Re: [lmi] Request for code review

[Greg Chicares]

On 2017-01-23 22:30, Greg Chicares wrote:
> On 2017-01-23 21:12, Vadim Zeitlin wrote:
>> On Mon, 23 Jan 2017 20:11:00 +0000 Greg Chicares <address@hidden> wrote:
>> 
>> GC> Two (tested) changes here. The "auto" change seems obvious.
>> 
>>  Yes, except that I'd use "auto const" but I am, of course, have somewhat
>> of an obsession with const local variables...
>> 
>> GC> The other is Meyers's "Avoid Duplication in const and Non-const Member
>> GC> Function" technique. It's already used elsewhere in this file, and I
>> GC> kind of like it.
>> 
>>  I can't say I like it but I definitely dislike it less than duplicating
>> code and personally I think it's as good an idea as it ever was. But why
>> use static_cast<> instead of (strictly less powerful and hence less
>> dangerous, without mentioning that it's also much clearer) const_cast<>?
> 
> Meyers says: "The cast that adds const is just forcing a safe conversion
> (from a non-const object to a const one), so we use a static_cast for
> that." I thought he might have had some profound reason, but I guess he
> might be avoiding const_cast because some people think of it as a code
> smell, or perhaps he liked to reserve const_cast for removing constness.
> Whatever his reason, I find your arguments persuasive. And of course if
> we wait until 2017 we can use std::as_const(). C++2017, that is.

Wait...const_cast<> is not strictly less powerful than static_cast<>,
because const_cast<> can cast away volatility but static_cast<> cannot.