I have got following code:

________________________________________________________
module main;
import std.stdio;
import std.process;

struct SSome{
  long a = 0;
  long b = 0;
}
class CFoo{
  private SSome m_u;
  private SSome m_v;

  public SSome u(SSome o_u){ return this.m_u = o_u; }
  public SSome u(         ){ return this.m_u;       }

  public SSome v(SSome o_v){ return this.m_v = o_v; }
  public SSome v(         ){ return this.m_v      ; }
}

class CBar{
  private CFoo m_foo;

  public CFoo foo(CFoo o_foo){ return this.m_foo = o_foo; }
  public CFoo foo(          ){ return this.m_foo        ; }

  this()
  body{
    this.m_foo = new CFoo;
  }
}

void
main(char [][] args){
  SSome l_some = SSome(1, 2);
  CFoo l_foo = new CFoo;
  writefln("l_foo.u.a: ", l_foo.u.a);
  l_foo.u = l_some;
  writefln("l_foo.u.a: ", l_foo.u.a);
  CBar l_bar = new CBar;
  writefln("l_bar.foo.u.a: ", l_bar.foo.u.a);
  l_bar.foo.u.a = 5;   // no effect
  writefln("l_bar.foo.u.a: ", l_bar.foo.u.a);
  system("pause");
}
________________________________________________________
result:
  l_foo.u.a: 0
  l_foo.u.a: 1
  l_bar.foo.u.a: 0
  l_bar.foo.u.a: 0
________________________________________________________
How to easy change value of property l_bar.foo.u.a? Because the above way doesn't work.

"Zarathustra" <adam.chrapkowski@gmail.com> wrote in message news:g62dv7$gao$1@digitalmars.com...

>  l_bar.foo.u.a = 5;   // no effect

I really wish the compiler could catch stuff like this.  It has no effect because it's doing something like this:

auto temp = l_bar.foo.u;
temp.a = 5;

And then throwing away 'temp'.  That is, it's not modifying l_bar.foo.u, it's modifying a temporary copy of it.

It would work if u() returned an SSome* (that is, "return &this.m_u;"), but I don't know if that's acceptable for you.

On Mon, 21 Jul 2008 21:26:04 +0400, Jarrett Billingsley <kb3ctd2@yahoo.com> wrote:

> "Zarathustra" <adam.chrapkowski@gmail.com> wrote in message
> news:g62dv7$gao$1@digitalmars.com...
>
>>  l_bar.foo.u.a = 5;   // no effect
>
> I really wish the compiler could catch stuff like this.  It has no effect
> because it's doing something like this:
>
> auto temp = l_bar.foo.u;
> temp.a = 5;
>
> And then throwing away 'temp'.  That is, it's not modifying l_bar.foo.u,
> it's modifying a temporary copy of it.
>
> It would work if u() returned an SSome* (that is, "return &this.m_u;"), but
> I don't know if that's acceptable for you.
>
>

That's why we need C++-style references!

Koroskin Denis Wrote:

> On Mon, 21 Jul 2008 21:26:04 +0400, Jarrett Billingsley <kb3ctd2@yahoo.com> wrote:
> 
> > "Zarathustra" <adam.chrapkowski@gmail.com> wrote in message news:g62dv7$gao$1@digitalmars.com...
> >
> >>  l_bar.foo.u.a = 5;   // no effect
> >
> > I really wish the compiler could catch stuff like this.  It has no effect because it's doing something like this:
> >
> > auto temp = l_bar.foo.u;
> > temp.a = 5;
> >
> > And then throwing away 'temp'.  That is, it's not modifying l_bar.foo.u, it's modifying a temporary copy of it.
> >
> > It would work if u() returned an SSome* (that is, "return &this.m_u;"),
> > but
> > I don't know if that's acceptable for you.
> >
> >
> 
> That's why we need C++-style references!

Hmm it's realy strange and in my opinion it harms the properties idea. Maybe, one day it will be changed. I hoped that was only my mistake :(. Ok, thanks a lot for help and if anybody know other solution I would like to know it.

Zarathustra wrote:

> it harms the properties idea

What is your idea of a property?

-manfred

-- 
Maybe some knowledge of some types of disagreeing and their relation can turn out to be useful: http://blog.createdebate.com/2008/04/07/writing-strong-arguments/