function declarators

[Norman Diamond]

In article <14000 at darkstar.ucsc.edu> daniel at terra.ucsc.edu (Daniel Edelson) writes:
>I have tried unsuccessfully to find an unambiguous statement....
>in the standard that allows or prohibits the following cases:
>	1) a function declaration that defines a new 
>	   type in the return type, and,
>	2) a function declaration that defines a new
>	   type in a parameter
Most of the allowed cases do not have specific statements for them.
If all allowed cases were specifically described, the document would
be larger than the universe.

>For example:
>	/* The return type of foo() defines a new type, struct S */
>	struct S { int a; } foo(void) { }
>	/* The parameter of bar() defines a new type, struct T */
>	void bar(struct T { int b; } p) { }

>Are these strictly conformant ANSI C code?
Yes.

>If so, in the above example, must ``struct S'' be given file scope? 

Since this is a DEFINITION (not just decaration) of foo, indeed it must be
at file scope.  Therefore the structure tag S has file scope.

If the declaration were  struct S { int a; } foo(void);
then the scope of S would depend on where the declaration is located.

>The function ``bar'' is not particularly
>useful because the only type-safe calls to it can be
>recursive calls, since ``struct T'' has block scope.

If structural equivalence (rather than name equivalence) is the rule, then
bar could be called with any struct that contains a single integer.

It has been asserted in this newsgroup that structural equivalence is the
rule within a single translation unit.  I have not found a statement about
this in the standard.  (The rule for equivalence among separate translation
units is neither of the above, and is specified.)
--
Norman Diamond       diamond at tkov50.enet.dec.com
If this were the company's opinion, I wouldn't be allowed to post it.