void main(string[] args)
{
    int a = 1;
    int b = to!int(args[1]);
    uint c = 2;

    switch (a)
    {
        case b: break; // OK
        case c: break; // Error: variable c cannot be read at compile time

        default: break;
    }
}

Switch spec says:
The case expressions must all evaluate to a constant value or array, or a runtime initialized const or immutable variable of integral type. They must be implicitly convertible to the type of the switch Expression.

So, should the case b compile or not ? Is the spec too restrictive here, or is it a bug ?

I was initially working to fix the case c (to allow casts), but this should be clarified.

Martin Krejcirik:

> So, should the case b compile or not ? Is the spec too restrictive here, or is it a bug ?

Apparently it's a WONTFIX mess. The spec should be updated. Walter&Andrei refused to fix a design bug here.

Bye,
bearophile

On 4/23/15 4:25 PM, bearophile wrote:
> Martin Krejcirik:
>
>> So, should the case b compile or not ? Is the spec too restrictive
>> here, or is it a bug ?
>
> Apparently it's a WONTFIX mess. The spec should be updated.
> Walter&Andrei refused to fix a design bug here.

Source?

IMO, the OP code does not warrant the errors cited.

-Steve

On Thursday, April 23, 2015 16:29:03 Steven Schveighoffer via Digitalmars-d wrote:
> On 4/23/15 4:25 PM, bearophile wrote:
> > Martin Krejcirik:
> >
> >> So, should the case b compile or not ? Is the spec too restrictive here, or is it a bug ?
> >
> > Apparently it's a WONTFIX mess. The spec should be updated. Walter&Andrei refused to fix a design bug here.
>
> Source?
>
> IMO, the OP code does not warrant the errors cited.

Well, there's this mess

https://issues.dlang.org/show_bug.cgi?id=6176

but I don't see anything in there from Walter or Andrei, so I'm not sure what Bearophile is referring to. Personally though, I wish that case statements only allowed compile-time constants... :|

- Jonathan M Davis

On 4/23/15 5:04 PM, Jonathan M Davis via Digitalmars-d wrote:
> On Thursday, April 23, 2015 16:29:03 Steven Schveighoffer via Digitalmars-d wrote:
>> On 4/23/15 4:25 PM, bearophile wrote:
>>> Martin Krejcirik:
>>>
>>>> So, should the case b compile or not ? Is the spec too restrictive
>>>> here, or is it a bug ?
>>>
>>> Apparently it's a WONTFIX mess. The spec should be updated.
>>> Walter&Andrei refused to fix a design bug here.
>>
>> Source?
>>
>> IMO, the OP code does not warrant the errors cited.
>
> Well, there's this mess
>
> https://issues.dlang.org/show_bug.cgi?id=6176
>
> but I don't see anything in there from Walter or Andrei, so I'm not sure
> what Bearophile is referring to. Personally though, I wish that case
> statements only allowed compile-time constants... :|
>
> - Jonathan M Davis
>

Found it by following some references:

https://github.com/D-Programming-Language/dmd/pull/2887

Ugh...

at the VERY LEAST, the error message should be fixed, as "can't be read at compile time" is clearly not the issue.

-Steve

> https://github.com/D-Programming-Language/dmd/pull/2887
>
> Ugh...
>
> at the VERY LEAST, the error message should be fixed, as "can't be read at compile time" is clearly not the issue.

Looks like this issue surfaces periodically. Somehow I've missed it so far.
Anyway, either the spec is wrong or the implementation is wrong.

> at the VERY LEAST, the error message should be fixed, as "can't be read at compile time" is clearly not the issue.

This is auto solved by allowing cast in case expression.