Path: utzoo!attcan!uunet!decwrl!sgi!davea@quasar.wpd.sgi.com
From: davea@quasar.wpd.sgi.com (David B. Anderson)
Newsgroups: comp.sys.sgi
Subject: Re: Another C bug?
Keywords: C, MIPS, bug, void
Message-ID: <65335@sgi.sgi.com>
Date: 27 Jul 90 15:31:55 GMT
References: <1990Jul26.200143.1625@midway.uchicago.edu>
Sender: davea@quasar.wpd.sgi.com
Reply-To: davea@quasar.UUCP (David B. Anderson)
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 27

In article <1990Jul26.200143.1625@midway.uchicago.edu> langer@rainbow.uchicago.edu (Steve langer) writes:
>
> I've been having trouble with the MIPS C compiler on a
>Silicon Graphics 4D/220 (3.2). I think this is slightly different
>than the (void *) bug that was discussed here recently. The
	No, really the same.
>problem is that pointers to functions returning void aren't
>treated properly.
	True in 3.2.
>The following program doesn't compile if TYPE is defined to be
>"void". It works if TYPE is defined to be "int". It compiles
>(and works!) in both cases on a Sun.
[stuff deleted]
>ccom: Error: junk.c, line 13: operands of == have incompatible types
>         if(func == function) (*function)();
>      ----------------------^

Neither void * _nor_ void work dependably in 3.2.  The void type
was assigned the same type tag as ``undefined'' internally in ccom, which
lead to various peculiar error messages.

The problems are fixed in 3.3. In 3.3 ccom ``void'' is assigned its own
unique internal type tag.

Regards.
[ David B. Anderson  Silicon Graphics  (415)335-1548  davea@sgi.com ]
[``What can go wrong?''                          --Calvin and Hobbes]