In e8b6b7ad, we relaxed the cmake version check down to 3.1, the first
version to expose target_compile_features().
However, an error while testing that change improperly concluded that
the change was OK. While target_compile_features() was indeed intriduced
in cmake 3.1, the actual feature we use it to test, cxx_std_11, was
really introduced only with cmake-3.8, which explained the actual
version that was requested before e8b6b7ad.
As Patrick noted, however, we can still convince cmake-3.1 to test
for C++11, by testing for an actual feature introduced in C++11, like
testing for cxx_range_for, which will instruct cmake to set the
appropriate C++11 options for the current compiler, which for gcc would
be adding -std=gnu++11.
Reported-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Commit 73cc5089 (Using target_compile_features to specify C++ 11
standard) bumped the required cmake version, from 3.0 to 3.8, so
as to get the definition of target_compile_features().
However, target_compile_features() was introduced in cmake-3.1:
https://cmake.org/cmake/help/v3.1/command/target_compile_features.html
And using cmake-3.1 is indeed sufficient to properly build.
As such, relax the minimum required version down to cmake-3.1,
so we can build on oldish, entreprise-grade distributions that
only have cmake-3.1 (or at least, don't have up to cmake-3.8).
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
o For some unknown reason, the complexity of the benchmark platform
prevented some C++ compilers from generating optimal code, properly
reflective of the real performance in actual deployment.
o Added the json_benchmarks_simple target, which performs the same
suite of tests as json_benchmarks.
o Simplified the benchmark platform, and emit an "Average" TPS
(Transactions Per Second) value reflective of aggregate parse/output
performance.
