Re: libstdc++/5268: More fun with global stream destruction semantics

Re: libstdc++/5268: More fun with global stream destruction semantics

[bkoz]

Synopsis: More fun with global stream destruction semantics

Responsible-Changed-From-To: unassigned->bkoz
Responsible-Changed-By: bkoz
Responsible-Changed-When: Tue Apr  2 18:37:17 2002
Responsible-Changed-Why:
    Mine.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5268

Re: libstdc++/5268: More fun with global stream destruction semantics

[bkoz]

Synopsis: More fun with global stream destruction semantics

State-Changed-From-To: feedback->closed
State-Changed-By: bkoz
State-Changed-When: Wed Apr  3 12:35:49 2002
State-Changed-Why:
    Fixed....

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5268

RE: libstdc++/5268: More fun with global stream destruction semantics

[Simpson, Kenny]

The following reply was made to PR libstdc++/5268; it has been noted by GNATS.

From: "Simpson, Kenny" <Kenny.Simpson@gs.com>
To: "'paolo@gcc.gnu.org'" <paolo@gcc.gnu.org>,
	"'gcc-bugs@gcc.gnu.org'" <gcc-bugs@gcc.gnu.org>,
	"'gcc-prs@gcc.gnu.org'" <gcc-prs@gcc.gnu.org>,
	"Simpson, Kenny" <Kenny.Simpson@gs.com>,
	"'nobody@gcc.gnu.org'" <nobody@gcc.gnu.org>,
	"'theonetruekenny@yahoo.com'" <theonetruekenny@yahoo.com>,
	"'gcc-gnats@gcc.gnu.org'" <gcc-gnats@gcc.gnu.org>
Cc:  
Subject: RE: libstdc++/5268: More fun with global stream destruction semantics
Date: Wed, 9 Jan 2002 17:37:02 -0500 

 Ok, here is a better (compilable) test:
 
 #include <iostream> // for std::cout
 #include <ostream>  // for std::ostream
 #include <sstream> // for basic_stringbuf and stringbuf typedef
 
 int main()
 {
   std::stringbuf b1;
   std::cout.rdbuf( &b1 );
   std::cout << "hello\n";
   return 0;
 }

Re: libstdc++/5268: More fun with global stream destruction semantics

[paolo]

Synopsis: More fun with global stream destruction semantics

State-Changed-From-To: open->feedback
State-Changed-By: paolo
State-Changed-When: Wed Jan  9 14:21:32 2002
State-Changed-Why:
    Could you please attach to your report a testcase actually
    compilable?
    
    Thanks,
    Paolo.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5268

libstdc++/5268: More fun with global stream destruction semantics

[theonetruekenny]

>Number:         5268
>Category:       libstdc++
>Synopsis:       More fun with global stream destruction semantics
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 03 14:36:00 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     Kenny Simpson
>Release:        cvs - probably also in 3.0
>Organization:
>Environment:
source
>Description:
Similar to PR 3045, in fact same example used.
The problem is that in ios.cc lines [190, 192] and [197, 199]
the flush method is called on the global stream objects cout, cerr, clog, etc

The reason this is a problem is that this will call methods on the
underlying stream buffer, which may no longer exist at this point.
(e.g. is the stream were redirected)
>How-To-Repeat:
// mostly copied from PR 3045
#include <fstream>
#include <iostream>

std::basic_filebuf b1;

int main()
{
  b1.open("foo");
  cout.rdbuf( b1 );
  std::cout << "hello\n";
  return 0;
}
// order of events:
//  magic std::ios_base::Init object is constructed causing
//    std::buf_cout constructed - BTW should this even be in std?
//    std::cout constructed to point at std::buf_cout
//  b1 is created
//  main begins
//  std::cout is redirected to b1
//  main ends
//  b1 is destroyed
//  magic std::ios_base::Init object destroyed causing
//    std::cout.flush, which in turn calls pubsync on b1 - oops

>Fix:
Don't call .flush on cout, cerr, clog, etc..
>Release-Note:
>Audit-Trail:
>Unformatted: