Re: c++/3827: new int problem.

Re: c++/3827: new int problem.

[Gabriel Dos Reis]

The following reply was made to PR c++/3827; it has been noted by GNATS.

From: Gabriel Dos Reis <gdr@codesourcery.com>
To: fasbjx@free.fr
Cc: gcc-gnats@gcc.gnu.org, debian-gcc@lists.debian.org
Subject: Re: c++/3827: new int problem.
Date: 26 Jul 2001 09:16:52 +0200

 fasbjx@libertysurf.fr, Branjonneau@free.fr writes:
 
 |   template< typename T, int role, bool ndebug >
 |   struct init {
 |     static T * do_it() { 
 |       T* tmp(new T);
 |       std::cout << "/tmp/ point to: " << *tmp << "\n";
 |       return tmp;
 |     }
 |   };
 
 | /tmp/ point to: 0
 | /tmp/ point to: 1075812784
 
 Hi,
 
 This is not a bug in the compiler.
 
 For fundamental types, a direct-new-expression of the form 'new T'
 does -not- do default initialization.  If you hve to say
 
 	new T()
 
 to get default initialization.  Please note the parens.
 
 -- Gaby
 CodeSourcery, LLC                       http://www.codesourcery.com

Re: c++/3827: new int problem.

[Franck Branjonneau]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 508 bytes --]

The following reply was made to PR c++/3827; it has been noted by GNATS.

From: Franck Branjonneau <franck@tchume.net>
To: Gabriel Dos Reis <gdr@codesourcery.com>
Cc: gcc-gnats@gcc.gnu.org, debian-gcc@lists.debian.org
Subject: Re: c++/3827: new int problem.
Date: Tue, 31 Jul 2001 00:28:09 +0200

 	Bonjour,
 
 Le 26 juillet 2001, Gabriel Dos Reis a écrit :
 
 > This is not a bug in the compiler.
 
 Yes ! I apologize for the inconvenient.
 -- 
 Boujou t'chu tè.
 Franck Branjonneau <fasbjx@free.fr>

c++/3827: new int problem.

[fasbjx, Branjonneau]

>Number:         3827
>Category:       c++
>Synopsis:       No initialization of an int allocated with new int.
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Wed Jul 25 17:36:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     Franck Branjonneau
>Release:        3.0 (Debian) (Debian testing/unstable)
>Organization:
>Environment:
System: Linux alpha.tchume.net 2.4.4 #2 mar mai 22 08:16:07 CEST 2001 i586 unknown
Architecture: i586

	
host: i386-pc-linux-gnu
build: i386-pc-linux-gnu
target: i386-pc-linux-gnu
configured with: ../src/configure -v --enable-languages=c,c++,java,f77,proto,objc --prefix=/usr --infodir=/share/info --mandir=/share/man --enable-shared --with-gnu-as --with-gnu-ld --with-system-zlib --enable-long-long --enable-nls --without-x --without-included-gettext --disable-checking --enable-threads=posix --enable-java-gc=boehm --with-cpp-install-dir=bin --enable-objc-gc i386-linux
>Description:
	
I compile the following code:

#include <iostream>

namespace utility {
  template< typename T, int role, bool ndebug >
  struct init {
    static T * do_it() { 
      T* tmp(new T);
      std::cout << "/tmp/ point to: " << *tmp << "\n";
      return tmp;
    }
  };


  template< typename T, int role, bool ndebug >
  struct exit {
    static void do_it(T * instance) { delete instance; }
  };
} // namespace utility

int main() {
  {
    int * tmp(utility::init< int, 1, true >::do_it());
    utility::exit< int, 1, true >::do_it(tmp);
  }
  int * tmp(utility::init< int, 0, true >::do_it());

}

got an executable and call it. The ouput was:

/tmp/ point to: 0
/tmp/ point to: 1075812784

>How-To-Repeat:
	
It's a context-dependant bug, so I don't know.
As an example, if I comment out utility::exit< int, 1, true >::do_it(tmp) in main(), the behavior is correct. 


>Fix:
	
Customize operator new() ;-)
>Release-Note:
>Audit-Trail:
>Unformatted: