Re: Ordered keys patch for lua-5.1.4 (+extra junk)
[
Date Prev][
Date Next][
Thread Prev][
Thread Next]
[
Date Index]
[
Thread Index]
- Subject: Re: Ordered keys patch for lua-5.1.4 (+extra junk)
- From: Roberto Ierusalimschy <roberto@...>
- Date: Fri, 8 Jan 2010 11:54:34 -0200
> On Fri, Jan 08, 2010 at 11:16:58AM -0200, Roberto Ierusalimschy wrote:
> > That is why we never get rid of this warning. Any trick will only make
> > things worse. This is a Posix function used in the way it is intended to
> > be used. Too bad that C does not have proper support for it.
>
> For aranha, I used something along the lines of:
>
> void *symbol = dlsym(modhandle, funcname);
> if (symbol != NULL) {
> lua_pushcclosure(L, *(lua_CFunction*)(void *)&symbol, 0);
> lua_pushboolean(L, module_system_forked);
> lua_pcall(L, 1, 0, 0);
> }
>
> I'll admit that the casting chain isn't pretty; but it is reliable as far as I
> can tell.
Generally, this construction is not reliable:
type1 var1 = f()
type2 var2 = *(type2 *)&var1;
A compiler may detect that the function does not dereference any
pointer to objects of 'type1' and that 'var1' is never accessed,
and then optimize away the assignment to 'var1'.
-- Roberto