[perl #44009] src/dynext.c casting warnings

[via RT]

# New Ticket Created by  Mark Glines 
# Please include the string:  [perl #44009]
# in the subject line of all future correspondence about this issue. 
# <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=44009 >


[07:03] <@particle> there are different warnings on different processors, of 
course :)
[07:43] <@Coke> patches welcome.
[07:43] <+purl> That's swahili for "Shut the hell up."

So after searching through the build log for a little while, I found
some interesting warnings:

src/dynext.c: In function `run_init_lib':
src/dynext.c:315: warning: cast does not match function type
src/dynext.c:322: warning: cast does not match function type

The code looks nasty, but innocent:

315:    load_func = (PMC * (*)(Interp *))D2FPTR(Parrot_dlsym(handle,
                    cload_func_name));


322:    init_func = (void (*)(Interp *, PMC *))D2FPTR(Parrot_dlsym(handle,
                    cinit_func_name));

But looking at it in the preprocessor, I think gcc's just complaining
about casting from one kind of function pointer to another.

typedef void (*funcptr_t)(void);
...
        load_func = (PMC * (*)(Interp 
*))(funcptr_t)((UINTVAL)(Parrot_dlsym(handle, cload_func_name)));

Soo... I dove into the details, found that D2FPTR uses PTR2INTVAL and
INTVAL2PTR, so I tried doing the same thing, just passing a better
prototype in instead of funcptr_t.

        load_func = UINTVAL2PTR(PMC * (*)(Interp *), PTR2UINTVAL(
                    Parrot_dlsym(handle, cload_func_name)));

gets preprocessed to

        load_func = (PMC * (*)(Interp *))((UINTVAL)(Parrot_dlsym(handle, 
cload_func_name)));

But, of course, gcc didn't like that either.

Since Parrot_dlsym returns a void*, is there any particular reason why
we don't just do:

        load_func = Parrot_dlsym(handle, cload_func_name);
        init_func = Parrot_dlsym(handle, cinit_func_name);

...which quiets the warnings, and is a hell of a lot easier to read and
to maintain?

Mark