Path: utzoo!telly!ddsw1!lll-winken!killer!osu-cis!tut.cis.ohio-state.edu!ATHENA.MIT.EDU!wesommer
From: wesommer@ATHENA.MIT.EDU (Bill Sommerfeld)
Newsgroups: gnu.gcc.bug
Subject: undefined statics become externs?
Message-ID: <8812300147.AA11412@ANUBIS.MIT.EDU>
Date: 30 Dec 88 01:47:35 GMT
Sender: daemon@tut.cis.ohio-state.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 23

It seems reasonable (to me at least) for a C compiler to complain
about the following source file:

----
static int foo();

static int (*bar)() = foo;
----

since foo is never defined anywhere in the file.

GCC (1.32/VAX) doesn't complain.  But then again, neither does PCC, or
High C, or Domain/OS C; all these compilers merely treat "foo" as if
it were declared "extern" rather than "static".

For you language lawyers out there: does the dpANS say anything about
this?  Ken Raeburn looked at it (very briefly) but couldn't find
anything explicit about undefined static functions (it was mentioned
that declared but undefined static "objects" were initialized to zero,
but the definition of "object" used in the dpANS appears to exclude
function bodies).

					- Bill