Added new data handler "noforwardsecrecy".

[mkenne11]

The "noforwardsecrecy" data handler detects cipher suites negotiated between client and server (in the Server Hello message) which don't support forward secrecy i.e. Ephemeral Diffie-Hellman methods DHE or ECDHE.

[mkenne11]

If the TLS record contains no messages do I need to continue keep buffing the response, or can I simply "pass" the error and return the response object?

Note. The same comment applies to the exception handler in the SunsetSHA1 on_response method in PR #78.
https://github.com/mkenne11/nogotofail/blob/sunsetsha1/nogotofail/mitm/connection/handlers/data/ssl.py#l231

[mkenne11]

Just checking if this PR and #78 are ok to merge? No worries if you are busy :)

[chadbrubaker]

(Sorry for the late response, holidays are busy :().

This would probably be easier to do using a subclass of the _TlsRecordHandler, that way you don't have to worry about the buffering and state

[mkenne11]

Thanks Chad that's good idea to subclass the _TlsRecordHandler handler. It will make the code much cleaner also.

[mkenne11]

I made the changes recommended:

• The NoForwardSecrecy handler now inherits from _TlsRecordHandler
• INFO items are added in the event log when alerts for this handler are raised

I am still raising notifications in the Android client when alerts for this handler are raised. Do you think this OK? (Not sure if it's too spammy/ or if it's application policy to notify on INFO events).

[mkenne11]

Should I return the TLS record (byte array) each time the method is run?