Fixing spurious "statement is not reachable" in template code

[Steven Schveighoffer via Digitalmars-d]

On 10/28/15 6:02 AM, Daniel Murphy wrote:
> On 28/10/2015 4:29 PM, tsbockman wrote:
>>
>> I would say none, since *the template* contains no unreachable code, and
>> the compiler can easily trim unreachable code from any *instantiation*
>> which needs it, without bothering me about it.
>
> If it's unreachable or not depends on what the template is instantiated
> with, there is no clear concept of unreachable code without knowing the
> template parameters.

You're not thinking from the user's point of view. It's a statement. I 
wrote it because I want it to work in certain instantiations. The 
compiler is incorrectly complaining that it can't be reached, because I 
can reach it by calling it a different way (easy to prove).

The warning should either be eliminated, or fixed so it's accurate. This 
reminds me of people adding empty try/catch clauses to Java to shut the 
compiler up about the checked exceptions.

>
> bool func(T)(T value) if (isUnsigned!T)
> {
>      if (value < 0)
>          return true;
>      return false;
> }
>
> Here the first return is definitely dead code for any instantiation, but
> to know this the compiler would have to reverse-engineer properties from
> the template constraints, which is not generally possible.

I'm OK with the compiler giving up and not warning here. This warning is 
an optimization, and a *warning*, not an error.

>> If most people actually *want* these warnings, then great - there's no
>> bug.
>> But, if most find the warnings conflict with how they want to use
>> templates, as I do - why not just change it?
>
> I don't want these warnings, so I don't generally build with warnings
> enabled.

As will others who do want warnings, but don't want to deal with the 
compiler making false accusations :)

>> The "reality" of D templates is whatever the D community chooses to make
>> it, subject to technical feasibility.
>
> As one of the core compiler devs, I'm saying it sounds infeasible.  I
> don't think either of your suggested solutions are implementable.
> Templates just do not work that way.

What about disabling unreachability detection if a static branch (either 
explicit or implied by const folding) depends on a template variable?

If we can't fix this, I think we're better off without the warning.

-Steve