These past few days have been restructuring the theoretical query language and simplifying documentation, as well as adding small components and their tests.

I had a debacle with deferreds and yield, I can’t honestly know what that loop is doing but personally, it feels wrong. Shouldn’t I have a class variable holding the redis connection?

class RedisFuncts():
	__init__
		self.redis_conn = new_redis_stuff
	get()
		self.redis_conn.get()

VS

@deferred.ignore_inlines
def get():
	conn = yield new_redis_stuff
	conn.get()

At least, it seems to violate what my database friend thinks. But, hey, the later code works, I can improve on something that works.

Other than that, played around with getting unified diffs outside of libgit/pygit, when I clearly should be using those… It’s a question of being able to revert from one diff to the next, which leads to patronizing discussions about ndiff vs udiff like this one: diff stackoverflow

Please note since unified diffs usually drops lines that compared equal to begin with it would be almost impossible to restore to the original string. ndiff shows everything, so it would be straight forward to restore the original.

Other than that, tried a bit of compression, and proved that my encoding, language, and entropy sensing (not histograms) could be bypassed as a method of censorship with a simple hex conversion and then using the alphanumeric “Alpha Bravo Charley” system.

No working code because I’m scared of working code.

The lint tests can now have a couple of sed commands run before them for correcting most of the styling. I’m not going for pep8 in this case, but close.

The client is using an example of an async terminal, though this makes it problematic because I won’t know how to recombine the client or the server to do the opposite of what they’re doing, establishing a connection to another host and asking something.

The Cryptography library in python is getting to feel a bit limiting, it was good when experimenting with an XOR padding, but not as convenient when I needed encryption (there seems to only be signing on the EC primitive?), honestly it’s probably just me.

That XOR went over here in crypto/padding.py, by the way.