Re: a valid statement??

[Dyreatnews]

"jalqadir@gmail.com" <jalqadir@gmail.com> writes:

> In my program I have the following statements
>
> void Person::setTitle(const Glib::ustring& str) {
>     switch ( str ) {
>     case "None" : {
>         title = title_t::None;
>         break;
>     }
>     case "Dr"   : {
>         title = title_t::Dr;
>         break;
>     }
> ......
>     case "Rabbi" :{
>         title = title_t::Rabbi;
>         break;
>    }
>    case "Shaykh" :{
>         title = title_t::Shaykh;
>         break;
>    }
>     default     : {
>         title = title_t::None;
>         break;
>     }
>     }// switch
> }//Method
>
> when compiling I get an error code that reads:
> ==== error: switch quantity not an integer  ====
> Which is true, but according to
>
> http://newdata.box.sk/bx/c/htm/ch07.htm#Heading54
>
> the error is unwarranted, since the parameter received by the method
> and later on given to the switch statement for scrutiny is a valid C/C+
> + statement [setTitle(const Glib::ustring& str)]

Not true. The 2003-standard says "The condition shall be of integral
type , enumeration type, or of a class type for which a single
conversion function to integral or enumeration type exists" (6.4.2.2)

>
> Can anyone shed some light to this problem?

It is also a problem that you are using string literals in your case
labels. The case labels need to be an "integral constant expression"
(6.4.2.2). And (5.19.1) spells out that a pointer is not.

-- 

Dyre Tjeldvoll