Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
branch: master
Commits on Aug 22, 2014
  1. splitting route configuration into a separate file

    as would be a best practice in the real world, since globalizing the
    routing module is the easiest way to use it
    also there are some sketches in there.  comments really
  2. errant sysout

  3. experimenting with resource routing

    actually all routing.  need to understand how i want this stuff to hook
    up - hewing as closely as possible to existing ideas but also keeping in
    mind the strengths of this system

  5. java8ifying, + comments

  6. removing a comment

    it no longer applies
Commits on Aug 20, 2014
  1. errant sysouts

  2. one must test one's classes

    most of the test just comes from Servable, because that was the home of
    the "standard" RequestProcessor implementation, most of which now lives
    in SimpleResourceServer
Commits on Aug 19, 2014
  1. actually, we don't want to pull classes here

    it makes a lot more sense to just get the server out, right? it can
    inject whatever it needs, which is stuff to bridge into the resource
    system, most likely. whatevs
  2. adding a helper to expose servable info

    this is the bridge to allow users to control how to route incoming
    requests to particular resources.  it's also the path to killing off the
    servables structure.  need to work out some details, like route
    contribution (word) and the actual connection of the routes but this is
    pretty much the design
  3. servable resource configuration

    need to bind up quite a few pieces of info to correctly serve a
    resource.  for now, we need a friendly name, and a processing class that
    can handle firing it off to the http world
    also i am sweating like i have a terrorist cousin.
Commits on Aug 16, 2014
  1. using URIMatch to find routes, so SHAs are handled uniformly

    this is step one.  step two is to make extension handling part of the
    trie structure. the plan is to refactor SeparatorNode logic into a base
    class and make a new subclass that essentially acts the same, but with
    special rules - only one per path through the trie, and it disallows
    SeparatorNodes to follow it
  2. clarifying rules about URI matching

    including some new tests! and comments! OH MY
  3. renaming URIMatch#baseName to #path

    for clarity.  baseName was a relic of an old, old version
Commits on Aug 15, 2014
  1. iterating! on a way to identify ServableResource paths

    but i think just a Pattern isn't going to do it, there will need to be a
    class that implements some interface
  2. fixing an order problem with a test

    silly me, unhooking a latch then flipping a flag.  amateur hour!
    also making the timing a little more realistic. 3000 seconds might have
    been a little long to let it sit
  3. binding a set of ServableResource classes

    also renamed the binder helper to more accurately reflect the purpose
  4. figure out the repl test failure

    naturally, it was a problem of determinism
    this test actually needs some expansion, but that's true of the REPL
    system in general.  the fact that it runs at all is pretty neat,
  5. adding system properties

    name clean-up on the environment exposure as well
Commits on Aug 14, 2014
  1. adding an API to read environment variables

    in a javascripty way that is.  noticed a weird issue - continuations
    inside an array forEach fail.  this will be painful! i'm going to have
    to investigate that
  2. testing resource configuration

    and the corollary, fixing the bugs it uncovered.  lol
  3. making the resource watch service configurable/restartable

    welcome to the 21st century, watch service!  mainly this is to support
    the idea that configurations will be per-environment, and
    programmatically updateable in that regard, i envision something like
    var env = ('jj/env');
    var rc = require('jj/resource-configuration');
    rc.watchFiles(env.STAGE == 'Development');
    or similar.
    this doesn't include the actual script piece yet!  haha
  4. adding an interruption facility to tasks

    task interruption can be used to break out of loop tasks that are
    waiting on some sort of JDK API, which generally speaking cannot be
    worked around.
  5. removing the first shot at per-resource configuration

    while i think about it. it definitely isn't right yet
Commits on Aug 13, 2014
  1. reading the resource configurations from a single source

    DRY IT UP! woo woo.
  2. adding the notion of JSON modules

    for one, apparently node does this.  for another, it's a simple way to
    get to raw data for the resource configuration. i'm sure it'll be handy
    in other ways later.
Commits on Aug 12, 2014
  1. tightening up the API

  2. further refactoring

    loading resource settings logic is now correctly centralized and tested,
    and the system is ready to accept overrides.
    however first I need to dry out the default configurations so they are
    defined in one place and used for both config-free and
    config-with-default scenarios
  3. drying up the ResourceSettings determination

    which is good! cause the rules are about to get more complicated
  4. using the configured resource settings

    needs a little more refactoring but the system basically works.
    basically.  however it's not very DRY and not super well tested.  yet.
    but I have... a... plan
Commits on Aug 11, 2014
  1. adding a way to configure resource properties via the configuration.

    not totally hooked up yet, but the mechanism works and is tested.  next
    step is to use it!
Commits on Aug 10, 2014
  1. removing an errant sysout

Something went wrong with that request. Please try again.