+
+def check_dsc_files(dsc_filename, dsc=None, dsc_files=None):
+ """
+ Verify that the files listed in the Files field of the .dsc are
+ those expected given the announced Format.
+
+ @type dsc_filename: string
+ @param dsc_filename: path of .dsc file
+
+ @type dsc: dict
+ @param dsc: the content of the .dsc parsed by C{parse_changes()}
+
+ @type dsc_files: dict
+ @param dsc_files: the file list returned by C{build_file_list()}
+
+ @rtype: list
+ @return: all errors detected
+ """
+ rejmsg = []
+
+ # Parse the file if needed
+ if dsc is None:
+ dsc = parse_changes(dsc_filename, signing_rules=1, dsc_file=1);
+
+ if dsc_files is None:
+ dsc_files = build_file_list(dsc, is_a_dsc=1)
+
+ # Ensure .dsc lists proper set of source files according to the format
+ # announced
+ has = defaultdict(lambda: 0)
+
+ ftype_lookup = (
+ (r'orig.tar.gz', ('orig_tar_gz', 'orig_tar')),
+ (r'diff.gz', ('debian_diff',)),
+ (r'tar.gz', ('native_tar_gz', 'native_tar')),
+ (r'debian\.tar\.(gz|bz2)', ('debian_tar',)),
+ (r'orig\.tar\.(gz|bz2)', ('orig_tar',)),
+ (r'tar\.(gz|bz2)', ('native_tar',)),
+ (r'orig-.+\.tar\.(gz|bz2)', ('more_orig_tar',)),
+ )
+
+ for f in dsc_files.keys():
+ m = re_issource.match(f)
+ if not m:
+ rejmsg.append("%s: %s in Files field not recognised as source."
+ % (dsc_filename, f))
+ continue
+
+ # Populate 'has' dictionary by resolving keys in lookup table
+ matched = False
+ for regex, keys in ftype_lookup:
+ if re.match(regex, m.group(3)):
+ matched = True
+ for key in keys:
+ has[key] += 1
+ break
+
+ # File does not match anything in lookup table; reject
+ if not matched:
+ reject("%s: unexpected source file '%s'" % (dsc_filename, f))
+
+ # Check for multiple files
+ for file_type in ('orig_tar', 'native_tar', 'debian_tar', 'debian_diff'):
+ if has[file_type] > 1:
+ rejmsg.append("%s: lists multiple %s" % (dsc_filename, file_type))
+
+ # Source format specific tests
+ try:
+ format = get_format_from_string(dsc['format'])
+ rejmsg.extend([
+ '%s: %s' % (dsc_filename, x) for x in format.reject_msgs(has)
+ ])
+
+ except UnknownFormatError:
+ # Not an error here for now
+ pass
+
+ return rejmsg
+
+################################################################################
+
+def check_hash_fields(what, manifest):
+ """
+ check_hash_fields ensures that there are no checksum fields in the
+ given dict that we do not know about.
+ """
+
+ rejmsg = []
+ hashes = map(lambda x: x[0], known_hashes)
+ for field in manifest:
+ if field.startswith("checksums-"):
+ hashname = field.split("-",1)[1]
+ if hashname not in hashes:
+ rejmsg.append("Unsupported checksum field for %s "\
+ "in %s" % (hashname, what))
+ return rejmsg
+
+################################################################################
+
+def _ensure_changes_hash(changes, format, version, files, hashname, hashfunc):
+ if format >= version:
+ # The version should contain the specified hash.
+ func = check_hash
+
+ # Import hashes from the changes
+ rejmsg = parse_checksums(".changes", files, changes, hashname)
+ if len(rejmsg) > 0:
+ return rejmsg
+ else:
+ # We need to calculate the hash because it can't possibly
+ # be in the file.
+ func = create_hash
+ return func(".changes", files, hashname, hashfunc)
+
+# We could add the orig which might be in the pool to the files dict to
+# access the checksums easily.
+
+def _ensure_dsc_hash(dsc, dsc_files, hashname, hashfunc):
+ """
+ ensure_dsc_hashes' task is to ensure that each and every *present* hash
+ in the dsc is correct, i.e. identical to the changes file and if necessary
+ the pool. The latter task is delegated to check_hash.
+ """
+
+ rejmsg = []
+ if not dsc.has_key('Checksums-%s' % (hashname,)):
+ return rejmsg
+ # Import hashes from the dsc
+ parse_checksums(".dsc", dsc_files, dsc, hashname)
+ # And check it...
+ rejmsg.extend(check_hash(".dsc", dsc_files, hashname, hashfunc))
+ return rejmsg
+
+################################################################################
+
+def parse_checksums(where, files, manifest, hashname):
+ rejmsg = []
+ field = 'checksums-%s' % hashname
+ if not field in manifest:
+ return rejmsg
+ for line in manifest[field].split('\n'):
+ if not line:
+ break
+ clist = line.strip().split(' ')
+ if len(clist) == 3:
+ checksum, size, checkfile = clist
+ else:
+ rejmsg.append("Cannot parse checksum line [%s]" % (line))
+ continue
+ if not files.has_key(checkfile):
+ # TODO: check for the file's entry in the original files dict, not
+ # the one modified by (auto)byhand and other weird stuff
+ # rejmsg.append("%s: not present in files but in checksums-%s in %s" %
+ # (file, hashname, where))
+ continue
+ if not files[checkfile]["size"] == size:
+ rejmsg.append("%s: size differs for files and checksums-%s entry "\
+ "in %s" % (checkfile, hashname, where))
+ continue
+ files[checkfile][hash_key(hashname)] = checksum
+ for f in files.keys():
+ if not files[f].has_key(hash_key(hashname)):
+ rejmsg.append("%s: no entry in checksums-%s in %s" % (checkfile,
+ hashname, where))
+ return rejmsg
+
+################################################################################
+