URL Parsing

[dlecocq]

This is essentially a translation of urlparse.py

@b4hand @martin-seomoz @seomoz/bigdata

[dlecocq]

In case it's not painfully obvious, I'm C++ rusty. Please feel free to tear this apart. With that in mind, while building up functionality, the focus will be on correctness and cleanliness. Only once we've gotten it "correct" will we focus on improving performance (though it's already like 15-20x faster than url-py).

[tammybailey]

The linkscape code standards are here: https://github.com/seomoz/linkscape/blob/master/CodingStandardCPP.md. If you want to make the PR easier on yourself ;)

[dlecocq]

PTAL -- I think I addressed most of the comments, but I imagine this first one will take a few rounds.

[dlecocq]

@b4hand -- I updated this to not do any unnecessary copies / reassignments in the constructor and as you might expect, performance got better. That said, in general, the porting (and tests) have been the main focus in the work done so far (including PRs not yet submitted).

In a benchmark on my vagrant machine, parsing the string "http://moz.com/some/kind/../of/weird/path;with=parameters????????and=a&&&query=value&&#fragment bar" (and provided as a command-line argument to benchmark) 1M times:

• With copies: 797ms average of 5 runs
• Without copies: 542ms average of 5 runs

[dlecocq]

@tammybailey -- is this more in line with our C++ conventions?

[b4hand]

LGTM