Implemented recursive lock using pthread mutex

[ilyapuchka]

Would be cool if Dip will not depend on Foundation. The only dependency now is NSRecursiveLock. So I grabbed implementation from here

Still needs to be tested. Will try it later on virtual machine, but need to set it up first.

[AliSoftware]

Love the idea 👍

[ilyapuchka]

Managed to implement thread safety tests for Linux 😫 I think it will be not very hard to port all other tests. Will do that later in a separate PR.

["Test Case \'ThreadSafetyTests.testSingletonThreadSafety\' started."]
["Test Case \'ThreadSafetyTests.testSingletonThreadSafety\' passed (1.004 seconds)."]
["Test Case \'ThreadSafetyTests.testFactoryThreadSafety\' started."]
["Test Case \'ThreadSafetyTests.testFactoryThreadSafety\' passed (1.007 seconds)."]
["Test Case \'ThreadSafetyTests.testCircularReferenceThreadSafety\' started."]
["Test Case \'ThreadSafetyTests.testCircularReferenceThreadSafety\' passed (2.008 seconds)."]
["Executed 3 tests, with 0 failures (0 unexpected) in 4.019 (4.027) seconds"]
["Total executed 3 tests, with 0 failures (0 unexpected) in 4.019 (4.029) seconds"]

[AliSoftware]

why?

[ilyapuchka]

Just to let al the threads finish. Of course joining each thread would be better (then dispatch_xxx should return thread's pid). Again it's only because of limitations of swift-core libs, in this case xctest does not provide async utils yet.

[AliSoftware]

I'll trust you on this one and let you merge this when you're ready, because I won't have time to install a Linux VM on my Mac to test the code in there 😉

[ilyapuchka]

Ok =) I really hope it will work. At least with tests I was able to fix the issue with lock vs tryLock (first one was causing deadlock when switching threads in resolveDependencies block in the last test).

I'm still struggling to understand how to organize Linux tests better, for instance how to avoid putting all of them in one file, but I will address that in a separate PR.