simple-httpd – serving HTTP made simpler.
This ruby gem wraps around sinatra to provide an even simpler way of setting up http based backends. It is especially helpful to:
- bind loosely related pieces of code together:
simple-httpd
lets a developer lay out their code and assets in directories and trees of directories and can then serve these via HTTP. - have an easy way to serve static assets via HTTP.
- allow existing applications, especially CLI tools, to easily start HTTP servers.
In some ways one might be reminded of the web's old days where one would throw a bunch of php scripts into a FTP location, and then an appache webserver (but, really, its php integration) would start serving requests via HTTPS. simple-httpd is not like that. This gem still supports the notion of an application; source files typically rely on other source files' existence and functionality.
Also, at least as of now, simple-httpd does not dynamically reload code parts on request. This might change in the future.
Is it useful?
At this point I don't know yet. We'll see. In any case this gem is used to test rspec-httpd (a rspec extension helping with testing HTTP endpoints), and is used within postjob-httpd where it is configured to glue HTTP endpoints to the postjob job queue system.
It has proven useful so far - but as it is a really lean wrapper around sinatra one might probably also use sinatra in most cases.
Mounting directories
simple-httpd
lets a user of the gem "mount" directories onto "mount points". A "mount point" describes the location of the actions or static assets at the HTTP endpoint. Note that two or more directories can be mounted at the same mount point.
Files in a mounted directory fall into different categories:
Static assets
Static assets are files with a predefined set of file extensions, including .txt
and .js
. (compare the static_mount.rb
source file for a full list.)
They become available at the location specified by their filename and extension.
Dynamic assets
Each mounted directory might contain a "routes.rb" ruby source file which is loaded in the context of a Sinatra controller to set up routes. Example:
# in v2/routes.rb
get "/queue/:id/events" do
events = [
{ job_id: params[:id], id: "event1" },
{ job_id: params[:id], id: "event2" }
]
json events
end
Ruby files below ./helpers
in the dynamically loaded directory is loaded as a helper file when setting up a dynamic mount, e.g. examples/ex1/helpers/ex1_helpers.rb
are executed in the context of a directory tree's root controller and provide functionality
available in the routes.rb
file.
[TODO] describe service usage.
Command line usage
simple-httpd
comes with a CLI tool, which lets one assemble multiple locations into a single HTTP backend: the following command serves the ./ex1 and ./ex2 directories at http://0.0.0.0:12345
and the ./v2 directory at http://0.0.0.0:12345/api/v2
.
simple-httpd --port=12345 ex1 ex2 v2:api/v2
The v2:api/v2
argument asks the v2
directory to be mounted into the web endpoint at /api/v2
. All relevant content is therefore served below http://0.0.0.0:12345/api/v2
.
The arguments ex1
and ex2
serve at the /
location. This notation really is a shorthand for ex1:/
Integration
simple-httpd
can be integrated into other ruby scripts. Example:
require "simple-httpd"
httpd_root_dir = File.join(__dir__, "httpd")
port = 12345
app = ::Simple::Httpd.build("/" => httpd_root_dir)
::Simple::Httpd.listen! app, port: port, logger: ::Logger.new(STDERR)
The example application
An example application is contained in ./examples. (Well, this example is probably not as useful for any purpose, but I hope it demonstrates all simple-httpd use cases, also it is used during tests.)
See its readme for more details.