Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.c++.moderated > #1422 > unrolled thread

See if this usage about initializer_list is conforming to the final draft

Back to article view | Back to comp.lang.c++.moderated

Contents

  See if this usage about initializer_list is conforming to the final draft  hayate <hayatelee@gmail.com> - 2011-06-09 07:02 -0600
    Re: See if this usage about initializer_list is conforming to the final draft  Daniel Krügler <daniel.kruegler@googlemail.com> - 2011-06-09 15:22 -0600
    Re: See if this usage about initializer_list is conforming to the  final draft Helmut Jarausch <jarausch@igpm.rwth-aachen.de> - 2011-06-09 15:54 -0600
      Re: See if this usage about initializer_list is conforming to the  final draft  Daniel Krügler <daniel.kruegler@googlemail.com> - 2011-06-10 03:13 -0600

#1422 — See if this usage about initializer_list is conforming to the final draft

code:
   http://ideone.com/T6GFO

#include <array>
#include <initializer_list>

using namespace std;

class A {
    public:
        A(array<int, 4>) {};
};

int main() {
    A a({1,2,3,4});
}

Its compilation fails on GCC 4.6
I checked n3291 at 8.5.4/1, it seems here for constructing a is a
direct-initialization, thus array's aggregation can be invoked.
Am I right?


-- 
      [ See http://www.gotw.ca/resources/clcm.htm for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]

[toc] | [next] | [standalone]

#1429

Am 09.06.2011 15:02, schrieb hayate:
>
> code:
>     http://ideone.com/T6GFO
>
> #include<array>
> #include<initializer_list>
>
> using namespace std;
>
> class A {
>      public:
>          A(array<int, 4>) {};
> };
>
> int main() {
>      A a({1,2,3,4});
> }
>
> Its compilation fails on GCC 4.6
> I checked n3291 at 8.5.4/1, it seems here for constructing a is a
> direct-initialization, thus array's aggregation can be invoked.
> Am I right?

You are right that std::array's aggregate initialization can be invoked,
but you are overlooking that (at least a non-zero-length) std::array
conceptually is an aggregate containing another aggregate (the actual C
array) as member. In this context the so-called brace-list elision
cannot be applied, you need to provide the correct number of {} levels,
which is in this case 2, thus

A a({{1,2,3,4}});

is necessary here. This is described in [over.ics.list] p4 of the FDIS:

"Otherwise, if the parameter has an aggregate type which can be
initialized from the initializer list according to the rules for
aggregate initialization (8.5.1),[..]"

HTH & Greetings from Bremen,

Daniel Krügler


-- 
      [ See http://www.gotw.ca/resources/clcm.htm for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]

[toc] | [prev] | [next] | [standalone]

#1434 — Re: See if this usage about initializer_list is conforming to the final draft

On Thu, 09 Jun 2011 07:02:48 -0600, hayate wrote:

> code:
>    http://ideone.com/T6GFO
> 
> #include <array>
> #include <initializer_list>
> 
> using namespace std;
> 
> class A {
>     public:
>         A(array<int, 4>) {};
> };
> 
> int main() {
>     A a({1,2,3,4});
> }
> 
> Its compilation fails on GCC 4.6
> I checked n3291 at 8.5.4/1, it seems here for constructing a is a
> direct-initialization, thus array's aggregation can be invoked. Am I
> right?

I've been bitten by this myself. 'array' is an aggregate, there is a long 
thread on comp.std.c++ why, thus you have to replace the declaration of 
'a' by
  A a({{1,2,3,4}});   //  note the double braces

Helmut.




-- 
Helmut Jarausch
Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany


      [ See http://www.gotw.ca/resources/clcm.htm for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]

[toc] | [prev] | [next] | [standalone]

#1438 — Re: See if this usage about initializer_list is conforming to the final draft

Am 09.06.2011 23:54, schrieb Helmut Jarausch:

[..]

> I've been bitten by this myself. 'array' is an aggregate, there is a long
> thread on comp.std.c++ why, thus you have to replace the declaration of
> 'a' by
>    A a({{1,2,3,4}});   //  note the double braces

This is a false herring: The problem is not because array is an
aggregate, but because array is an aggregate *containing* an aggregate.
This becomes obvious, if we look at a slightly different example:

struct aggr {
  int i1, i2, i3, i4;
};

struct B {
  B(aggr){}
};

int main() {
  B({1,2,3,4}); // OK
}

In this case, we have no wrapped aggregate, therefore we need exactly
one level of braces.

HTH & Greetings from Bremen,

Daniel Krügler


-- 
      [ See http://www.gotw.ca/resources/clcm.htm for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.c++.moderated

csiph-web