NOTES

TODO
  - better error handling + logging
  - url parsing
  - concurrency (again)
  - some sort of versioning / release system
  - remove timeout from error codes

NOTES

  - Ah, managing directory objects is annoying. maybe there's a better way

  - the way uri are mapped to filenames works, but it feels pretty hacky

  - we use lots of static buffer when string parsing, this gets kind of messy because
    at any given moment there are a ton of static buffer lying around not being used,
    wasting memory.... would be nice to have a single buffer where all string manipulation happens,
    potentially make this its own library?

  - learned a few things today:

    - Unless you set the `Connection` header to `close`, HTTP clients will attempt to reuse TCP connections.
      - see: RFC 2616 section 8.1 https://tools.ietf.org/html/rfc2616#section-8.1

    - If `connection` IS set to close, clients will properly indicate an EOF (writing 0 bytes), and we haven't handled this yet
      - TODO :)

    - EOF is now handled (when read(2) returns 0)

    - logs now follows logstalgia format: 1371769989|127.0.0.1|/index.html|200|1024

    - see: http://logstalgia.io/