Re: [Dotgnu-libjit] [PATCH] Enable constant folding

[Aleksey Demakov]

Hi Alexey,

As far as I see _jit_opcode_apply() currently does not properly
support opcode errors such as division by zero and integer
overflow. So your patch might break things in certain cases.

The opcode stuff was written by Klaus and now I need a little
bit more time to fully understand where to go with it.

Regards,
Aleksey

On Tue, Feb 26, 2013 at 7:19 PM, Alexey Galakhov <address@hidden> wrote:
> Signed-off-by: Alexey Galakhov <address@hidden>
> ---
>  jit/jit-insn.c |   18 ++++++++++++++++++
>  1 file changed, 18 insertions(+)
>
> diff --git a/jit/jit-insn.c b/jit/jit-insn.c
> index bbd60b6..baf438c 100644
> --- a/jit/jit-insn.c
> +++ b/jit/jit-insn.c
> @@ -288,6 +288,15 @@ static jit_value_t apply_unary
>         {
>                 return 0;
>         }
> +       if(value1->is_constant &&
> +          !jit_context_get_meta_numeric(func->context, JIT_OPTION_DONT_FOLD))
> +       {
> +               dest = _jit_opcode_apply(func, (short)oper, result_type, 
> value1, 0);
> +               if (dest)
> +               {
> +                       return dest;
> +               }
> +       }
>         insn = _jit_block_add_insn(func->builder->current_block);
>         if(!insn)
>         {
> @@ -322,6 +331,15 @@ static jit_value_t apply_binary
>         {
>                 return 0;
>         }
> +       if(value1->is_constant && value2->is_constant &&
> +          !jit_context_get_meta_numeric(func->context, JIT_OPTION_DONT_FOLD))
> +       {
> +               dest = _jit_opcode_apply(func, (short)oper, result_type, 
> value1, value2);
> +               if (dest)
> +               {
> +                       return dest;
> +               }
> +       }
>         insn = _jit_block_add_insn(func->builder->current_block);
>         if(!insn)
>         {
> --
> 1.7.10.4
>
>
> _______________________________________________
> Dotgnu-libjit mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/dotgnu-libjit