> You can leave get_inner_reference in its place then ... or move it. It's
> hardly only used by expansion now.
Yes, but it's a hot function and I'd rather preserve its SVN history so I'm
leaving it in its original place. Thanks for the review.
Here's the patch I have installed after having te
On Wed, May 27, 2015 at 12:50 PM, Eric Botcazou wrote:
>> No. Prototypes of functions defined in A.c should be in A.h, not in some
>> other header. We've been (slowly) moving to that. You should have moved
>> them all to expr.h instead, or move the implementations to tree.c.
>
> The former is s
> No. Prototypes of functions defined in A.c should be in A.h, not in some
> other header. We've been (slowly) moving to that. You should have moved
> them all to expr.h instead, or move the implementations to tree.c.
The former is simply not possible since expr.h is poisoned for FEs... I can
On Wed, May 27, 2015 at 12:00 PM, Eric Botcazou wrote:
> Hi,
>
> a few functions manipulating generic trees from expr.c are useful for FEs too
> and some of them (array_ref_{low,up}_bound, get_inner_reference) are already
> declared in tree.h instead of expr.h. This patch moves 3 similar function
Hi,
a few functions manipulating generic trees from expr.c are useful for FEs too
and some of them (array_ref_{low,up}_bound, get_inner_reference) are already
declared in tree.h instead of expr.h. This patch moves 3 similar functions
(array_ref_element_size, array_at_struct_end_p, component_re
