boot-static

a static web server for boot

[tailrecursion/boot-static "0.2.0"] ;; latest release

overview

this task uses Vert.X to create a standalone static webserver. the server in this project will eventually be factored out and moved into a separate boot-vertx task. in this future configuration, the purpose of boot-static will to to inject an application into the build pipeline that serves static resources, while the role of boot-vertx will be to isolate the app in a prewarmed pod fresh from the pod-pool and serve it.

rationale

serving artifacts such as a single page app and its associated resources is a concern separate from the job of the web service, which is responsible for returning data instead of files. consequently, it's a good practice for the spa to live in its own project where boot orchestrates the compilation to a separate artifact to be served from its own origin. this configuration, however, necessitates a local development server to approximate the behavior of the deployment environment (such as an s3 bucket).

you can currently accomplish the same thing using boot-jetty, although you may find this task a bit snappier. this is the first step along the path to a more ambitious undertaking that will utilize boot's pods and pod-pools to maintain a clean, interactive environment without the need for annoying jvm restarts or libraries that compromise lisp's ability to clearly express the logic of the problem your application is solving.

usage

(require
  '[adzerk.boot-cljs          :refer [cljs]]
  '[adzerk.boot-reload        :refer [reload]]
  '[hoplon.boot-hoplon        :refer [hoplon]]
  '[tailrecursion.boot-static :refer [serve]])

(deftask develop
  [o optimizations OPT kw "Optimization level of the closure compiler (:none :simple :advanced)"]
  (let [o (or optimizations :none)]
    (comp (watch) (speak) (hoplon) (reload) (cljs :optimizations o) (serve))))

notes

when serving files at an origin different from the webservice (such as a separate port on localhost), it is also necessary to handle cross-origin requests. a cors middleware can be added to the server for this purpose. the same holds true in production, although a different approach (such as uniting the file and data services behind the same origin, perhaps through cloudfront if using aws) might be advisable, particularly if support for the ms browser spoofers (ie 8 and 9 don't support cors) is required. serving from the same origin also prevents the client from making additional preflight requests.