Re: [3.3 PATCH] Fix C++ ICE on designated initializer (regression from 3.2.x)

[Jakub Jelinek <jakub at redhat dot com>]

On Wed, Feb 18, 2004 at 05:10:13PM -0500, Jason Merrill wrote:
> On Tue, 17 Feb 2004 11:46:29 -0500, Jakub Jelinek <jakub@redhat.com> wrote:
> 
> > On Tue, Feb 17, 2004 at 11:38:01AM -0500, Jason Merrill wrote:
> >> Is this not needed for 3.4/trunk?
> >
> > No, the new parser doesn't recognize any form of designated initializers
> > (neither the old GNU style, nor the ISO C99 ones).
> 
> Hmm, that sounds like a regression.

Well, the array designators in G++ 3.3 and earlier were completely useless
anyway (only the old syntax, one couldn't specify a different index than
what it actually had and only the first item could have designator):
char x[4] = { [0] 1 };
used to be accepted, but
char y[4] = { [1] 2 };
char z[4] = { 2, [1] 3 };
did not.

But certainly support for at least ISO C99 designated initializers would be
very good thing to have, they are used in many projects these days.

	Jakub