I want to test a function which is using Task.async
In order to make my test pass, I need to make it sleep for 100ms before assertions, otherwise the test process is killed before the async task is executed.
Is there a better way?
EDITED, adding code samples :
The code I would like to test (roughly) :
def search(params) do
RateLimiter.rate_limit(fn ->
parsed_params = ExTwitter.Parser.parse_request_params(params)
json = ExTwitter.API.Base.request(:get, "1.1/search/tweets.json", parsed_params)
Task.async(fn -> process_search_output(json) end)
new_max_id(json)
end)
end
And the test I already wrote (working only with the call to sleep)
test "processes and store tweets" do
with_mock ExTwitter.API.Base, [request: fn(_,_,_) -> json_fixture end] do
with_mock TwitterRateLimiter, [rate_limit: fn(fun) -> fun.() end] do
TSearch.search([q: "my query"])
:timer.sleep(100)
# assertions
assert called TStore.store("some tweet from my fixtures")
assert called TStore.store("another one")
end
end
end
Since the question is a bit vague, I will give the general answer here. The usual technique is to monitor the process and wait for the down message. Something like this:
task = Task.async(fn -> "foo" end)
ref = Process.monitor(task.pid)
assert_receive {:DOWN, ^ref, :process, _, :normal}, 500
Some important things:
The fifth element of the tuple is the exit reason. I am asserting the Task exit is :normal
. Change that accordingly if you are expecting another exit.
The second value in assert_receive
is the timeout. 500 miliseconds sounds like a reasonable amount given you currently have a 100 ms sleep.