Hi Paul, Paul Eggert wrote: > If so, it sounds like I should install the second > attached patch into Gnulib.
Nope, this patch introduces way too much code for something that should be simple: testing whether a certain value is NULL or not. The conditional use of API introduced in newer Windows versions is already verbose enough; we don't need yet another macro just for silencing a warning on a secondary platform. As I said in another mail: The only acceptable silencing would be to disable '-Waddress' for the scope of the entire source file. Like this: #if _GL_GNUC_PREREQ (4, 2) # pragma GCC diagnostic ignored "-Waddress" #endif Bruno
