[Tuples] Cleaned up accessor patch.
Christopher Matthews
chrismatthews@google.com
Mon Jun 25 21:20:00 GMT 2007
All Diego's suggested changes.
On 6/25/07, Diego Novillo <dnovillo@google.com> wrote:
> On 6/25/07 3:37 PM, Christopher Matthews wrote:
>
> > typedef struct {
> > - gimple *ptr;
> > + gimple ptr;
>
> s/ptr/stmt/
>
> > +/*static inline gimple *
> > gsi_stmt_ptr (gimple_stmt_iterator i)
> > {
> > return i.ptr;
> > -}
> > +}*/
>
> Just remove this function and the associated comment. We can always
> revive it, if needed.
>
> > -/* Construct a GS_OMP_CONTINUE statement for BODY. */
> > +/* Construct a GS_OMP_CONTINUE statement.
> > + FIXME tuples: BODY. */
>
> No, much like OMP_RETURN. OMP_CONTINUE has no body. They are both IL
> markers used when lowering OMP.
>
> > +static inline gimple
> > +gs_seq_last(gs_seq s)
>
> space before '('
>
> > +
> > +static inline bool
> > +gs_seq_empty_p(gs_seq s)
>
> Likewise.
>
> > +/* The positions for each operand type for an assignment. */
> > +#define GS_ASSIGN_OP_LHS 0
> > +#define GS_ASSIGN_OP_RHS1 (size_t) 1
> > +#define GS_ASSIGN_OP_RHS2 (size_t) 2
>
> Well, I'm not sure about this. This is only accessed by the internal
> handlers. Optimizers will always access the operands via the operand
> interfaces, so it won't be needed outside of this. Let's not do these
> symbolic names for now.
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: accessors-v4.diff
Type: text/x-patch
Size: 31521 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20070625/2d49f4bd/attachment.bin>
More information about the Gcc-patches
mailing list