unit testing - Replacing a C function from a common .o file for the scope of a single test executable -

Then, I got a library call init_foo () , and a function bar () which calls it live with some other useful stuff in the library.o , both are needed.

I have to write some code, bar_init_failure.tc , what happens to test when init_foo () fails, install failure Without doing In Pearl, a large part of our codebase, I launch and replace the library call with a stub that returns the failure code. To accomplish something similar in C, I have the impression that I can redefine the init_foo in the source of bar_init_failure.tc and still For the rest of the code in library.o , but complain about the GCC and ld duplicate symbols (instead of choosing the first one), then I suppose Is that I should be wrong about something (and I'm quite rusting on these kinds of things, so I More confidence in Nniti'm not).

Is there any way to please Linkkar, or is there another strategy that I should use? (I do not have to hack library.c code if I can help it.)

If you can recompile the library to a shared library, you can redefine the init_foo () in unit tests. UDD will not complain about duplicate symbols, nor will the time be compiled, will not be executed, and in the executable which is a dynamic library, at least it will call a function on Unix-like systems - I never try on Windows Of

This can prevent you from examining the bar () because you will not be able to call the actual init_foo () in other tests; Unless you apply it to your own init_foo () , for example when a global variable is set