This makes it easier to use GTest in projects that build with the
-Wmissing-declarations warning. This fixes the warning in headers and
source files, though not GTest's own tests as it is rather noisy there.
This is part (hopefully all) of the fixes for #776. The top-level
configure.ac configures googletest first and then googlemock.
With this changes it is possible to embed googletest into another
project that uses autoconf. For an example (though it is WIP), see
the commits (and soon PR) referenced from google/protobuf#236.
If the user's cmakelists.txt first look for threads using
find_package(Threads), then set(gtest_disable_pthreads ON),
and then include googletest. GoogleTest will not look for
threads. But since they have already been found before in
user's cmakelists, it will use them regardless.
This helped me fix build issue in darktable-org/rawspeed
on windows/MSYS2, even though there are threads, and they
are usable, googletest build was failing with issues
about AutoHandle. I was first looking for threads, and only
then including googletest, so no matter the value of
gtest_disable_pthreads, it failed.
The other obvious solution is for user to first include
googletest, and only then look for threads by himself.
VS2010 solution only to simplify old users (who used these solutions) upgrading to new gtest/gmock, new users should use CMake generated solutions. VS2010 solution can be opened in any new VS.
Modify library install destinations to install .dll's to the correct
location (`bin`, not `lib`), and to install other artifacts to the
correct platform-dependent location by using GNUInstallDirs. This is
required for some distributions (e.g. Fedora) and will fix an issue that
otherwise requires those distributions to patch the upstream sources.
Also, add options to suppress installation, which may be useful for
projects that embed Google Test.
Since Google Test is trying to support archaic versions of CMake, a
brain-dead fallback (which requires that the user set either LIB_SUFFIX
or CMAKE_INSTALL_LIBDIR themselves) is included for versions that
predate GNUInstallDirs.
Fixes#1161.
Co-Authored-By: d3x0r <d3x0r@users.noreply.github.com>
The documentation of former versions is available in the history (e.g.
by checking out the appropriate git tags), so there is no need to keep
them in parallel to the current documentation.
The document itself uses lambdas later, all the scaffolding to
work around lack of lambdas should be considered for removal, but
that is much larger an effort than I can commit to.
This merges a Google-internal change (117235625).
Original CL description:
This CL was created manually in about an hour with sed, a Python script
to find all the places unqualified 'string' was mentioned, and some help
from Emacs to add the "std::" qualifications, plus a few manual tweaks.
Adds instructions for how to add gtest and gmock to
another CMake project directly. Downloading of the
googletest sources happens as configure time, allowing
it to be added to the main build directly via the
add_subdirectory() command. This ensures googletest is
built with the same compiler settings, etc. and will
typically result in a more robust and more convenient
build arrangement.
Put occurrences of "#include" in a code span so they are not interpreted
as headers.
Other documents were not broken because the #include was not at the
start of the line, but put them in code spans anyway just in case the
text gets refilled in the future.
The helper needs to be templated because its argument type can’t be
known. FloatingPointTest is instantiated with RawType = float and
RawType = double, so Bits will be an unsigned 32-bit or 64-bit type.
size_t will be either 32 or 64 bits depending on the system’s
definition, typically based on pointer size.
This fixes these warnings:
…\gtest\googletest\test\gtest-port_test.cc(78) : error C2220: warning treated as error - no 'object' file generated
…\gtest\googletest\test\gtest-port_test.cc(78) : warning C4309: 'static_cast' : truncation of constant value
…\gtest\googletest\test\gtest-port_test.cc(79) : warning C4309: 'static_cast' : truncation of constant value
…\gtest\googlemock\test\gmock-matchers_test.cc(2712) : error C2220: warning treated as error - no 'object' file generated
…\gtest\googlemock\test\gmock-matchers_test.cc(2706) : while compiling class template member function 'testing::gmock_matchers_test::FloatingPointTest<float>::FloatingPointTest(void)'
…\gtest\googlemock\test\gmock-matchers_test.cc(2896) : see reference to function template instantiation 'testing::gmock_matchers_test::FloatingPointTest<float>::FloatingPointTest(void)' being compiled
…\gtest\googlemock\test\gmock-matchers_test.cc(2896) : see reference to class template instantiation 'testing::gmock_matchers_test::FloatingPointTest<float>' being compiled
…\gtest\googlemock\test\gmock-matchers_test.cc(2712) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2714) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2716) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2717) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2718) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2721) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
…\gtest\googlemock\test\gmock-matchers_test.cc(2723) : warning C4267: 'argument' : conversion from 'size_t' to 'const unsigned int', possible loss of data
These tuples are created with make_tuple, which is given temporaries.
The make_tuple is in a function argument list.
A possibly overzealous static_assert in libc++'s std::tuple ctor
is firing in our 'Perform(make_tuple("hi"))' calls, so
we can't use its make_tuple here. Instead we will use
explicitly-constructed tuples constructed from non-temporary strings.
Workaround for llvm bug:
https://llvm.org/bugs/show_bug.cgi?id=20855
An alternative to https://github.com/google/googletest/pull/580 .
