Re: c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

Re: c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

[bangerth]

Synopsis: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

State-Changed-From-To: open->feedback
State-Changed-By: bangerth
State-Changed-When: Mon Apr 28 14:32:44 2003
State-Changed-Why:
    As you mention, this report is basically a duplicate of the
    ones you cite. What is different from the information given
    in those that would merit keeping it open?
    
    W.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10512

Re: c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

[Wolfgang Bangerth]

The following reply was made to PR c++/10512; it has been noted by GNATS.

From: Wolfgang Bangerth <bangerth@ices.utexas.edu>
To: Joachim Eibl <joachim.eibl@gmx.de>
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't
 set breakpoints
Date: Tue, 29 Apr 2003 12:01:03 -0500 (CDT)

 > I mainly posted this in the gcc-bug-database because the mention of the
 > gdb-problem wasn't there yet. IMHO the gcc-developers should be aware
 > that their changes also can break gdb-functionality.
 
 I bet they are :-)
 
 > As a user I expect that these tools work together.
 
 I think we are all aware of this and agree that it is a bug. Given limited 
 resources in a volunteer project, it seems as if just nobody ever got 
 around to fixing this. You are invited to generate a patch!
 
 
 > I didn't really understand this at first. Probably I should rather have
 > added a note instead of opening a new report - sorry for any
 > inconvenience. (Why does adding a note require the use of email, when
 > starting a new report does not?)
 
 There's a button "Send mail to interested parties" in GNATS, that allows 
 you to attach a note to an existing report.
 
 I would suggest the following: Can you please send a brief summary of this 
 bug report to the audit trail of one of the other two PRs you cite (the 
 one that matches best)? I'll close this one, since there is really no good 
 reason to force those who can fix bugs to search through the entire 
 database.
 
 W.
 
 -------------------------------------------------------------------------
 Wolfgang Bangerth              email:            bangerth@ices.utexas.edu
                                www: http://www.ices.utexas.edu/~bangerth/
 
 
 

Re: c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

[bangerth]

Synopsis: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

State-Changed-From-To: feedback->closed
State-Changed-By: bangerth
State-Changed-When: Tue Apr 29 17:01:57 2003
State-Changed-Why:
    Duplicate of two other reports.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10512

c++/10512: g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints

[joachim.eibl]

>Number:         10512
>Category:       c++
>Synopsis:       g++ 3.2 creates multiple constructors: gdb 5.3 can't set breakpoints
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Apr 27 12:16:01 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     joachim.eibl@gmx.de
>Release:        gdb 5.3 in combination with  g++ 3.2
>Organization:
>Environment:
Suse 8.2, i586
>Description:
This is not new, but I want to point out that these two bugs are related:

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=3187
http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gdb&pr=267

In the little program below I want to place a breakpoint in the constructor via gdb. But it will never be called. The reason is that the binary contains the constructor function twice. The breakpoint will be set in the first function. But the second function is called.

// File simpletest.cpp: g++ -g3 simpletest.cpp
class A
{
public:
   A();
private:
   bool bFlag;
   int  num;
};

int main(int argc, char *argv[])
{
  A* pA = new A;
  return 0;
}

A::A()
{
   bFlag = false;   // Here I want to place a breakpoint
   num = 9;
}

It is visible even in the assembler output, created via 
g++ -S simpletest.cpp
Both constructors are identical and follow behind each other.
...
.globl _ZN1AC2Ev
        .type   _ZN1AC2Ev, @function
_ZN1AC2Ev:
.LFB6:
        pushl   %ebp
.LCFI4:
        movl    %esp, %ebp
.LCFI5:
        movl    8(%ebp), %eax
        movb    $0, (%eax)
        movl    8(%ebp), %eax
        movl    $9, 4(%eax)
        popl    %ebp
        ret
.LFE6:
        .size   _ZN1AC2Ev, .-_ZN1AC2Ev
        .align 2
.globl _ZN1AC1Ev
        .type   _ZN1AC1Ev, @function
_ZN1AC1Ev:
.LFB8:
        pushl   %ebp
.LCFI6:
        movl    %esp, %ebp
.LCFI7:
        movl    8(%ebp), %eax
        movb    $0, (%eax)
        movl    8(%ebp), %eax
        movl    $9, 4(%eax)
        popl    %ebp
        ret
...

(I've also posted this message on the gdb-bugs-site.)
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: