Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 1 | # Copyright (C) 2001-2006 Python Software Foundation |
Barry Warsaw | bb11386 | 2004-10-03 03:16:19 +0000 | [diff] [blame] | 2 | # Author: Barry Warsaw |
| 3 | # Contact: email-sig@python.org |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 4 | |
Barry Warsaw | 61e5616 | 2004-05-13 22:50:12 +0000 | [diff] [blame] | 5 | """Encodings and related functions.""" |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 6 | |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 7 | __all__ = [ |
| 8 | 'encode_7or8bit', |
| 9 | 'encode_base64', |
| 10 | 'encode_noop', |
| 11 | 'encode_quopri', |
| 12 | ] |
| 13 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 14 | import base64 |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 15 | |
Barry Warsaw | bb11386 | 2004-10-03 03:16:19 +0000 | [diff] [blame] | 16 | from quopri import encodestring as _encodestring |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 17 | |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 18 | |
| 19 | |
Barry Warsaw | bb11386 | 2004-10-03 03:16:19 +0000 | [diff] [blame] | 20 | def _qencode(s): |
| 21 | enc = _encodestring(s, quotetabs=True) |
| 22 | # Must encode spaces, which quopri.encodestring() doesn't do |
| 23 | return enc.replace(' ', '=20') |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 24 | |
Barry Warsaw | 6f70c41 | 2001-09-26 05:26:22 +0000 | [diff] [blame] | 25 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 26 | def _bencode(s): |
| 27 | # We can't quite use base64.encodestring() since it tacks on a "courtesy |
| 28 | # newline". Blech! |
| 29 | if not s: |
| 30 | return s |
| 31 | hasnewline = (s[-1] == '\n') |
| 32 | value = base64.encodestring(s) |
| 33 | if not hasnewline and value[-1] == '\n': |
| 34 | return value[:-1] |
| 35 | return value |
| 36 | |
| 37 | |
Barry Warsaw | e968ead | 2001-10-04 17:05:11 +0000 | [diff] [blame] | 38 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 39 | def encode_base64(msg): |
| 40 | """Encode the message's payload in Base64. |
| 41 | |
Barry Warsaw | 12272a2 | 2002-10-01 00:05:24 +0000 | [diff] [blame] | 42 | Also, add an appropriate Content-Transfer-Encoding header. |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 43 | """ |
| 44 | orig = msg.get_payload() |
| 45 | encdata = _bencode(orig) |
| 46 | msg.set_payload(encdata) |
| 47 | msg['Content-Transfer-Encoding'] = 'base64' |
| 48 | |
| 49 | |
Barry Warsaw | e968ead | 2001-10-04 17:05:11 +0000 | [diff] [blame] | 50 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 51 | def encode_quopri(msg): |
Barry Warsaw | 12272a2 | 2002-10-01 00:05:24 +0000 | [diff] [blame] | 52 | """Encode the message's payload in quoted-printable. |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 53 | |
Barry Warsaw | 12272a2 | 2002-10-01 00:05:24 +0000 | [diff] [blame] | 54 | Also, add an appropriate Content-Transfer-Encoding header. |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 55 | """ |
| 56 | orig = msg.get_payload() |
| 57 | encdata = _qencode(orig) |
| 58 | msg.set_payload(encdata) |
| 59 | msg['Content-Transfer-Encoding'] = 'quoted-printable' |
| 60 | |
| 61 | |
Barry Warsaw | e968ead | 2001-10-04 17:05:11 +0000 | [diff] [blame] | 62 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 63 | def encode_7or8bit(msg): |
Barry Warsaw | 12272a2 | 2002-10-01 00:05:24 +0000 | [diff] [blame] | 64 | """Set the Content-Transfer-Encoding header to 7bit or 8bit.""" |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 65 | orig = msg.get_payload() |
Barry Warsaw | 409a4c0 | 2002-04-10 21:01:31 +0000 | [diff] [blame] | 66 | if orig is None: |
| 67 | # There's no payload. For backwards compatibility we use 7bit |
| 68 | msg['Content-Transfer-Encoding'] = '7bit' |
| 69 | return |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 70 | # We play a trick to make this go fast. If encoding to ASCII succeeds, we |
| 71 | # know the data must be 7bit, otherwise treat it as 8bit. |
| 72 | try: |
| 73 | orig.encode('ascii') |
| 74 | except UnicodeError: |
Barry Warsaw | 61e5616 | 2004-05-13 22:50:12 +0000 | [diff] [blame] | 75 | # iso-2022-* is non-ASCII but still 7-bit |
| 76 | charset = msg.get_charset() |
| 77 | output_cset = charset and charset.output_charset |
| 78 | if output_cset and output_cset.lower().startswith('iso-2202-'): |
| 79 | msg['Content-Transfer-Encoding'] = '7bit' |
| 80 | else: |
| 81 | msg['Content-Transfer-Encoding'] = '8bit' |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 82 | else: |
| 83 | msg['Content-Transfer-Encoding'] = '7bit' |
| 84 | |
| 85 | |
Barry Warsaw | e968ead | 2001-10-04 17:05:11 +0000 | [diff] [blame] | 86 | |
Barry Warsaw | ba92580 | 2001-09-23 03:17:28 +0000 | [diff] [blame] | 87 | def encode_noop(msg): |
| 88 | """Do nothing.""" |