Issue 5198 - workaround for race condition in tests

It's merely a workaround to have tests running until it's properly solved.

The issue currently is that adding of a subscription triggers timeouts with zero seconds timeout. As the result when the test finishes JsEngine is not referenced anymore but the callback for zero timeout is being still executed and tries to destroy JsEngine and we hit the throwing of an exception from destructor (std::thread::join does it to avoid dead-lock). The workaround gives some time to finish these timer callbacks. So far, having such additional sleep only in that place is enough to run tests.

# this change is also required for https://issues.adblockplus.org/ticket/5182

[hub, 2017-04-28 13:07:23]

LGTM with a 'nit

https://codereview.adblockplus.org/29424663/diff/29424664/test/FilterEngine.cpp
File test/FilterEngine.cpp (right):

https://codereview.adblockplus.org/29424663/diff/29424664/test/FilterEngine.c...
test/FilterEngine.cpp:60: void TearDown() override
I would add a comment here to point at the actual issue we are working around so
that we know this is to be properly fixed.

[sergei, 2017-04-28 13:14:33]

https://codereview.adblockplus.org/29424663/diff/29424664/test/FilterEngine.cpp
File test/FilterEngine.cpp (right):

https://codereview.adblockplus.org/29424663/diff/29424664/test/FilterEngine.c...
test/FilterEngine.cpp:60: void TearDown() override
On 2017/04/28 13:07:23, hub wrote:
> I would add a comment here to point at the actual issue we are working around
so
> that we know this is to be properly fixed.

Agree, I thought that one will later trace it using blame from a version control
but it seems impractical because firstly one should really question that line,
having a comment in the code will help here.

[hub, 2017-04-28 14:24:33]

LGTM