Sha256: f4af9dacc033e2361fc4595c98508363a42c08fddc0fa3750a80c3f7e8fb8503
Contents?: true
Size: 1.61 KB
Versions: 9
Compression:
Stored size: 1.61 KB
Contents
# frozen_string_literal: true
require 'async'
require 'async/http/internet'
# Experimental.
require 'kernel/sync'
class Search
def initialize(app)
@internet = Async::HTTP::Internet.new
@app = app
end
# This method uses the `Async` method to create a reactor if required, and then executes the contained code without waiting for the result. So even if the search query takes a long time (e.g. 100ms), it won't hold up the request processing.
def async(close: !Async::Task.current?)
Async do
response = @internet.get("https://google.com/search?q=async+ruby")
puts response.inspect
ensure
response&.finish
end
end
# This method uses the experimental `Sync` method to create a reactor if required. If the code is already running in a reactor, it runs synchronously, otherwise it's effectively the same as `Async` and a blocking operation. This allow you to write efficient non-blocking code that works in both traditional web application servers, but gains additional scalability in `Async`-aware servers like Falcon.
# You can achieve a similar result by calling `Async{}.wait`, but this is less efficient as it will allocate a sub-task even thought it's not needed.
def sync(close: !Async::Task.current?)
Sync do
response = @internet.get("https://google.com/search?q=async+ruby")
puts response.inspect
ensure
response&.finish
end
end
# The only point of this is to invoke one of the above two methods.
def call(env)
case env['PATH_INFO']
when '/sync'
self.sync
when '/async'
self.async
end
return @app.call(env)
end
end
use Search
run lambda{|env| [404, [], []]}
Version data entries
9 entries across 9 versions & 1 rubygems