SDL build fails under Ubuntu 20.04 with default gcc 9.3.0 compiler

[dboltovskyi]

Bug Report

Within introduction of Ubuntu 20.04 support SDL should be able to be built with default compiler (gcc 9.3.0) However there is an error during build.

Note: SDL build succeeded if compiler downgraded to gcc 8.3.0

Reproduction Steps

1. Setup clean Ubuntu 20.04 environment
2. Clone sdl_core repository
3. Create build folder and try to build project with the commands:

   cmake ../sdl_core && make install-3rd_party_logger && make install -j4

Expected Behavior

SDL is built successfully

Observed Behavior

There is a build error:

/home/developer/sdl/b/3rd_party/include/boost/beast/core/impl/file_stdio.ipp: In member function ‘void boost::beast::file_stdio::seek(uint64_t, boost::beast::error_code&)’:
/home/developer/sdl/b/3rd_party/include/boost/beast/core/impl/file_stdio.ipp:185:15: error: comparison of integer expressions of different signedness: ‘uint64_t’ {aka ‘long unsigned int’} and ‘long int’ [-Werror=sign-compare]
  185 |     if(offset > (std::numeric_limits<long>::max)())
      |        ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1plus: all warnings being treated as errors
make[2]: *** [src/components/hmi_message_handler/CMakeFiles/HMIMessageHandler.dir/build.make:89: src/components/hmi_message_handler/CMakeFiles/HMIMessageHandler.dir/src/mb_controller.cc.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:2425: src/components/hmi_message_handler/CMakeFiles/HMIMessageHandler.dir/all] Error 2

OS & Version Information

• OS/Version: Ubuntu 20.04
• SDL Core Version: https://github.com/smartdevicelink/sdl_core/tree/develop (d8cd806)
• Testing Against: N/A

Test Case, Sample Code, and / or Example App

• Docker environment: 20.04_gcc_9.tar.gz

[dboltovskyi]

After analysis it was found out that if THIRD_PARTY_INSTALL_PREFIX and THIRD_PARTY_INSTALL_PREFIX_ARCH environment variables are not defined build succeeded.

Usually we define these ones in order to have all 3rd party libraries built into defined folder (not to default /usr/local/lib)

[jacobkeeler]

@dboltovskyi can you verify if #3651 fixes this issue on your end? I was able to install using a custom install directory without any issues using those changes

[dboltovskyi]

@jacobkeeler Sure. Thank you for the fix. I have already started testing it. Will get back with results.

[jacobkeeler]

@dboltovskyi It's looking like some of the boost websocket functionality (including the HMI message handler) is broken in that PR, so we're debugging that at the moment. That said, it would still be helpful to know if the CMake changes are correct for fixing the build on your end

[dboltovskyi]

@jacobkeeler Please notice fix you provided works fine on our environments

[jacobkeeler]

closing with the merge of #3651