> But I wonder at the moment why it does not work (as expected) for the original > complete source file. I discovered that a diff hunk (or usable patch?) is generated if the return statement is deleted (or commented out) before the jump label which refers to a potentially unwanted function call at the mentioned place. How will the support evolve for automatic adjustment of such source code combinations by the semantic patch language (Coccinelle software)? Regards, Markus