template foo(int i)
{
  void foo() { }
}

void main()
{
  struct S { int bar; }
  foo!(S.bar.offset)(); // error: identifier 'offset' of 'S.bar.offset' is not defined
}

Sark7 wrote:
> template foo(int i)
> {
>   void foo() { }
> }
> 
> void main()
> {
>   struct S { int bar; }
>   foo!(S.bar.offset)(); // error: identifier 'offset' of 'S.bar.offset' is  not defined
> }

I know this bug report has been out a while (so you might have already solved your problem), but I just looked at this.

What is "offset"? I initially thought it was a property of int, but I can't find it on that page of the spec. It's not a keyword. I think that's why the compiler is confused. I'm confused, too.

If I'm wrong and there is an offset built into the language, please tell me what it is (and where I can find out about it). It might come in handy.

-- 
Justin (a/k/a jcc7)
http://jcc_7.tripod.com/d/

you're using a property without knowing what it means?  that's a new one to me ;)

offset gives you the offset in bytes of a member variable from the start of a struct.  because structs are left in the order they are, and assuming the default alignment is 4 bytes, in struct S { int a,b; }, a.offset=0 and b.offset=4.

Jarrett Billingsley wrote:
> you're using a property without knowing what it means?  that's a new one to
> me ;)
> 
> offset gives you the offset in bytes of a member variable from the start of
> a struct.  because structs are left in the order they are, and assuming the
> default alignment is 4 bytes, in struct S { int a,b; }, a.offset=0 and
> b.offset=4.

Who said I was using it? I'm sometimes careless, but I'm never psychic. ;) I didn't know there was such a thing, and I still can't find a reference to it. Here's what the spec lists as Struct Properties:

  .sizeof	Size in bytes of struct
  .size		Same as .sizeof
  .alignof	Size boundary struct needs to be aligned on


As far I can tell there isn't an "offset" property (although "alignof" may be what is desired).

import std.stdio;

struct S
{
	int bar;
}

void main()
{
	S myStruct;
	writef(myStruct.bar);
//	writef(myStruct.offset);
// offset2.d(13): no property 'offset' for type 'S'

	writef(myStruct.bar.offset);
// offset2.d(17): no property 'offset' for type 'int'
}


-- 
Justin (a/k/a jcc7)
http://jcc_7.tripod.com/d/

odd.  this is from the D docs, in the "converting C to D" section.
Getting the offset of a struct member.
The C Way
Naturally, another macro is used:
           #include <stddef>
           struct Foo { int x; int y; };

           off = offsetof(Foo, y);
The D Way
An offset is just another property:
           struct Foo { int x; int y; }

           off = Foo.y.offset;
but the compiler doesn't recognize it?

Jarrett Billingsley wrote:
> odd.  this is from the D docs, in the "converting C to D" section.
> Getting the offset of a struct member.
> The C Way
> Naturally, another macro is used:
>            #include <stddef>
>            struct Foo { int x; int y; };
> 
>            off = offsetof(Foo, y);
> The D Way
> An offset is just another property:
>            struct Foo { int x; int y; }
> 
>            off = Foo.y.offset;
> but the compiler doesn't recognize it?

Okay. I finally got the compiler to recognize offset. So I agree now that the original post does correctly identify a bug. I'm not sure if the bug is a poor error message or if the code is legal, but I'd call it a bug.

I'm going to post a new bug complaining about how offset is essentially undocumented (yes, it's on ctod.html, but I think it should be on the struct page, too).

import std.stdio;

struct S
{
	int bar;
}


void main()
{
   S myStruct;

// writef(myStruct.offset);
// offset2.d(13): no property 'offset' for type 'S'

// writef(myStruct.bar.offset);
// offset2.d(16): no property 'offset' for type 'int'

    writefln(S.bar.offset);
    writefln(typeof(myStruct).bar.offset);
}

-- 
Justin (a/k/a jcc7)
http://jcc_7.tripod.com/d/