在python中从ActionScript读取base64 / zlib编码的字符串

Question

I have a string saved to a database that is encoded by Actionscript by base64ing it and then zlib compressing it.

An example string is this: "eNrj4mZkrShgWdHBx1mUmJfNosQIACUhBCI="

If I unzip and un base64 this via Actionscript, I get what I expect:

{"xp": 656398, "rank": 34}

But, I need to be able to also read this server-side. For now I'm using Python but I'd be open to a working PHP solution or similar.

So far in Python I have tried this:

import base64
import zlib

s = 'eNrj4mZkrShgWdHBx1mUmJfNosQIACUhBCI='

print s.decode("base64").decode("zlib")

It looks like Actionscript adds some extra bits into the header, but my Python is not strong enough to defeat this :) Any help would be much appreciated!

EDIT: Actionscript first takes an AS Object and converts it to a ByteArray, before zlib compressing it and base64ing it. It looks like this is what is generating the extra header/mangled data info.

Answer 1

If the string is first base64-ed and then zipped, decoding it should be the other way around!

Your example and output strings don't match;

In [1]: t = '{"xp": 656398, "rank": 34}'

In [2]: t.encode('base64')
Out[2]: 'eyJ4cCI6IDY1NjM5OCwgInJhbmsiOiAzNH0=\n'

In [3]: t.encode('zlib').encode('base64')
Out[3]: 'eJyrVqooULJSMDM1M7a00FFQKkrMywbyjU1qAVupBsE=\n'

In [4]: t.encode('base64').encode('zlib')
Out[4]: 'x\x9cK\xad\xf42Iv\xf64\xf3t\x894\xf4\xcb\xf25\xf5w.O\xf7\xcc\xf3\xcaH\xca-\xce\xf4\xcft\xac\xf2\xf30\xb0\xe5\x02\x00\xe3E\x0b\xd7'

The given input string '{"xp": 656398, "rank": 34}' does not produce the example output (see Out[3] and Out[4]).

You should also note that in this case the base64 encoded string is longer than the original, and the additional zlib encoding is longest. Compressing strings this short is usually not worth the overhead.

If we take your example output and process it, this is what we get;

In [5]: s = 'eNrj4mZkrShgWdHBx1mUmJfNosQIACUhBCI='

In [6]: s.decode('base64')
Out[6]: 'x\xda\xe3\xe2fd\xad(`Y\xd1\xc1\xc7Y\x94\x98\x97\xcd\xa2\xc4\x08\x00%!\x04"'

In [7]: s.decode('base64').decode('zlib')
Out[7]: '\n\x0b\x01\x05xp\x04\xa8\x88\x0e\trank\x04"\x01'

You can clearly see the texts 'xp' and 'rank' in the output. And the '"' near the end could be interpreted as the integer 34.

It seems that your actionscript does some mangling of the data before encoding and compressing it.

Answer 2

To me it looks like your Python script is decoding in the wrong order. If what you say about the encoding order is correct, first base64, then zlib, you must decode in the opposite order:

print s.decode("zlib").decode("base64")

Answer 3

Tosh, you say that you first base64-encode data and then zip it before storing it to the database. This is fine so far. After zipping the data, it is binary data. Hence, I assume that you have stored this data in binary format in the database. Now you show us a string:

An example string is this: "eNrj4mZkrShgWdHBx1mUmJfNosQIACUhBCI="

This, according to you, represents the zipped data. Now, there is a contradiction. On the one hand zipped data is binary data, on the other hand you clearly show us a string here that resulted from base64-encoding (as you can infer from the '=' in the end).

You confuse something.

Answer 4

Flash could be mangling this, but you should share an example from Actionscript that creates and checks this data.

If you're using a 3rd party library, it might be using an alternate characterset than Python is expecting.

if you're using the official library, it seems that Actionscript implements UTF8 and ascii differently. ( there are different methods here - http://help.adobe.com/en_US/FlashPlatform/reference/actionscript/3/mx/utils/Base64Encoder.html )

it's much easier to just see the actionscript and note if there's an issue with the library or your code, than try to reverse engineer this.

I think it also might be worth trying to bugfix this in reverse...

import base64
import zlib

flash =  "eNrj4mZkrShgWdHBx1mUmJfNosQIACUhBCI="

original = '{"xp": 656398, "rank": 34}'
encoded = original.encode('zlib').encode('base64')
decoded = encoded.decode('base64').decode('zlib')
print original
print encoded
print decoded

Can I also ask why you're even doing this ?

1- unless your packets are much bigger, you're not going to be saving much bandwidth. in fact, compression on small payloads can actually increase the size.

2- assuming this is web based, you should be able to have the server do this on the fly.