base64.encode and decode not correct

[jepler at unpythonic.net]

It seems unlikely that there is a bug in Python's base64 encode/decode
functions, but it is possible.

For me, a long string of the same length you mention survives an encode/decode
pair (a round-trip):
    Python 2.3.3 (#1, May  7 2004, 10:31:40) 
    [GCC 3.3.3 20040412 (Red Hat Linux 3.3.3-7)] on linux2
    Type "help", "copyright", "credits" or "license" for more information.
    >>> b = "x" * 1728512
    >>> b.encode("base64").decode("base64") == b
    True
    >>> import base64
    >>> b == base64.decodestring(base64.encodestring(b))
    True

I also ran a program which tests this round-trip property for random data
of all lengths from 0 to 4095, and it ran until I stopped it, rather than
printing an AssertionError:
    import os, base64
    f = open("/dev/urandom")

    while 1:
	for l in range(4096):
	    b = f.read(l)
	    assert b == base64.decodestring(base64.encodestring(b))
	    os.write(2, ".")
	print "!"

Of course, Python itself has a test-suite for base64 which tests strings of
several lengths against the values they should give after encoding or decoding.
    $ python /usr/lib/python2.3/test/test_base64.py
    test_decodestring (__main__.Base64TestCase) ... ok
    test_encodestring (__main__.Base64TestCase) ... ok

    ----------------------------------------------------------------------
    Ran 2 tests in 0.001s

    OK

If your 1728512-character string doesn't pass the simple round-trip test, then
you've uncovered a latent Python bug that my tests didn't demonstrate.  If it
does, then the problem lies somewhere else in your code, possibly in the part
that transmits the encoded message.

Jeff
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-list/attachments/20050816/079eaa56/attachment.sig>