base64 decode of the JWT failing

[sid3windr]

Thanks for this library.

When trying to auth against the CZDS service I run into a Python exception: binascii.Error: Incorrect padding (even if www.base64decode.org did decode my JWT base64 string just fine into JSON).

I'm not a pro in Python or this way of authentication at all, but at least padding the base64 string with == made the code work for me.

I patched https://github.com/mdiez/pyCZDS/blob/main/src/pyczds/authentication.py#L33

+        s = base64.b64decode(payload_b64)
-        s = base64.b64decode(f"{payload_b64}==")

I can make a pull request if you prefer, but I'm not sure my hack is the right solution - let me know.

[mdiez]

Hi Tom,

I really appreciate you using this library, and many thanks for your report.

I have just released version 1.6 which I believe fixes your problem. It was related to the JWT payload's length not being a multiple of 4, and thus lacking sufficient padding (as you laid out already).

Hope this works -- let me know if not!

Regards, Max

[sid3windr]

Thanks for the fix, this is working well. Also thanks for fixing the importlib stuff, I wasn't sure where the fault was with that!

Hang on, false alarm: listing zonefiles works (might have worked before?), but downloading still yields the same binascii error.

[sid3windr]

I think I see the issue, you're comparing the full token length % 4 in _authenticate() but the code itself is using only the second part: payload_b64 = self._token.split('.')[1] in _get_token_jwt_payload()

It's that one that isn't a multiple of 4 even if the full token is (which it is, there's no padding added on my system - padded_token and token are identical in _authenticate())