[Latent-issue] Avoid duplications when handling the dependencies for static and shared libraries routines

[pavly-gerges]

I've noticed that the front-end bash build script for electrostatic-core which provides the build routines for the generic CMakeLists.txt script, has a potential issue which is a latent one (i.e., not manifested) while declaring the dependencies to be linked with output binaries; that is this line:

dependencies=$(find "$(pwd)/${source_dir}/dependencies/libs/"  \
                        -name *.a -o -name *.so -o -name *.ar | tr '\n' ';')

This line entails finding all the dependencies whether static archives or shared libraries that matches the pattern ending in *.so or *.a, and then using them for both routines of building the static and the shared libraries in the CMakeLists.txt script:

target_link_libraries(${library_a} PUBLIC "${DEPENDENCIES}")
target_link_libraries(${library_so} PUBLIC "${DEPENDENCIES}")

So, this introduces a potential of duplication of function tables if not handled internally by GCC which is a scripting anti-pattern anyway (i.e., Merging of mutually exclusive or disjoint routines that are rather incompatible).

Reasons for latency and no error reports:
The main reason for latency is entailed to the fact that the electrostatic-core module haven't got any external dependencies, yet. While the internal dependencies linked (including the math and the pthread libraries) are handled internally by the linker at compile-time.