Best way to use a compiled function in findroot

Hi,

suppose I make a compiled function

f(x):=(mode_declare(x,float),sin(x));
compile(x);

[obviously here the function is quite naive, but this is just for 
example sake]

let us suppose that I want to use find_root on it (just suppose it is a 
more complicated function)

find_root(f(x),x,-1,1);

does not work... find root tries to evaluate its first argument and I get

Maxima encountered a Lisp error: The value $X is not of type 
DOUBLE-FLOAT.Automatically continuing.To enable the Lisp debugger set 
*debugger-hook* to nil.

fine, so let me try

find_root('f(x),x,-1,1);

this works! So far so good.

Let me now try to use it to define an inverse function

g(y):=block([x:gensym()],findroot('f(x)-y,x,-1,1));

no... this does not work...

g(0);

gives:

findroot(f(g2097),g2097,-1,1)

however, also

g(y):=block([x:gensym()],findroot(f(x)-y,x,-1,1));

does not work.

So what is the correct approach?

Thanks in advance,

Sergio