Re: [PATCH 6/8] const-fy regcache::m_aspace

[Simon Marchi <simon.marchi@polymtl.ca>]

On 2017-11-01 05:42, Yao Qi wrote:
> Simon Marchi <simon.marchi@ericsson.com> writes:
> 
>> I don't really understand what this patch tries to achieve.  From the
>> description above, I thought you wanted to make the m_aspace field 
>> const,
>> not the pointed object.
>> 
> 
> I want to achieve both, the field m_aspace is a const, and the pointed
> object is const too.
> 
>    /* The address space of this register cache (for registers where it
>       makes sense, like PC or SP).  */
> -  struct address_space *m_aspace;
> +  const address_space * const m_aspace;

Ah ok I had missed that there was two consts added.

>> If constifying the pointed address_space object is really what you 
>> meant to
>> do, I find having the const_cast more confusing than anything else.  I 
>> think
>> we should constify all the way (removing const_casts, putting more 
>> consts
>> where needed) or not at all.
> 
> OK, const_cast is removed in the updated patch, what do you think?

LGTM!