Prefer -pthread to -lpthread flag in INTERFACE_LINK_LIBS for better compatibility when cross-compiling

[wesm]

Linking with -lpthread instead of -pthread can cause failures of the sort like

/home/travis/build/apache/arrow/cpp-toolchain/bin/ld: /home/travis/build/apache/arrow/cpp-toolchain/lib/libbenchmark.a(benchmark.cc.o): in function `benchmark::RunSpecifiedBenchmarks(benchmark::BenchmarkReporter*, benchmark::BenchmarkReporter*)':
benchmark.cc:(.text+0x2235): undefined reference to `std::thread::_M_start_thread(std::unique_ptr<std::thread::_State, std::default_delete<std::thread::_State> >, void (*)())'
/home/travis/build/apache/arrow/cpp-toolchain/bin/ld: /home/travis/build/apache/arrow/cpp-toolchain/lib/libbenchmark.a(benchmark.cc.o): in function `std::thread::_State_impl<std::thread::_Invoker<std::tuple<void (*)(benchmark::internal::Benchmark::Instance const*, unsigned long, int, benchmark::internal::ThreadManager*), benchmark::internal::Benchmark::Instance const*, unsigned long, int, benchmark::internal::ThreadManager*> > >::~_State_impl()':
benchmark.cc:(.text._ZNSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEED2Ev[_ZNSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEED5Ev]+0x10): undefined reference to `std::thread::_State::~_State()'
/home/travis/build/apache/arrow/cpp-toolchain/bin/ld: /home/travis/build/apache/arrow/cpp-toolchain/lib/libbenchmark.a(benchmark.cc.o): in function `std::thread::_State_impl<std::thread::_Invoker<std::tuple<void (*)(benchmark::internal::Benchmark::Instance const*, unsigned long, int, benchmark::internal::ThreadManager*), benchmark::internal::Benchmark::Instance const*, unsigned long, int, benchmark::internal::ThreadManager*> > >::~_State_impl()':
benchmark.cc:(.text._ZNSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEED0Ev[_ZNSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEED5Ev]+0x14): undefined reference to `std::thread::_State::~_State()'
/home/travis/build/apache/arrow/cpp-toolchain/bin/ld: /home/travis/build/apache/arrow/cpp-toolchain/lib/libbenchmark.a(benchmark.cc.o):(.data.rel.ro._ZTINSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEEE[_ZTINSt6thread11_State_implINS_8_InvokerISt5tupleIJPFvPKN9benchmark8internal9Benchmark8InstanceEmiPNS4_13ThreadManagerEES8_miSA_EEEEEE]+0x10): undefined reference to `typeinfo for std::thread::_State'
collect2: error: ld returned 1 exit status

when using libraries cross-compiled on one host for use on another. CMake version 3.1 and higher has the THREADS_PREFER_PTHREAD_FLAG flag which toggles the behavior of ${CMAKE_THREAD_LIBS_INIT}.

See https://cmake.org/cmake/help/v3.1/module/FindThreads.html

[googlebot]

Thanks for your pull request. It looks like this may be your first contribution to a Google open source project (if not, look below for help). Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please visit https://cla.developers.google.com/ to sign.

Once you've signed (or fixed any issues), please reply here (e.g. I signed it!) and we'll verify it.

---

What to do if you already signed the CLA

Individual signers

• It's possible we don't have your GitHub username or you're using a different email address on your commit. Check your existing CLA data and verify that your email is set on your git commits.

Corporate signers

• Your company has a Point of Contact who decides which employees are authorized to participate. Ask your POC to be added to the group of authorized contributors. If you don't know who your Point of Contact is, direct the Google project maintainer to go/cla#troubleshoot (Public version).
• The email used to register you as an authorized contributor must be the email used for the Git commit. Check your existing CLA data and verify that your email is set on your git commits.
• The email used to register you as an authorized contributor must also be attached to your GitHub account.

Googlers can find more info about SignCLA and this PR by following this link.

[googlebot]

CLAs look good, thanks!

Googlers can find more info about SignCLA and this PR by following this link.

[dmah42]

Thanks!

[LebedevRI]

Is this actually correct/sufficient though?
https://cmake.org/cmake/help/v3.1/module/FindThreads.html

If the use of the -pthread compiler and linker flag is prefered then the caller can set
  THREADS_PREFER_PTHREAD_FLAG
Please note that the compiler flag can only be used with the imported target. Use of both the imported target as well as this switch is highly recommended for new code.

And yet benchmark uses CMAKE_THREAD_LIBS_INIT, not the Threads::Threads import target,
as far as i can see:

benchmark/src/CMakeLists.txt

Line 30 in 505be96

target_link_libraries(benchmark ${BENCHMARK_CXX_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})

[wesm]

@LebedevRI if you use Threads::Threads then this is what appears in benchmarkTargets.cmake instead of -pthread. So consumers of the installed library would need to also use FindThreads in order to use the CMake config

[wesm]

I haven't tested this extensively on other platforms so would be interested to get feedback from experts who understand the nuances more deeply

[jcfr]

(Background: @wesm I am commenting here following up you last tweet.)

So consumers of the installed library would need to also use FindThreads in order to use the CMake config

The key here is to update cmake/Config.cmake.in so that it has:

include(CMakeFindDependencyMacro)
set(THREADS_PREFER_PTHREAD_FLAG 1)
find_dependency(Threads)

Doing so would ensure that consumer do not have to worry about the private dependencies of the benchmark project and would allow to use Thread::Thread target.

Note that CMakeLists.txt would have to be updated to also set THREADS_PREFER_PTHREAD_FLAG

[wesm]

Seems reasonable. @jcfr are you able to submit a PR for this?