error: unterminated raw string

[brianbreitsch]

G++ 4.8.2

I'm not 100% sure that I'm doing things correctly--nevertheless, the first error I get when running make is:

g++ -std=c++11  -Wall -Wextra -pedantic -Weffc++ -Wcast-align -Wcast-qual -Wctor-dtor-privacy -Wdisabled-optimization -Wformat=2 -Winit-self -Wmissing-declarations -Wmissing-include-dirs -Wold-style-cast -Woverloaded-virtual -Wredundant-decls -Wshadow -Wsign-conversion -Wsign-promo -Wstrict-overflow=5 -Wswitch -Wundef -Wno-unused -Wnon-virtual-dtor -Wreorder -Wdeprecated -Wfloat-equal  -I src -I test test/unit.cpp  -o json_unit
test/unit.cpp:9675:32: error: unterminated raw string
             CHECK_NOTHROW(json(R"(

[nlohmann]

I'll check this. If I remember correctly, this error occurs in the following constellation:

• g++-4.8
• Catch (the unit test framework) we use
• C++11 raw strings

If you cannot compile the unit tests, the library will still work. I'll check if there is anything I can do to fix this.

[benloong]

this is a gcc bug Preprocessor macros with C++11 raw string literals fail to compile

no error using gcc 4.9 with c++11

[nlohmann]

I close this issue, because it does not affect the code as such. I also added a note to the README file that the unit tests cannot be compiled using GCC 4.8.

[goswamig]

@nlohmann What is the way to bypass the unit test compilation ?

[nlohmann]

You can pass -DJSON_BuildTests=OFF to CMake.

[goswamig]

Thanks.

[goswamig]

?Can I try this in bazel build

[nlohmann]

I do not know what you mean.

[goswamig]

I am running build like below

export TF_NEED_GCP="0"
TF_NEED_HDFS="0"
GCC_HOST_COMPILER_PATH="/usr/bin/gcc"
TF_NEED_OPENCL="0"
CUDA_TOOLKIT_PATH="/usr/local/cuda-9.0"
TF_CUDA_COMPUTE_CAPABILITIES="3.0,3.5,5.2,7.0"
CUDNN_INSTALL_PATH="/usr/local/cuda-9.0"
TF_NEED_CUDA="1" TF_ENABLE_XLA="0"
PYTHON_BIN_PATH="/home/ec2-user/anaconda3/envs/tensorflow_p27/bin/python" TF_CUDNN_VERSION="7"
TF_NEED_JEMALLOC="0"
TF_CUDA_VERSION="9.0"
TF_UNOFFICIAL_SETTING="1"

bazel build -c opt --copt=-msse4.1 --copt=-msse4.2 --copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-O3 --verbose_failures tensorflow_serving/...

[nlohmann]

Sorry, I am not familiar with bazel. But the library is a single header, so you just need to include it in your code - there is no need to compile the unit tests at all.