Re: [PATCH] Improve debug info for Ada (3/4)

[Richard Henderson <rth at redhat dot com>]

On Mon, Feb 12, 2007 at 10:17:27PM +0100, Eric Botcazou wrote:
> > Is there some way to avoid adding the extra block if the
> > function already has one?  Or is this moot because we
> > wind up stripping empty BLOCKS later?
> 
> Are you talking about the extra BIND_EXPR or really about the new block?

The new BLOCK.  I'm certain the BIND_EXPR goes away; I have
no idea if the new BLOCK goes away before debug info is 
emitted (for those cases in which nothing is added to it).

In particular, I think I'd prefer to see an existing top-level
BLOCK be reused here.  I'm pretty certain that we'd been using
that top-level block for other variable additions during 
gimplification.


r~