test: remove timer from test-http-1.0

[santigimeno]

It's possible that the end event is emitted after the timeout fires
causing the test to fail. Just remove the timer. If for some reason the
end would never fire, the test will fail with a timeout.

It tries to fix the issue described here: #4600 (comment), that I also have rarely experienced on Debian Jessie 64.

[Trott]

Can we change wrap the callback function in c.on('end', function() { in common.mustCall() to guarantee that the validation code runs?

[Trott]

I think one thing the setTimeout() and process.exit() were there for was to prevent side effects with other tests. They try to guarantee that the server gets closed so that subsequent tests don't fail with EADDRINUSE if this test fails.

[jasnell]

LGTM if CI is green

[santigimeno]

PR update with @Trott suggestion.

[santigimeno]

@Trott Do you think I should close the PR then? Or can we try how it goes without the timer?

[Trott]

@santigimeno Would a decent middle ground be to leave the process.on('exit', ...) but get rid of the setTimeout()?

(Or maybe we can confirm that the test runner on CI will close the server if the test leaves it open? Maybe create a test that runs before all others that opens a server on common.PORT and leaves it open and see what happens when you run make test-ci?)

[santigimeno]

@Trott Sorry for the delay answering this, I've been very busy.
In order to confirm if we can remove the setTimeout I've done the following: I have created 64 copies of the new test-http-1.0.js test (without the Timeout). I've modified 2 of them: one would throw as soon as the server starts listening, the other would never close the server. Then I've run: make test-ci and make test. In both cases, the 2 tests failed but there were no issues with the port being in use.

[Trott]

CI: https://ci.nodejs.org/job/node-test-pull-request/1679/

[jasnell]

Build bot failure on Windows in CI, running again: https://ci.nodejs.org/job/node-test-pull-request/1685/

[jasnell]

@Trott ... any further issues with this?

[Trott]

@jasnell No issues from me. LGTM

[jasnell]

Landed in 6a3d38f