""" Representation of ar archives for use with Debian binary packages These classes are primarily intended to be used with the :class:`debian.debfile.DebFile` class for working with Debian binary packages. """ # Copyright (C) 2007 Stefano Zacchiroli <zack@debian.org> # Copyright (C) 2007 Filippo Giunchedi <filippo@debian.org> # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, but # WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. from __future__ import print_function import sys try: # pylint: disable=unused-import from typing import ( Any, BinaryIO, Dict, Generator, IO, Iterator, List, NoReturn, Optional, ) except ImportError: # Missing types aren't important at runtime pass GLOBAL_HEADER = b"!<arch>\n" GLOBAL_HEADER_LENGTH = len(GLOBAL_HEADER) FILE_HEADER_LENGTH = 60 FILE_MAGIC = b"`\n" class ArError(Exception): """ Common base for all exceptions raised within the arfile module """ class ArFile(object): """ Representation of an ar archive, see man 1 ar. The interface of this class tries to mimic that of the TarFile module in the standard library. ArFile objects have the following (read-only) properties: - members same as getmembers() """ def __init__(self, filename=None, # type: Optional[str] mode='r', # type: str fileobj=None, # type: Optional[BinaryIO] encoding=None, # type: Optional[str] errors=None, # type: Optional[str] ): # type: (...) -> None """ Build an ar file representation starting from either a filename or an existing file object. The only supported mode is 'r'. In Python 3, the encoding and errors parameters control how member names are decoded into Unicode strings. Like tarfile, the default encoding is sys.getfilesystemencoding() and the default error handling scheme is 'surrogateescape' (>= 3.2) or 'strict' (< 3.2). """ self.__members = [] # type: List[ArMember] self.__members_dict = {} # type: Dict[str, ArMember] self.__fname = filename # type: Optional[str] self.__fileobj = fileobj if encoding is None: encoding = sys.getfilesystemencoding() self.__encoding = encoding if errors is None: if sys.version >= '3.2': errors = 'surrogateescape' else: errors = 'strict' self.__errors = errors if mode == "r": self.__index_archive() # TODO write support def __index_archive(self): # type: () -> None if self.__fname: fp = open(self.__fname, "rb") elif self.__fileobj: fp = self.__fileobj else: raise ArError("Unable to open valid file") if fp.read(GLOBAL_HEADER_LENGTH) != GLOBAL_HEADER: raise ArError("Unable to find global header") while True: newmember = ArMember.from_file(fp, self.__fname, encoding=self.__encoding, errors=self.__errors) if not newmember: break self.__members.append(newmember) self.__members_dict[newmember.name] = newmember if newmember.size % 2 == 0: # even, no padding fp.seek(newmember.size, 1) # skip to next header else: fp.seek(newmember.size + 1, 1) # skip to next header if self.__fname: fp.close() def getmember(self, name): # type: (str) -> ArMember """ Return the (last occurrence of a) member in the archive whose name is 'name'. Raise KeyError if no member matches the given name. Note that in case of name collisions the only way to retrieve all members matching a given name is to use getmembers. """ return self.__members_dict[name] def getmembers(self): # type: () -> List[ArMember] """ Return a list of all members contained in the archive. The list has the same order of members in the archive and can contain duplicate members (i.e. members with the same name) if they are duplicate in the archive itself. """ return self.__members members = property(getmembers) def getnames(self): # type: () -> List[str] """ Return a list of all member names in the archive. """ return [f.name for f in self.__members] def extractall(self): # type: () -> NoReturn """ Not (yet) implemented. """ raise NotImplementedError # TODO def extract(self, member, path): # type: (str, str) -> NoReturn """ Not (yet) implemented. """ raise NotImplementedError # TODO def extractfile(self, member): # type: (str) -> Optional[ArMember] """ Return a file object corresponding to the requested member. A member can be specified either as a string (its name) or as a ArMember instance. """ # TODO(jsw): What is the point of this method? It differs from # getmember in the following ways: # - It returns the *first* member with the given name instead of the # last. # - If member is an ArMember, it uses that ArMember's name as the key. # The former just seems confusing (and this implementation less # efficient than getmember's - probably historical), and I'm having a # hard time seeing the use-case for the latter. for m in self.__members: if isinstance(member, ArMember) and m.name == member.name: return m if member == m.name: return m return None # container emulation def __iter__(self): # type: () -> Iterator[ArMember] """ Iterate over the members of the present ar archive. """ return iter(self.__members) def __getitem__(self, name): # type: (str) -> ArMember """ Same as .getmember(name). """ return self.getmember(name) class ArMember(object): """ Member of an ar archive. Implements most of a file object interface: read, readline, next, readlines, seek, tell, close. ArMember objects have the following (read-only) properties: - name member name in an ar archive - mtime modification time - owner owner user - group owner group - fmode file permissions - size size in bytes - fname file name""" def __init__(self): # type: () -> None # member name (i.e. filename) in the archive self.__name = None # type: Optional[str] # last modification time self.__mtime = None # type: Optional[int] # owner user id self.__owner = None # type: Optional[int] # owner group id self.__group = None # type: Optional[int] # permissions as octal bytes self.__fmode = None # type: Optional[bytes] # member size in bytes self.__size = None # type: Optional[int] # file name associated with this member self.__fname = "" # type: Optional[str] # file pointer self.__fp = None # type: Optional[BinaryIO] # start-of-data offset self.__offset = 0 # type: int # end-of-data offset self.__end = 0 # type: int # current position self.__cur = 0 # type: int @staticmethod def from_file(fp, # type: BinaryIO fname, # type: Optional[str] encoding=None, # type: Optional[str] errors=None, # type: Optional[str] ): # type: (...) -> Optional[ArMember] """fp is an open File object positioned on a valid file header inside an ar archive. Return a new ArMember on success, None otherwise. """ buf = fp.read(FILE_HEADER_LENGTH) if not buf: return None # sanity checks if len(buf) < FILE_HEADER_LENGTH: raise IOError("Incorrect header length") if buf[58:60] != FILE_MAGIC: raise IOError("Incorrect file magic") if sys.version >= '3': if encoding is None: encoding = sys.getfilesystemencoding() if errors is None: if sys.version >= '3.2': errors = 'surrogateescape' else: errors = 'strict' else: # Unused parameters for Python 2 encoding = "" errors = "" # http://en.wikipedia.org/wiki/Ar_(Unix) # from to Name Format # 0 15 File name ASCII # 16 27 File modification date Decimal # 28 33 Owner ID Decimal # 34 39 Group ID Decimal # 40 47 File mode Octal # 48 57 File size in bytes Decimal # 58 59 File magic \140\012 # XXX struct.unpack can be used as well here f = ArMember() name = buf[0:16].split(b"/")[0].strip() if sys.version >= '3': f.__name = name.decode(encoding, errors) else: f.__name = name # type: ignore f.__mtime = int(buf[16:28]) f.__owner = int(buf[28:34]) f.__group = int(buf[34:40]) f.__fmode = buf[40:48] # XXX octal value f.__size = int(buf[48:58]) f.__fname = fname if not fname: f.__fp = fp f.__offset = fp.tell() # start-of-data f.__end = f.__offset + f.__size f.__cur = f.__offset return f # file interface # XXX this is not a sequence like file objects def read(self, size=0): # type: (int) -> bytes if self.__fp is None: if self.__fname is None: raise ValueError("Cannot have both fp and fname undefined") self.__fp = open(self.__fname, "rb") self.__fp.seek(self.__cur) if 0 < size <= self.__end - self.__cur: # there's room buf = self.__fp.read(size) self.__cur = self.__fp.tell() return buf if self.__cur >= self.__end or self.__cur < self.__offset: return b'' buf = self.__fp.read(self.__end - self.__cur) self.__cur = self.__fp.tell() return buf def readline(self, size=None): # type: (Optional[int]) -> bytes if self.__fp is None: if self.__fname is None: raise ValueError("Cannot have both fp and fname undefined") self.__fp = open(self.__fname, "rb") self.__fp.seek(self.__cur) if size is not None: buf = self.__fp.readline(size) self.__cur = self.__fp.tell() if self.__cur > self.__end: return b'' return buf buf = self.__fp.readline() self.__cur = self.__fp.tell() if self.__cur > self.__end: return b'' return buf def readlines(self, sizehint=0): # type: (int) -> List[bytes] # pylint: disable=unused-argument buf = None lines = [] while True: buf = self.readline() if not buf: break lines.append(buf) return lines def seek(self, offset, whence=0): # type: (int, int) -> None if self.__cur < self.__offset: self.__cur = self.__offset if whence < 2 and offset + self.__cur < self.__offset: raise IOError("Can't seek at %d" % offset) if whence == 1: self.__cur = self.__cur + offset elif whence == 0: self.__cur = self.__offset + offset elif whence == 2: self.__cur = self.__end + offset def tell(self): # type: () -> int if self.__cur < self.__offset: return 0 return self.__cur - self.__offset def seekable(self): # type: () -> bool # pylint: disable=no-self-use return True def close(self): # type: () -> None if self.__fp is not None and self.__fname is not None: self.__fp.close() self.__fp = None def next(self): # type: () -> bytes return self.readline() def __iter__(self): # type: () -> Iterator[bytes] def nextline(): # type: () -> Generator[bytes, None, None] line = self.readline() if line: yield line return iter(nextline()) name = property(lambda self: self.__name) mtime = property(lambda self: self.__mtime) owner = property(lambda self: self.__owner) group = property(lambda self: self.__group) fmode = property(lambda self: self.__fmode) size = property(lambda self: self.__size) fname = property(lambda self: self.__fname) if __name__ == '__main__': # test # ar r test.ar <file1> <file2> .. <fileN> a = ArFile("test.ar") print("\n".join(a.getnames()))