[comp.os.minix] What's the future of VOIDSTAR et.al. in 1.4b?

cagney@chook.ua.oz (Andrew Cagney - aka Noid) (11/10/89)

I'm compiling some minix tools using an ansi compiler and hitting
compability problems. Specifically in dd.c there is the declaration
and call:

	int over();
	...
	signal(SIGINT, over);

Which in ansi-c, I understand to be invalid as the second parameter to signal
should be declared as:

	void *over();

Minix 1.4a, in several places, uses the symbol VOIDSTAR in an attempt
to get arround this incompability. So it would be declared as

	VOIDSTAR over();

And when compiling under ansi-c you would define VOIDSTAR=void* etc

Is this going to become the accepted way of doing these declarations or
will some other method be adopted?

					Andrew Cagney

cagney@chook.ua.oz (Andrew Cagney - aka Noid,285,5585,3362395) (11/10/89)

From article <647@augean.OZ>, by cagney@chook.ua.oz (Andrew Cagney - aka Noid):
> 	void *over();
I should have written
	void over();
and be refering to void. Regardless the basic question holds. Sorry.

					Andrew Cagney

ast@cs.vu.nl (Andy Tanenbaum) (11/11/89)

In article <647@augean.OZ> cagney@chook.ua.oz (Andrew Cagney - aka Noid) writes:
>Is this going to become the accepted way of doing these declarations or
>will some other method be adopted?

I haven't really decided.  In any case, the MINIX compiler fully accepts
void, so I won't fudge VOIDSTAR.  I will try to get the MINIX 1.4b headers
fully ANSI/POSIX conformant, which should reduce problems of the type you
mention.

Andy Tanenbaum (ast@cs.vu.nl)