fix: correct connect timeout setting for ApacheHttpRequest #803
Conversation
googlebot
added
the
cla: yes
|
Same comment as in #804: The Apache HttpClient implementation does a great job and making itself uninspectable once initialized - we would need to test its behavior (or use reflection to look at internals). For these transport adapters, we will need to set up an integration test that somehow ensures the desired behavior. |
There was a problem hiding this comment.
This one initially seemed harder to test, but I think @chanseokoh had a really good idea in the bug report that will enable us to test this one by setting a low timeout and connecting to a non-existent site.
chingor13
added
the
kokoro:force-run
kokoro-team
removed
the
kokoro:force-run
chingor13
added
the
kokoro:force-run
kokoro-team
removed
the
kokoro:force-run
chingor13
added
the
kokoro:force-run
kokoro-team
removed
the
kokoro:force-run
|
I'm not sure why this fails on windows, but it seems to be within the Apache adapter. We should be able to skip this test on windows. |
| @@ -183,6 +187,23 @@ public void process(HttpRequest request, HttpContext context) | |||
| assertTrue("Expected to have called our test interceptor", interceptorCalled.get()); | |||
| } | |||
|
|
|||
| @Test(timeout = 10_000L) | |||
| public void testConnectTimeout() { | |||
| // Apache HttpClient doesn't appear to behave correctly on windows | |||
There was a problem hiding this comment.
Is there a bug filed against Apache we could reference here?
Fixes #799