[PATCH] [1/2] value_fetch_lazy - ensure parent is not lazy before accessing.

[Andrew Burgess]

On 30/07/2013 3:45 PM, Pedro Alves wrote:
> On 07/11/2013 03:34 PM, Andrew Burgess wrote:
> 
> The:
> 
>> -      if (!value_bits_valid (val,
>> +      if (!value_bits_valid (parent,
> 
> change gave me pause, as it wasn't mentioned in the email.
> It just looked like a mistake in your patch, as one would
> only expect the new value_fetch_lazy call, and nothing else.
> But, looking deeper, AFAICS, that change is correct, we were
> looking at the wrong value, even if the lazy issue didn't exist.
> Please correct me if I'm wrong.  (I'd have preferred that bit to
> have been a separate, preparatory change with its own
> justification, or at least that this had been mentioned in the
> email.  Don't assume what's obvious to you is obvious to others.)

Apologies for the poor description in the email.  You did figure out
what I intended.

I committed the fix, with a slightly revised ChangeLog message to
emphasise the two aspects of the patch:

gdb/ChangeLog

	* value.c (value_fetch_lazy): Ensure parent value is not lazy
	before checking which bits of the parent, not the child, value are
	valid.


Thanks,
Andrew