PR 2765

[Joseph S. Myers]

On Sat, 19 May 2001, Mark Mitchell wrote:

> Either that, or we sacrifice C99 on the altar of C89 regression
> elimination.
> 
> I completely agree (see original posting) that this code is a tangle.
> 
> Why is this one clause treated differently than all those above it,
> which still get errors?

I *think* the code that is meant to trigger this is (with -std=c89 
-pedantic)

void foo(int x, int y) { struct {int x; int y;} a = { x, y }; }

which used to generate a hard error, and now generates a warning - since
-pedantic should never cause errors, only warnings.

-- 
Joseph S. Myers
jsm28@cam.ac.uk