The behavior of get_payload() is different when the quoted-printable text has \n line endings or \r\n line endings. If it's \n and the last byte of a line in that file is 0x0D it confuses them for a \r\n line ending and strips both bytes.
This behavior does not occur if the same file has \r\n line endings because the confusing line will end in \r\r\n; then \r\n will be correctly stripped, leaving a \r at the end of a line, as it should. More graphically: Quoted-printable encoded file with \n newlines: 'Byte number 13 is =0D\n' <--- get_payload(decode = True) will strip the last 0x0D byte, leaving a short by 1 file. Quoted-printable encoded file with \r\n newlines: 'Byte number 13 is =0D\r\n' <--- get_payload(decode = True) will leave the 0x0D byte in the file. I think this is a bug? This has bitten me in a script run from procmail; postfix hands \n line endings and the bug happens about once in four months, I can work around it substituting \n with \r\n ensuring that the \r that are part of the file and happen to be in the last position of a line will never be stripped but this is unelegant. Thanks Lucia -- http://mail.python.org/mailman/listinfo/python-list
