Re: [rfc] patch for pr8880

[Sami Wagiaalla <swagiaal@redhat.com>]

On 02/08/2010 12:30 PM, Tom Tromey wrote:
>>>>>> "Sami" == Sami Wagiaalla<swagiaal@redhat.com>  writes:
>
> Sami>  This was a long chase but an easy fix.
> Sami>  Thoughts ?
>
> Lots of formatting nits.
>
> Also I have a few more substantive comments.
>
> Sami>  +struct value *
> Sami>  +value_at_value (struct value *value)
> Sami>  +{
> Sami>  +  return value_at(TYPE_TARGET_TYPE (value_type(value)),
> Sami>  +                  value_as_address(value));
>
> Why not just use value_ind?
>

value_ind works. I just missed it :)

> Sami>  @@ -2104,6 +2111,9 @@ value_struct_elt (struct value **argp, struct value **args,
>
> Modifying value_struct_elt seems dangerous.  This is called from many
> places in the code, including some which don't expect ADL to be used --
> e.g., Java, Ada.
>
> Can this be done by the caller somehow instead?
> Or perhaps some refactoring is needed.
>

I could put this code in a function to be called from value_x_binop and 
value_x_unop. That would at least avoid the awkward argument counting 
but not earlier than that since the arguments need to have been 
evaluated and/or add a check for la_language == language_cplus.

> Also, ADL should only be done for unqualified names.
> It isn't clear to me that this change satisfies that requirement.
>

Let me look into this. It might be a general problem I don't think gdb 
ever differentiates between qualified and unqualified names