If a class is defined as auto, I am unable to keep an instance alive longer than its scope. I would like there to be a way to specify non-auto. Most of the time I want to immediately destruct a memory mapped file, but sometime I might want to keep it around for awhile; I'd have to rewrite or modify the code!

--
Christopher E. Miller

Vathix wrote:
> If a class is defined as auto, I am unable to keep an instance alive longer than its scope. I would like there to be a way to specify non-auto. Most of the time I want to immediately destruct a memory mapped file, but sometime I might want to keep it around for awhile; I'd have to rewrite or modify the code!

You're right, they are.  On the bright side, you don't have to use auto that way.  Just write the class itself as non-auto, and use auto instances as needed.

This is perfectly legal:

    class A {
        ...
    }

    A foo() {
        auto A a = new A();
        A b = new A();

        // a should be cleaned up when the function is done, b will not,
        // as it was not declared as auto
        return b;
    }

 -- andy

In article <c7ggfk$v9r$1@digitaldaemon.com>, andy@ikagames.com says...
>This is perfectly legal:
>
>     class A {
>         ...
>     }
>
>     A foo() {
>         auto A a = new A();
>         A b = new A();
>
>         // a should be cleaned up when the function is done, b will not,
>         // as it was not declared as auto
>         return b;
>     }
>
>  -- andy

Right, but I mean when using someone else's code; like std.mmfile.MmFile is auto, so I'm stuck with an auto class, or I have to use my own code. (By the way, when I click std.mmfile at the top of phobos.html, it doesn't scroll down to it because the <A> tag is incorrect.)


-- 
Christopher E. Miller

Vathix wrote:

> In article <c7ggfk$v9r$1@digitaldaemon.com>, andy@ikagames.com says...
> 
>>This is perfectly legal:
>>
>>    class A {
>>        ...
>>    }
>>
>>    A foo() {
>>        auto A a = new A();
>>        A b = new A();
>>
>>        // a should be cleaned up when the function is done, b will not,
>>        // as it was not declared as auto
>>        return b;
>>    }
>>
>> -- andy
> 
> 
> Right, but I mean when using someone else's code; like std.mmfile.MmFile is auto, so I'm stuck with an auto class, or I have to use my own code. (By the way, when I click std.mmfile at the top of phobos.html, it doesn't scroll down to it because the <A> tag is incorrect.)

I don't see why that particular class has to be auto either.  Why should an mmfile have to have function-local scope whether the programmer likes it or not?

I suppose a possible workaround is to wrap the mmfile in a thread....

Stewart.

-- 
My e-mail is valid but not my primary mailbox, aside from its being the unfortunate victim of intensive mail-bombing at the moment.  Please keep replies on the 'group where everyone may benefit.

"Stewart Gordon" <smjg_1998@yahoo.com> wrote in message news:c7o59r$2u8c$1@digitaldaemon.com...
> Vathix wrote:
>
> > In article <c7ggfk$v9r$1@digitaldaemon.com>, andy@ikagames.com says...
> >
> >>This is perfectly legal:
> >>
> >>    class A {
> >>        ...
> >>    }
> >>
> >>    A foo() {
> >>        auto A a = new A();
> >>        A b = new A();
> >>
> >>        // a should be cleaned up when the function is done, b will not,
> >>        // as it was not declared as auto
> >>        return b;
> >>    }
> >>
> >> -- andy
> >
> >
> > Right, but I mean when using someone else's code; like std.mmfile.MmFile is auto, so I'm stuck with an auto class, or I have to use my own code. (By the way, when I click std.mmfile at the top of phobos.html, it doesn't scroll
down
> > to it because the <A> tag is incorrect.)
>
> I don't see why that particular class has to be auto either.  Why should an mmfile have to have function-local scope whether the programmer likes it or not?

That was the way Walter wanted it, and it seemed ok to me at the time. I concur that it should probably not be.

Perhaps one of you could post is as a bug on the bugs NG, and we'll fix it for next time?

>
> I suppose a possible workaround is to wrap the mmfile in a thread....
>
> Stewart.
>
> -- 
> My e-mail is valid but not my primary mailbox, aside from its being the unfortunate victim of intensive mail-bombing at the moment.  Please keep replies on the 'group where everyone may benefit.