public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug c++/43330]  New: trivial types are not being statically initialized
@ 2010-03-11 11:24 eric dot niebler at gmail dot com
  2010-03-11 11:41 ` Andrew Pinski
                   ` (3 more replies)
  0 siblings, 4 replies; 5+ messages in thread
From: eric dot niebler at gmail dot com @ 2010-03-11 11:24 UTC (permalink / raw)
  To: gcc-bugs

According to the C++0x status page
(http://gcc.gnu.org/gcc-4.5/cxx0x_status.html), "Standard Layout Types"
(http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2342.htm) have been
implemented in gcc-4.5. A quick test shows that trivial types are not being
statically initialized, as the paper requires. The following struct S is
trivial and the non-local constant s should be statically initialized, IIUC:

  struct S
  {
    S() {}
  };

  S const s{};

However, when compiled with -O2 -std=c++0x -S using the latest 4.5 snapshot, I
see the following assembly:

        .file   "main.cpp"
        .text
        .p2align 4,,15
        .def    __GLOBAL__I_main.cpp;   .scl    3;      .type   32;     .endef
__GLOBAL__I_main.cpp:
        pushl   %ebp
        movl    %esp, %ebp
        popl    %ebp
        ret
        .section        .ctors,"w"
        .align 4
        .long   __GLOBAL__I_main.cpp

Unless I'm mistaken, the global s is still being dynamically initialized.
Apologies in advance if I'm reading this wrong.


-- 
           Summary: trivial types are not being statically initialized
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: eric dot niebler at gmail dot com
 GCC build triplet: gcc-4.5-20100304


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Bug c++/43330]  New: trivial types are not being statically initialized
  2010-03-11 11:24 [Bug c++/43330] New: trivial types are not being statically initialized eric dot niebler at gmail dot com
@ 2010-03-11 11:41 ` Andrew Pinski
  2010-03-11 11:42 ` [Bug c++/43330] " pinskia at gmail dot com
                   ` (2 subsequent siblings)
  3 siblings, 0 replies; 5+ messages in thread
From: Andrew Pinski @ 2010-03-11 11:41 UTC (permalink / raw)
  To: gcc-bugzilla; +Cc: gcc-bugs



Sent from my iPhone

On Mar 11, 2010, at 3:24 AM, "eric dot niebler at gmail dot com" <gcc-bugzilla@gcc.gnu.org 
 > wrote:

> According to the C++0x status page
> (http://gcc.gnu.org/gcc-4.5/cxx0x_status.html), "Standard Layout  
> Types"
> (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2342.htm)  
> have been
> implemented in gcc-4.5. A quick test shows that trivial types are  
> not being
> statically initialized, as the paper requires. The following struct  
> S is
> trivial and the non-local constant s should be statically  
> initialized, IIUC:
>
>  struct S
>  {
>    S() {}
>  };

The above is not a trivial type though. It has a non trivial default  
constructor. If you want a trivial constructor use either "= default"  
or delete it all together.


>
>  S const s{};
>
> However, when compiled with -O2 -std=c++0x -S using the latest 4.5  
> snapshot, I
> see the following assembly:
>
>        .file   "main.cpp"
>        .text
>        .p2align 4,,15
>        .def    __GLOBAL__I_main.cpp;   .scl    3;      .type    
> 32;     .endef
> __GLOBAL__I_main.cpp:
>        pushl   %ebp
>        movl    %esp, %ebp
>        popl    %ebp
>        ret
>        .section        .ctors,"w"
>        .align 4
>        .long   __GLOBAL__I_main.cpp
>
> Unless I'm mistaken, the global s is still being dynamically  
> initialized.
> Apologies in advance if I'm reading this wrong.
>
>
> -- 
>           Summary: trivial types are not being statically initialized
>           Product: gcc
>           Version: 4.5.0
>            Status: UNCONFIRMED
>          Severity: normal
>          Priority: P3
>         Component: c++
>        AssignedTo: unassigned at gcc dot gnu dot org
>        ReportedBy: eric dot niebler at gmail dot com
> GCC build triplet: gcc-4.5-20100304
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330
>


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug c++/43330] trivial types are not being statically initialized
  2010-03-11 11:24 [Bug c++/43330] New: trivial types are not being statically initialized eric dot niebler at gmail dot com
  2010-03-11 11:41 ` Andrew Pinski
@ 2010-03-11 11:42 ` pinskia at gmail dot com
  2010-03-11 11:54 ` eric dot niebler at gmail dot com
  2010-03-11 11:56 ` eric dot niebler at gmail dot com
  3 siblings, 0 replies; 5+ messages in thread
From: pinskia at gmail dot com @ 2010-03-11 11:42 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #1 from pinskia at gmail dot com  2010-03-11 11:41 -------
Subject: Re:   New: trivial types are not being statically initialized



Sent from my iPhone

On Mar 11, 2010, at 3:24 AM, "eric dot niebler at gmail dot com"
<gcc-bugzilla@gcc.gnu.org 
 > wrote:

> According to the C++0x status page
> (http://gcc.gnu.org/gcc-4.5/cxx0x_status.html), "Standard Layout  
> Types"
> (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2342.htm)  
> have been
> implemented in gcc-4.5. A quick test shows that trivial types are  
> not being
> statically initialized, as the paper requires. The following struct  
> S is
> trivial and the non-local constant s should be statically  
> initialized, IIUC:
>
>  struct S
>  {
>    S() {}
>  };

The above is not a trivial type though. It has a non trivial default  
constructor. If you want a trivial constructor use either "= default"  
or delete it all together.


>
>  S const s{};
>
> However, when compiled with -O2 -std=c++0x -S using the latest 4.5  
> snapshot, I
> see the following assembly:
>
>        .file   "main.cpp"
>        .text
>        .p2align 4,,15
>        .def    __GLOBAL__I_main.cpp;   .scl    3;      .type    
> 32;     .endef
> __GLOBAL__I_main.cpp:
>        pushl   %ebp
>        movl    %esp, %ebp
>        popl    %ebp
>        ret
>        .section        .ctors,"w"
>        .align 4
>        .long   __GLOBAL__I_main.cpp
>
> Unless I'm mistaken, the global s is still being dynamically  
> initialized.
> Apologies in advance if I'm reading this wrong.
>
>
> -- 
>           Summary: trivial types are not being statically initialized
>           Product: gcc
>           Version: 4.5.0
>            Status: UNCONFIRMED
>          Severity: normal
>          Priority: P3
>         Component: c++
>        AssignedTo: unassigned at gcc dot gnu dot org
>        ReportedBy: eric dot niebler at gmail dot com
> GCC build triplet: gcc-4.5-20100304
>
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330
>


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug c++/43330] trivial types are not being statically initialized
  2010-03-11 11:24 [Bug c++/43330] New: trivial types are not being statically initialized eric dot niebler at gmail dot com
  2010-03-11 11:41 ` Andrew Pinski
  2010-03-11 11:42 ` [Bug c++/43330] " pinskia at gmail dot com
@ 2010-03-11 11:54 ` eric dot niebler at gmail dot com
  2010-03-11 11:56 ` eric dot niebler at gmail dot com
  3 siblings, 0 replies; 5+ messages in thread
From: eric dot niebler at gmail dot com @ 2010-03-11 11:54 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #2 from eric dot niebler at gmail dot com  2010-03-11 11:53 -------
You're right. Let me amend my example.

  struct S
  {
    S() = default;
    S(S const &) = default;
    S(int i) : i(i) {}
    int i;
  };

  S const s {1};

S would seem to satisfy the requirements for a trivial type now, correct? N2342
states:

"When an aggregate with static storage duration is initialized with a
brace-enclosed initializer-list, if all the member initializer expressions are
constant expressions, and the aggregate is a trivial type, the initialization
shall be done during the static phase of initialization"

GCC initializes s dynamically. That doesn't seem right. I see that constant
expressions are not yet implemented in gcc-4.5. Am I correct in assuming that
once constant expressions are implemented, this initialization will be static?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [Bug c++/43330] trivial types are not being statically initialized
  2010-03-11 11:24 [Bug c++/43330] New: trivial types are not being statically initialized eric dot niebler at gmail dot com
                   ` (2 preceding siblings ...)
  2010-03-11 11:54 ` eric dot niebler at gmail dot com
@ 2010-03-11 11:56 ` eric dot niebler at gmail dot com
  3 siblings, 0 replies; 5+ messages in thread
From: eric dot niebler at gmail dot com @ 2010-03-11 11:56 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #3 from eric dot niebler at gmail dot com  2010-03-11 11:56 -------
Oops, it's not an aggregate. Disregard.


-- 

eric dot niebler at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43330


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2010-03-11 11:56 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-03-11 11:24 [Bug c++/43330] New: trivial types are not being statically initialized eric dot niebler at gmail dot com
2010-03-11 11:41 ` Andrew Pinski
2010-03-11 11:42 ` [Bug c++/43330] " pinskia at gmail dot com
2010-03-11 11:54 ` eric dot niebler at gmail dot com
2010-03-11 11:56 ` eric dot niebler at gmail dot com

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).