Dimitrie O. Paun wrote:
OK, so I'm confused: if we both agree that we should be passing .a and .so stuff to gcc, why are we ignoring these libs? (At this point, I'm not sure at all of the context of the patch, it may be that just the comment is a bit confusing).
The patched code looks like ...
---------------------------------------------------------------- static void sort_lib_file(const char* library) { char *lib;
if (find_dll(library)) { strarray_add(dll_files, strmake("-l%s", library)); } else if ((lib = find_lib(library))) { /* winebuild likes the full path for .a files */ strarray_add(arh_files, lib); } else { /* Probably one of: * .def - which winebuild won't find * .so - which winebuild doesn't support * .a - which gcc won't find */ fprintf(stderr, "Ignoring library %s\n", library); } } ----------------------------------------------------------------
... isn't it clear that the comment refers to libraries that couldn't be found (in the -L library search path).
If you prefer some other comment (or none ;-)), feel free to suggest it.
Perhaps you were deceived by the following line?
- strarray_add(lib_files, strmake("-l%s", library));
The code was just adding stuff to lib_files and then not using it.
-- Richard