Hello,
thanks for provide the parser for Radiotap
I have a question that in radiotap.py line 320-321
https://github.com/radiotap/python-radiotap/blob/master/radiotap/radiotap.py#L320
why pass the pad!=0 case for radiotap and return the 0,{}
In Wire-Shark , the packet with pad!=0 still has radiotap field and can parse the information the same as your function do.
So I wonder why pad!=0 can't parse in this case?

Thanks!

Because the radiotap module uses some implicit relative imports (e.g. from vht import * where vht.py is a file in the same directory), it does not run under Python 3.

Python 3 has completely removed support for such implicit relative imports (see PEP 404), relative imports must be explicitly specified to avoid shadowing names.

Otherwise, everything works fine in Python 3.

I will be submitting a PR shortly which changes the imports to be explicitly relative. This is backwards compatible with Python 2.7.

Hi again,

following up on #5, I realized I had forgotten to commit changes to tests/parse.py to make it run properly under Python 3.

I've cleaned up the changes and made sure they work correctly in Python 2.7 as well as 3.
I grabbed (with attribution) the iterbytes function from the six module, which is MIT licensed so should be compatible licensing-wise.

Submitting PR shortly.

The _parse_xchannel routine is not aligning to the required 4-octet alignment per the XChannel documentation. I fixed mine by:

def _parse_xchannel(packet, offset):
    mod_4_offset = offset - offset % -4
    xchannel_flags, xchannel_freq, xchannel_num, xchannel_maxpower = \
        struct.unpack_from('<IHBB', packet, mod_4_offset)
    return mod_4_offset + 8, {
        'xchannel_flags' : xchannel_flags,
        'xchannel_freq' : xchannel_freq,
        'xchannel_num' : xchannel_num,
        'xchannel_maxpower' : xchannel_maxpower
    }

The improper alignment was causing IndexError exceptions in the _parse_mcs due to it starting at an improper offset.

Hi,
may I am wrong, but as I read the ieee802.11 standard, the fragment number should be 4 bit of the sequence control field, as far as I understand this line it is currently only 2 bit. Or am I wrong?

Thx for the library anyway.

Kind regards
Alex

Hey,

is it possible to make a pypi package for this library?

I could do a PR if needed.