Shellcode encoders seem broken

[zachriggle]

The encode() routine should not ever emit bytes in avoid.

>>> encode(asm(shellcraft.sh()), avoid='\x01')
b'jhh///sh/bin\x89\xe3h\x01\x01\x01\x01\x814$ri\x01\x011\xc9Qj\x04Y\x01\xe1Q\x89\xe11\xd2j\x0bX\xcd\x80'

[zachriggle]

It looks like this is a Python3-ism, and the general function is broken if you use the correct b prefix.

>>> encode(b'\x01', avoid=b'\x01')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-2-f36d34f8cc24> in <module>
----> 1 encode(b'\x01', avoid=b'\x01')

~/github.com/pwntools/pwnlib/context/__init__.py in setter(*a, **kw)
   1440                 raise AttributeError("Invalid arch/endianness combination: %s/%s" % (arch, endian))
   1441
-> 1442             return function(*a, **kw)
   1443     return setter
   1444

~/github.com/pwntools/pwnlib/encoders/encoder.py in encode(raw_bytes, avoid, expr, force, pcreg)
     81
     82         try:
---> 83             v = encoder(raw_bytes, avoid, pcreg)
     84         except NotImplementedError:
     85             continue

~/github.com/pwntools/pwnlib/encoders/i386/xor.py in __call__(self, raw_bytes, avoid, pcreg)
     67     def __call__(self, raw_bytes, avoid, pcreg=''):
     68         while len(raw_bytes) % context.bytes:
---> 69             raw_bytes += '\x00'
     70
     71         a, b = xor_pair(raw_bytes, avoid)

TypeError: can't concat str to bytes
>>>

[heapcrash]

Yep, it's broken because set('foo') == {'f', 'o'} while set(b'foo') == {102, 111}.

This is a Python3-ism that needs to be worked around.

[Arusekk]

Encoders need more doctests.